Products > Test Equipment
FeelTech FY6600 60MHz 2-Ch VCO Function Arbitrary Waveform Signal Generator
Johnny B Good:
--- Quote from: Noy on January 23, 2019, 07:55:33 pm ---Ist it worth to get a FY6800 if you already have a Toellner 7401 and the build in AWG from Rigol MSO5000?
Or better a old wavetek Model 275? I think the wavetek is to old or? Waveform has to be tipped in over the buttons because i don't have a GPIB..
--- End quote ---
Hi, Noy,
The answer to that question rather depends on what you need. Not being familiar with those items of test gear, I had to search for information and landed up spending pretty well all day looking not only at youtube vids and other eevblog postings about the gear you mentioned but also distractions on frequency standards kit such as gpsdo projects and kits, taking note along the way that I might want to buy myself a u-blox NEO‑M8N gps module and have a go at making my own gpsdo reference or else blow 149 euro on a GPSDO2 from SV1AFN. Anyhow, all that aside, the conclusion I came to was that an FY6800 would, in several respects be a significant upgrade over all of the test gear you mentioned. Considering the cost of the MSO alone, an extra 100 dollars US on Feeltech's offering seems to be well worth considering if you want to go beyond the limits of a single channel basic 2MHz sig genny or the 5v p-p limits of the MSO's built in dual channel AWG.
Regarding my own speculation on the possible need to replace the front panel of my much modded 6600 by cannibalising the front panel from a brand new FY6800, I finally got round to taking a closer look at the front panel which revealed actual soldered in switches and the possible reason for my losing the digit cursor highlighting in the shape of a badly soldered in encoder module.
I re-sweated the solder joints (3 stand off pins top and bottom of the rotary encoder board which joined it to the front panel pcb) to get rid of the downward droop and reflow a suspect joint. Then I removed the offending right arrow switch and swapped it with the "CH2" switch where the extra stress would be better tolerated away from the extreme end of the panel's PCB and the better action of the "CH2" switch, now sat in the right arrow switch location would reduce the forces being applied to the extreme RHS of the front panel PCB.
I can no longer see any misbehaviour no matter how hard I try to provoke it, so it seems the switch swap (and resoldering the encoder connections) has nicely solved this issue. The original right arrow switch works just fine in its new "CH2" better supported location under the larger and easier to press square shaped button dome. Also, I noticed the use of only two fibre washers between the display/switch panel and the cpu board so fitted fibre washers to the other 3 or 4 fixings to try and balance it up. I used thinner fibre washers by way of a compromise in case there was an actual reason for Feeltech choosing only to fit two washers.
Although I'm not overjoyed by the use of membrane operated switches, I'm happy enough that the right arrow button does now operate tolerably well (by the standards of membrane switches) and no longer suffers random loss of the digit cursor which is a relief seeing as how frequently the cursor positioning switches will be operated compared to all the others.
The reason for my taking the front panel apart was on account of my stripping the whole contraption down in order to improve soldering iron access to modify the output attenuator from an 85ohm pad into something approximating a 50 ohm pad. I shunted the 100 ohm shunt elements with 120ohm smd resistors and wired some ancient multiturn pots across the 510 ohm series elements to let me trim the unterminated attenuation to match the original calibration Feeltech had applied to hide their goof.
It seems I miscalculated the new resistor network values and found myself needing a much lower value across the 510 ohm series elements than my 8K2 estimate, landing up with trimpot values of 339 and 350 ohms, resulting in a 44.5 ohm pad rather than the desired 50 ohm pad. Still, whilst not perfect, it'll do for now (a 5.5 ohm discrepancy is a considerable improvement over the original 35 ohm discrepancy). I've subsequently had two more goes at calculating replacement resistor values and I'm just getting more confused as to which two sets of values are closest to being correct, so I'm putting further refinement effort to one side for now.
After using my HF transceiver to listen to the 20MHz carrier being produced by the AWG, I'm once more looking for a better PSU solution since I'm convinced that what others have described as 'appalling sideband/phase/jitter noise is nothing more than modulation from voltage rail ripple. I need to make up a battery supply (two 12v SLAs with a 5v regulator to provide the three rails) to test this hypothesis before I spend Ghod knows how long to track down a suitable well screened three rail power supply solution. All I know at this stage is that there actually are suitable "designed for test gear" ultra low noise dc-dc switching converter modules available - I just haven't dared to seek a price quote. :scared:
One final point. When I replaced the crappy C8 mains connector with a C6 last week, I used a two pole rocker switch. Unfortunately, it wasn't one rated to survive soldering wires to its tags and I damaged it. It still worked at that time but I wasn't entirely happy about it. Nevertheless, I'd spent enough time sorting it out so wasn't in the mood to wire up another switch so accepted it as the best of a bad job.
A couple of days ago, that came to bite me in the backside when, after unusually having left it switched off overnight, I came to turn it on early in the morning to check out its frequency drift from cold, it didn't power up until I held the rocker switch pressed in the on position. Having proved my doubts about its long term survival of the soldering iron abuse, I had no choice but to pick out another rocker switch replacement from my assortment of salvaged single pole rocker switches (that damaged switch had been the only two pole example of the breed I'd had in my possession).
This meant I had to recover the original plug in switch flylead from its repurposing as a jumper plug courtesy of a crimp on silicone grease filled connector and hope it was still long enough for the job (it was, but only just!) and thin out the lower flange of the switch so it could overlap the top of the slide in C8 connector without fouling by use of a junior hacksaw and a small file (hence my original reluctance to replace the damaged switch to begin with).
Anyhow, I successfully soldered the flylead onto the replacement switch's solder tags without damage this time (I guess the plastic used by this one must have been rated for soldering temperatures - I'd had to unsolder the cut off remains of the original wires before I could repurpose it and that went ok). Thankfully, the flylead was just long enough to reach the psu connector so all was well with my repair and I was finally able to start my "Frequency Drift from Cold" test of the 0.1ppm 50MHz TCXO module using the BBC Radio 4 LW transmission on 198KHz as my Rubidium frequency reference.
It does have a strange power on drift characteristic from an ambient of 17deg C where it starts some 30ppb off frequency transiting the 0ppb mark some 5 minutes or so later before settling on a 15 to 20 ppb error some 30 to 40 minutes after that. Once the ambient reaches 21 to 22 deg C it settles to within +/- 5ppb of the 198KHz Rubidium reference frequency. This behaviour (settling to within +/-5ppb at 21/22deg C at any rate) seems consistent from day to day which is quite gratifying for a 100ppb rated TCXO module which had cost me less than 16 quid delivered. :)
For me, the Feeltech FY6600 has been a very cost effective investment indeed. The only other T&M kit I've ever possessed have both been boat anchors of 1950s vintage. A Marconi signal generator and a 5MHz 'scope which I last used over quarter of a century ago, now languishing in my rather neglected basement radio shack come workshop.
Indeed, it's been so long since I last used them, I can't even recall the frequency range of the sig genny nor whether the 'scope was single or dual beam. Furthermore, I'm not even sure I have enough strength to manhandle the 'scope off the dexion shelving to set it up on a bench for testing. The Marconi sig genny I think is a little lighter and also sat on a lower shelf so I think I could just about manage to drag it out and onto a test bench... but only if I had a mind to do so.
Quite frankly, as much as I'd like to, I would have to be out of my mind to even attempt such a strenuous exercise just to satisfy casual curiosity, besides which, I'd have to find somewhere else to stack the stuff already piled up on my basement test bench just to make room for said boat anchors - I wouldn't want to be lugging them about any more than I'd absolutely have to if curiosity were to ever overcome my sense of self preservation.
Anyhow, now that I've upgraded those boat anchors to a Siglent SDS1202X-E and the FY6600, they're now well and truly relegated to the status of vintage T&M curiosities. There's been considerable development in T&M kit over the past half century or so since that vintage T&M gear was last manufactured, thank Ghod! :)
Regards, Johnny B Good.
Noy:
Hi, got a FY6800 now..
Managed somebody to run the software in Linux? Maybe with wine? Which special things do i have to install in wine(winetricks)?
Is there another Linux software around?
Which hacks are worth to do on the fy6800?
Adding THS3095 / THS3491 ? Although switching to another 50MHz TCXO? Which one will fit directly in the pcb? There are some 0.28ppm 50MHz available at mouser? Or a bit cheaper 2.5ppm. Both better than 100ppm
Johnny B Good:
--- Quote from: Noy on January 29, 2019, 10:08:27 pm ---Hi, got a FY6800 now..
Managed somebody to run the software in Linux? Maybe with wine? Which special things do i have to install in wine(winetricks)?
Is there another Linux software around?
Which hacks are worth to do on the fy6800?
Adding THS3095 / THS3491 ? Although switching to another 50MHz TCXO? Which one will fit directly in the pcb? There are some 0.28ppm 50MHz available at mouser? Or a bit cheaper 2.5ppm. Both better than 100ppm
--- End quote ---
Hi, Noy
Can't help with the software since I still haven't gotten round to setting up the USB pass through for my winXP VM (VBox) and, afaiaa, there isn't any *nix version of the Feeltech SW being offered so it's either wine or a windows VM solution.
The opamp upgrade is worth doing. Even the THS3001 opamps would be an improvement over the 3002i originally fitted. The 3095, if still available will be fine otherwise use the latest 3491 chip which, TBH, is a little over-specified for this application.
Fitting a drop in TCXO replacement for the XO chip is not a good idea since your replacement is likely to land up running at 50 deg C just like the original chip did on account of its proximity to the three analogue voltage regulators raising the PCB temperature to 70degC! You really do need to install a cooling fan before upgrading the XO chip or else the 20 to 30 deg C delta from switch on will make a mockery of your 0.1ppm 50MHz TCXO upgrade by hiking its temperature well beyond the 0 to 25 deg C operating range typically specified in order to meet the 0.1ppm stability rating. Typically, the spec degrades to 0.5ppm in the sub zero and the 25 to 65 deg C temperature ranges (mine only claimed a degradation of 0.2ppm for this extended range as it happened).
A much better option is to buy a TCXO clock module, similar to this 100MHz one here, <https://tinyurl.com/ycsqvlml> (it seems this seller has now sold all of his stock of 50MHz oscillator modules) so you can mount it well away from the heat. I've got mine mounted at a jaunty 45 degree angle above the 50mm fan I'd installed into the base of the case between the back panel and the psu board, adjacent to the main board. This not only acts as a deflector to direct the airflow towards the front of the case, it also receives the full benefit of the incoming cooling airflow to isolate it from the heat within. Even so, there is still a small amount of heat pollution to raise the oscillator board temperature a degree or two above room ambient but at least it's a vast improvement over the more likely 10 to 20 degree rise in even the coolest part of one without the fan modification.
I chose the 50MHz board simply because it was the cheapest way to get hold of a 0.1ppm 50MHz TCXO. I was just going to remove the TCXO from the board and transplant it directly to the original XO chip location but, when I measured the temperatures on and around the XO with my IR thermometer, I had a complete change of mind and decided to make use of the whole oscillator board as supplied.
In retrospect, it would have been better to have used a 10MHz TCXO and replace the XO chip with an NB3N502 14 MHz to 190 MHz (output) PLL Clock Multiplier chip wired up to multiply it back up to 50MHz. The 10MHz TCXO modules are more available than the rarer 50MHz ones so can often be had below the 20 dollar price I paid for mine, plus, you then have the option of fitting an external 10MHz clock socket and a switch giving you the option to feed it from an external GPSDO or Rubidium clock reference.
If you're not planning on replacing the smpsu board with something better but want to raise the 11.5 ish volts on the nominal +/-12 volt rails to a more useful 13.5 to 14 volts to make more effective use of an opamp upgrade and give you a couple of volts margin on the DC offset range (literally none whatsoever at the 20Vp-p output setting!), then the most efficient way to achieve this is to wind a single turn secondary on the transformer (there's ample space to wind 3 or 4 such turns in parallel with the existing transformer using ordinary 0.6mm solid core hook up wire and it should be do-able without unsoldering the transformer as I had to when I added two turns aiding to each end of the 12v windings). This then can be used to buck the 5v winding, saving the need to raise the 4.95 volts to the 5.49 volt mark just to increase the 12v rail voltages.
You can lift the anode end of the upgraded diode used for the 5v rail so as to wire in this single turn in series with the original transformer connection to said diode. If you connect it series aiding, the 12v rails will drop by a couple of volts, indicating the need to reverse the connection. This provides a less lossy way to raise the 12v rails to a more useful voltage level and reduces the extra heat generated on the main board with the 10% boost on the 5v rail applied only for the sake of raising the 12v rail voltages. Unlike the "Add two turns onto each end of the 12v windings which involved unsoldering the windings from the tags and soldering the extra wire onto the released ends of those windings, this transformer mod completely eliminates the risk of transformer damage and is readily reversed. Regardless of this, it's still worth upgrading the original crappy rectifier diodes with better suited types which makes the additional single turn buck winding mod a trivial side task to the work involved in replacing the diodes.
Of course, you can simply replace the whole PSU with a ready made unit of better specification since, even with the fitting of a cooling fan in the base to the rear of the original PSU, there's still ample space to fit a larger PSU module. I have a suspicion that the switching ripple noise of the original psu board is the real culprit behind the claims of 'horrendous jitter' when using it as an RF signal generator with HF radio equipment but I haven't yet gotten round to making up a + and - 12v battery supply feeding a 7805 regulator to provide the 5v rail entirely free from any such PSU noise to verify this suspicion that what I hear using an HF transceiver on a 20MHz sine wave output setting is simply amplitude modulation from the psu ripple voltage. Hopefully, this noise is due to supply ripple voltage rather than something more fundamental to the mainboard itself.
IOW, I'm now reconsidering the analogue PSU option as a possible requirement to eliminate this defect in spite of my reservations over the additional waste heat such a supply will inject into the box. There are, apparently, such things as ultra low noise DC-DC switching converter modules designed with the needs of T&M gear specifically in mind. These can be used with a conventional transformer and rectifier setup in place of the heat generating anaolgue voltage regulators, neatly avoiding the regulator waste heat issue but I haven't yet dared to get any quotes on pricing since I've yet to prove my suspicions about the existing smpsu board. In any case I may yet be able to get hold of a three rail screened smpsu with low noise and ripple specifications similar to those dc-dc converter modules.
Although the FY6800 is already blessed with an IEC C13/14 connector to start with, neatly addressing the thorny issue of half live mains leakage voltage, I'm not impressed with the way they gone about connecting the safety earth from what I've seen of published photos of the vandalising of the PSU to main board connecting cable where they've rather pragmatically simply chopped one of the two ground wires to join the safety earth connection to, leaving the other end disconnected, thus compromising the ground connection between the psu and the main board. I suggest you join that loose grounding wire back onto the joint used by Feeltech to bodge the safety earth connection.
Such bodgery by Feeltech isn't unusual, witness the firmware bodgery employed to correct the mistake made in the board's manufacture where they placed 100 ohm shunt resistors in error for the (most likely) 56 ohm resistors required to create a 20dBish 50 ohm attenuator pad in place of the 85 ohm pad they landed up with as a result of that cockup. Since the FY6800 is supposed to be an improved version of the FY6600, I'm wondering about whether or not that extends to correcting this manufacturing cock up.
It's a fairly trivial matter to test what happens to the output voltage when the relay clicks in at the 500mV setting under terminated and unterminated conditions. There might be a small discontinuity in the change of voltage level in each case but if they've corrected this issue, it should be similar in both cases (open circuit or terminated). With the Fy6600, the discontinuity in the open circuit case was imperceptible but repeating this test with a 50 ohm termination in place produced a massive level change as the relay switched the attenuator in and out of circuit. Perhaps you can repeat this test and let us know your findings?
Anyway, that's quite a chunk of 'things to do' so I'll summarise the list here and in order of priority as I see it:-
1: Install a small cooling fan. Natural convective cooling with the current vent slots arrangement does nothing for the cooling, especially when tilted up on its prop stand where you'll get just about the same cooling effect as if no vents whatsoever had been provided. Those vents will only be a benefit with the addition of forced air cooling.
2: Upgrade the PSU (modify or replace) and, in the case of the FY6600 (not needed for the FY6800), fit an earthed mains socket (preferably an IEC C6 if you want to reduce the "Tail Wags Dog" effect). You can use a 10K resistor to link the 0v rail to the safety earth pin if you just want to kill the half live mains voltage without introducing troublesome mains earth grounding loops and the need to incorporate a "Grounded/Floating" option switch.
3: Once you've dealt with points 1 and 2 above, you can upgrade the rather weedy THS3002i opamp with a pair of THS 3001/3095/3491 opamps (take your pick, any of them will be better than what it was cursed with).
4: Remove the rather execrable XO chip from its rather toasty 50 deg C location and install a 0.1ppm 50MHz TXCO module in the coolest part of the case or use a 10MHz TCXO module with an NB3N502 PLL Clock Multiplier chip soldered onto the original XO location (it won't mind the 50 deg C temperatures - just make sure the TCXO module is kept away from the heat!).
[EDIT 2020-04-14]
5: Depending on whether there is still the issue of the 85 ohm attenuator pad to contend with in the 'Improved' FY6800 model, you might find yourself impelled to repopulate with resistors recalculated to create a 50 ohm pad that has the same unterminated voltage attenuation as the 85 ohm one to match the correction applied by the firmware fix.
Working out the required resistor values isn't quite so easy to do, even with the assistance of on line dB ratio and attenuator resistor network calculators. I landed up with a 45 ohm attenuator pad after adjusting the series element resistance with multiturn trimpots after putting 120ohm resistors across the 100ohm shunt elements to get a precise match to the Hi Z impedance attenuation of the original. I'm still confused by the results of my two subsequent attempts to calculate the required resistor values so I've left that as a minor annoyance to be dealt with at a later time. The resultant 45 ohm attenuator is at least some improvement over the original (but it's still not ideal).
If you do test the output levels around the 500mV setting where you can hear the relay click in and out, please let us know your results.
5: Ignore all the tosh above and replace those cockamamie 85 ohm resistor networks with a bog standard 20dB 50 ohm pad by replacing the 100 ohm shunt resistors with 61.2 ohm resistors and the 510 ohm series elements with 249 ohm resistors (both from the more expensive E192 preferred values 0.5% range). You can use an on-line Pi attenuator calculator to try alternative preferred values such as this one here:-
https://chemandy.com/calculators/pi-attenuator-calculator.htm
Regards, Johnny B Good.
Noy:
Thx for the summary. I will test it and give a Report.
Johnny B Good:
--- Quote from: Noy on January 30, 2019, 08:25:53 pm ---Thx for the summary. I will test it and give a Report.
--- End quote ---
Thanks for the swift response and I look forward to your report.
Regards, Johnny B Good.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version