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

[norms]

Is there some place which has waveforms that can be downloaded for this generator, or is there a site which has user uploaded waveforms which may be compatible. Would be nice to have wide selection to choose from.

[cybermaus]

I just want mine to work!

A very valid goal.

And I think it is OK to inform new users by posting every so often that these mods are not strictly needed to have a nicely working $90 AWG signal generator with a good cost/benefit ratio out of the box. But that does not mean we cannot play. 

And also, every so often some handy piece of information may fall out that even the more conservative owners can use.

But yes, adding a €45 TCXO on a $90 device also took my attention. Only do that if you enjoy doing that.

[paulca]

Any tips on getting this to produce bode plot?

I tried earlier today and I don't see a way to output a square wave for the trigger on channel 2.  Also is there a way to one shot the sweep?

Granted it would do without those, but I think at least the trigger would make it easier to read on the scope.

[cybermaus]

What I did though is I made a custom waveform, 0 to max, for VCO input to sweep, and a small -max pulse at the beginning to trigger the scope. And I fed that into the VCO in the back. So no need for a square wave.

If you are doing this at higher freuqencies, the pulse at the beginning may need to be more than just one sample, because we may have 8192 samples, but above 30KHz not all samples make it to the signal (250MHz sample rate / 8192 samples)

How to one-shot it though? not sure, I guess that would be a feature request to fremen67.
But why not simply repeat your bode plot?

[spec123]

Buying and modifying a FY6600 is not as bad as it sounds. My FY6600 cost $98 with free shipping, the DJ75J-050 TXCO cost $11.44 and the two THS3095Ds cost $19.84. My total: US$129.28  Assuming you don't fry your FY6600 making the modifications, you will have a FG/ARB SG that compares in many features to some of the higher priced ARB/SGs from Rigol, Hantek and Tek.

For example, the Tek AFG1062 has many of the same features but costs over $840. Granted the AFG1062 is better built and has much better support. But in my case I wanted an ARB/SG for doing some limited experiments with variable phase output and internal direct modulation. For this it works good.  Its also a good tool for demonstrating various waveforms and phase shift to science classes.

BTW, on another thread here on EEVblog it was mentioned that a new FY6800 model will be released next week. It is supposed to fix many of the problems discussed here. It will be interesting to see what it looks like.

 

[cybermaus]

Assuming you don't fry your FY6600 making the modifications, you will have a FG/ARB SG that compares in many features to some of the higher priced ARB/SGs from Rigol, Hantek and Tek.

Arguably, without mods we have that.

Please note these specs from the Rigol DG1062Z, comparible dual 60MHz for $950

   Range
   ?10MHz: 1.0mVpp to 10Vpp
   ?30MHz: 1.0mVpp to 5.0Vpp
   ?60MHz: 1.0mVpp to 2.5Vp
   Offset (into 50 ?)
   Range (Peak ac+dc) ±5Vpk ac+dc

So, we may complain about distortion into 50 Ohm load above 10Vpp, but the Rigol does not even go there.
And we had some High-Z SA images that proved to be very good without 50 Ohm
To compare we should check 50 Ohm only to 5Vpp, and the others only to 10Vpp

Interesting about the FY6800 though. Likely going to be the better choise going forward. But hacking, modding, tweaking aside, unless I brick it, I am not in need of a new one.

[Lanvernazal]

Hello, does anyone know what was supposed to improve version 3.0 compared to version 2.9 (which I own and which works well for the moment)?

[cybermaus]

We know they de-proved the lifespan. Significantly. V3.0 will die on you.

But it seems a moot question, we don't have the firmware files, so you cannot upgrade.
Unless you are thinking of buying another one, but if you see a 3.0 in a shop, leave it well alone, only 3.2 or higher is advised for new purchases.

[paulca]

What I did though is I made a custom waveform, 0 to max, for VCO input to sweep, and a small -max pulse at the beginning to trigger the scope. And I fed that into the VCO in the back. So no need for a square wave.

If you are doing this at higher freuqencies, the pulse at the beginning may need to be more than just one sample, because we may have 8192 samples, but above 30KHz not all samples make it to the signal (250MHz sample rate / 8192 samples)

How to one-shot it though? not sure, I guess that would be a feature request to fremen67.
But why not simply repeat your bode plot?

Thanks.  To be honest I was just experimenting, never tried to make a bode plot before, I figured out the concept one rainy morning but then found Dave's video on the same.

My scope is a bit limited and I found it was tricky without the trigger and with the sweep repeating to set the time base and pause the scope at the right point.

If I understand you correctly, your custom waveform sets the frequency of the sweep?  Like the remote VCO input on an old synth?   Could you not add 0Hz portions to either end to at least delimit the bode plot clearly.

Anyway, I think for the level of testing I am doing, which is purely indicatively verifying the real circuit does roughly what Spice said it would it's probably fine with what I have, though I might experiement with the VCO waveform at some point.

[spec123]

Paulca, expanding on the suggestion of Cybermaus, you can get a triggered sweep for your scope by feeding a low frequency negative ramp from CH2 to the VCO terminal in back. Make sure your ramp goes from 5V to 0 V, using offset and amplitude adjustments. Set your start and stop frequencies, then use"Forth" in Sweep selection to get a forward frequency plot. Your scope can be triggered directly from the positive edge of the ramp using a second scope channel or ext trigger.

While this does work, it is cumbersome and fraught with possible errors. For more serious work with frequency response plots I suggest you get a low cost scalar network analyzer. There are plenty of low cost ones on eBay.

[plb]

Paulca, expanding on the suggestion of Cybermaus, you can get a triggered sweep for your scope by feeding a low frequency negative ramp from CH2 to the VCO terminal in back. Make sure your ramp goes from 5V to 0 V, using offset and amplitude adjustments. Set your start and stop frequencies, then use"Forth" in Sweep selection to get a forward frequency plot. Your scope can be triggered directly from the positive edge of the ramp using a second scope channel or ext trigger.

While this does work, it is cumbersome and fraught with possible errors. For more serious work with frequency response plots I suggest you get a low cost scalar network analyzer. There are plenty of low cost ones on eBay.

My FY6600 (V3.2) has quite significant variation of amplitude with frequency. While I have used the CH2 ramp to sweep, the result is approximate at best.

[soundtec]

So ,today I took out the original op-amp and soldered in the pair of THS3095's , worked out fine ,slight clipping on the upper sideband of the sine at 20V p-p , as soon as I dialed in .1 volt offset negative the clipping appeared on the lower side instead , shows that balance is fairly good at least . My 12 volt rails are around 11.5 volts so I reckon a couple of extra volts will sort out the clipping issue.The other thing I did was add two layers of thermally conductive insulation on top of each 3095 ,that made up the height difference between the 4558 and the new chips nicely ,I put a small amount of thermal paste on the 4558 also ,even without the new op amps I think sorting the height difference makes sense and allows better conduction to the heatsink . I didnt get a chance to check to see if the op amps made any difference to THD ,but  I'll do that at a later date . It would be nice to be able to bring THD down from the .6% figure quoted in the specs ,would definately make it more suitable for audio testing .Id be interested if anyone else has measured distortion with a linear psu also.

[KD4PBS]

For what it's worth as another data point for discussion on this topic, I'll add my experiences with our shop purchasing the Koolertron version of this generator a couple of months ago. 
The 60MHz unit arrived from the Ebay vendor's warehouse in Walton, KY quickly.  Upon initial unboxing and examination of the unit, it was discovered that several waveforms were missing from the generator, right out of the box: Lorenz, Sinc (I assume they meant "Sync"?), Multi-Tone, Exp-Decay, Exp-Rise, and Noise.  No waveforms were output, and the representative waveform on the display was missing; nothing but a flat-line.  Immediately I filed a return.  The vendor was quick to answer back with accusations that I didn't know how top operate the generator, that I should watch a video, and that obviously the generator was fine - the PEBKAG.  They demanded me to post a video of the problem before they would do anything about it.  I wasn't going to stand for that crap and having to jump through their hoops (I can't easily post a video for their amusement) and let them know exactly what was going to happen; send the defective device back for a full refund.  First they stalled a few days.  Then their shipping label that they uploaded to Ebay was irretrievable.  Upon investigation, it was discovered that this is a frequent ploy that some less-than-honest sellers will do; upload an invalid .pdf file as a return shipping label, and in the process, shift all of the legwork of doing the return to the poor soul who bought the defective item in hopes that they'll eventually just give up, or perhaps stall things long enough that the return period will expire.
Long story short, I ended up having to drive tens of miles to get the thing shipped, using my own money.  Their idea of "reimbursement" for the shipping fee was to short change me by exactly $1.  They told me that since they accepted return on my device (yeah, after more than 2 weeks of back-and-forth ridiculousness!) they expected me to reverse the negative feedback given for the transaction (the first and ONLY time I've ever given such in 15 years of Ebaying).  And their response to that negative feedback was that I didn't know how to operate the device.  Uh... yeah.  That's it... I knew how to get sine, square, triangle, and all kinds of other waveforms out of the device but for some reason I just didn't know how to operate the unit well enough to get those other waveforms out of it.
At least the ARB waveforms would work, and I left within the generator a nice little hand-drawn "hand with middle finger saluting them" waveform.
After all that fiasco, I bought the Feeltech model.  The buttons aren't as nice, and I'll have to deal with the voltage on the ground lead, but other than this, so far so good...  I just need to check what version of firmware.
And if anyone needs to know the vendor name, let me know.

[cybermaus]

No, it is spelled Sinc alright.
Interestingly, you leaving a "middle finger" waveform will probably be interpreted by them as "you see, the Sinc function does work"
Because Sinc sort of looks like a middle finger: http://mathworld.wolfram.com/SincFunction.html

Though, honestly? No one will ever look at it, and thus not see your middle finger at all.

As to your ebay return experience: Pretty much how would I expect it to be, except for the end result: you actually managed to return it at their expense (minus $1)

[rhb]

I just want mine to work!

A very valid goal.

And I think it is OK to inform new users by posting every so often that these mods are not strictly needed to have a nicely working $90 AWG signal generator with a good cost/benefit ratio out of the box. But that does not mean we cannot play. 

And also, every so often some handy piece of information may fall out that even the more conservative owners can use.

But yes, adding a €45 TCXO on a $90 device also took my attention. Only do that if you enjoy doing that.

I'd be far more interested in feeding my GPSDO from Leo Bodnar to it as the clock.  If I need higher output to drive a mixer, I think an outboard LNA that takes the cleanest signal it produces is a lot simpler to do.  And it has other uses as well.  I think the same startegy is more appropriate for the TCXO as well.

I'm still waiting on my STM32F103 boards and have been busy with an LM399 voltage reference and an HP 3478A DMM that came yesterday.  What's the status of just getting usable FW in my 3.1 unit?  Are we there yet?   Recent posts have been on replacing op amps and such which I'm not rally interested in doing.  I've got an STM32 JTAG so I'm ready to reflash the unit once we have a FW candidate.

[KD4PBS]

Ahh!  Thanks!  Silly me thought it was supposed to be "sync", as in a synchronization pulse of some kind.
Little did I know it was a math function, especially since I never saw the waveform 

Honestly?  Yes.  The poor dude that gets this generator re-sold to him from this shady vendor might see it and realize that they're selling B (and possibly C) stock. I don't care how it makes someone else feel... it sure made me feel better!  I will say though, my artwork made it obvious what it was, and it looked nothing like a sinc waveform  .  I won't even go into what the other arbitrary waveform I drew up resembled...

We'll see how long the Feeltech replacement works, but as often as we use something like this in the shop, I'm betting it will be fine for what it is.  First impressions?  What an excellent little generator, with a few limitations and gotchas, but a good bang-for-the-buck.  Does exactly what we need it to do around the shop.

[cybermaus]

What's the status of just getting usable FW in my 3.1 unit?  Are we there yet?

While true that things are a bit quiet around a usable custom firmware, I'd argue the V3.1 is pretty usable already. It is V3.0 that crashes and hangs.
Pretty happy with my own V3.1

[rhb]

Thanks.  I was a bit concerned that it might get worse.  Also I've been buying new toys faster than I can try them all out.  Mostly I'd like to get the FY6600 fixed and off my "to do" list.

[Insatman]

For what it's worth as another data point for discussion on this topic, I'll add my experiences with our shop purchasing the Koolertron version of this generator a couple of months ago. 
The 60MHz unit arrived from the Ebay vendor's warehouse in Walton, KY quickly.  Upon initial unboxing and examination of the unit, it was discovered that several waveforms were missing from the generator, right out of the box: Lorenz, Sinc (I assume they meant "Sync"?), Multi-Tone, Exp-Decay, Exp-Rise, and Noise.  No waveforms were output, and the representative waveform on the display was missing; nothing but a flat-line.  Immediately I filed a return.  The vendor was quick to answer back with accusations that I didn't know how top operate the generator, that I should watch a video, and that obviously the generator was fine - the PEBKAG.  They demanded me to post a video of the problem before they would do anything about it.  I wasn't going to stand for that crap and having to jump through their hoops (I can't easily post a video for their amusement) and let them know exactly what was going to happen; send the defective device back for a full refund.  First they stalled a few days.  Then their shipping label that they uploaded to Ebay was irretrievable.  Upon investigation, it was discovered that this is a frequent ploy that some less-than-honest sellers will do; upload an invalid .pdf file as a return shipping label, and in the process, shift all of the legwork of doing the return to the poor soul who bought the defective item in hopes that they'll eventually just give up, or perhaps stall things long enough that the return period will expire.
Long story short, I ended up having to drive tens of miles to get the thing shipped, using my own money.  Their idea of "reimbursement" for the shipping fee was to short change me by exactly $1.  They told me that since they accepted return on my device (yeah, after more than 2 weeks of back-and-forth ridiculousness!) they expected me to reverse the negative feedback given for the transaction (the first and ONLY time I've ever given such in 15 years of Ebaying).  And their response to that negative feedback was that I didn't know how to operate the device.  Uh... yeah.  That's it... I knew how to get sine, square, triangle, and all kinds of other waveforms out of the device but for some reason I just didn't know how to operate the unit well enough to get those other waveforms out of it.
At least the ARB waveforms would work, and I left within the generator a nice little hand-drawn "hand with middle finger saluting them" waveform.
After all that fiasco, I bought the Feeltech model.  The buttons aren't as nice, and I'll have to deal with the voltage on the ground lead, but other than this, so far so good...  I just need to check what version of firmware.
And if anyone needs to know the vendor name, let me know.

Interesting story and I think unfortunately all to common on Ebay.   As a side note; besides owning a FY6600 I also bought a JDS6600 unit a few months before buying the FY6600.   Your Koolertron unit appears to be a re-brand of the JDS unit (or perhaps visa-versa).   Out of curiosity, I checked the waveform types that were missing on your unit and they all worked on mine.  I don't use the JDS unit much anymore preferring the Feeltech unit after several modifications/upgrades.   Incidentally, the JDS (and likely Koolertron) units also suffer from the AC leakage current even though they use an external DC power supply.   This is due to the cheap type of switching supply used and lack of any earth ground connection.  I measured similar values of ac leakage on both the JDS and FeelTech units prior to adding ground references to them.

[ArcticGeek]

Could someone who owns one of these FY6600 signal generators do a simple test - could you set the unit to output a pulse waveform at 10MHz with a ~40% duty cycle and see if it has any jitter as described in this thread:

https://www.eevblog.com/forum/testgear/rd-jds6600-signal-generator-jitter-on-pulse-output-when-duty-cycle-is-not-50/

If the Feeltech is clean I might return my unit, otherwise I suppose I might live with it *sigh*.

Thanks much.

[Insatman]

Could someone who owns one of these FY6600 signal generators do a simple test - could you set the unit to output a pulse waveform at 10MHz with a ~40% duty cycle and see if it has any jitter as described in this thread:

https://www.eevblog.com/forum/testgear/rd-jds6600-signal-generator-jitter-on-pulse-output-when-duty-cycle-is-not-50/

If the Feeltech is clean I might return my unit, otherwise I suppose I might live with it *sigh*.

Thanks much.

My modified FY6600 has jitter on 50% duty cycle but not at 40%...go figure.   Mods to my unit are; upgraded clock (DJ75), replaced output caps in PS, added ceramic caps and choke to output of PS.  Added replacement opamps (THS3095).   I don't think any of these mods affects this particular problem for good or bad.   I'm guessing it has to do with hitting the duty cycle time close to the clock transition of your particular unit regardless of manufacturer.  A slight change in the duty cycle fixes it.  For instance on my FY6600 going from 50% to 48% eliminates the jitter. 

[cybermaus]

I am assuming that the 5ns jitter comes from the 200MHz sample rate that the unit has.  I don't really know how DDS signal generators work, but I am assuming this 5ns is an artifact of the 200MHz base clock.

For the same reason, the 250MHz DAC in the FY6600 has 4ns. So a hair less, but pretty much the same.

My modified FY6600 has jitter on 50% duty cycle but not at 40%...go figure.

Did you per chance use the frequency calibration? If you set that different then 10000000, then your 10MHz is no longer a perfect divider and the points where jitter happen shift and are no longer predictable.

[Insatman]

My modified FY6600 has jitter on 50% duty cycle but not at 40%...go figure.

Did you per chance use the frequency calibration? If you set that different then 10000000, then your 10MHz is no longer a perfect divider and the points where jitter happen shift and are no longer predictable.
[/quote]

Yes I did use the frequency calibration to correct for the error in my TCXO so I get a very accurate output frequency.   I think you are correct in that it will change where the jitter shift happens.

[ArcticGeek]

Thanks for the replies, that tells me that this issue is inherent in the design.  I guess I'll live with it on my unit.

[SMB784]

Well I replaced the old stock version 3.1 chips (winbond flash and stm front panel chip) with the version 3.2 chips that my eBay seller sent via the slow boat from China and they worked like a charm. My device is now upgraded to version 3.2 and everything still works swimmingly.

If there was any doubt, this transaction has convinced me that the sellers know that all the problems are with the front panel programming. Anyway, I have an extra working 3.1 chipset that I will be sending out to Fremen67 for further analysis.