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

[jdmoranii]

Hi, Dave. Sorry to resurface a long-dead thread, but…

I recently discovered that my FY6600 was dead. The -12V DC supply was being loaded down. I connected a (-) dc supply to that rail, limited current to 1A and cranked up the voltage until 750mA flowed. C81 got toasty hot. So I removed it. Searching this thread for “C81” and “C82”, I found your comment that C81/C82 were 100nf. But they look physically too big, the smaller caps in parallel with them are 100nf. So I removed C82 and measured it. It was 10u. I replaced C81 with a 10u.

[DaveR]

By chance, I just happened to look in on this thread for the first time in several months!

I remember the C82 incident, and have to say that I had no idea what its value was at the time (it blew on 10 Aug 2018, according to my notes!), but 100nF has done the job for the last four years, so it can't be too critical.  Physical size is, of course, no indication of value, and C53, 54, 56 and 57, which are all supposed to be 10uF, are massive in comparison to the original C81 and C82 on my PCB (as per some photos I took when did the 3095 opamps upgrade).  However, looking again at the photos, it would appear that C81 and C82 were actually different beasts (at least in colour and shape - although you couldn't tell that from the frazzled remains of C82 I removed), so my original assumption that they were the same value may have been incorrect - although why the capacitors on the 4558 VCC+ and VCC- should be different, I have no idea.  Thanks for your message, though: if C82 ever blows again I'll consider replacing it with a 10uF (if I can find one with a small enough footprint to fit in its place on the board).

Regards,
Dave

[Johnny B Good]

 Hi Dave 

 If you don't want to clutter your web browser with a tabbed link to a moribund topic thread or have to check your bookmarks' folder every so often just to take look for fresh postings, why not click on the "Notify" button? 

==============================================================================================
[EDIT 2022-09-03]

 To add to the above advice, that the key benefit to clicking the "Notify" button is that you can delete the tab any time thereafter and de-clutter the web browser. I usually wait a day or three to see how many times the attached images have been viewed before doing that, even longer if I think I may have more to add. If the viewed count rate drops right off after a week or three and I'm still planning on adding another post, I really aught to add it to my bookmarks folder and close the tab but I'm not that organised so just tend to leave the tab open.
==============================================================================================

 Another trick you can use when trying to resurrect a thread that hasn't seen a new post in several months (or even a year or three!) is to attach an image or three to your post. That way, you can gauge the level of passive interest from the number of views they get even if no one bothers to post a reply.

 Here's an example of the effectiveness of that trick where the latest post had been two and a half years prior to my own contribution. Sadly, no actual replies but the view count against each attached image does at least reveal a level of interest (the initial 4 views in the space of only 5 minutes strongly hinting at the use of the notification button).

https://www.eevblog.com/forum/metrology/i-just-bought-an-efratom-100318/msg4176703/#msg4176703

 By way of an on topic update, I'm pleased to able to report that my much modded FY6600-60M is still going strong. Indeed, its upgrade replacement, a Siglent SDG2042X (120MHz version ), has effectively been relegated to acting as a support plinth for the 6600, mainly it has to be said, for its inability to offer the advertised 1 microHertz tuning resolution 'feature' beyond 100KHz for an insufficiency of digits in the frequency display

 I've attached an image which I think might be of some interest (after all, this is the latest round in my search for a better frequency reference than the 6600 had originally been cursed with).

 BTW, the gps disciplining wobble accounts for some 6 or 7 ns of the band of persistence recorded in that screen grab. The remaining 29 ns comes from what appears to be a diurnal temperature influence on the ruby's frequency stability (along with a tiny calibration error and a smidgen of ageing drift).

[DaveR]

Hi John,

Nice to see you're still active!

Thanks for the tips, but, as someone who keeps the most cluttered browser in the universe (three of them, in fact, all multi-windowed and multi-tabbed), I don't intend to spend half of what's left of my life resorting tabs or converting them into bookmarks just to save a few seconds every six months or so .  Some of them are several years old, but I know where they are and I can easily find them thanks to the ColorfulTabs addon.  A number are linked to topics on here that I've participated in over the years, and I keep them open because it's an amusing diversion to look in occasionally to see how many more times the wheel has been reinvented since I became a lapsed member of the thread.

The FY6600 was a nice project to be involved in, and it eventually evolved to a level of sophistication which more than met my user needs (and many others', it would appear), and your efforts to squeeze even more performance from it have taken it to another level again.  It's amazing what can done with a bit of cheap kit and some ingenuity, isn't it?

Regards,
Dave

[Johnny B Good]

 Hi Dave,

 I've just edited that post to clarify the benefit of clicking the notify button as per the following.

 To add to the above advice, that the key benefit to clicking the "Notify" button is that you can delete the tab any time thereafter and de-clutter the web browser. I usually wait a day or three to see how many times the attached images have been viewed before doing that, even longer if I think I may have more to add. If the viewed count rate drops right off after a week or three and I'm still planning on adding another post, I really aught to add it to my bookmarks folder and close the tab but I'm not that organised so just tend to leave the tab open.

 As for the venerable FY6600, that proved to be a surprisingly excellent performer once you'd voided its warranty with several relatively trivial modifications to fix its seemingly beancounteritus inspired penny pinching induced shortcomings. Sure, the penny pinching did save a few cents on the dollar in production costs but I think the motivation to even risk such a ludicrously foolish penny pinching strategy was more to do with reducing their costs on handling warranty returns down to a nice fat zero by taunting their target market demographic into turning this Sow's Ear into a silken purse at the expense of a voided warranty.

 I think Feeltech knew their target demographic far better than it knew itself and were maximising the PRC's government subsidies on export shipping costs to make any warranty return attempts a high risk exercise if the AWG had been 'fiddled with' in any way since the return shipping charges would be an upfront cost to the customer, only refunded if Feeltech deemed the warranty hadn't been voided.

 When I'd offered this hypothesis to the group two or three years ago, it was discounted as just another "conspiracy theory" but the fact that their latest 6900 version has an unpopulated fan cooler aperture moulded into its back panel only lends further credence to this hypothesis imho.

 The original motivation for such nonsensical penny pinching may well have been simply that of reducing production costs which, after monitoring this topic thread, would have revealed the warranty voiding efforts of its target demographic soon enough that it seems they'd decided to pretend ignorance of the more egregious defects, electing to apply a cack handed half live mains leakage fix in the 6800 (and the 6900!) models and leave the remaining defects unaddressed to ensure a minimum of valid warranty returns where they would not only have had to deal with the expense of any repair (in reality more likely just take another unit from their stock of spares set aside for such warranty replacements) but also that of the customer's warranty return shipping costs.

 Looking at the way FeelTech have failed to respond to their customers' requests to fix blatant shortcomings, looks very much like a strategy to cut overheads by taking advantage of their customers' propensity to void their warranties (deemed worthless by most anyway) by having a go at fixing the issues for themselves, and thereby reduce the number of warranty returns they would otherwise be obliged to deal with.

 However, despite what I think of the situation, I'd still be inclined to replace the 6600 with a 6900 should it ever develop an irreparable fault. After all, for the money, it still represents an excellent bargain in AWG technology for what most of us would consider "chump change".

 It's certainly cheap enough to risk improving to a higher standard with an OCXO lockable to an external 10MHz reference and an upgrade of the 85 ohm attenuator to a real 50 ohm 20 dB resistor network plus replacing the earthing wire link to the PE pin of the C14 socket with a 4.7K resistor to eliminate random dc offsets and noise coming in from the mains earth wiring and, of course, fitting a fan into the case for a modicum of active cooling.

 Incidentally, the abysmal lack of passive convective cooling airflow proved inspirational in the ventilation arrangements I'd used in my rubidium frequency standard. I needed the cooling fan to be the sole arbiter over the LPRO's base plate temperature so any natural convective flow would have been counterproductive, so it's hats off to FeelTech for a real tour de force lesson on how to achieve virtually zero convective cooling.   

[DaveR]

Hi John,

I think the warranty thing is a bit of overthinking by you.  Like all Chinese companies at the lower end of the market, warranty doesn't enter their minds: they just know that they can build something for x dollars and sell it for y dollars, and if (y-x)*n sales is enough to set them on the road to riches then they are happy.  No need to follow international safety standards or awkward things like that, and they write the specs themselves and sometimes even manage to meet them: if they don't, so what?  Caveat emptor!  If they've done a bit of research, they'll know that there's absolutely no chance of getting a faulty item returned to them, because sending it back is likely to cost more than the item is worth, and if, by chance, something is sent back, they can just throw it in the skip and say they never received it!  It's all in the Chinese Business Manual.  That's not to say that they all follow the Manual to the letter, because I've dealt with some really good sellers in the past, who went way beyond expectations to provide a good service, but you just have to weigh the chances of buying a pig in a poke before you part with your money, and be prepared to write it off if the item turns out to be porcine.

As for the actual AWG itself, I knew what I was getting after the FY6600 and bought a 6800 because I saw a good deal from some seller, knowing I could easily upgrade it in a couple of hours - although I stopped after the earthing correction and the TCXO change because the newer firmware meant that I didn't see the need to do anything else to it - and even went on to buy a 6900 a couple of years ago just because curiosity got the better of me.  (I'm actually not that keen on the 6900 because the keypad buttons are too firm, and are a definite downgrade from the 6800.)  Having just looked at the 6900 case, I find that I did actually fit the "missing" fan (I can't remember doing it, but I'll have had one or two suitable ones lying around, no doubt), but that's probably all I've done to it, apart from the obligatory earthing fix.

Out of interest, do you actually NEED a rubidium standard, or is that just another obsession?  (Says I, with two Thunderbolts next to the PC keeping my feet warm, four more GPSDOs in various places, and half a dozen home-made 10MHz OCXOs sitting in a box upstairs!) 

Regards,
Dave

[Johnny B Good]

Hi Dave.

 Very nicely put. That's a less contentious version than mine but the conclusions reached are the same. I knew I was buying a "Pig in a poke" (more accurately, a "Sow's Ear") when I bought mine from a UK based Ebay seller as I mentioned in my very first eevblog post way back mid November 2018 to which my first reply had come from your good self! 

https://www.eevblog.com/forum/testgear/feeltech-fy6600-60mhz-2-ch-vco-function-arbitrary-waveform-signal-generator/msg1973777/#msg1973777

 I was interested to note your comments on the 6900's overly stiff buttons. Obviously stiff compared to the 6800 but surely an improvement over those bumps used on the 6600 front panel? These operate actual momentary push switches soldered onto the PCB. Those bumps are the one remaining annoyance only tempered by the fact that it's at least possible to replace any switches that may fail or start behaving erratically.

 Answering your question over my need of a rubidium standard, I have to offer an unqualified yes. I had been tempted to claim you'd answered your own question but although there is an element of want over need, in this case it's a vital tool in assessing other frequency standards (tcxo, ocxo, docxo, gpsdo and so on). In any case, I'll eventually be converting it into a GPSDRO to get rid of its ageing drift.

[DaveR]

Hi John,

The buttons on the 6900 are truly awful, at least on the model I bought, although they may have been changed again to something more acceptable on the latest 5v incarnations.  The buttons on mine are soft silicone rubber on a very stiff spring or switch; there is no tactile feedback, and it takes a lot of effort to compress the wobbly blobs enough to move the spring and activate the switches underneath - solid plastic buttons would have been far better.  The membrane buttons on the 6600 are a pleasure to use by comparison - at least you can feel the click when the switch operates.  The 6800 buttons are light touch, and feel just right: I couldn't believe how far backwards Feeltech had gone with the 6900, because you need two hands, or have it braced against something immovable, to use it!

Thanks for that link!  It's almost four years ago now, but reading a few pages from the thread brought it all back to me as if it were just last week.  I see I was stalling on the 6800 opamp upgrade even then, and never did carry it out .

Regards,
Dave

[Johnny B Good]

 Hi Dave,

 Thanks for giving me a "heads up" on the overly stiff buttons used by the (hopefully) early versions of the 6900. If they're so stiff as to make those used in the 6600 models a joy to use by comparison, they must be very stiff indeed (BTW "a joy to use" isn't a phrase that immediately springs to mind when using the 6600 buttons).

 If I ever need to replace the 6600 with another cheap AWG capable of 1uHz resolution from DC to 60 or 100MHz and lower sine wave jitter than that of an SDG2000X model, the latest 6900 models look (stiff button issues aside) to be an ideal replacement since most of the hard work, with the exception of replacing the resistors in the 85 ohm 20 dB attenuator, has already been done.

 They carried on using those cheap dual opamps but doubled them up to reduce sine wave distortion at the 20v pk-pk and 20MHz limits when driving 50 ohm loads. The improvement falls a little short of the opamp upgrade in the 6600 and 6800 models but still sufficient to make the expense of a pair of THS3091s or whatever, questionable so I'd be inclined not to bother.

 The current models using a single rail smpsu board makes for a simpler upgrade to a quieter 5v psu (it also simplifies the battery powered test setup somewhat to check whether such an upgrade is even worth the trouble of purchasing or making up a replacement psu).

 The change from a 50MHz XO to a 10MHz XO eliminates any need to add a 3N502 clock multiplier chip onto the OCXO upgrade's BOM list, saving both time and costs all of which make for a relatively easier DIY upgrade path. What's not to like (stiff buttons aside)?

 With regard to unused component purchases (that still to be done opamp upgrade you mentioned), that's not so unusual with DIY projects. The simplified battery power test I mentioned was prompted by the fact that the two 6v lantern batteries I'd purchased almost three years ago now, are still gathering dust on a shelf. I never did get around to testing whether a psu upgrade would be worth the effort.  

 I think the reason for putting this "battery test" exercise off indefinitely was due to having an epiphany involving the use of an HF transceiver with a short wire aerial to monitor the quality of the AWG's output into a metre or so of wire plugged into its BNC socket and being left a little bit bemused by the apparent high levels of mains frequency sideband noise which would vary in strength depending on how I laid out the AWG's transmitting antenna.

 Suspecting that the mains wiring was re-radiating the 10MHz test signal and mains current was modulating this re-radiated signal via some non-linearity mechanism (the "rusty bolt" effect), I connected the transceiver directly to the AWG through a 2 metre BNC patch lead and an SO239(?) adaptor, having already enabled its built in 20dB front end attenuator (no pre-amp required for the HF bands - just a bandpass filter feeding the 1st mixer directly). Lo and behold! A nice clean carrier sans the mains frequency harmonics. It turns out that relying on an "over - the - air" link path with an HF receiver to monitor the signal quality is not really a good idea (it's a quick and (very) dirty test with the "quick" of that description being its only redeeming quality).

 After belatedly running a proper test, I realised that what I'd initially concluded to be a psu noise issue was in fact just the result of a badly organised test that had allowed a polluted version of the test signal to be re-radiated from the mains wiring. The "psu noise issue" had proved to be nowhere near as bad as I'd assumed and quite possibly not even an issue at all so my planned battery test lost all sense of urgency and was literally left to gather dust.

 I might eventually get round to running this battery test, most likely when I get round to adding more padding to the trimpot on the OCXO board to make it less 'touchy' to adjust. The main impediment to running the battery test had been the lack of a suitable 6 pin connector to plug into the main board power socket. However, since I'd now like to be able to switch between battery and mains power at the flick of a switch to run the comparison as a "Blink Test", I'll probably just (temporarily) solder the wires straight onto the cct board (psu or main). One additional 'complication' in my case being the need to include the OCXO's 12v rail in this scheme since it's powered independently from its own smpsu (another possible noise source) regardless of the rear panel on/off switch state.

[tzok]

Any chance that the Front Panel firmware will be finished, or at least it will reach full functionality, like ability to turn the device off

[adeuring]

Got recently a second hand FY6800. One channel worked fine, the other had a terrible positive offset that made is practically unusable. (Forgot unfortunately to make a DSO screenshot before repairing the thingy...)

I suspected that the opamp driven by the DAC, an AD8009, (IC1 in DerKammi's excellent schematics, https://github.com/DerKammi/FY6600-15-30-50-60M/blob/master/Hardware/FY6600_Main_A0.pdf ) was broken and replaced it; now both channels work fine. Attached is photo of the old AD8009 (left) and the replacement (right). Note that the removed chip does not have the Analog Devices logo...



But before anybody concludes that Feeltech uses fake chips: (1) The device has an unknown history – I do not know who soldered the left opamp in the photo onto the board; (2) I am not familiar with the "pattern" used by Analog devices to label their chips – perhaps the left chip is genuine; (3) clones do not have to be worse the original…

Another oddity: The counterpart of the chip I replaced for the other channel (U10 in DerKammi's schematics) is an LMH6702MA, not another AD8009… But again: I do not know if the previous owner(s) of the device or Feeltech soldered this chip.

Thanks to DerKammi and everybody else who contributed to the schematic. Having it available made the repair really simple.

PS: I suspect that this thread and DerKammi's Github repo for the FY6600 can lead to a sort of addiction to the FX66xx. I bought it mainly for some audio measurements but after reading this thread I'm wondering if I shouldn't start too to tinker with this "sow's ear" .

[rhb]

Hi,guys.  I finally wandered back for a visit after a long hiatus and am happy to see my favorite energizer bunny, the F***tech thread is going strong :-)

Naturally, I had to get an eBay update on FY6900.  One seller in the US is offering a 2 year warranty and “100% satisfaction”.

Naturally, rather suspect.

Is there any hope yet of replacement FW to fix my twice self-borked 6600?

HaveFun!
Reg

[kaldach79]

Hi,
Do you know what driver is used in the generator display?
My LCD is damaged - grey backround, or only backlight. After freeze works fine by short time.
The original display has an 18 pin, but i have problem identifying the driver ILI9341 or ST7789.
Thanks for the help!

[kaldach79]

Am updating my post. Display containing ST7789 driver chip, not working properly.The image is inverted, and the blue and red color are reversed.
I will try display with driver ILI9341.

[hlj.pycad]

when I try to turn on the FY-6600 , it will show the brand icon "FeelTech" for a brief moment,  then the screen turn white. The keys seem still fuction. Any clue where I should look into?

[hlj.pycad]

I fixed the last problem. It's PSU output caps' problem.

But some how my "sinewave" became 3/4 sine+1/4square.... it's there a fix for this?

[Atlan]

Repair sine with program?

[hlj.pycad]

Yes. I just found that. Thank you.