FeelTech FY6600 60MHz 2-Ch VCO Function Arbitrary Waveform Signal Generator

[Miti]

O dear... I thought American people have no clue they were doing this... Considering they keep doing it after they realize the mess they're in...  It's sad.

[Zenith]

 I bought my Siglent SDS 1202X-E from their UK agent (Labtronix, if anyone's interested) a lttle over 3 months ago as a result of my attention being drawn to this thread by a usenet posting into SED. It was the startling revelation of the amazingly low prices for brand new modern T&M kit that prompted me to replace my very neglected "Boat Anchor" 'scope and sig generator with high performance replacements I could carry around one handed.

I hadn't taken much interest in electronics for about five years. The scope I used most was a Tek 475 which is a beautiful instrument, but it's failed and I've had to fix it a couple of times and it's a bit big. I really ought to get a scrapper as a spares mine, as some of the parts are becoming hard to find. The same as you, I looked around and found that DSOs had come of age were quite cheap, also CRT based scopes are disappearing from the market.

I bought the Siglent for about £275 because 100MHz is enough for most things, and the Tek goes to 200MHz if needed. The Siglent offers a lot for the money, but most importantly, they seem to have a good reputation for reliability. I might have spent more and bought one with higher bandwidth, protocol decoding and four channels, but I don't find myself hankering after that. The only real problem is that sometimes it becomes confused and only shows what looks like a distorted 50Hz signal, but the Default Setup button clears that.  The manual could easily be better.

A week after getting hold of the 'scope, I realised I needed a signal generator to complement it so seriously contemplated Siglent's SDG1032X (just a fiver shy of the 365 quid I'd already shelled out on the 'scope). When I compared the specs, I realised I could get a poor man's version (the FY6600) which offered similar features (some better, others not quite so good) for just 21% of the SDG1032X's asking price, so re-read the whole of this thread... (twice!) before placing an order with a Chinese supplier who held stock in the UK and less than a week later, I had myself an AWG of my own. From then on it was downhill into the depths of DIY electronics improvement. 

I have a number of signal generators and a couple of analogue function generators. I bought the FY6800 largely because I wanted a sweep generator for the visual alignment of the IF stages of valve comms receivers. I've never seen much in the way of sweep generators on the second hand market. I've come across various wobbulator projects, but they seem limited and a lot of trouble for not much. The analogue function generators won't cut it. I could just cut to the chase and buy a spectrum analyser with tracking generator, but it seems a lot to pay for the amount I'd use it. Apart from that the FY6800 is a nice, compact and versatile signal source. Function generators are things I've always regarded as handy, but rough and ready. The FY6800 by comparison with analogue function generators, is very accurate and very stable.

The sweep function on the FY6800 looks like a marketing feature. No output sweep waveform, not even a sync pulse and you can't supply a sweep. Not much use really, unless I've overlooked something.

The VCO function is getting there, but it's not quite linear, so it would involve a custom non-linear sweep to compensate, which can be done. Here, the desperately bad PC software has enough functionality to be able to load a text file created by a program and load it as an arbitrary waveform. Also the VCO, on my FY6800 does not work from 0 to 5V, it works from around -0.9V to about 5.3V. A lot of effort to get it to work, but practical.

The FM function seems to work extremely well and is the answer. Using an 11 step staircase sweep arbitrary waveform created via a C program and the ropy PC software, and a frequency counter says that that for a sweep from 400KHz to 500KHz, all the points are where they should be to within close limits. Furthermore the modulation source can be Ch2 routed internally, so there isn't the slight nuisance of using a cable to connect to the VCO connector. I think it will do what I want.

It's this thing all over, like the 85 Ohm pad, the rubbish software and all the rest. It's a mixture of the outstanding for the price, and ridiculous failings caused by extreme penny pinching or thoughtlessness.

As you've noted with the supplied 'scope probes, they leave a little to be desired (certainly with regard to their BNC adapter tips) but aside from that niggling shortcoming, they (and the 'scope itself) appear to meet their specifications just fine. Of course, as I've pointed out several times here and elsewhere, all test kit has limitations, even the expensive kit costing ten to a hundred times more.

The probes are justifiably described as 100MHz and they were thrown in with the scope.  There are limits to what you can reasonably expect for a certain price. I don't know what Keysight, LeCroy and Co., throw in as probes. A pair of decent quality leads for an inverter welder will cost £40 and up. Cheap to moderately priced welders come with leads which are grudgingly up to the job. Expensive models often come with no leads, because "our customers are very discerning as to leads and electrode holders and prefer to source their own". Penny pinching IMHO.

With regard to the limited resolution of the voltmeter function, that's inherent to the 8 bit high speed ADCs used in these DSOs to capture the signal waveforms. With only 8 bits resolution, you're only going to see a maximum of 256 steps (twos complement signed integer - -128 to +127) in the Y axis of the 'scope's display. The ENOB setting using oversampling can extend the resolution by another 1 to 3 bits but that introduces a compromise which may or may not be acceptable depending on which aspects of the waveform you're trying to examine.

 Luckily, for the purpose of displaying a waveform on the screen, this is generally ample resolution. Even luckier in the case of DSOs, if you need to see finer detail, you can alter the gain and dc offset so as to concentrate the view over a limited Y axis range of interest in the waveform. However, as I've already mentioned, this technique has its limits and is yet another compromise (one, however, which does extend the capability beyond that of the traditional CRT based scope).

Scopes don't make great voltmeters, despite what teachers said in A Level Physics years ago,  but the feature on digital scopes is damned convenient and by playing with the offset as you describe, they can produce better resolution. For instance, I have a Li-ion battery that a meter says is at 3.98V. Casually checking with the scope gives between 3.92 and 4.00. Averaging gives 4.00. increasing V/Div and putting in an offset of 3.0V and the scope says 3.98V, but it's a bit of messing around. If you used an analogue scope to do the same, you'd probably say it's at 4V, rather than 3.8 or 4.2 and leave it at that, as a quick and dirty measurement. One feature I'd like to see on this scope is a coarse vertical shift control setting.

The number of digits after the decimal point offered by digital scopes and instruments like the FY6800 tend to seduce you into taking them at face value, but when you look closer, it's different story.

 Although my FY6600 has some DC offset, it's mercifully only a matter of a millivolt or so in the sub 500mV range (the range where that silly 85ohm pad is switched into the output circuit) and still low enough to be insignificant at the higher output ranges so I've never been tempted to tweak the trimmers (let sleeping dogs lie undisturbed lest they bite you in the bum). I may have a go at trimming out such residual offsets later on when I'm better tooled up to make sense of any such adjustments but, for now, they're not a problem and I don't see any benefit in taking the risk of making things worse. 

A 100mV ptp unterminated signal is reduced to about 39mV when a 50 Ohm terminator is added. That's consistent with an 85 Ohm output impedance. A relay cuts in at over 500mV and the ptp on the scope jumps up to half the ptp given by the FY6800, so the silly 85 Ohm pad is still there.

It's my experience that messing about trying to get rid of something like a 1mV offset, which you can live with, is likely to end in tears. You adjust the pot and get 15mV, then adjust it and get -11mV, then with safecracker fingers you can dot around zero, but never hit zero, and eventually settle for something over -2mV, which is worse than when you started. Then you can break something, involving much more work, or you find that the offset drifts over time because of thermal effects etc.

Being mindful of what I'd said in my previous post about the impedance of the generator being immaterial when driving a load matched to the impedance of the connecting transmission line (in this case, a 50 ohm dummy load hanging off the BNC T adapter plugged into the 'scope's Y channel input socket using the very short 50 ohm BNC patch lead supplied with the FY6600), I've just checked the frequency response of my FY6600 and see 74 and 75 percent of the amplitude (or, in this case, the Peak to Peak) at the 60MHz limit compared to the amplitude produced at 1MHz (the 100KHz amplitude was just a tiny fraction higher by way of a sanity check).

I have a collection of BNC patch leads, terminators and I usually terminate things properly. Are we going to see much in the way of transmission line effects with a 100KHz signal and about 1 metre BNC cables?

This is less than the 29% drop associated with the -3dB bandwidth point of 70.70707% voltage amplitude. Since this frequency is well within the 'scope's claimed 200MHz bandwidth, it would seem that my FY6600 example comfortably meets its own bandwidth specification. However, I'm just waiting for the half hour screen saver setting to time out before running the self calibration routine to repeat those tests (which event has occurred just now). I'm now waiting for the self calibration to complete and I can then repeat the bandwidth test...

 I've gotten pretty well the same figures on both channels of the generator and ditto for the 'scope channels, doing only one channel at a time. If anything, there seems to be a slight discrepancy with the 'scope's frequency response where it rises slightly below the 1MHz mark on the Y2 channel compared to the Y1 channel. Just a small fraction of a dB, nothing to get worked up about.

 All in all, the FY6600 seems to be performing to its specification on sine wave bandwidth at least. Since the PSU, opamp and TCXO mods, it now exceeds its original DC offset/signal clipping limits and frequency stability capabilities (and the impedance mismatch below the 500mV setting isn't quite so shabby as it originally was - 45 ohm versus the original 85 ohm, whilst still not ideal, is a noticeable improvement just the same).

 Since the FY6800 is essentially a revamped FY6600, I'd expect that to likewise exceed its sine wave output bandwidth claims, barring any out of the ordinary manufacturing defects (they have enough ordinary cost cutting induced deficiencies to begin with).

As far as I can tell after checking with an HP8640A sig gen, which has not been calibrated for many years, this scope exceeds the maker's bandwidth claim and I recall was pretty much flat up to about 60MHz. According to the scope the FY6800 is 1.7dB down at 60MHz and 1.1dB down at 40MHz. I find that remarkable for a unit at this price.

I must look into the the PSU, opamp and TCXO mods. There's something about this contraption that starts the urge to tinker. I think it's because in some ways it's outstanding, but it's got so many daft faults. It's a bit like the £3 radio sold in SuperDrug a few years back. People started reverse engineering it and fitting product detectors and so on. There was also an astronomical telescope sold by Lidl for about £50, which had people doing detailed investigations and finding various ways to improve it.

I've only addressed the matter of the frequency bandwidth aspect, ignoring the question of output voltage level calibration which seems to be within +/- 10% (according to my 'scope which, for the moment I'm assuming to be within its own +/-3% tolerance range - ICBA to knock up a simple peak voltage detector using a point contact diode to sanity check the voltage levels right now).

 It's all too easy to obtain misleading results when checking out such kit when you overlook the transmission line effects of the connecting cables at radio frequencies and neglect the need to terminate the cables with a resistive impedance matching that of the cable (50 ohm dummy loads with 50 ohm BNC cables). It's also worth keeping in mind the limitations of your test and measuring equipment (DSOs don't, for example, make a very good substitute as a digital voltmeter).

 If you're seeing unexpected results when checking out new kit, it's more than likely you have overlooked something vital like properly terminating a test cable. Take a pause from what you've been doing and consider the possibility that you've overlooked something in your test setup before concluding you have been 'sold a pup'.  If that doesn't resolve the issue, you can always seek advice from any technical discussion group, whether it's EEVBlog or a usenet group just as Andreax1985 has done in this case. There's nothing wrong in seeking expert advice or a second opinion from your peers.

JBG

I started messing with this after seeing Andreax1985's problem with the offset, which seemed a lot to me. I'm fairly new to DSOs and looked into it in a fairly lazy way, using the cursors, forgetting to force the scope to recalibrate first. The DVMs with a DC waveform (which I've found useful) showed an offset of under 1mV. I assume the offset is independent of the waveform. Returning to it with the scope and proceeding more carefully and using the voltage measurements, it seemed to be about 1mV. I learned something about the scope and the FY6800, and I hoped my observations were useful to Andreax1985, at least in that he didn't send back a sig gen which was behaving as expected.

It's my experience that when investigating things with a scope on its highest sensitivity range, you can see strange things, like differences between two BNC patch cables of the same length. That's at DC. It wouldn't take much say, by way of a voltaic effect to do it. Cheap Chinese BNC cables are definitely suspect.

EEVBlog looks like a thriving community full of knowledgeable people and I shan't hesitate to either ask advice or give advice when I think I have something to contribute.

[Andreax1985]

Thanks all for your suggestions, a lot of food for thought. So I did some further testing and I gathered the following evidences:

FY6800 set to DC, 10.0mV output. Here my measurements:

i) My cheapo digital multimeter: 9.5mV
ii) My DSO using a (cheap) direct BNC-BNC coaxial cable: 11.49mV
iii) My DSO using a (cheap) 1X probe, clipping the probe ground lead to FY6800 CH1 BNC ground and touching the FY6800 CH1 BNC signal lead with the probe tip: unable to get a stable measure, I read from 16 to 23mV, depending on where exactly I clip the probe ground lead on FY6800 CH1 BNC ground.

FY6800 set to DC, 0.0mV output. Here my measurements:

i) My cheapo digital multimeter: 0.2mV
ii) My DSO using a (cheap) direct BNC-BNC coaxial cable: 1.70mV
iii) My DSO using a (cheap) 1X probe, clipping the probe ground lead to FY6800 CH1 BNC ground and touching the FY6800 CH1 BNC signal lead with the probe tip: unable to get a stable measure, as before

Questions:

1) Method iii) appears totally unreliable: why?
2) First thing that comes to my mind is that if my multimeter is right, my scope has around 2mV DC offset (which, btw, is within the specs). But then why, if I switch off FY6800, the scope correctly measures around 0.1mV (whereas when FY6800 is on and set to DC 0.0mV the scope reads 2.10mV)??

[Zenith]

I see much the same. Try lengthening the earth lead from the probe with a jumper and you get some really strange results. The answer is that although it's easy to assume that ground is ground and all at the same potential, but conductors act as antennas and the longer they are, the more RFI and induced voltages they can pick up. If you are looking at the waveforms produced by a 555 oscillator at a few volts it doesn't matter, but if you are looking at low level signals, you have to pay attention to earthing and ground loops. There's a reason why they include the little spring earth connector with scope probes and there are Tektronix technical notes on probing.

As for using a DVM with unshielded leads connected to a BNC/croc clip cable, that's not really ideal for looking at voltages of around 1mV. We have to work with what we have and it seems to work fairly well, but if there are a few inconsistencies we shouldn't be surprised.

I've also noticed that it's as well to self-calibrate the scope regularly. Every time it recalibrates the no signal levels can be very slightly different. Don't forget that we are talking about measuring voltages towards the limits of what the scopes and normal, cheap DVMs can do.

What I have seen is that sometimes the FY6800 gives about a 13mV offset according to the DVM. The scope shows the same as a mean of about 13mV with a noisy waveform superimposed. The noisy waveform has spikes at around 65KHz. Powering off the FY6800 for a minute then powering it up clears it and it returns to having an offset of around 1mV. It reminds me of your original post about a 20mV offset and it looks like a definite flaw in the product. I wonder if it's been covered somewhere in this long thread.

[Andreax1985]

I see much the same. Try lengthening the earth lead from the probe with a jumper and you get some really strange results. The answer is that although it's easy to assume that ground is ground and all at the same potential, but conductors act as antennas and the longer they are, the more RFI and induced voltages they can pick up. If you are looking at the waveforms produced by a 555 oscillator at a few volts it doesn't matter, but if you are looking at low level signals, you have to pay attention to earthing and ground loops. There's a reason why they include the little spring earth connector with scope probes and there are Tektronix technical notes on probing.

As for using a DVM with unshielded leads connected to a BNC/croc clip cable, that's not really ideal for looking at voltages of around 1mV. We have to work with what we have and it seems to work fairly well, but if there are a few inconsistencies we shouldn't be surprised.

I've also noticed that it's as well to self-calibrate the scope regularly. Every time it recalibrates the no signal levels can be very slightly different. Don't forget that we are talking about measuring voltages towards the limits of what the scopes and normal, cheap DVMs can do.

What I have seen is that sometimes the FY6800 gives about a 13mV offset according to the DVM. The scope shows the same as a mean of about 13mV with a noisy waveform superimposed. The noisy waveform has spikes at around 65KHz. Powering off the FY6800 for a minute then powering it up clears it and it returns to having an offset of around 1mV. It reminds me of your original post about a 20mV offset and it looks like a definite flaw in the product. I wonder if it's been covered somewhere in this long thread.

I think I got around the 13mV offset, try if this works for you:

i) Make sure that in your "SINE" waveform settings, the default value for amplitude is 5V.
ii) Switch off FY6800 using the main switch on the rear (not the standby button on the front).
iii) Switch on FY6800, then select the "DC" waveform (using the wave button). Set "volt" to 0.000mV.
iv) Read the actual voltage with your multimeter: you should read the infamous 13mV offset.
v) Now select the "SINE" waveform (using the wave button). Set "amplitude" to zero (and while you are at it also set "offset" to zero).
vi) Go back to "DC" waveform (using the wave button). Set "volt" to 0.000mV.
vii) Read the actual voltage with your multimeter: the abnormal offset should be gone and you should read around 0mV as expected.

So, this seems a firmware flaw to me (I'm using 1.7.1 fw). Can anyone confirm?

[Zenith]

It didn't work at first. Then I remembered my FY6800 does not power up with factory defaults, so I set it to do that and it's as you describe, apart from being an 8mV offset on the DVM and I recall a 9mV offset on the scope. The rest of it is exactly as you said. However, there isn't the noise superimposed on the offset as I saw before. Firmware is V 1.7.1.  An offset which can appear depending which settings the system boots with? It has to be a firmware bug.

If I remember correctly, nothing but power cycling got rid of the 13mV noisy offset, so it could be two different fw problems.

[Johnny B Good]

To Zenith

 I chose the 1202 since, for my immediate needs, it seemed the "Best Bang for my Buck" option. The 1204 was just 30 quid shy of double the price at 700 quid and a 4 channel 'scope was one luxury I could happily forego for my immediate needs. If I needed a better 'scope in a year or three's time, I'd probably be considering something much better than the 1204 anyway.

 I ran into the "'scope gets all confused issue" too. The first few times, I just switched it off and back on again (windows PC syndrome - usually my customers', rarely my own win2k setup) before I noticed the big blue "Auto Setup" button.    Strangely, "confusion" doesn't seem to happen quite so often these days. TBH, I'm not sure whether or not I've tried the default button. Possibly I have but hitting the blue button usually sorts things out.

 All manuals, as a rule. could be better written but with something as complex and fully featured as a DSO, it's no easy task for the author to get a nice balance between simple and complex. As technical manuals go, the Siglent one isn't so bad. I've seen far worse with other (even less complex) kit.

 As for my own "Boat Anchors", they're both languishing in my basement workshop come radio shack. The scope, as you can see from the attached photos, is parked on a currently inaccessible shelf (I can't even gain enough access to unclip the cover to reveal the CRT face and controls, let alone haul it out to look for a make/model name plate).

 AFAICR, it's a 5MHz BW 'scope, possibly even dual beam - it's been more than a decade since it was last used). The Marconi signal generator has probably not been used in over two decades (but at least it's more accessible) so these recent T&M kit purchases have been a very welcome upgrade on what may well be non functional kit, never mind an upgrade specifications wise (apart from the 310MHz top end of the Marconi sig genny - assuming it could still stretch itself to such a giddy frequency limit).

 Whilst the FY66/6800 generators are, in many ways, a vast improvement over older kit that's even younger than those boat anchors of mine, I rather think you'll find the frequency stability rather leaves a lot to be desired. Before I turfed out the rather cheap 'n' nasty commodity CD player 50MHz reference oscillator chip it had been cursed with, I tested several  of my stock of standard DIL packaged oscillator modules recovered from various scrapped electronic kit over the past three decades (mostly PC boards), comparing their frequencies against the FY6600 using the siglent 'scope.

 Trying to get the 6600 to match frequency was like chasing "Wil 'o' the Wisp". When I did manage to get a meaningful display, it only took a gentle breath (literally!) of wind on the Oscillator under test to send the waveform racing ahead/behind, neatly demonstrating just just how precise you had to be to match them up.

 When I repeated this exercise after the TCXO upgrade, there was far less chasing "Wil 'o' the Wisp" and the effects of small temperature changes on my oscillators under test were far more predictable and stable. Now that I've gotten hold of a u-Blox M8N Arduino/RasPi GPS module to generate a 10MHz reference directly from its PPS pin, I can see that the TCXO stays within 20 ppb from cold startup to a nice toasty room temperature of around 23 deg C and settles to within 10 ppb of frequency ( it's currently just over 7.5ppb in error at 22 deg C at 0100 hours UTC).

 No doubt the fact that the TCXO module is in the direct firing line of the small cooling fan to hold it to within two or three degrees of room temperature (instead of running at 50 deg C like the original oscillator chip did) contributes to its rapid "warm up" since it doesn't have to do very much warming up to reach thermal equilibrium. Even so, there is still the issue of 'self heating effect' to account for but this is negligible compared to what the original oscillator chip was subjected to by being within 1cm of a 70 deg C heat source (the three analogue voltage regulators or "Stooges" as they might best be named in this joke of a main board layout).

 I tend to leave the signal generator running 24/7 (it only consumes 8 watts at its current settings) and for the past three "mornings", my initial frequency check has shown it unaccountably running within 0.5 ppb of the GPS reference. However, within minutes of taking stopwatch timings, it will start to drift off, typically showing all the signs of passing through the exact frequency of 30,000,000.000Hz (I'm monitoring the third harmonic) before it drifts to an initial offset some ten to fifteen minutes later of around 4 to 5 ppb where it will then very slowly drift by another 3 to 5 ppb over the rest of the day.

 This 'morning behaviour' is a little 'spooky' since it implies it can sense my mere presence in the room. My best guess is that it's not so much my presence as the effect of powering up the adjacent TS140S HF rig and the 10A 13.8v psu made up from a couple of 40 year old 5v smpsus strapped together in series and maladjusted to 6.9v each. The HF rig consumes 14W in Rx mode with the PSU raising the 'from the mains socket' load to 25 to 30 watts. It seems unlikely to be convective heat polluting the fan intake and I can't see the slightly increased loading on the long but heavy duty mains extension cable having any effect on the FY6600's smpsu voltages sufficient to upset the 3.3v regulator on the oscillator board so my best guess is that it must be responding to infra red energy from the 13.8v PSU and the HF rig (not forgetting my own small contribution, assuming this IR hypothesis has any validity).

 I can now see the motive behind adapting these generators to utilise an external GPSDO 10MHz reference. A 0.1ppm TCXO whilst undoubtedly a vast improvement over the commodity XO chip, is still some three or four orders of magnitude short of GPSDO perfection. Anyone motivated enough to put Feeltech's 'wrongdoings' to rights is going to find the challenge to achieve the best possible frequency accuracy and stability irresistible.

 Until I got my hands on the u-Blox module and active antenna, I was still thinking that Arthur Dent's posting a year or so back regarding his replacement of the XO chip with a 10MHz OCXO and a low jitter PLL chip to mulitply it up to 50MHz, along with a change over switch and external BNC socket to plug an external 10MHz reference source into the FY6600 was a little OTT to say the least.

 I can see his point and I'm now sorry I didn't use an easier to obtain 10MHz TCXO to use with a PLL chip rather than aim for a 'drop in replacement' 50MHz TCXO. Never mind, I daresay I'll be able add the PLL chip and automate the changeover on sensing the presence of a valid 10MHz reference on the external reference socket to minimise the switching glitch between local and external references. I suspect any nasty glitching on the 50MHz clock signal might crash the main board so this might be more complicated to achieve, especially in the event of loss of the external reference, whether by unplugging it or just simply due to a loss of the feed for any other reason.

 Anyhow, that's a new adventure for my own satisfaction. Returning to the business of the 85 ohm attenuator pad still being present on the FY6800 main board, that does rather suggest a deliberate action on Feeltech's part since they've had ample opportunity to put this particular "Skoolboy Howler" well and truly behind them with the new iteration of the main board.

 Either they're still leaving temptation to tinker so as to minimise after sales service costs (voided warranties) or else are bowing to pressure from the likes of Rigol and Siglent to retain such annoying defects so as not to seriously dent their own sales, otherwise I can't see how they could have missed the opportunity to correct this error with the reworked board designs. They must have become aware of the problem straight from day one since they've obviously applied a fix in the firmware - admittedly one that could only be effective in the Hi Z condition they elected to correct for.

 It's not a fatal flaw so much as an egregious annoyance. You can get round the issue of generating sub 500mV signals by never going below the 500mV setting and using an external 20dB 50 ohm pad instead. Aside from such limitations and seemingly deliberately included annoyances, these Feeltech products are otherwise excellent value for money test gear.

 Sure, the penny pinching effects are there but I don't think all of the deficiencies are down to random penny pinching alone. A lot of it seems to be down to 'defective by design' since not all defects are the result of saving a penny or two. That 85 ohm pad cost them just the same to implement as a proper 50 ohm pad would have.

 When it comes to accessories like 'scope probe leads and welding cables, it is just a matter of 'penny pinching'. The decisions taken by manufacturers are rather dependant on what their customers' needs are. For a budget item where expectations are low, the manufacturers see a chance to bundle cheap but serviceable accessories to provide 'instant gratification'. With high end kit, the manufacturers will often take the view that it's better to let their customer decide what to spend on their own accessories which may be nothing at all if they already own high quality accessories to begin with rather than spend a small fortune supplying accessories of a quality that matches that of the product which their customers may not need or even wish to use.

 Regarding the lack of a coarse voltage offset adjustment, I think you'll find it in the form of a rate control where it jumps in bigger increments if you twist the control fast enough. However, it tends to result in overshooting the mark. I guess familiarity from lots of practice in its use helps mitigate this problem. 

 I guess a firmware update might tame this "All or Nothing" effect of the rate control. Unlike the volts per division setting control where the push action toggles between coarse and fine, the push action on the offset adjustment is used to reset back to zero offset, leaving rate control as the only convenient alternative way to increase the step size of the increments. Short of adding a separate push button or adding it as "pull" action on these knobs, there's really no easy solution. Although it is a little hit and miss, I don't find this is that much of a problem.

 As for the staggeringly high digit count on the FY66/6800 frequency setting (14 digits worth! And all of them valid), not even the best frequency reference standards (Rubidium or GPSDO) can match that amount of resolution. They might all be valid but even with the best available atomic clock derived laboratory frequency reference, the last digit or two will be meaningless or, at best, only meaningful for frequency settings lower than 1MHz.

 You can take some comfort from the fact that a lack of digits won't limit its ultimate frequency setting precision, not even when fed from a Rubidium or GPSDO reference clock. The fact that its original XO chip makes a complete mockery of the nine least significant digits in the display is neither here nor there. Once you're familiar with your test gear's limitations, such 'anomalies' cease to be an issue. However, in this case, the desire to upgrade the XO to a decent TCXO or OCXO (with or without an external atomic frequency standard clock drive option) can become an irresistible challenge. 

 Regarding the various mods to the FY6600, most of the more interesting ones started to appear around page 8 of this 73 page forum. I've already mentioned most of them in recent posts in the last three pages or so but if you're interested, it's well worth a skim through the early pages to pick them up. Be warned though, you could land up spending a solid couple of evening's worth trying to plough through them all. Obviously, you can skip the discussions and progress reports on fixing the firmware issues that sadly afflicted some of the early adopters stuck with bricked firmware (versions 3.0 AFAICR). 

 Mention of the 3 quid radio and the Lidl astronomical telescope modding exercises, that reminds me of my own, paused project to add an electric starter function to the Parkside PGI 1200 B2 inverter generator which Lidl were selling last April and May for just £99.95. I bought one, returned as faulty for an exchange for one that also had a fault and landed up with a refund since they had no others in stock to exchange.

 Frustratingly, I'd discovered by then that both faults were the result of handling 'damage' from the endless motorway road trips they were routinely being subjected to whilst being shunted from store to store or central depot because the shop floor also had to double as the shops' stock rooms and each new "This Week's specials" which had to have space cleared to display them, they'd be shipped out of the store.

 These 'stock faults' (quite aptly named in this case) were quite trivial to fix since it was either a stuck oil level float or else a dislodged connector on the inverter module both of which could readily be fixed. About a week after that episode, I tried a less likely Lidl shop in the unlikely hope that they'd also received a delivery of these excellent little generators and very much to my astonishment, they had three in stock.

 I picked the two with the least scuffed boxes to take to the checkout. I just wanted to do any returns wholesale rather than piecemeal one at a time. As it happened, both proved fault free so I had to choose which one to return to the store for a full refund. Unfortunately, I hadn't thought to verify the 1KW overload limit settings and I landed up with the one that regarded 980W as the trip point instead of the expected 1030 or 1050 watt setting (whether the one I returned did have a higher set point remains an unknown).

 Unfortunately, unlike every other inverter module I've ever seen pictures or videos of, these Parkside ones don't have any of the otherwise ubiquitous trim pots to adjust such niceties as overload current trip points and voltage settings - not one single trim pot! I guess the adjustments are flashed into the controller's firmware, never to be accessed with something as simple as a trimtool. 

 Never mind, when these come back into stock again, as they must, I'll snag a couple more to use as donors to replace the ungenerously set unit in mine, assuming at least one is more generously configured with regard to the overload trip point. If both are good in this respect, I'll keep one and return the donor machine for another full refund. At just £99.95 each, it's worth keeping an extra one just as a spare.

 Anyway, after struggling with the recoil starter to get it up and running on account the fuelling system not being the simple gravity feed type where you can use a float tickler to flood the carb to ease starting but one relying on an engine vacuum driven fuel pump which after running the float bowl dry to prevent it gumming up means you have to crank it several times to prime the carb float bowl, I considered how I might be able to add some sort of electric starter to it. There was no easy way to add a conventional starter motor and since the inverter module blocked access to the end of the crankshaft where the common trick of using a cordless electric drill with a suitable socket spanner to engage the retaining bolt so as to spin it up, I started to think 'outside the box' when I realised that the 6 or 7 pole pair three phase permanent magnet 380v alternator was nothing more than a BLDC motor in disguise simply waiting to be turned into a starter motor by a genius like myself. 

 I have to say, I've managed to get as far as testing with a 60v rated BLDC scooter motor controller and 48 volt's worth of battery, proving the concept. Unfortunately, sensorless control option just doesn't cut it with such mechanical loadings. I could get it to spin up with the spark plug removed just fine but even just blocking the spark plug hole with a finger tip was all that was needed to confuse the sensorless control algorithm and stall it.

 However, fitting an angular position sensor to the end of the crankshaft will overcome that problem once I finally get around to figuring out the pole pair count so I can order a unit programmed to emulate a hall sensor drive signal more accurately than even the best conventionally hall sensored BLDC motors can achieve. The project is paused because I'm waiting for another Round Tuit. Besides which, I'm still busy fettling the test gear needed to test and configure my electric start system.

 Harking back to Andreax1985's problem with the offset voltages he's seeing, I suspect it's probably more to do with thermocouple effects or galvanic corrosion within the hard mains earth loop between the 'scope and generator grounds. Compared to such a low impedance ground loop circuit, the scope lead's shielding resistance and that of the crocodile clip contact resistance is far too high to shunt such low impedance voltage sources effectively. If he can bring himself to disconnect the safety earth and cross bond any of the FY6800's BNC grounds to his scope, he may well find a significant reduction in this offset voltage. This was one of the reasons why I used an 11K resistor in series with the signal ground and the mains protective earth on the C6 socket (although, to be fair, I hadn't considered DC offsets, just ac induction voltages).

 Regarding the photos. Apologies for the low resolution but I'd forgotten about the 640 by 480 setting from when I hadn't wanted to exceed the 2MB limit in a previous posting. Also the two offset angle views of the Marconi signal generator were required to avoid reflection from the on camera flash burning the image out.

 The scope shot is the best I could do on account of the foreground clutter as shown in the next image (the scope is at the top in that picture). I thought the Min/Max thermometer would give you some idea of the storage environmental conditions. Whether any of this kit is still serviceable after spending some three decades in my basement remains an unanswered question although the remarkably good condition of my HF transceiver, which had spent just two decades there, does offer some hope.

 JBG

[Andreax1985]

I decided to float my scope (ok, I know it's not advisable, but considered the particular application I felt safe in doing so) in order to break possible ground loops. I can confirm that the residual dc offset was due to ground looping and now my readings are perfect. My device is working properly and I'm not going to return it.

[Johnny B Good]

I decided to float my scope (ok, I know it's not advisable, but considered the particular application I felt safe in doing so) in order to break possible ground loops. I can confirm that the residual dc offset was due to ground looping and now my readings are perfect. My device is working properly and I'm not going to return it.

 Excellent news! Thank you for that feedback. I'm so glad my 'educated guess' actually "nailed it" for you. 

 Since you're planning on keeping hold of the generator, does that mean you're going to modify its earthing circuit with a 10K or so resistor to avoid the low impedance earth loop problem but still suppress the half live mains touch voltage issue (that so many of us here were bitterly complaining about with the FY6600) to a more acceptable half volt or less? The PSU is still essentially a class II double insulated type intended to be used without any need for a safety earth connection.

 Looking at this, it seems Feeltech had a just, if flawed, reason not to encumber the FY6600 and its predecessors with a safety earth connection. The flaw being the penny pinching profit motive and the justness being the avoidance (more by accident than design) of this earth loop problem.

 In responding to all the demands to eliminate the half live mains touch voltage issue, it seems they've created another issue by hard wiring the internal ground rail to the safety earth connection rather than choose the sensible compromise of a resistive connection to reduce the touch voltage to an acceptably low value without introducing a low impedance ground loop problem.

 It would seem that FY6800 owners can add yet another modification to their "To Do List", adding a 10K or so resistor to the earth connection being the second of the earthing related mods with the first being to rejoin the cut off  0v rail wire in the ribbon cable linking the PSU to the main board to double up the 0v rail wires again as per the original FY6600 arrangement.

 I do have to wonder at this demonstration of utter ineptitude on Feeltech's part. After all, a safety earth connection isn't actually required with the cheap commodity three rail class II smpsu board they're using so a 10K resistive link to avoid earth loop issues yet reduce the touch voltage from a high of 90 odd volts to an insignificant half volt wouldn't have been deemed inappropriate on the grounds of safety.

 There's clearly three hierarchical levels within the T&M ecosystem with the likes of Keysight occupying the top slot, Siglent the middle and Feeltech the bottom and I can't help imagining conversations between Keysight and Siglent over the fine art of not treading on toes with Siglent then passing on this favour to their own underlings, Feeltech being one of them in this case. Such an imagined scenario offers a compelling rationalisation of the apparent "Shooting oneself in the foot" exercises that Feeltech seem to be so fixated on repeating at every opportunity that comes their way to improve their products for very little to zero out of pocket expense. The persistence of the 85 ohm pad into the FY6800 model being a prime example of such missed golden opportunities to correct the mistakes of earlier production runs.

 Quite frankly, it's impossible to shake off the feeling that there's an element of design to the penny pinching practices at Feeltech since most of the resulting shortcomings are amenable to relatively easy fixes on the part of a reasonably skilled end user prepared to void the warranty.

 If you include enough such warranty voiding enticements in your product, you may never have to deal with a single warranty return. This is a prospect made all the more likely by the rock bottom pricing and the one way export shipping subsidy adding the barriers of return shipping costs and delays to the headaches of collecting on a Chinese warranty. I do have to wonder whether anyone has ever managed to collect on Feeltech's warranty without being left out of pocket.

JBG

[Zenith]

@ Johnny B Good,

I've always thought specsmanship crept into scope choices far too much, especially when a company is paying for them.

The scope has gone out to lunch problem caused some worry at first, What was going on? Would there be the hassle of sending it back? And so on. It seems to happen after using slow sweep speeds, but it's not a frequent thing and when it dawned that pressing the Auto button or the Default button clears its head, it's an occasional minor nuisance.

My Boat Anchors are something like 15 valve comms receivers, and a dozen valve scopes. There are also about 20 or 30 items of test equipment mainly from the 70s and 80s. I don't recognise that Marconi sig gen. I have a Marconi TF144 which I eventually started taking to bits and it's now in the garage. It was was one problem after another, dead meters, sealed and not easily fixed with similar meters, a dead double tetrode, which took ages to find a replacement for, and a host of other things. It really was a boat anchor. I have an HP8640A, which is excellent, apart from the silly plastic gears.

The FY6800 had been running for a day or so. I have a couple of Racal frequency counters with overnised crystals, which haven't been calibrated for probably 20 years and so probably wouldn't agree closely with a precision standard, but I'd bet are stable. I connected the FY6800 to one of the frequency counters and the scope, terminated. I set the FY6800 to a 10MHz sine at 2 volts. The frequency counter said 9.999923MHz , a couple of hours later, 9.999922MHz and ten minutes later, 9.999924MHz. A couple of hours after that it was reading 9.999917MHz. The scope said 9.99996/7MHz. I see the frequency measurement on the scope as a convenience more than anything else.

So an accuracy of maybe 100Hz in 10MHz setting and drifting through less than 10Hz in 10MHz over a few hours - this is assuming that the Racal wasn't drifting and isn't wildly out. I'd expect that from a decent quality, but not special crystal module in the FY6800. I'd say it's enough for most needs, but I'm inclined to get hold of a  u-Blox M8N Arduino/RasPi GPS module you mention and do the frequency module upgrade. As you said previously, the FY6800 is likely to be aimed at the hobbyist.

What's behind some of the strange decisions in evolving the FY6800 from the FY6600 is open to all sorts of theories. Some bizarre culture in FeelTech. They got it working but it was all so skin of the teeth that they were frightened to modify anything too much. The lead engineer who came up with the design got another job, or left to form his own company, and they were left with no one who really knew what they were doing.  One aspect I find odd is the PC software where some features don't appear to work at all. If you enter data in the wrong way in some fields, the program crashes and there's the almost unusable feature for drawing an arbitrary waveform. I get the impression that someone with no knowledge of, or enthusiasm for, software was presented with a PC with a VB system and told to write the software. They cobbled up something which could be described as the software, Feeltech had something they could include as the software, and whoever wrote it went back to their regular job with a sigh of relief. Of course, it could be as you say; they've generated a following and a lot of publicity for the product by its defects creating so much interest in retrofitted improvements.

OTOH, my TTI 5MHz function generator, calibrated according to the manual, has a very inaccurate scale and can only be set to 455KHz by using a frequency counter and gentle caressing the frequency control to get close to 455KHz. Half an hour later it can be at 457, half an hour after that, 454.  I came across a Maplin catalogue from 2011/12. They had two function generators, both up to 3MHz. One was £145 and the other £205. I doubt either would bear comparison with the FY6800, for all its little faults.

I saw the generators in Lidl but I have no real need for one. Aren't they 2 stroke?  My impression of Lidl and Aldi specials is that they are always interesting and sometimes remarkable for the price. I had a need for some 1 metre SDS drills a few years back. Wickes were £15 and upwards each. Aldi were selling three for a tenner. I thought we'd do the job with the Aldi drills, after which they'd be knackered, but we'd be a fiver in, so why not? As it was there was about five times as much work as we thought. We mainly used one drill and when we finished it was as good as when we started. It's a strange thing to find 1 metre SDS drills in a grocery store, but they do them quite often and they always seem to sell.

If you have equipment which has been kept in the damp for a time, it's very good idea to leave it in the airing cupboard, or somewhere similarly warm and dry, for at least a fortnight. Then it's worth reforming or replacing electrolytic caps and being generally cautious about powering it up. Bunging it on the mains can cause some unpleasant surprises.

[beanflying]

From memory mine was always always about 30+ Hz high at 10Meg against my GPSDO and I think early on in the thread there is some mods done on the time base if you look for them with different TCXO's fitted.

What all of this shows is that they were chasing the bottom line against 'their' competitor in JDS instead of focusing on being best in class. What they failed to see is there is a huge gap to anything better they could have played in.

Better Power Supply
More accurate Time base
Better Op-amps and trimming in the factory
Upgradable Firmware
Maybe $20-30 retail needed to get it done right?

Compared to most likely hundreds of hours of 'fun', speculation and tinkering by us all. Still great bang for your $ 

[Zenith]

What all of this shows is that they were chasing the bottom line against 'their' competitor in JDS instead of focusing on being best in class. What they failed to see is there is a huge gap to anything better they could have played in.

Better Power Supply
More accurate Time base
Better Op-amps and trimming in the factory
Upgradable Firmware
Maybe $20-30 retail needed to get it done right?

Yeah. Conspiracies exist, and a few conspiracy theories have been proven true, but I think the cock up theory usually wins out over the conspiracy theory.

So thinking about the history of the instrument business, I don't see JBG's view of the hierarchy ringing true.

Look you lot are educational and skint hobbyist, see? A nice little function generator you've got here. Too nice. You're getting above yourselves. So we'll have a crappy 85 Ohm pad in here. Your choice of op amps is unfortunate, and could prove unhealthy. Change it. As for that TXCO and the option for an OCXO and an external reference, err, no. Now if you can't see the logic of these design improvements we have associates who will explain them more clearly. They are persuasive if not altogether civiliised.

This is a cock up.

Compared to most likely hundreds of hours of 'fun', speculation and tinkering by us all. Still great bang for your $ 

Better value than going to the pictures.

[Johnny B Good]

@Zenith

Specifications are an important marketing tool, especially in such a competitive market as this where the feature set can be overwhelmingly complex with modern DSO and MSO kit so 'Specmanship' is going to be inevitable.

 My first experience of the Siglent going into a fugue was a little disconcerting at first but bearing in mind all the microprocessor control and digital processing used in a modern DSO, I instinctively applied the 'switch it off then on again' technique the first two or three times this happened before it occurred to me to try the 'Default' and 'Auto setup' (amongst other) buttons to avoid the protracted power cycling method which provided reassurance that these episodes were more a case of 'Operator Error' than any indication of something more serious. These buttons are obviously provided to back out of any conflicted settings that may have been applied by operator abuse.

 I had hoped to be able to discern the model number of the Marconi signal generator from the photos, forgetting I still had the camera set to the 640 by 480 setting from a previous photo shoot to minimise the risk of going over the 2MB attachment upload limit when posting images of the u-blox GPS module to [urlhttps://www.eevblog.com/forum/projects/ublox-neo-m8n-gps-navigation-signal-amplify-module-for-arduino-rasppery-pi/][/url]

 Hopefully, that link hasn't been mangled too badly (if at all - I'm not quite sure how you're meant to use the "Insert Hyperlink" button in the compose a reply window). It looks ok here so that will take you to a thread I started some 2 1/2 weeks ago trying to get some advice and opinions over its suitability as the basis for a homebrewed GPSDO.

 I think I must have bought the sig genny at an amateur radio rally some thirty years ago. The scope was an earlier purchase from a local government surplus shop afaicr some time around 1972/73 to let me record oscillographs of test waveforms of a 200W per channel (RMS per 4 ohm loads) PA amp I'd been commissioned to design and build by a friend who played in a local band. The need to photograph the CRT display led me into the amateur photography hobby, something I'd previously never had any interest in. It's surprising how one thing leads to another.     Anyway, the one of each item, 'scope and signal genny, was all I needed and could afford so I neatly avoided accumulating any more 'boat anchors'.

 Your frequency calibration checks of the FY6800 appear, on the face of it, to be fairly respectable considering how Feeltech cheaped out on the XO chip. You'd expect OCXO s to be pretty stable even if they're in need of recalibration after twenty years. Mind you, if they've not been used very often, they might not be all that far off frequency even twenty years later. The hardware frequency counters built into these Siglent 'scopes aren't remarkably accurate. Mine, after 8 or more hours of warm up was adrift by about 4.33ppm (it read high on frequency to the tune of 130Hz on a 30MHz signal afaicr). Not very impressive at all imo but, as you've observed, more a courtesy feature to get a ball park figure.

 Your own thoughts on what's going on at Feeltech are possibly closer to the truth of the matter. The cynic in me tends to nurture conspiracy theories since I prefer to look for less chaotic reasons for such otherwise inexplicable behaviour. Whatever is going on, they're certainly not doing themselves any favours, more's the pity.

 There's no doubt that the Feeltech generators, as cheap as they are, are in a totally different league to even modestly vintage test gear that wasn't made by Hewlet Packard.

  If the generators you saw in Lidl last year were either the very retro styled PGI 1200 A1 (still surprisingly carrying their original £129.95 price tag of the previous year) or the more suitcasey styled PGi 1200 B2 at just £99.95, they're both 4 stroke powered in spite of the A1's rather retro two stroke styling with its top mounted gravity feed fuel tank and open frame with plastic panels slung around the sides to disguise its open framed nature.

 It's rather surprising that an earlier version of the B2 model should be placed on sale about a month after the B2s had been on sale with a 30 quid higher price tag. Hopefully, the B2 models will reappear again come the Spring and I'll be able to pick myself a couple of donor machines to resolve the irritatingly low overload trip point on my current acquisition.

 My interest is in their use as an alternative to investing several hundred quid in SLA battery packs just to extend the autonomy of my APC SmartUPS2000 an extra 6 to 8 hours. Basically, it's a backup to the UPS. As an autonomy extender, it's a far less expensive option to installing a bigger battery pack which, at the end of the day, is just an expensive consumable.

 With only a 1KW rating (30 seconds surge rating of 1200W), it's a little on the marginal side but the 2KW rated Workzone inverter gensets that Aldi had on sale about a month or so after I'd acquired the Parkside unit proved to be total shite, to put it mildly (and I should know since, optimist that I am, I tried three of these before giving them up as a lost cause).

 Compared to the Workzone units, the Parkside ones are inverter genset perfection, not the slightest hint of engine revs dropping on application of full load in the eco idle setting, unlike the Workzone units which threatened to completely stall out when subjected to full load from the half loaded state. AFAICR, the Workzone units were permanently in eco-mode with no option for 'normal mode' so there was no way to mitigate the problem like there was with the Parkside units to reduce the risk of a stall whilst they were warming up from stone cold (the only time I managed to stall a Parkside unit was switching straight to eco-mode after startup followed immediately with the application of full load - restarting it and giving it a couple of minutes to warm up solved the problem).

 Take it from me, those little Parkside PGi1200 B2 suitcase gensets are well worth the money, even if they hike the price up to that of its predecessor. The Aldi Workzone units otoh, aren't worth touching with a ten foot bargepole so take note.

 I hear what you're saying about the effects of prolonged damp on electronic kit, particularly when bias voltages of two or three hundred volts are involved. As it happens, I do have a small 180VA Variac but it's a a panel mount unit and I don't have a suitable box to mount it in. Also, I'm not sure whether a 180VA rated unit would be up to the demands of an ancient scope or even the Marconi sig genny but I guess it might be able to cope with the overload for the few minutes it would take to bring the voltage up slowly enough for the electrolytics to reform.

 Otoh, I might just say "Bugger it! I'll just switch them on and take my chances.". If either or both go bang, then so be it. That'll just be a boat anchor or two less to clutter up my storage space. They're so ancient that at best they'll simply be electronic curios to admire for their sheer chutzpa at even pretending to be test gear. Life's just too short to be trying to repair such ancient kit of limited utility, hence the fatalistic attitude to the problem. Apologies if that sounds a little cavalier but that's just the way things are right now.

JBG
 

[Noy]

So today i put the THS3491 into the generator and put a new proper heatsink pad below the Heaspreader.
Now i can get 22 Vpp with 20 MHz sine.

But i get now an overshoot with square wave If i go above 500mV Independent of high impedance or 50 Ohm termination...

[Johnny B Good]

So today i put the THS3491 into the generator and put a new proper heatsink pad below the Heaspreader.
Now i can get 22 Vpp with 20 MHz sine.

But i get now an overshoot with square wave If i go above 500mV Independent of high impedance or 50 Ohm termination...

 Remind me please, Noy. Is this the FY6600 or the FY6800? I've lost track of who has what.    Also, can you provide some details about the 'scope and the way you've got it connected up to the generator. If you can post a screenshot or two of the relevant traces, that'll help us to work out what's actually happening.

 I'm a little puzzled as to why you're getting 22v p-p. This is 10% above the 20v p-p limit of the generator's amplitude range which begs the question as to whether this is a calibration issue or a voltage magnification effect due to the use of a few metres of unterminated 50 ohm BNC patch cable.

 Out of curiosity, I repeated your tests using the half metre BNC patch lead that was supplied with my FY6600. I wanted to verify whether the square wave was limited to a 5 or 10 volt p-p level or not at 20MHz and lower frequencies since, if it had been (it isn't - the 5v p-p limit only kicks in for frequencies above 20MHz), that would have excluded the THS opamp from the circuit (that only gets switched into circuit for p-p settings above the 5 volt mark).

 That reads like you've installed just the one THS3491 (U21 and U22 showing the THS3001 locations) and left the original THS3002i chip (U5A and U5B) in place. The reason why I'm querying this is because the reverse engineered circuit so kindly provided by ebel0410 last year (which he uploaded to the Github repository)  implies the simultaneous presence of the original dual opamp along with the additional single opamps. That's a little confusing to say the least although most would realise that it's either just one or the other arrangement (just the original 3002i in the U5A/U5B locations with the U21/22 left empty or else the original 3002i removed with either 3001, 3095 or 3491 ICs fitted in the U21/22 locations.

 Hopefully, what you actually meant was that you'd removed the original THS3002i chip and fitted a pair of THS3491 chips into the vacant U21/22 locations. Assuming this to be the case, one possibility for the overshoot could be the need to tie the Vref pins to a voltage 5 volts or lower than the V+ rail voltage, typically either to ground or else the V- rail. Provided this is the case, the /PD pin can be left to float as with the earlier 3495 chips. The easiest way to meet this requirement, is simply to strap pins 1 and 4 together (the data sheet insists that the Vref pin (pin 1) should not be allowed to float).

 I've attached some screen shots for you to compare with yours. The first two are from the end of last year just after I'd upgraded the opamp. The second pair were done just now, using the half metre of supplied BNC patch cable, without termination (the terminated traces looked virtually the same only, of course, at half the p-p voltage).

 JBG

[Zenith]

@Johnny B Good

The link was garbled but I found the thread with a search. Thanks for trying.

Boat anchor receivers are sought after and some people like to restore them to as issued in WWII condition. Boat anchor test equipment has much less of a following and people don't seem sentimental about old sig gens. If they work, fine, but the size and weight is a pain. It they break and can be fixed easily, fine. If they are too much of a PITA, they go. It's a shame, because the standard of construction can be very impressive.

My valve scopes are from the 60s mainly. They are not quite boat anchors, they are usable as scopes and occasionally I fire them up and use them. I've really enjoyed sorting out their problems and getting them working properly. I think the interest there is that I wanted one as a kid in the 60s, but they were completely unaffordable.

Feeltech may have put in a slightly better crystal lately, or maybe I got a good one. The accuracy is fine for what I bought the FY6800 for, sweeping IF stages for visual alignment, and for that purpose the drift is essentially non-existent. I'm still tempted to tinker 'though.

I've always found the Racal frequency counters very useful and very reliable, and they're quite small. Looking on ebay, I saw a couple for about £30.

From what you say, you have a perfectly good reason for buying the Parkside generator. I'm sure a lot of people buy the Lidl/Aldi specials because they can't resist a bargain and wonder whether they need it afterwards. I've fallen into that trap once or twice. You're probably aware of this, but petrol can go stale and it's hard to start an engine with stale petrol. I think what happens is that volatile fractions, which allow  a spark to ignite the fuel, oxidise. This can be a real problem with engines which aren't used for months, but you expect to start with little trouble when they are needed. However, you can buy fuel stabilisers which stop the fuel going stale. I believe a lot of two strokes oils contain some stabiliser.

It's just possible that your boat anchors have some very desirable parts, such as the rectifier and possibly a KT66 used in a stabilised power supply.

[Johnny B Good]

@ Zenith

 By the time I could have afforded any of those boat anchor radios, I had other priorities, namely the acquisition of a decent HF transceiver, hence the TS140S (the first of Kenwood's modern synthesiser tuned budget HF transceiver models). I bought it at a radio ham rally nearly thirty years ago. It had a 750 quid price tag but I managed to do a part exchange deal which reduced the damage to a mere 500 quid. The strange thing is that I can't recall just what it was I'd offered in part exchange (and I've tried my best to recall just what it had been over the past year or so).

 I suspect it may have been a customer return on account half of the commands in the PC control software didn't work. Weird things like the RIT would only tune upwards, the downwards RIT tuning command being ignored amongst similar oddities.

 Luckily for me, it used the Z80 cpu in its controller board and I was able to pull the eprom out and copy it to a file which I could run my home brewed Z80 disassembler on to get a full listing from which I was able to track down the command interpreter routine and its look up table. The table search loop counter value had been corrupted to half its intended value, probably by a single bit error in the eprom so was a very easy fix - just a matter of burning another eprom (or two just to make sure of having at least one good spare) and dropping it in as a replacement, after which, the radio worked perfectly with the PC controlling software.

 As for the excellent standard of construction in such vintage kit (radio and test gear), it had to be in order to achieve the best possible frequency/timing stability out of the analogue tuning systems employed. Nowadays, it's really only the TCXO or OCXO reference oscillators that need to be protected from mechanical stresses and only when ppb accuracies are demanded. When the requirements demand just a mere 1ppm level of accuracy, the XO module can even be placed directly on the main board, right next to a bunch of analogue voltage regulators if that's a manufacturing convenience where heating it up to 50 deg C isn't deemed any great detriment. 

 As for the desire to "Tinker", even now, I'm wondering where it'll all end. The GPS module was acquired simply because the R4LW transmitter is at too low a frequency for attempting to trim the TCXO (it's good enough for checking the calibration directly but I'd have to build a 10MHz disciplined reference receiver to turn it into a useful calibration aid) and the 10MHz WWV broadcasts can't be received due to propagation conditions being at an all time low (we're currently in a sunspot minima right now). As things stand, with talk of the R4LW losing its NPL reference status in the next few years, the GPSDO project seemed to be the better option than making a special frequency reference receiver based on the 198KHz broadcast.

 Just as it stands, without adding a disciplined 10MHz OCVCXO to it, the 10MHz output on the PPS pin provides a far more stable reference calibration source than the WWV broadcast, free as it is of fading and phase distortion from skywave propagation effects. Indeed, programming it to output 198000Hz to beat against the R4LW transmission as a sanity check that this still justifies its status as an NPL grade reference, revealed no discernible phase shift over the space of 90 minutes. As a sanity check that this zero beat effect was actually real, I'd actually programmed the gps module to output a 1Hz offset each side of the frequency (it can only be programmed in 1Hz steps) to time 60 beats in exactly 1 minute each time (to within 50ms according to my Smiths stopwatch with its 30 seconds per sweep of the dial calibration and high speed escapement).

 For the past week or so, I've been monitoring the frequency drift of the signal generator against the third harmonic of the GPS 10MHz output which reveals a small drift with temperature somewhere in the region of plus or minus 5ppb with the odd random shift of one or two ppb every so often (and a spooky reaction to my presence in the morning as I check its overnight drift).

 The stability of the TCXO was threatening to become an obsession but I think I've collected enough data to convince myself that it's just about as close to calibrated as a 0.1ppm TCXO can get. The next obvious step is to turn the GPS module into a self contained reference (with or without a 10MHz VCOCXO for it to discipline) with buffered outputs from which I can feed to the generator as a frequency reference locked to the GPS atomic clock standard.

 The TCXO module is at least two and half orders of magnitude more accurate and stable than the original XO chip so I'm satisfied that its self contained reference source easily exceeds the accuracy of any other kit I have which might be in need of a calibration check. Any later kit that might boast sub 0.1ppm frequency accuracy will almost certainly be equipped with an external 10MHz reference input socket of its own anyway (as will the FY6600 when I've finished modding it).

 To be fair, we've only suffered one 4 hour prolonged outage of supply in the thirty odd years we've lived at our current address and that was just about thirty years ago. I did happen to have my first Emerson30 UPS back then (I still have it but it's currently without a battery pack - two 7AH SLAs).

 The power went out in the early evening just as I was about to head off to the radio club so, after shutting down my PC, I moved it onto the first half landing to power a floor standing lamp with a Philips SL13  to keep the hallway and stairs illuminated as a safety measure. I had to leave it running since it couldn't be "Black Started". I don't think we had a downstairs toilet back then, just the one toilet on that half landing. It kept that lamp burning right through the outage, during which the children had used it to do their homework by. The power was back on by the time I returned and we've never suffered any further outages since (just one or two briefly lived brown out events and the odd brief dropout over the years).

 On the face of it, there seems very little justification to even keep a UPS up and running but I prefer to have such protection in place even against all perceived need. I know that Murphy will do his best to make my investment in UPS battery packs seem like a complete waste of money but the cost of a set of four 7AH SLAs, especially when the float charging voltage is turned down to the life prolonging 13.5v per 12v SLA from the life sacrifing for maximum autonomy of a brand new battery pack 13.8 volts that not just APC inflict as a default float charging voltage level is a small price to pay.

 Looking at how tight the reserve margin has become for winter demand on our national grid due to wilful neglect by recent governments since before the millennium, I'm thinking about the very real possibility of load shedding during our next severe winter. Whilst the SmartUPS2000 can handle extended battery run times given a larger battery (unlike the tiny 500VA rated units which rely on the small battery capacity to save them from burning out), the investment required in a battery capable of providing autonomy for 5 or 6 hours is way too high for my liking when taking into consideration their status as a "Consumable", even when the float charging voltage is chosen to extend the battery life.

 A 17AH battery pack as specified by APC for this UPS, would cost 150 quid or more so at just under a hundred quid, that inverter genset was an absolute bargain since it can run for over 5 hours on just a gallon of fuel. TBH though, I'd bought it to prove a point, the point being that it would be free of the overvolting effect on the AVR of a standard 2.8KVA unit by the pair of 4.7μF capacitors effectively in parallel across the mains input of the UPS which gets switched out when running on battery power.

 On the face of it, the 2.8KVA genset should have been the solution to providing longer term backup power via the UPS instead of being the problem it turned out to be. All this nonsense about "Poor Quality Power" from a small petrol powered generator was just total and utter bullshit. The real problem was this gross overvolting effect (north of the 275v mark) when presented with any loads possessed of the slightest hint of leading current (a single 4.7μF fluorescent lamp PFC capacitor was all it needed to induce overvolting). The excuse of off frequency and distorted waveforms was total bollix. I discovered just how tolerant of "Poor Quality" these UPSes can be. When you think about the horribly distorted mains waveform, this should really come as no surprise - the sine inverter output of a UPS or an inverter genset is purity personified compared to the mains supply.

 Having eventually proved my point (that only the inverter type of emergency genset can be trusted to be free of this overvolting issue), I realised that my underpowered generator may not be all that underpowered after all since virtually all of the lighting load had been upgraded to LED to such an extent that not only could I keep my IT kit running, along with the main TV set, I could light up every light in the house with power to spare for the central heating circulation pump.

 The only fly in the ointment was that the fridge would probably trip the generator, leaving the UPS to handle the surge and then start draining the below spec battery pack, necessitating a quick dash to the genset to trick it out of reset by killing the engine until it was almost stopped before unkilling it in the hope the inverter module had reset itself before the engine had actually come to a standstill.

 The trick works but it's a close run thing and hard to get right. If it does actually stop, it just means another go on the recoil rope starter. I've considered making up a current sensing relay that sheds any overload that risks an inverter shut down for the 20 seconds or so for the UPS to deal with the fridge compressor startup surge before restoring the power. Assuming such trickery will work, that just leaves the chest freezer to be managed if we have to ride out any prolonged outages lasting more than a day.

 A neat way to solve the fridge compressor startup surge issue is to upgrade to one using a VFD compressor but I haven't seen such technology used in anything but high end overpriced fridge/freezers. Since I'm not ever going to give house room to any such atrocity as a Fridge/Freezer of any sort, it looks like I won't be solving this little problem quite so neatly as the VFD compressored fridge solution. 

 Those 2KW rated Workzone inverter gensets that Aldi started selling last year for nearly 300 quid looked like a more suitable alternative despite them costing some 50% more per watt of output. However, they turned out to be utter shite as I was finally forced to conclude after returning the third unit as "Not Fit For Purpose". Aside from the very serious issue of almost stalling when the 50% loading is increased to 95%, the other failing is that the fuel shut off is linked to the ignition kill switch so short of modifying the ignition kill switch wiring, you have no means of running the float bowl dry prior to putting it back into storage.

 The Parkside units very sensibly kept the ignition kill switch separate from the fuel shut off valve which endowed it with two features absent in the Workzone units. The first being the obvious benefit of running the float bowl dry with the less obvious one being that you could crank the engine on the pull cord with the engine kill switch turned off without fear of snatch back as you primed the float bowl using the vacuum operated fuel lift pump driven from the engine. Mind you, a fuel priming bulb to activate the fuel lift pump pneumatically independently of engine vacuum impulses would have been a nicer touch but at least you had a snatch back free option to do this on the pull cord.

 Mention of the fuel going stale and the need to dose it with a product like "Stabil" to extend its life, reminds me that it's been several months now since I last fired the Parkside up. Presently, it's sat in the unused front room with an almost full tank of fuel. There isn't anywhere else where I'd care to store it right now. Luckily, there's very little venting of petrol vapour so there's just the smallest hint of the smell of petroleum spirit to suggest the presence of some 4 litres of petrol in its tank. If the deterioration is due to the loss of the more volatile fractions, that bodes well for its long term fuel storage.

 Normally, I'd prefer to run it up for half an hour per month to run fresh fuel through the carb as well as make sure that it's ready to go and not developing any hidden problems through lack of use. I've been a little pre-occupied of late but I'll take the next opportunity to fire it up when the missus can't complain that I'm disturbing our neighbours' peace and quiet on a Saturday or Sunday afternoon. A weekday afternoon seems the least likely time to offend any of our immediate neighbours who may be cursed with overly sensitive hearing. That won't stop 'Her Indoors' from protesting, unasked, on their behalf, it just means I can better demolish her objections just as I have on previous occasions.

==========================================================================================

[EDIT 2020-04-15]


 I've been browsing these ancient posts trying to track down the point in time when I first mentioned the 85 ohm attenuator issue. Just like "Cat" in "Red Dwarf", I find myself distracted by these old posts of mine in much the same way that Cat gets distracted by every mirror reflection of himself, hence this edit for the benefit of others who may have recently discovered this topic thread and only now, some 14 months late, are acquainting themselves with the gory details of this "Fix-It-Yourself" function generator project so kindly provided by Feeltech.

 Today has been a rather pleasant sunny spring Wednesday (the third week , iirc - I've lost track- of the imposition of the Covid19 lockdown/social distancing measures here in the UK). This seemed an ideal opportunity to exercise both myself and that aforementioned Parkside emergency inverter generator (the very same unit - Lidl still haven't gotten round to offering these again as has been their normal annual practice).

 It must have been over six months since it was last given an outing and with a refuelling some 6 to 9 months earlier. Mindful of its nasty "Kickback" habit, as all such recoil started gensets are prone to, I turned the fuel on, set the 'eco' switch to 'non-eco', left the kill switch in the off position (no such luxury with that Workzone bag o' shite I'd sampled three of from Aldi before giving up my attempt at a cheap upgrade to an 1800W / 2000W peak output inverter genset) before turning it over a few times on the pull cord, ready for an actual starting attempt with the kill switch in the run position.

 It failed to fire up. Only then did I think to check if the choke lever had been moved back to the cold start position. Naturally enough (this is The Lord Murphy's playground, after all!) it hadn't. I moved the choke lever and turned the switch to "kill" before giving the starting cord another two yanks for good measure before setting the kill switch back to the run position for a second attempt. This time (and rather gratifyingly) it fired up on that "very first pull". It had proved no harder to startup than when I'd last run it up from a dry carburettor bowl state, indicating that the fuel hadn't suffered much (if any) deterioration whilst stored in the genset's tank (possibly a testament to how well sealed against loss by evaporation the filler cap provides)

 After a couple or three seconds, I took the choke off and left it to warm up for another half minute or so before setting it into 'eco' mode to enjoy a quieter no-load "idle" (a comparative term - full bore or non-eco mode is a matter of 4,600rpm versus the eco idle speed of 3,100 or so rpm) before carrying it to its allocated spot between the rear of the garage and the back wall.

 Surprisingly, as is typical of mid-week in normal times (these aren't normal times), none of our neighbours happened to be out and about in the surrounding back gardens to be slightly disturbed by what little noise could hop over the brick boundary wall from my strategically located 'generator slab'. Despite this, and as I've already mentioned, "Her indoors" played "Devil's Advocate" on behalf of those absent neighbours, suggesting that I relocate it to our front garden which, in order to avoid a pointless argument (men rarely never win such arguments), I did.

 To be honest, this was as much to satisfy my curiosity as it was to satisfy "'Er Indoors"'s demands since this was the one location I hadn't already tested out. This proved convenient for passing an extension lead through the basement window to reach my earth and neutral bonding point (an ex-mains filter plug fitted with a C14 socket repurposed to accept a C13 ended feed cable and pass the generator live and the local supply's ground and neutral connections cross bonded to those of the genset to the 13A "Filtered output" socket.

 This 'generator access point' is kept plugged into a handy unswitched twin outlet socket with the mains plug of the SmartUPS2000 plugged into the remaining socket to allow a swift manual transfer from mains supply to emergency generator supply. Having got everything set, I took a deep breath and transferred the UPS plug across to the genset access point socket and then strolled upstairs to check that this procedure hadn't crashed my desktop PC in my first floor office/workshop. The transfer from mains power to genset power had gone as smoothly as I'd expected with the UPS in this mix specifically to make such transitions seamless (the protection against supply outages is a given).

 APC specify a 17 or 18 AH 48v SLA battery for their SmartUPS2000 and I had recommissioned it back into 'Protected Supply' service almost two years ago with a set of 7AH SLAs simply as the cheapest possible way to run my inverter genset test to prove the absence of overvolting effect from capacitive (leading current) loadings that afflicts conventional gensets which boast the advantage of AVR.

 Despite my taking the precaution of reducing the life shortening float charging voltage setting of 13.8v per 12v SLA to a life prolonging 13.5 volt setting, I wasn't sure how well this 7AH battery pack had held up in its past two years of service. After strolling across the road to experience just how quietly the genset was running whilst supplying a 200W load, followed by a stroll up and down the road to look for any possible 'hot spots' noisewise, I had the pleasure of hearing a distant petrol chainsaw (maybe 100 or more yards distant) drowning out the genset noise at a distance of just 30 to 40 yards.

 When I got back to my office/workshop, I decided to try plugging in a 3x150W tungsten filament lamp load whilst listening to its effect on the generator through an opened window. Initially, I added each lamp load one at a time and heard the generator getting a little louder each time. This test confirmed that the mains extension feed to my 'workbench' was powered off of the 'protected supply' after all (I hadn't been sure since the tangle of cables had made it difficult to identify what was plugged into what by visual examination alone).

 A while later, I repeated this test with all three lamps plugged into their sockets which caused the UPS to sense the drop in supply voltage from the genset and switch to battery power, shedding its loading on the supply before the genset could react and sabotage the whole enterprise. The UPS would then restore the emergency genset connection after a few seconds delay to verify a restored supply and sync itself to the incoming supply frequency to effect a glitchless change-over.

 This was an unexpected bonus arising out of the use of a 2KVA 1.5KW UPS between the genset and its load. Not only did it afford protection against the genset giving up and falling over from a modest overload that exceeded its 1200W for 30 seconds rating overload trip out point, it also prevented the genset reacting immediately to greater transient overloads that would otherwise have immediately shut the inverter down, leaving the prime mover idling with no load.

 Unfortunately, this doesn't protect against modest sustained overloads in the 980 to 1180 watt range which would cause the genset to disable its inverter after 30 seconds, allowing the prime mover to keep running indefinitely and the UPS to exhaust its battery pack if the situation isn't remedied within a matter of minutes.

 That behaviour might well overcome the fridge's and the freezer's transient compressor starting overloads inducing a genset shutdown. The normal power demand of the fridge (and also very likely that of the chest freezer) is only some 70W or so at a 50% duty cycle (2 hours run time with 2 hours idle time having been noted for the fridge).

 The freezer can be manually managed by scheduling when it's allowed to run so we can shut everything off in the daytime for the brief period required to allow the freezer to successfully start up and then run it for the few hours required to bring the temperature down to -28 deg C so we can leave it disconnected for the rest of a 24 or possibly 48 hour period between scheduled restarts. The fridge is a little trickier to manage since it has less temperature margin to play with (1 to 8 deg C range at most if we want to avoid freezing stuff that should not be allowed to freeze - normally a 5 to 8 deg range).

 I have the impression that our fridge is near to the end of its life. I might replace it with a modern design that uses a VFD compressor that completely lacks this compressor startup surge current demand. Unfortunately, when I last looked at VFD based fridges just over a year ago, there were none. You had to resort to overly priced fridge/freezer models to take advantage of this technology (you pay one hell of a premium over and above the additional manufacturing costs + rational profit margin for this privilege).

 The main point of this edit is that provided you shut off the fuel feed and let the generator run the carburettor float bowl dry before storing it away for another 6 to 12 months, you're less likely to be faced with a 'gummed up' carburettor next time you come to use it. Also, the issue of fuel left in the tank going stale appears to be non-existent over periods of 12 months or so (at least in this case).

 As for the 20 litre jerrycan of unleaded, I'm hoping that'll remain viable for several more years yet. The seal on the filler seems to be gastight - there's not the slightest hint of petroleum vapour fumes to be had which bodes well for a long shelf life.

 As for our hoped for house move, we're still looking to relocate to a nicer locale and property. However, this Covid19 business has rather put a stop to casual house hunting excursions - we could well land up leaving our current property in a box or two for all we know.    

[END_EDIT 2020-04-15]  I now return you back to the distant past  ======================================


 If it comes to the matter of scrapping my boat anchors, I'll certainly be retrieving the valves to add to my existing, though modest, collection (along with anything else that might come in handy). However, that may be some years off. We're currently considering a move from this six bedroom semi-detached Victorian house into a more modestly sized house or bungalow (empty nest syndrome) with a garage/shed suitable for conversion into a workshop.

 The problem is that we're in a reasonably good area with good access to local amenities and we're undecided whether it's worth all the hassle and expense of moving to a smaller property in a nicer but less well served area or stick with the Devil we know and replace the garage (which was already a ramshackle affair when we moved in over thirty years ago and now in danger of collapsing in the next big winter gale) with a more substantially constructed garage come workshop that's actually fit for purpose.

 It's a rather frustrating position to be in but staying or leaving is a major lifestyle decision worthy of more than just casual thought. We've looked at several properties, mostly from the outside it has to be admitted - but there seems little point in proceeding to the guided tour unless you're happy with the surrounding area and the property's setting. Further developments along the line of fitting out a decent workshop are basically on hold for the time being.

 I'm reduced to tinkering about in a spare bedroom I'd commandeered as an office come computer repair workshop just over twenty years ago when going it alone as a sole trader in the business of assembling, selling and repairing desktop computers seemed such a good idea. Since I retired from this business some two years ago to collect my pension, the room has now become my "Man Cave". Hopefully, we'll arrive at a decision on whether to stay put or move onto greener pastures in the next few months to break the impasse we're currently facing.

 JBG

[Zenith]

@ Johnny B Good,

The boat anchor thing started over 20 years ago when a work colleague had to have his mother move in with him and had to get rid of much of his junk. He knew I had an oscilloscope and liked to mess about, so I ended up buying an AR88, spare valves and manuals for £50. After that the temptation was too much and came all the British WWII stuff, Eddystone, Hammarlund, Racal. It was mostly about £50 a pop and then came the test equipment and chasing down spare valves at radio rallies.

The test equipment followed and was fun to fix and use - for the most part. I have a Schlumberger sweeper which I've had a long rest from, but it will not defeat me.

There's a vintage domestic radio scene which seems healthy and organised, but I find the radios boring. It and the amateur radio scene fit into a boat anchor radio scene. There's a vintage telly scene, which I can see the point of, but don't want to be involved in. There's a vintage radar scene, which apparently involves big bucks. There's a juke box scene, I can see the point of, but it's big bucks and you need a lot of room. Test equipment is generally regarded as functional and complicated, so it's easily junked when something better comes along, which it does. Look at these DSOs and the FY6800. So it goes quite cheap, which suits me.

Then of course, there's the vintage audio scene, which rapidly merges into audiophoolery and nonsense on stilts. I despise it.

I have a couple of Z80 boards and they remind me of a kinder, gentler time, although it didn't seem like that then. A nice one fixing that 'though.

Yes, the old high performance stuff had to use expensive precision construction. A good engineer can do for a shilling what any damned fool can do for ten quid, but these people were good engineers. Times have moved on and we've found better ways of doing things, but what they did is still magnificent. HP and Tek in particular, seemed to go a bit overboard when it came to doing things right, but their gear usually seems to stand the test of time best.

As for the desire to tinker and time standards. Some amateurs have spent a lot of time and money on this. I'd like an off-air standard good for 0.1ppm at least. Don't R4LW receivers need a 198KHz crystal, which can probably be custom made for surprisingly little (£40???) and then you may as well use a published design and buy the circuit boards. A ton at least? I think I'll look into the GPSDO option.

As for where it will all end, I can't see it matters as long as you enjoy it and it doesn't impact others in your life; kids going barefoot to school because daddy has blown the loot on a caesium beam standard.

An energy policy constructed without a power generation engineer in sight is worrying. Also, energy policy is long term and so can't very well accommodate the fads of the hour. Most shops these days couldn't open without electricity. Your comments on this have made me reconsider things. In January, with a blocking high for a couple of weeks, things could get miserable. I suppose the least I'd need is 240V AC to produce  a maximum 50W to keep the central heating system going for a few hours. It could come from an inverter and the car battery. The Lidl generator is well worth a thought.

The stale fuel and generator idea came from a friend whose daughter ran a chicken farm. It had an emergency generator which was petrol fired, but when the power went off, always in the winter, daddy was called out to fix it, and the problem was stale fuel.

As for the house moving business, I've gone through the same process and have decided to go home, where property is cheaper, I can be with the folks I dig and can catch up on some fishing and shooting. It's a lot of work and a big decision, but the older you get the harder it gets.

It's been a pleasure to correspond with such a thoughtful person.

For anyone else following this long and possibly completely irrelevant post I'll bring it back to the FY6800.

IMHO it's a remarkable product for the money. It's a shame Feeltech haven't chased up some of the observations and improvements made and maybe incorporated them in a deluxe version or options at extra cost, then they could have had a killer product.

I'd recommend it without hesitation for anyone wanting a function generator and judging it by the standards of function generators. Features such as the burst function and arbitrary functions I can see being useful in sorting out knotty problems. It wouldn't cut the mustard compared to a halfway reasonable RF or decent low distortion AF sig gen. If you are interested in low level signals it's questionable.

Generally well worth the money.


 

[Johnny B Good]

@ Johnny B Good,

The boat anchor thing started over 20 years ago when a work colleague had to have his mother move in with him and had to get rid of much of his junk. He knew I had an oscilloscope and liked to mess about, so I ended up buying an AR88, spare valves and manuals for £50. After that the temptation was too much and came all the British WWII stuff, Eddystone, Hammarlund, Racal. It was mostly about £50 a pop and then came the test equipment and chasing down spare valves at radio rallies.

The test equipment followed and was fun to fix and use - for the most part. I have a Schlumberger sweeper which I've had a long rest from, but it will not defeat me.

There's a vintage domestic radio scene which seems healthy and organised, but I find the radios boring. It and the amateur radio scene fit into a boat anchor radio scene. There's a vintage telly scene, which I can see the point of, but don't want to be involved in. There's a vintage radar scene, which apparently involves big bucks. There's a juke box scene, I can see the point of, but it's big bucks and you need a lot of room. Test equipment is generally regarded as functional and complicated, so it's easily junked when something better comes along, which it does. Look at these DSOs and the FY6800. So it goes quite cheap, which suits me.

Then of course, there's the vintage audio scene, which rapidly merges into audiophoolery and nonsense on stilts. I despise it.

I have a couple of Z80 boards and they remind me of a kinder, gentler time, although it didn't seem like that then. A nice one fixing that 'though.

Yes, the old high performance stuff had to use expensive precision construction. A good engineer can do for a shilling what any damned fool can do for ten quid, but these people were good engineers. Times have moved on and we've found better ways of doing things, but what they did is still magnificent. HP and Tek in particular, seemed to go a bit overboard when it came to doing things right, but their gear usually seems to stand the test of time best.

As for the desire to tinker and time standards. Some amateurs have spent a lot of time and money on this. I'd like an off-air standard good for 0.1ppm at least. Don't R4LW receivers need a 198KHz crystal, which can probably be custom made for surprisingly little (£40???) and then you may as well use a published design and buy the circuit boards. A ton at least? I think I'll look into the GPSDO option.

As for where it will all end, I can't see it matters as long as you enjoy it and it doesn't impact others in your life; kids going barefoot to school because daddy has blown the loot on a caesium beam standard.

An energy policy constructed without a power generation engineer in sight is worrying. Also, energy policy is long term and so can't very well accommodate the fads of the hour. Most shops these days couldn't open without electricity. Your comments on this have made me reconsider things. In January, with a blocking high for a couple of weeks, things could get miserable. I suppose the least I'd need is 240V AC to produce  a maximum 50W to keep the central heating system going for a few hours. It could come from an inverter and the car battery. The Lidl generator is well worth a thought.

The stale fuel and generator idea came from a friend whose daughter ran a chicken farm. It had an emergency generator which was petrol fired, but when the power went off, always in the winter, daddy was called out to fix it, and the problem was stale fuel.

As for the house moving business, I've gone through the same process and have decided to go home, where property is cheaper, I can be with the folks I dig and can catch up on some fishing and shooting. It's a lot of work and a big decision, but the older you get the harder it gets.

It's been a pleasure to correspond with such a thoughtful person.

For anyone else following this long and possibly completely irrelevant post I'll bring it back to the FY6800.

IMHO it's a remarkable product for the money. It's a shame Feeltech haven't chased up some of the observations and improvements made and maybe incorporated them in a deluxe version or options at extra cost, then they could have had a killer product.

I'd recommend it without hesitation for anyone wanting a function generator and judging it by the standards of function generators. Features such as the burst function and arbitrary functions I can see being useful in sorting out knotty problems. It wouldn't cut the mustard compared to a halfway reasonable RF or decent low distortion AF sig gen. If you are interested in low level signals it's questionable.

Generally well worth the money.

@ zenith

 Apologies for the tardy response but I've had several distractions consuming my 'spare time' of late.

 I never became a 'collector' since I didn't have the space for more than the most basic of signal generator and 'scope kit, hence my only owning just two boat anchors. Of course, now that modern T&M kit is not only so relatively cheap but compact and lightweight, there is now a real danger of expanding my T&M gear into a collection that goes beyond my cheap (but excellent value) Siglent DSO and the even cheaper (and supremely excellent value) FY6600.

 Perhaps I will eventually land up investing in an SA and perhaps even a LA, keeping the existing T&M kit as backup to yet more highly specified versions that I might feel justified in splashing out on in the event that the needs of future projects will justify such extra expenditure. It's never a good idea to invest in T&M upgrades whilst the existing T&M kit can still serve its purpose even if not so conveniently as an overdue upgrade would do. Like desktop PCs, it's always best to wait until you can feel a tangible benefit from the upgrade rather than doing such purely for the 'bragging rights'. The longer you can put off an upgrade, the more value you'll get from the investment.

 As for vintage audio gear, whilst I respect it for the qualities that old fashioned valves (tubes) could offer (600v/microsecond slew rates versus the 2 to 10 volts per microsecond slew rates of 2N3055 power transistors) which belatedly revealed itself as SID in the first of the "modern solid state" power amps (the electronic equivalent of tracking distortion with pick up cartridges that were either set up with too little down force or else simply had too large an effective tip mass for even down force settings at the highest allowable limit could fix), the time has long since passed by where this is true. I don't see too much virtue in using such ancient valved gear other than as expensive (to both own and run) space heaters.

 I must admit to hankering after the simpler times of coding in Z80 assembler but I'm faced with a broken Transam Tuscan S100 bus home built from a kit desktop computer that had fallen victim to the damp basement environment I'd kept it in for two decades. I've got a spare main board which had been stored in a dryer and warmer part of the same basement which I'm planning on using, assuming I ever get round to testing it before applying all the necessary hardware mods required to undo the many "Skoolboy Howler" design errors that had been inflicted upon it (address decoding faults and tuning the modem chip used for the original tape interface to the 300 baud rate rather than the 1070/1270 Hz carrier frequencies used in the echo mode it was set up for (echo mode being required for tape storage use) being just two examples that come to mind right now - I'm sure there were other similar design errors).

 I don't think there are any cost effective Z80 versions of the cheap as chips Arduino Nano board I shelled a whole 3 quid out on just on the off chance that I might actually get to use it with my u-blox GPS module. If there were, I'd be into it in a shot since I'm so familiar with Z80 assembly, it's like an old friend. I guess I'll just have to resume my studies into C (actually open and read the "Beginning C" book I bought nearly three decades ago). They say, "You're never too old to learn.". Perhaps the Arduino Nano might give me enough motivation to demonstrate whether that saying has any validity or not (the three quid investment should, in my case, provide ample motivation!). 

 You don't need a 198KHz crystal to make up a 198KHz R4LW locked reference. You can use an old transistor radio with LW and a bunch of cmos ttl logic (divider chips) to generate a 2KHz tone locked to the 198KHz to mix with the 198KHz signal to generate a 200KHz reference from which to phase lock your 10MHz OCVCXO module at a 50:1 ratio as per the original 200KHz frequency standard that was used prior to the 9KHz channelising of the long wave broadcast band in (iirc, the late 60s) which so inconveniently resulted in the 2KHz drop in frequency that applies to this day.

 You can search for the original construction articles which will provide the necessary BoM by which to construct such a frequency reference but if your time is precious, you might do better to go the GPSDO route and create a more futureproof solution. The 10MHz OCVCXO will still be in the BoM list if you want to create a nice clean and steady jitter free 10MHz reference. Otoh, it is possible to use the 10MHz output from the u-Blox NEO M8N module directly as a frequency calibration source if you don't mind the low level of noise modulation (weird beeps) and the periodic 'ticking' at just under 1Hz which can be heard alongside of the 10MHz or its third harmonic when zero beating an HF rig to precisely the required frequency. Despite all these "Noises Off", it still makes a much better alternative to the 10MHz WWV broadcasts you just might be able to receive in around five years time.

 For the very first time today (yesterday in fact), I finally set my 'scope up on that pesky desk that sits in the window bay, close enough to let me connect to the GPS module's PPS output pin so as to actually get a picture of the waveform. It's a surprisingly accurate square wave shape at 10MHz allowing for the inevitable jitter you get from dividing the module's own internal 48MHz TCVCXO down to a non integer 10MHz output and the ringing created by using the longish "Ground Wire" connection to the probe tip circuit.

 I can eliminate the jitter by choosing frequencies of 12, 8, 6, 3 or 2 MHz so an obvious way to obtain a jitter free 10MHz reference would be to set it to output 2MHz and multiply back up to 10MHz with an ultra low jitter PLL chip programmed to give a 5 times multiplication. Indeed, there's little to stop me using the same jitter free PLL chip to multiply up from even lower frequency options which may well completely sanitise my 10MHz of all of the modulation noises including that weird not quite synchronised to once a second 'ticks' that appear in the modulation noise pollution. Using a PLL is a good way of filtering out such noise leaking onto the PPS output pin from the rest of the GPS module's digital processing activity.

 I managed to short out the gnd and 5v Vcc line with my 'scope probe which corrupted the antenna power settings, dropping the sat signals by some 30dB. The antenna settings, as usual, failed to be accepted until, as before, I programmed the module back to defaults, requiring me to reprogram its PPS line to 10MHz output and to reprogram the antenna power settings to get back to the 30 to 55 dB C/N levels I'd been getting prior to this little 'accident'. Just out of curiosity, I pulled the external antenna to see whether I could get usable signals from the module's built in patch antenna with microscopic LNA chip and found it was receiving almost as many satellites though barely strong enough to decode usable fix and time data. It seemed to be stable so I left it alone for a half hour or so only to find it had stopped receiving any satellites strongly enough get a lock on so had to reconnect the external antenna again.

 This issue with the built in patch antenna is a rather strange one which looks like it may just be a cock up in the reference circuit used with the microscopic LNA chip. I'm beginning to wonder whether a strategically placed resistor might restore a missing bias current/voltage to the LNA. This delayed failure has all the signs of a charge on a coupling capacitor biassing an ultra high impedance pin slowly decaying away, resulting in the LNA becoming disabled. You've read the other thread so you'll be aware that this is a problem that has been seen with at least one other similar M7N based u-blox module.

 Luckily, this isn't a "Show Stopper", just an irritating feature of a non functioning built in patch antenna that, by all accounts, could prove a serviceable alternative, given suitable reception conditions, to the use of an outside antenna. It would have been far better if the manufacturer had decided to forego the built in patch antenna option altogether and reduce their BoM costs (and, hopefully its final selling price).

 I've mentioned this before but I'd thought I'd raise the subject again. I've noticed (but only after doing the TCXO upgrade) that my FY6600 is sensitive to any HF transients causing it to effectively disable its outputs. Initially, almost every time I connected a grounded DUT to the BNC ground shields as it shorted out the voltage from the Y cap. Fitting the C6 mains connector and wiring the ground rail to the mains earthing pin via an 11K resistor cured this particular issue but I've recently discovered that it's still possible to recreate this effect by switching a lamp plugged into the same mains extension lead 4 way socket that serves the FY6600 with mains power on and off several times until the signal generator finally succumbs, disabling both channel outputs.

 The sensitivity to such mains borne transients is still there, it just needs more effort to recreate an effect I could guarantee almost every time just by simply connecting a grounded DUT to the signal generator. I suspect it's an issue that would go away whenever the generator is plugged directly into a ring main fed wall socket (or at least plugged into a mains extension lead that's not also serving any other potential switching transient  sources). I'm just wondering whether anyone else has witnessed similar sensitivity to such transients or whether either my TCXO mod has introduced such sensitivity or I've been just plain unlucky in landing up with a unit that was cursed with this defect to begin with.

 With regard to the residual distortion of the AWG at audio frequencies, this isn't necessarily a show stopper when it comes to determining distortion levels in audio gear since you can use what was once a common technique of nulling out the signal generator's signal voltage waveform from that of the output from the amplifier under test and look at what's left over (which is essentially the distortion signals introduced by the amplifier itself). Whilst a distortion free test signal is nice to have, it's not an essential requirement for measuring distortion levels in an amplifier.

 AFAICR, the distortion levels at audio frequencies are less than half a percent with these signal/function generators which is better than a lot of the older and venerated HP signal generator kit and more than adequate to get high quality measurements with. I wouldn't be too concerned with the distortion figures for a device intended to serve the needs of the hobbyist rather than that of a busy professional audio testing lab where time is money and speed of measurement is of the essence. The hobbyist user has the luxury of time by which to get results matching those of a professional testing lab which can justify the time saving costs of a 'distortionless' signal generator.

 Incidentally, when I came to run some tests with the Parkside genset a few days ago, involving the refrigerator as the test load, I discovered a fault that stopped it generating output from its inverter module. The symptoms are entirely indicative of a poor contact in the plug/socket connection between the PM generator output and and the inverter input. I've no doubt that this is simply the result of my 'starter motor' testing forcing the plug connectors open when I inserted the solid core wires obtained from a length of 2.5mm flat twin and earth mains cable to connect it up to the BLDC controller module I was using.

 I didn't bother to sort it out since I plan on doing yet more testing once I've determined the pole pair number of the PMG (which is where the 'scope earns its keep) and purchase a pre-programmed Adapter Board with AS5047P rotary encoder chip to provide the "Hall Effect" commutation signals to drive the BLDC motor controller with commutation signals that are hard locked to the rotor position (sensorless mode just doesn't cut the mustard in this application - the mechanical loading is far more difficult than that of a predictable fan or propeller load for a sensorless control algorithm to handle). The refrigerator load test can hang fire for the time being whilst I get things organised to verify the PMG pole pair number and place an order for that adapter board with pre-programmed chip.

 I had, rather optimistically, hoped I could get away with sensorless operation for my add on BLDC controller but tests revealed the stark truth that in this usage case, I have no choice but to use commutation signals that are positively locked to the rotor position after the fashion of a purpose made BLDC motor with Hall Effect sensors built in by design.

 The only consolation with retrofitting the PMG with HE signalling using the programmable AS5047P rotary encoder chip is that I get a superior position sensing system that outperforms that of even the best purpose made BLDC motor with HE sensors built into it by design. It involves a little more modification than that of merely adding extra electronic components, without gross mechanical modifications as I'd originally hoped, in that I now have to attach a magnetic disk to the end of the crankshaft and attach the sensor PCB to the crankcase mountings, positioned so as to detect the magnetic field from the sensor magnet.

 Sensorless operation only worked when the sparkplug was removed by way of testing that it was actually possible to run the PMG as a motor but as soon as I blocked the sparkplug hole with a finger, the extra load was enough to confuse the sensorless algorithm and put the controller into a stalled state. The test confirmed that, given a reliable position only dependant commutation feedback system, there was every chance that I could utilise the PMG as a starter motor - it just wasn't going to be quite so simple a task as merely adding a bunch of electronic components alone in order to achieve this. 

JBG

[d3m0]

Hi everyone, I found this post looking for a cheap DDS. I read a few pages but my lack of knowledge makes me lost sometimes. My short question is: should I go for 6800 or 6600 model? both are kind same price....

Regarding to improving them What are necessary / recommended mods?

I catch:

  - Linear 317 337 PSU
  - Heatsink every chip gets hot
  - adding / changing opamps  <-- is there a specific post that explains it for a kid? 


anything else? can be upgrade max frequency? I think 60Mhz for me is enough but if it's free..... 


Thank you very much, great effort everybody.
 

[rhb]

Don't get the 6600 as it has mains leakage voltages on the BNCs.

The sine wave performance of my 6600 at low output voltages compares very favorably to my Keysight 33622A which is a 120 MHz, 1 GSa/S AWG with a list of around $6500.  I got a used one from the Keysight eBay store for 1/2 that, but still my most expensive piece of T&M kit.  Ironically, I got a replacement front panel to fix my borked V3.0 unit just after I got the 33622A.  After 9 months of trying to get F***Tech to replace it.

At the moment I've got my 6600 which only has the grounding mod generating a 10 nS pulse at 1 V (2 V setting) into 50 ohms on a 1.5 GHz Lecroy DDA-125.  It's got a little bit of ringing, but the pulse is clean enough to use for time domain reflectometry and vector network analysis using a DSO with 10 Mpts or better sample depth.

That setup will give about 70-80 dB of dynamic range with an 8 bit DSO and suitable software which I plan to write.  It will provide VNA functionality up to the 6 m ham band at 50-54 MHz.  The same software will provide similar quality spectrum analysis.  It will only run on a PC, but it will do the job, albeit somewhat awkwardly.

A more ambitious project of mine is to develop FOSS DSO FW for COTS DSOs based on the Zynq and Cyclone V chips.  That will take a couple of years.  Right now I'm working on getting the data from the BNC to  memory with pipelined math operations in the FPGA PL on the way.

[beanflying]

There is still some very good deals on what I assume is the last of the 50MHz 6600's on evilbay if your budget is lean. Forget the Linear supply upgrade and earth it like I did on about page 6 of this thread (I think about there). Treat it as a 'reasonable' Signal Generator and don't waste to much time and money on trying to make Champagne from Pig Swill.

Otherwise head for the 6800 as it seems to be firmware upgradeable at least 

[d3m0]

Treat it as a 'reasonable' Signal Generator and don't waste to much time and money on trying to make Champagne from Pig Swill.

Yeah of course I'm conscient of what I'm buying. I don't want to waste time and that's why I'm asking here. I think changing opamps for a few bucks is worth the time and money. The improvement looked in pictures I think is great.

So... going for a 6800 and only change opamps? What about firmware upgrade and open  bandwith?


Thanks again!

[DaveR]

Neither the 6600 nor the 6800 is a waste of time or money (negative comments on here usually relate to problems with the earliest 6600s from 2017), but both can be improved quickly and cheaply.  Go for the 6800 (it has a better front panel) and fit new opamps.  Also fit a TCXO to give much better frequency accuracy and stop drift.  If you want to push the boat out, modify the existing power supply or fit a linear.  The upgradability is nice, but it's only relevant as long as Feeltech bother to produce new firmware for it, which probably won't be for very long - but there's nothing wrong with the existing firmware, so why worry about it?  That's really all there is to consider, unless you want to pay several times as much for a "branded" item which does much the same things.

[Johnny B Good]

Treat it as a 'reasonable' Signal Generator and don't waste to much time and money on trying to make Champagne from Pig Swill.

Yeah of course I'm conscient of what I'm buying. I don't want to waste time and that's why I'm asking here. I think changing opamps for a few bucks is worth the time and money. The improvement looked in pictures I think is great.

So... going for a 6800 and only change opamps? What about firmware upgrade and open  bandwith?


Thanks again!

 It all depends on what you're planning to use it for. If it's just for audio work or you're hardly going to push it beyond the 5v p-p level, upgrading the opamp and the PSU will largely be wasted effort. The dual THS3002i opamp only gets switched into circuit at settings above the 5v p-p "amplitude" mark and the 11.7ish volts on the +/- "12 volt" rails only becomes an issue with amplitudes close to the 20v p-p limit when you need to apply DC offsets of more than half a volt or so.

 At audio frequencies, the existing 50ppm rated XO chip is more than adequate for the job so upgrading that to a 0.1ppm TCXO would offer very little tangible benefit. However, regardless of all this, what is worth doing is the fitting of a small cooling fan to reduce the rather high internal temperatures due to the vent slots not being placed to make use of passive convective cooling to any effect (tilting it up on its kick stand actually makes the situation worse in this case).

 Although Feeltech seem to have taken notice of all the vociferous complaints about the issue of the half mains live 'touch voltage' and "done something about it" by upgrading the mains socket from the cheap 'n' cheerful IEC C8 unpolarised two pole connector, I'm afraid to say they've been a little slapdash over inflicting a heavy duty IEC C13/14 3 pole earthed connector on this 'tiny little doggy', giving it a tail that will wag it rather than it doing the tail wagging.

 Worse still, not only have they introduced a low impedance earth loop with its stray interference voltages and DC offsets from galvanic and thermocouple effects at millivolt output settings, they've compromised the ground return connection between the main board and the PSU zero volt rail by simply chopping one of the two ground wires to divert to the safety earth pin of the mains connector. It's as if they have no understanding of the pros and cons of the original floating class II arrangement, concentrating on fixing the "con" at the expense of the "pro".

 What this means for someone purchasing an FY6800 is that they need to undo the damage inflicted by Feeltech's bodging of the IEC C13/14 connector if they want to avoid seeing the mystery DC offsets (and ground loop interference) that Andreax1985 had reported when checking out his recently acquired FY6800.

 In short, this means disconnecting the safety earth wire from its connection into the chopped wire in the ribbon cable connector between the PSU and the main board, making good the damage inflicted by Feeltech by rejoining the cut ends. At this point, you can either add a 10K resistor between this joint and the disconnected earth wire from the IEC C13/14 connector's earth pin or else, find a better way to connect this resistor to the zero volt/ground rail to free this ribbon cable of Feeltech's bodgery.

 The PSU is still a class II item as before so doesn't require such a safety earth connection. All that was ever required of a polarised 3 pole mains connector upgrade with the FY6600 (and the FY6800) was a means of eliminating the half mains 'touch voltage' leakage via the EMC mandated Y class capacitor.

 In this case, a 10K resistive connection serves this function very nicely, attenuating as it does, the unwanted stray galvanic dc voltages and mains borne interference by some 60 to 80 dB, neatly limiting the half live mains voltage to, worst case on 240v mains, just half a volt ac which neatly addresses the issue of damage to any devices being tested through high voltage discharge from the class Y emc bodge capacitor in the PSU.

 If you feel that even half a volt ac is still too much of a risk, you can use a 1K resistor instead to get this down to 50mV ac and still attenuate the earth loop induced galvanic and noise voltages by some 40 to 60 dB. This resistive earth connection is essentially a very low impedance static discharge drain connection rather than a class I (III?) protective earth.

 Of course, the problem of doing any mods like this, no matter how urgent or essential, is the issue of "Voiding The Warranty". Unless you've bought this via a local Agent bound by your country's consumer protection laws, even when you have a valid warranty claim, the extra pain of dealing directly with Chinese suppliers can be sufficient to put you off making a claim when the original cost is so low as to make any out of pocket expenses (whether temporary or not) look out of all proportion to the benefit.

 Although the cost of these products is 'chump change' compared to the cost of  just the next quality level up (Siglent territory), it's right on the edge when it comes to deciding whether  to "Collect on the warranty or just write it off?". The sensible approach in this case would be to allow enough time for any manufacturing defects to show themselves within the 30 or 60 day return period before embarking on any warranty voiding modification work even if you've already decided that you'd write it off rather than face the hassle of trying to collect on a "Chinese Warranty" in the event it fails outside of that initial period.

 Although the earth loop and fan cooling mods are urgent and necessary mods, you can provide non invasive fixes by making up a special mains (extension) lead with a 10 or 1 K resistor in series with the earth wire and use a small desk fan to blow cooling air over the back of the generator to mitigate the excessive temperatures until you're ready to take the plunge into more elegant, yet warranty voiding, fixes once the honeymoon period is over.

 All of the improvements that were applied to the FY6600, are all equally applicable to its successor (including the half live mains leakage earthing mods - the bulk of the effort has already been applied by Feeltech, it just needs to be refined to eliminate the mains earth loop issue they'd introduced). It's basically just a question of deciding which, if any, of those mods are worth doing.

 Regarding the possibility of firmware updates/hacks to expand the frequency range past the current 60MHz limit, that seems rather doubtful since beyond 60MHz, you'll need to modify the 5th order anti-aliasing LPFs on the main board to nudge them closer to the 125MHz Nyquist limit.

 Being an arbitrary wave function generator, there's nothing to stop you loading a sine wave that's all 2nd harmonic and no fundamental to double up the output frequency to test the response of the on board LPF to identify exactly where the cut off frequency has been set to. ISTR a figure of 75MHz being mentioned a year or so back in this thread.

 This trick of loading a custom sine waveform into one of the arbitrary wave slots might be a handy work around to the 60MHz limit to get a useful signal output in the 4 metre amateur radio band (and possibly even into the VHF/FM broadcast band if all you need are a few tens of millivolts of signal output).

 The chances of Feeltech ever doing a firmware upgrade to raise the frequency limit to 75 or 80MHz seem pretty slim in view of the existing LPF's turnover frequency somewhere around the 75MHz mark. If they CBA to fix their 85 ohm attenuator goof, what chance is there that they'd modify the LPF filter components? In any case, such a raising of the turnover frequency would also require a steeper roll off to avoid aliasing problems above the 125MHz Nyquist limit so they'd need to upgrade it from its current 5th order design to a 7th or 9th order design which would probably only be practical using separate filter modules in place of the existing filter network components.

 Experimenting with a custom arbitrary sine waveform to explore the frequency response of that filter should give you some hint as to whether Feeltech are likely to extend the frequency limit with a firmware change in a later version of the FY6800 as they've done in the past with the FY6600 models. The existing filter might just allow Feeltech to offer a 75 or 80 MHz option but don't expect to see a 120MHz version in the same sort of price bracket any time soon.

 At the end of the day, these Feeltech signal generators are what they are, damn good value for the money 'wartified' versions of their more expensive cousins with the bonus of being amenable to several cost effective enhancements to cure it of the worst of its wartiness.

 The biggest danger for most of its users is that of getting carried away with the scale of these attempts to turn this particular "Sow's Ear" into a "Silk Purse". If you can dial your ambitions down from "Silk Purse" to "Rayon Purse" in this particular pursuit of perfection, that should save you from losing all sense of proportion and getting sucked into a never ending upgrading project that's taken on a life all of its own.

 The trick to this is knowing when to call time on the project and say, "Enough is enough! My work here is done.". For something as cheap as this, I'd say a good rule of thumb to apply is that once you realise that you've spent as much on the upgrade components as you've done on the whole generator itself, you've gone about as far as is sensible, if not further, and it's now time to insert those long case screws (mine have been a permanent workbench exhibit for the past 4 months now) for the very last time and be done with it.

 If your requirements for perfection increase with more ambitious projects later on, let the quote from the "Jaws" movie, "We're going to need a bigger boat" be your guide as to the solution and move upmarket to something better again. Your old,Feeltech AWG can always be boxed up and put into storage as an emergency generator - after all, it's not exactly a "Boat Anchor" that's likely to give you a hernia in ten years' time when you're eventually forced to have a clear out. :-)

 BTW, I thought I'd attach a picture or two of interest. The first one shows my permanent bench exhibit of the four case screws to the bottom left in the foreground. The u-Blox NEO M8N GPS module is just to the right, powered from a 5v wallwart with a 14cm  3/4 wavelength wire antenna plugged into the SMA socket (the built in Patch antenna only works sporadically due to a suspected error in the LNA 'reference circuit' used by the manufacturer) to receive some 6 to 12 satellites whilst a good 16 feet away from the window.

 It's set up this way because the eight metre (correction: 5 metre) active antenna cable won't stretch to the workbench. The scope is displaying a perfectly jitter free 2MHz squarewave on the PPS line. The 'ringing' as anyone observant enough to spot, arises out of the excess 10 inches' worth of earth clip lead between the probe and the zero volt rail of the GPS module.

 The second photo is a screen capture showing the rather horrific level of jitter you get when dividing a 48MHz TCVCXO clock  down with a non-integer division ratio to obtain a 10MHz clock. This demonstrates why I have three ON SEMICONDUCTOR - NB3N502DG - PLL CLOCK MULTIPLIER, 8SOIC chips winging their way to me as I type.

 The plan being to multiply the already jitter free 2MHz back up to 10MHz with just a mere 40ps p-p jitter on the clock multiplier's output (interestingly, with rise and fall times specified at just 1ns) to make a "Poor Man's GPSDO frequency calibration reference source" that's good enough to use as an external 10MHz clock supply for anything that has such an external 10MHz reference socket (currently nothing right now until I modify the FY6600 with such an option).

JBG