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

[DaveR]

We can decide all we want, but the thread isn't going to be found by people interested in the FY6800 very quickly if they weren't already familiar with the thread.

Eh???  Just try typnig "FY6800" in the search box and see how many hits to this thread come up!

It's the most elegant solution.

Solution to which problem, exactly?

[GregDunn]

Im wondering is the limits in terms of distortion in the FY 66/6800 due to the dac/filter and the bits , in which case Im not sure if theres much benefit in distortion terms from changing out the op amps , surely though if you wanted the highest quality signal from the unit ,say for audio testing which often will require smaller signal voltages ,you'd be better off using the sig gen to make a large amplitude signal ,then using a passive attenuator network to attenuate that to a level suitable for your purpose , in otherwords does producing a smaller output voltage in the FY limit the bit rate of the output signal ?. Could a specially made switched attenuator, optimised for 50 ohms, used after the generator  allow an improvement in distortion ? I know that in the case of using a sound card for audio testing a couple of the smarter people in another forum I visit made up a special attenuator allowing attenuation in Db ,the signal generator is then run at 0db Fs digital for minimum digital artifacts. If at higher levels in the 66/6800distortion of the output stage is the limiting factor then power supply and op amp upgrades are worthwhile. For most audio testing the small distortion in the output signal of this unit is of no consequence ,it only comes into play really if its distortion  your trying to measure . I guess a computer and high quality dac /attenuator is a better option for producing signals in the audio range with very low distortion in any case .

That's right, the DAC is used to generate the output and it's just buffered by the op amps;  using it to generate a lower level signal will effectively use fewer bits.  Running the output at max level and putting a post attenuator on the FY6800 ought to lower the distortion quite a bit due to (1) reduced quantization errors and (2) linear scaling of the signal.  I should have thought of that, and I will try it.

[bitseeker]

I just did a search in the forum for "FY6800" and none of the top six results are this thread, five of which have "FY6800" in the thread title.

Anyway, that's all. Carry on.

[GregDunn]

OK, that worked well.  I maxed out the FY6800 at 20V, ran it into the 331A through the attenuator, and measured a THD+N of approximately .03% - close to the theoretical limit for a 12-bit system.  That's more than sufficient for anything I'm doing, and not far from my IG5218.  Color me suitably impressed.

[soundtec]

I like the filter approach suggested by JS ,you needen't butcher the internals of the 66/6800 ,it  could easlity be made into a passive box and placed on the output of the sig gen with a known amount of attenuation of signal .into a given load impedence ,maybe with switchable centre frequencies , Bruel & Kjaer make some very nice passive attenuators/filters for audio measurement work ,I wonder if one of these would be a good off the shelf add on to a sig gen setup , the units themselves seem to be for interface with certain other B&K gear ,but adding some BNC's would surely be an option .


https://www.ebay.de/itm/Bruel-Kjaer-1616-1-3-Octave-Filter-Set/123213621878

https://www.ebay.de/itm/Bruel-Kjaer-Logarithmic-Potentiometer-ZR-0003-92C-3/140942018492

Im not sure about the impedences these B&K units are designed for .



Maybe an even simpler solution would be a series of inline filters with set frequencies,  a bnc female to male adapter could be used to house the components ,no switches no pots ,a handfull of preset values might have you covered .We know that at certain whole mulptiples or divisors of the clock frequency the performance of this generator is better ,so why not create a set of filters that corresponds to these sweet spots on the 66/6800, ensuring best possible performance from each range .


Is there software that would help calculate component values for audio filters with defined requirements in terms of impedence and insertion loss in db etc? Ive seen a few of these type programs but never got the hang of them ,or they were not dedicated to audio specifically. Im also not MSoffice aware so I didnt try spread sheets

From what I can gather the 3491 is capable of swinging a handfull more volts at the output for a given rail-rail supply than 3091/95 , all the better if we need to drive passive filters with attendant losses . With the stock switcher psu and the stock output ic ,I see around 600millivolts drop across the rails at full output into 50 ohms both channels driven , thats pretty bad regulation by anyones standards ,simply bolting on the extra horsepower of the 3491 will cause an even larger modulation of the powersupply rails . Peak current of two 3491's is likely to be in the region of 1A and to get best performance your going to want good voltage regulation right through the current range .

I know someone who designed an attenuator specifically for sound measurement ,designed to be placed on the output of the soundcard ,therefore completely avoiding digital attenuation of any kind ,digital out is always hitting 0db fs ,you then have a coarse and fine resistor attenuation network say 5 coarse steps of 10db and fine scale of +/-5db ,using selected or hi-stability resistors very close tollerance attenuation can achieved over a wide range. I'll look up the post and make a link to it

[GregDunn]

That would be interesting, thanks!  There are some online filter calculators which work nicely; I've used a couple of them to design both active and passive crossovers for my stereo.

For what I'm doing, .03% is perfectly fine - I'm just using it as a quick performance check on amplifiers when I'm setting bias; and none of the amps requires an input voltage of more than 1.5V for full output, so attenuator losses could be pretty severe and not hurt the measurements.  But ultimately I'll need a separate box because all I did here was run directly into the HP331A and it has a full range attenuator on its input already.

[soundtec]

Thanks for testing the theory Greg ,my suspicions were correct ,
.03%, thats impressive , As you suggest thats an equivalent figure for a 12 bit system.  Maybe distortion in the 3002 op amp is the limit your seeing at 20 volts swing ,are you running standard parts,voltages in your unit ?If it is, my guess is the 3491's distortion should be an order of magnitude lower than what were seeing now and maybe the full 14 bit resolution is just achievable .

With the use of the afore mentioned attenuator ,mods and maybe JS's filtering idea its seems likely the 66/6800 could generate very high quality audio waveforms ,good enough for distortion measurements in other equipments . Its kind of interesting in that we spent a good deal of time thinking about ways of improving the signal quality  in the digital domain ,in fact the answer was make best use of the bits we have  .I wonder with a filter to supress the harmonics could we exceed the theoretical limits of the 14 bit convertors .

[GregDunn]

Yes, though that's at full output, 10 KHz only, it's a completely standard unit as far as I'm aware.  Latest firmware. 

Keep in mind that I don't have a super-quiet bench and that 331A is only spec'ed to a residual THD+N of 70 dB at audio frequencies, so I'm likely just measuring the meter's residuals here.  That's why super-low distortion is not a big concern to me at this time.  Now maybe when I find that minty Sound Technology 1700B at a hamfest... 

[bdivi]

Where did you buy in BG with free shipping, my suplyer of choice (TME) doesn't seem to stock it?

Direct order to TI - store.ti.com

[danco]

Thank you, I'll try it myself.

Where did you buy in BG with free shipping, my suplyer of choice (TME) doesn't seem to stock it?

Direct order to TI - store.ti.com

Sent from my MI 5s Plus using Tapatalk

[soundtec]

I see in the THS3491 specific resistor combinations are recomended for  gain setting. What is the gain in the original fy6600's final op amp output? ,surely if we want the display to show a meaningfull output voltage  the right resistor combination needs to be used to set the gain the same as the original op amp set up ,simply leaving the originals in place will no doubt  means the gain wont be a whole number  .It seems setting the gain higher impacts on the distortion produced by the output stage in any case .Maybe someone can test the output stage gain on a machine with 3491's installed , I have a vague recollection the original setting on the unit is x3 gain for the 3002 op amp .

The spec sheet for the 3491 shows Rg and Rf to be 200 and 798 ohms for a gain of 5 ,2.1k and 2.1k gives a gain of 2 .
On the FY66/68 we have 499 and 1.5k in these positions . Tables are available for each op amp giving the optimum resistor values and gain .

[soundtec]

THS3491:

Gain           rg            rf   
   2             2.1k      2.1k
   5             200       798 ohms
  10            78.7      704 ohms

THS3095:
Gain           rg            rf
   1              -             1.78k
   2            1.21k        1.21k
   5             249          1k     

THS3002
Gain          rg             rf
   1                            1k
   2                           680ohm
   5                           560ohm

So with 1.5k rf on the unit, a gain of less than 1 would be expected  ,3095 users would have a gain somwhere around 1.5 and 3491 would have a gain of around 3 .

So I setup the generators for 5.1volts 1khz sine,this should be 1.803  vrms

Fy6800(standard 3002 output stage)
Ch-1  1.7966 vrms
Ch-2  1.7815 vrms
Fy6600 (3095 x2,same Rf Rg)
Ch-1  1.7750 vrms
Ch-2  1.7735 vrms
at 0.51 volts (opa-686)
Ch-1  178.53mVrms
Ch-2  179.45mVrms

So not a big descrepency between the two opamps . I set the 6600 to .51volts output and used trimmers W3 and W5 to bring my output levels on both channels  to .1803 volts.
Now when I engage the 3095 at 5.1volts I measure 1.7935  on channel one and 1.7822 volts on the other channel so 10mV low on channel one and 21mV low on ch2 ,maybe its normal for the opamps to differ to that degree. I might check tomorrow  if the tollerances of the impedence/feedback setting components on the output play a part in the difference between channels .

[soundtec]

SO a couple of ohms difference between the main 50 ohm output series resistor after the 3095 ,into 1Mohm meter input I cant see this making much or any difference, the output attenuation network is matched very precisely between channels  , so maybe theres either a difference between op amps or a slight variation on the feedback resistances between stages . Maybe a small gain control preset pot on the final opamp  output stage feedback network in order to match the gains exactly might be a good plan for the next revision of the machine from  Feeltech ,it would also facillitate setting up alternative op amps.It would certainly allow for much more precise calibration of the output amplitude at higher levels . As we stand all the amplitude decimal places below .00 are worthless ,an ability to calibrate the outputs would make lots of sense for the precision of the machine .The way things are you can calibrate either low or high range output precisely ,but once you change to the other range  its no longer calibrated

Thanks for the link to Robert Cordell JS ,bound to be some good stuff in there ,just downloaded and getting stuck in ,

[DPA31]

Soundtech, If I may help :

 The gain of the output stage is 4
It is  G = 1+ rf/rg so G = 1 + R63/R45 on the original board gives 1 + 3

The difference you measure comes from resistors. But no need to adjust it alone as all the chain gain can adjusted by W05(CH1) and W03 (CH2).

I have replaced trimmers with Bourns SMD 1K 11turns 3224W-&-102E found on eBay
W26 : DC_AMPL_CH1_OFFSET
W05 : DC_AMPL_CH1_GAIN
W16 : DC_AMPL_CH2_OFFSET
W03 : DC_AMPL_CH2_GAIN
W01 : DC_OFFSET_CH2_CAL
W02 : DC_OFFSET_CH2_CAL

Dom.

[soundtec]

Hi Dpa31,

If you look at the schematic the gain control pot is in the feedback loop of the OPA686 ,when I calibrate the output below 5 volts with W3 and W5 I can get it very close ,but as theres no fine tune of the final output IC's your either calibrated to high or low level signals ,but it cant be right at the same time on both .

The idea of using ten turn pots came to mind with me a good while ago ,Im glad someone did it ,I'd imagine its a lot less fiddly now to set things up accurately .Was there any trouble fitting the ten turn pots? ,I didnt know there were smd types of ten turn pot available . Are they pin compatible with the existing presets?



 

[DaveR]

Soundtec - are you getting the values of R45 and R54 from the schematic, which shows them as 499 ohms?  On my two boards, and in the high res photos posted in May, the values are 68 ohms, which would give a gain of 23, not 4.  Either the schematic values are wrong, or there was a redesign carried out by Feeltech in late 2017.  I was thinking I might change R74 and R83 to 1k to give a more conservative gain of around 16 if I put 3491s in (or do one channel to see if there's any real difference).

Dave

[soundtec]

I didnt measure them on my board ,but just relied on the schematic ,

with either standard output stage or upgraded op amps Im getting more or less the same from both units within 10 to 20 mV, . A buddy just explained to me the gain is very little to do with the opamps themselves and more to do with the proportions of the feedback and other resistor than goes to ground . Id still like to see a pot on the feedback path to the final op amp ,just for fine tuning .

[DPA31]

Hi Sountech, Hi Dave,
 
I verified R45 & R54 resistors on the SB: they are coded 68A that is 499 ohms.
just in case : http://www.resistorguide.com/resistor-smd-code/

@Sountech

 Yes trimmer footprints are compatible with the pads on the PCB. see attached picture.

Can you describe the measure you would like me to do : 5V sine both channels, frequency ? You measure mV on both channels with ?

Cheers.
Dom.

[soundtec]

They fit in nice Dom , I'll have to do that mod myself for extra tweekage.

What you could try is dial in 5.1 p-p (sine)on the unit ,now with your RMS multimeter measure the output voltage , it should  read approx 1.803 v rms.
Now if you repeat the same test but with an output voltage of .51 volts ,this will mean only the opa686 is engaged ,now measure again your rms volts at the output ,you should be seeing .1803 volts rms . On the lower output range you could tweek the adjust pots on the opa686  to get exactly .1803 volts ac ,but this inevitably means your high range (over 5 volts output) will be off by a margin . Having adjust pots on the final high level output stage would allow an extra measure of accuracy to be tweeked in . So it looks like our feedback network on the high level output stage uses 1% tollerance components ,maybe upping these to .1% would be a worthwhile exercise .

[DaveR]

Thanks for that info and link, Dom - I thought it was 68R I was seeing, so just assumed it was 68 ohms.  You live and learn!! 

Regards,
Dave

[DPA31]

I got it  , Thks Soundtec.

I forgot the switch relay effect on range over 5V

@ 50Hz I measure 0.181 Vrms for 0.5 Vpp  but 0.177 Vrms for 0.51Vpp
                           1.769 Vrms  for    5 Vpp but 1.766 Vrms for  5.1 Vpp

I envisage to put a  piggyback SMD resistor in parallel either on R63 & R74 (1.5 K) or R45/R54 to adjust THS3091 outpout stage gain. But it will not correct the DAC non-linearity. So the question is at what frequency and what amplitude to adjust the final stage ?

BTW I 've got an adjustment range of 24 mV for half a turn on the gain trimpot on a 5V amplitude signal. I was able to adjust equally the gain of both channel.

Another point : what is the purpose of DC_OFFSET_Channel_CAL pot ? Does anyone know or test it ?

Cheers.

Dom.

[DaveR]

Hi Dom,

As far as I could tell when I adjusted mine a few weeks ago, the DC_OFFSET_Channel_CAL pot adjusts the vertical symmetry of the sine or square waves.  I used it in conjunction with the GAIN pot to get the waveforms on both channels identical by overlaying them on the scope.  Only the top half of the waveform was affected by adjusting it, and its range was quite small, but it served its purpose (if that's what it's meant for).  The name on the schematic is rather cryptic, but it's safe to give the pot a wiggle and watch what happens.

Regards,
Dave

[soundtec]

Ok thats all good information to get in ,
Word from the mouth of someone with 60 years experience in electronics is you use finer tollerance resistors to make up the potential divider across the output IC . Each IC has a range of specific resistor values ,not simply any resistors of a certain ratio,that ensure best performance ,generally between gains of about 2 and 5 best performance is had . 

When the drive to the 3095 is engaged the signal is fed from the opa686 outputs , my guess is theres an offset built into the software calibration which tries to make up for the slight gain differences as you switch between high and low range ,if we have calibration 'flat' and can adjust amplitudes independantly between opa and ths op amps externally we could essentially do away with using some calibration features in the digital domain ,maybe this is the issue as when you switched to 5.1 volts the output level actually went down DPA31.

So were essentially getting 12 bit performance , good op amp technique should be able to squeeze those last two bits out easy ,better sorting the amplitude differences between ranges in the analog domain as much as possible and not relying on bit munching internal calibration .

Theres a few nice theories for you to mull over in any case ,

Your idea to piggy back two resistors has merrit too Dom , if you picked x2 the required values ,then selected from a bunch of 0.1% tollerance in both rg and rf , you could get the resolution of the display to read bang-on to better than what we see now , if the range was a few mV off you could simply select from your bag of components a marginally lower value for each channel individually until its right 'on the money' ,in the words of Zappa .

Kindest regards Gents ,
Its Friday and the beer is flowing ,I hope your taking a dose of your fav poison and kicking back like me ........



as is as we saw its maybe a few tens of mV off across the ranges at a volt or two amplitude .

[DPA31]

Hi,

BTW, another mean for adjusting the gain had been described before in the blog:

"With regard to the circuit FY6600_Main_A0.pdf  at the THS3002 amp, the gain figure is 3.986 with a 10K input termination resistor and 4.006 without this resistor; changing this resistor to 33K results in a 3.999951 gain factor which is very close to a perfect 4."

These 10K are R44 & R53.
They act as a voltage divider on the THS 3095 input. The effect is very light.

Dom.

[soundtec]

Yeah I can see the effect your talking about alright Dom . Thing is,  raising the input impedence of the 3002 wiring and all its associated components to 33k will most likely effect the high frequency performance to some degree ,higher load probably does reduce harmonic distortion from the opa686 to some degree but the effect of any stray capacitance in the wiring etc could be magnified .

My guess is pots over time and temperature are subject drift , probably a lot more than a 0.1% resistor , if we use the doubling up or paralelling idea for the resistors we might be able to select much much better than 0.1% tollerance .The two 100 ohm resistors in parralel at the output of the 3002 are pretty low spec ,into high impedence like a meter a few ohms difference here wont make any odds ,but say if your driving 50 ohms ,or the relay clicks in the output attenuator ,any minor discrepency in the 50 OHm value will cause an error in the values displayed .

the use of 499ohms and 1.5k seems like a reasonable compromise value that works with most op amps  ,but I wonder to get the very best performance  in 3095 or 3491 op amps do we need other values with a 3:1 ratio .

What tollerance are you guys seeing on the 499, 1.5k ,10kand 50(2x100ohm)  parts ?