Products > Test Equipment
FeelTech FY6600 60MHz 2-Ch VCO Function Arbitrary Waveform Signal Generator
Johnny B Good:
--- Quote from: MikeLud on August 04, 2019, 02:35:17 am ---JBG,
Thanks for your advice, I will be adding a cooling fan. Once I finish the PS upgrade I will be upgrading to a D75J and THS3091.
Thanks,
Mike
--- End quote ---
You might want to rethink the D75J option when you consider the fact that where the original crappy smd XO lives may well be running at 45 to 50 deg C as a consequence of all the heat being put into the board by the three LDO regulators located just one centimetre away. Those LDO regulators won't cool themselves you know, something has to handle the heat and that something is the PCB, including that smd XO. I'd measured 50 deg C on the XO with my IR thermometer with a 70 deg reading coming from the LDOs nearby.
Mind you, this was whilst I still had the wick turned up by 10% on the 5 volt rail and fortunately before I extracted the 50MHz TCXO oscillator can from the "50MHz TCXO Power Board" I'd ordered as a cheap way to get hold of an oscillator to directly swap out with the smd XO (albeit by having to resort to the 'mount on stilts' bodge).
In fact, I was still waiting for the oscillator board to be delivered when I made this discovery so was able to formulate a better use for the whole oscillator module to act as a deflector plate above the small 50mm square cooling fan I'd decided to fit into the base between the PS board and the back panel, adjacent to the main board where it could be held to within 2 or 3 degrees of room temperature rather suffer the 30 deg C temperature cycling of the original XO's location.
As I mentioned, these temperature readings had been obtained during my "5.49 volt period", so would likely have been a few degrees hotter than the standard (4.96 volts) setting and by few, I mean less than 5 degrees.
A couple of things that had put me off going the D75J route, aside from the delivered cost, was the absence of a calibration trimmer and the rather lacklustre +/-1ppm accuracy and stability. My motto being, "If a job's worth doing, it's worth doing well or not at all.", I felt the 20 fold improvement in frequency accuracy and stability being offered wasn't sufficient to justify the time and effort when, for pretty well the same money and just a modicum of extra effort, I could do ten times better on the stability front and somewhat better again on the calibration accuracy front, courtesy of the trimmer option gained by the purchase of that "oscillator power board".
The JYEC DIP oscillator on that oscillator power board boasts a stability of +/- 0.1ppm over a a 0 to 25 deg C range and +/-0.2ppm over the -10 to +50 or 60 deg C range and, of necessity for this sort of accuracy and stability, boasts a trimmer to allow calibration against a reference frequency source such as a GPSDO.
I've attached a photo of the oscillator board - it's now going spare, perhaps I can flog it as a CD player upgrade to some audiophool ;)
The only thing going for the D75J option was that it was a simple drop in replacement (assuming smd reworking holds no terrors). The 50MHz +/-0.1ppm oscillator board, otoh, did offer the opportunity to avoid 30degC temperature cycling, allow calibration against an NPL standard, and a 200 fold improvement in stability and accuracy over the original XO versus the marginal 20 fold improvement offered by the D75J. The choice really was a no brainer, imho. ::)
I was eventually able to calibrate that TCXO to within 30ppb before I gave up trying to improve my best efforts once I realised that it was destined to vary by around the same amount during each day's use. The only thing better than a TCXO being an OCXO but, at the time when I first read Arthur Dent's posts on his own OCXO upgrade project, my initial thought had been that he had gone just a little OTT in this upgrade exercise, especially considering the need to find an extra 3 or 4 hundred milliamps at 12 volts to power the OCXO's oven, necessitating either a PSU upgrade or else the addition of another 12v half amp or better psu board inside the box[1].
Strange to say, that's now where I'm at with my own pet FY6600 project having decided that such an OCXO upgrade wasn't so OTT after all! :) It's amazing the things you can observe with a function generator that boasts a frequency accuracy better than 100ppt with a day to day stability of 20 to 30ppt. :)
That's now more than sufficient to show the limitations of a basic PLL GPSDO based on a "cheap 'n' cheerful" navigation only GPS receiver module whose positional accuracy (and therefore timing data) is at the mercy of the vagaries of the ionosphere. I'm talking about the minute to minute nanosecond phase shifts rather than the long term precision which is not in any doubt.
Anyway, enough about the issue of exposing a D75J to such temperature extremes as are to be found inside these Feeltech function generators. When I was suggesting there was no need to bump the 10μF 450v smoothing cap up to a 68μF 400v one, nor indeed any need even when powered off a 117v 60Hz mains supply, suggesting there was little point in such an "Upgrade" if you were powering it from 220 or 240 volt 50Hz mains supplies, I'd overlooked the Stars and Stripes flag signifying your location as being somewhere in the US. I had fleetingly thought of suggesting that a more modest increase of capacity to 12 or 15 or even 22μF (even 33μF if that's what you happened to have in your parts bin) could be applied for perhaps some marginal benefit to those on 117v mains power where the main benefit would be more a matter of riding through brief drop outs rather than improved smoothing.
I'm not suggesting that you need to remove that 33μF cap: that's obviously fine in view of the fact that quite a few modders must have already proved that the 68μF has only harmed their bank balance with no ill effect to the PSU board itself. Of course, if you want to save that 33μF cap for 'better things', you might want to swap the original back in or select a 12 or 15μF cap instead. Choices, choices, choices! >:D
[NOTES]
[1] I had to add a 12v 500mA smpsu board I'd removed from a small Linksys psu I'd had going surplus to requirements to power the OCXO board since I'm still running with the original (though modified) smpsu board. That actually gives me an advantage over Arthur Dent's version where, afaicr, he relied upon an upgraded PSU to power the OCXO. The advantage being I can now shut the whole signal generator down on the rear panel switch without interrupting power to the OCXO since I wired its dedicated psu directly to the mains socket so that for as long as it's plugged into a live mains outlet, it can maintain the crystal oven temperature indefinitely.
The OCXO draws 5 watts from the mains when first powered up from cold but once up to temperature (after a 2 or 3 minute warmup time), this drops to a mere 1.3 watts (about half a watt lost in the psu with the OCXO taking some 700 to 800mW from the 12v supply). Putting the signal generator into standby using the front panel on/off button, only saves a watt at most with a typical standby consumption of around 5 watts so this is a very worthwhile economy and life extending feature in an item of test gear you'd otherwise tend to keep powered up 24/7 in the interests of maintaining frequency stability.
There is a tiny penalty in shutting it down overnight in that there seems to be a 30 to 60ppt 'warm up drift' (it's hard to quantify exactly with my current GPSDO lashup). For most practical purposes this is insignificant and later on, after I've installed an external reference input socket to feed it from a GPSDO, that isn't going to matter any more for any tests involving the need for such accuracy.
JBG
DaveR:
--- Quote from: Johnny B Good on August 03, 2019, 11:05:23 pm --- The analogue psu can be an effective solution but the downside is that you'll introduce greater thermal stress than the original (typically 70 to 80% efficient) smpsu board on the unit if you neglect to add a cooling fan to prevent an already hot running generator from overheating.
If you're going to build a custom analogue psu, you need to be very mindful of this thermal issue.
--- End quote ---
Use an R-core transformer instead of a standard EI one. They run very cool and heat from them isn't an issue. No fans necessary.
Regards,
Dave
aliaj00:
Guys i just bought the FY6800.
As i am new i could find is the specs the max current this one can provide and the single channel.
ArthurDent:
It’s been a while since I posted an update because basically I haven’t made any changes to my FY6600 for about one year. Here is a recap of what I did.
As you can see from the photos I’m still using the same FY6600 I bought originally with V3.1 firmware. I’m also using the same output amps that came with the unit. The high quality SMPS I settled on as a replacement for the crappy one is about 25W, is quiet, runs cool so no fan is required, and is quite small.
The timebase modification I made using the 5X PLL chip and the 10Mhz OCXO is working well and if I don’t mind waiting about ½ hour for it to warm up and stabilize, I don’t have to use the external GPSDO 10Mhz at all. This 15 year old OCXO is very stable and is good to about the first 10 digits. Keep in mind that this OCXO is far more accurate than the original and warms up fast so I could use it within minutes of powering it up but I prefer to wait until the drift rate is almost nil. When I do check it I use my scope with the time/division set to 2ns/div. The DDS circuit they use is damn good to another few digits as well if fed from the GPSDO or Rb standards.
Johnny B Good:
Thanks for posting those pictures, Arthur.
When I first read your exploits with the OCXO, my first thought had been that such a modification was a little OTT but that was before I'd bagged myself a bunch of CQE OCXOs for experimenting with building GPSDOs after realising that a decent OCXO was the starting point rather than just an inconvenient add on to deal with the jitter from a u-blox M8N GPS receiver module's programmed 10MHz output. :-[
I've long since come to properly appreciate the value of such an upgrade so my thanks to you for the inspirational use of a 3N502 clock multiplier chip. The initial choice I made in using a TCXO instead of the full fat solution of a TCXO, was driven by the fact that I wouldn't have to replace the existing PSU with a more powerful one for the sake of keeping an OCXO's oven up to temperature. In the end, I still avoided replacing the original psu by the expedient of adding a separate 12v 500mA smpsu board to power one those 12v CQE OCXOs I'd decided to put to good use as the generator's frequency reference.
I'm planning on adding an external 10MHz reference input socket and use the injection locking technique to avoid the risk of glitching the front panel micro controller from switching between the local and external references - it had become a little twitchy in that regard after I'd installed the 50MHz TCXO board. It seems to have lost that twitchiness since upgrading to the OCXO but I'm still mindful of the potential problem such mechanical switching of oscillator sources might cause. Also, of course, it's a neat way to avoid the need for an actual switch to be fitted on the back panel. :)
As for both yours and DaveR's view that there's no need to fit a cooling fan if either a more efficient smpsu is fitted or else an R type transformer is used for an analogue PSU replacement, I feel I must point out that you're overlooking the fact that both the FY6600 and Fy6800 models were already running uncomfortably hot before any additional heat sources were introduced into the box.
The 25W rated smpsu might be cooler running from its point of view but the fact remains that its efficiency at a relatively low loading may be no better than the crappy original and quite likely adding a little more heat into the box. R type transformers in an analogue PSU aren't the problem, it's the waste heat from the analogue regulators that poses the real threat. R transformers are a good way to go but I'm not so sure about the analogue aspect unless it's being used to eliminate the high frequency ripple of a switching psu with consideration to the penalty of additional waste heat this introduces.
At the time when such psu mods were a popular pastime, it seemed to be largely driven by the need to eliminate the half live mains 'leakage' esd risk inherent to all small class II smpsus through the mandated Y cap EMC bodge to reduce common mode conducted interference.
Unfortunately, without ready access to the safety earth, even a conventional mains transformer analogue supply will still represent some esd risk to a poor defenceless DUT. That being the case, the pragmatic solution to that particular problem is simply to replace the non polarised two pin mains lead connector with a three pin type to give access to the safety earth by which to short out the unwanted leakage with a 1 to 10k 'drain resistor'. You could hard wire the BNC grounds if you like but this introduces a whole new can of worms to the FY6600 in the form of an unwanted mains wiring routed grounding loop (along with undesired galvanic and thermocouple induced dc offsets when working with millivolt level signal voltages).
The main problem with the original PSU board after replacing the rectifier diodes with ones more suited to the task and doubling the 12v rail cap values to 16v 470μF types (the voltage rating isn't an issue here) and adding a single turn winding to buck the 5v secondary to get a better balance of power distribution to the 12v rails without making the three stooges (LDO regulators) get unnecessarily all hot and bothered, is the high level of HF ripple being both conducted and radiated to pollute the signal processing circuitry with noise. There's no doubt a much quieter PSU would improve the quality of the signals being generated whether it be an analogue or a switching psu.
Most all modern T&M gear these days relies on SMPSU technology so it's obviously possible to design and manufacture low noise smpsus - I've just been too frightened to ask the price of such specialised ultra low noise switching psus and dc-dc converter regulator modules so I've put the question of improving the PSU on the back burner for now.
I suspect it may be a lot easier to tame the problem of switching noise in DC-DC regulator converter modules than it is to deal with the high voltage switching noise in a conventional smpsu so a better approach might be to use an R transformer with conventional rectifier and smoothing to power switching DC-DC converter modules in the place of your typical 7805, 7812 and 7912 regulators.
As it happens, I've just ordered a couple of small DC-DC converters from BangGood (along with a KSGER V2.1S T12 Digital Temperature Controller Soldering Station and other bits 'n' pieces) to try out with my current GPSDO project as an alternative to using a 7805 bolted onto the aluminium case to handle the 9 or 12 volt wallwart voltage conversion to the single internal 5 volt rail supply (I'm using the original 5v 13MHz CQE OCXO I'd fortuitously spotted and purchased for just 4 quid at the Blackpool amateur radio rally last April) so I'll be able to check the <30mV ripple claim made for the fixed 5v output module (the other module does 3.3, 5, 6, 9 and 12 volt but is silent on the subject of output ripple).
A small efficient mains transformer using low noise switching converters may provide a more efficient solution to the problem of power supply ripple noise with the present PSU board. I don't really want to speed the fan up to compensate for an extra 5 to 10 watts heating load from an analogue PSU solution if I can help it.
JBG
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version