FeelTech FY3224S 24MHz 2-Channel DDS AW Function Signal Generator

[bitseeker]

Mike,

Try this:

1. Turn off the function generator.
2. Press and hold "Parm" and the left arrow together, then turn on and continue to hold for two seconds.
3. Press the blue knob/button to save
4. Press "Parm" button

Does the displayed frequency now match the output frequency?

[mikeekim]

WOW really fast reply and worked perfectly!!!
Thx for your help.
Thx, thx, thx.

[bitseeker]

Glad to hear that fixed it. There's a firmware bug that several members were discussing on the prior page. It appears you triggered it on your generator, too.

Here's the original post where the calibration reset procedure was outlined:
https://www.eevblog.com/forum/testgear/feeltech-fy3224s-24mhz-2-channel-dds-aw-function-signal-generator/msg1135001/#msg1135001

[laseractive]

Dear forum,

A question about the amplitude range of these cheap signal generators:

If you want to use this kind of signal generator for testing transistor amplifier circuits, it should also support amplitudes in the mV range.



Do these cheap DDS signal generators support these tiny amplitudes?
If their minimum amplitude is 1V, this sort of renders them useless for these kind of lab experiments.
Are both channel outputs identical in terms of precision and min/max amplitude range?

Regards,
Laseractive

[bitseeker]

On my FY3224S, the settable resolution for amplitude is 10mV. It will output down that far, but a sine wave that small has a lot of distortion and the amplitude isn't super accurate. 1V will produce a much cleaner signal.

If you need a good, small sine wave with a cheap generator, select an amplitude that produces the best waveform and then put an attenuator on the output.

[HoracioDos]

Just wanted to show @laseractive what he can expect from these units. I'm not using a BNC Feed-Thru 50 ohm adapter. It's still in the mail.

[bitseeker]

At 2mV/div you can probably see the distortion better (and get more accuracy on the measurements).

[acropolis]

Many thanks richie, bitseeker and claytonedgeuk for the tips on recalbration. Was pulling my hair out trying to figure why the generator said it was putting out a 10khz signal but my scope was saying 11.3Khz. This tip worked brilliantly, thank you so much!

[bitseeker]

Glad that worked for you. It's interesting that there are more posts about this frequency problem lately.

[ruiseixas]

On my FY3224S, the settable resolution for amplitude is 10mV. It will output down that far, but a sine wave that small has a lot of distortion and the amplitude isn't super accurate. 1V will produce a much cleaner signal.

If you need a good, small sine wave with a cheap generator, select an amplitude that produces the best waveform and then put an attenuator on the output.

That is the problem with the FeelTech, at very low voltages (mV) you get too much noise, so the solution of the attenuator sounds perfect however I don't know exactly where and what attenuator should I buy!

BTW, does a Linear Power Supply solves the low voltage noise issue?

[bitseeker]

Typical RF attenuators should work fine as long as they cover the frequency range that you require. For example, the HP 355C and 355D attenuators work from DC to 1GHz. There are fixed (one attenuation amount) and adjustable kinds.

If buying used, be sure that they still work. Often, they've had too much power put through them and their internal components have been fried.

[fonograph]

Can it make modulation? AM  or PWM?

[bitseeker]

Can it make modulation? AM  or PWM?

See: https://www.eevblog.com/forum/testgear/feeltech-fy3224s-24mhz-2-channel-dds-aw-function-signal-generator/msg741894/#msg741894

[Tegra]

Here is my power cord mod to ground the unit. (Plugs into one of the TTL rear panel bncs)
($19.95 free shipping

Mine was missing the CD.
Is there a link to the software and manual?

  T

[stuartmp]

Hi all. I just purchased 5he FY3224S and I am wondering if its possible to set the frequency / period of the square wave to greater than 1 second for instance. Can I set it to a 5 second period for example?

[bitseeker]

I've never tried it. See if you're able to. It won't break.

[classical]

yes, I used it over days and days. Turn frequency to 0.2Hz for 5 Seconds period or to 0.1Hz for 5 seconds on time. 

[masterx81]

Hi! For anyone that is interesetd, i've made some measurements of the on and off resistance of the altera chip. For reduce the dac error first i've tried to increase r-2r values from 499/1000 to 1000/2000, but someone here pointed out that the accuracy eventually will be reduced by this mod.
So i've measured the resistances, and i've found that when warmed up has an Ron of near 9,53ohm, and Roff of 7.86 ohm. Average near 8.7 ohm (rounded 9 ohm)
I've also glued (with apposite themal conducting glue) a big dissipator over the altera chip for mantain the temperature as stable as possible.
So, remaining around the same 500/1000 orginal range we need or 491/1000ohm, 500/1018ohm or similar values.
Now i need to try to find some hign precision resistors with this values...

[nctnico]

If you really want to improve things then you better find a DAC which doesn't glitch when changing between codes and wire that in. The current solution they used depends on the assumption resistors next to eachother on the same reel have the same error. Getting better than 8 bit resolution from an R2R network is pushing the limits so don't bother going into that direction.

[masterx81]

With 1k/2k solution i've got an accettable error (the signal is linear and the glitch is almost inexistent), so if i try to retain the (almost) 499/1000 r-2r solution, but with resistors matched with the ron/roff of the altera chip, i think that i can get the best of the both worlds without having to wire in a better dac. For the amplitude i've already substituted the original hc4053 analog switches with the low ron version adg733 (with the supporting mods).
If i'll have to substitute the dac, i thing that is better to simply throw away the feeltech and buy a better signal generator. I want only to make it "acceptable"...

[Tegra]

I received no disk, but was able to find the software online.

However, the "FY3200 PC Contrl Software" shows a version "(c)? ??"

Does anyone have a newer version?

   Tom

[mcmusic]

... I am wondering if its possible to set the frequency / period of the square wave to greater than 1 second for instance. Can I set it to a 5 second period for example?

Hi stuartmp,

the other day I was playing with an Arduino and wrote a few lines to measure some square waves of the FY3224S in the low frequency region. I was surprised to see that at 1 Hz and below the FY3224S begins to develop a noticeable deviation to the theoretical values.

At 1Hz, the measured period is 1.0053 secs, which is 0.53% too long. In the slowest position of 0.01Hz the period should be 100 secs, I measured 178.95 secs instead, which is 79% (!) too long.
So be aware if you want to use the very slow waves for quantitative measurements.

For anyones convenience I attach the results with other low frequencies.

Greetings from Germany
mcmusic

[rbm]

Sorry to bring up the power supply issue but just wanted a quick verification of my plan to modify my FY3224.

At the end of the day I used:

• 2 pieces  PRINTTRAFO (PCB Transformer) Gerth 2X30V 5,6VA #V050561 @ Voelkner.de (cheaper (5,47 EUR each) and I have shipping flat rate) or

I'm planning on using a single Hammond 160H120, [120V C.T., 200mA 24VA], transformer inside the enclosure to provide the same isolation as "classical" demonstrated.  The local electronics supply can order in this part but they have stock on a Hammond 160G120, [120V C.T. 85mA 10.2VA]. Is this smaller transformer sufficiently sized as to not overheat in the enclosed space or should I actually purchase and install the larger transformer?  My gut says it might be OK since the capacity is only slightly smaller than "classical's" transformers.

[classical]

I can report that I had the signal generator for several weeks in continuous operation (but with a high impedance load) without any thermal issue. I can check the real power consumption in W and VA under 50 Ohm load on both channels in a few hours.

[classical]

Under full load (both channels 20V@24MHz loaded by 50 Ohms) mine draws 8.5W // 14.2VA on the primary side. As far as I know the transformers are rated regarding the secondary side.