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

[s1ig0]

Well, whatever about the level of inaccuracy, I'm considering it defective.  So, I'm now considering my options and it's not looking good.  I could try returning it to the chinese ebay retailer, but I could end up with more losses.  If I had to keep it, I might try a repair.  Anyone have any ideas about where to start looking?  I assume it's clocked by a crystal, so I could start there?  I'm guessing it also has some kind of frequency multiplier, maybe a PLL?  Of course, that's likely to be embedded in the fpga - hmmm.  Not good.

[cdev]

Compare the signal to WWV or similar using an oscilloscope?

[Kleinstein]

The frequency should not be off by more than about 0.05% - so if 5 % off there is something wrong. Usually also a DSO and and most DMMs should not be off so much. A non locking PLL could be a reason - as this is likely inside the FPGA there is little on can do to fix this. If you are lucky it could be something like a wrong crystal of maybe an missing / not connected (e.g. bad solder) decoupling cap that brings the PLL out of function.

For the distortion at high frequencies close to there upper limit, I am not surprised to see a lot of spurious frequencies. This is about normal for an DDS without a good reconstruction filter and poor DAC. I have even expected more harmonics.

[mcmusic]

Hi everyone,

eventually I have discovered a bug in the Manual Trigger mode of the FY3224S. In order to not overload this topic, I opened a new one on
https://www.eevblog.com/forum/testgear/feeltech-fy3224s-24mhz-function-signal-generator-manual-trigger-has-bugs/

Greetings,
mcmusic

[rbm]

How complex of a waveform can be downloaded to the memory of the FY3224S, given it 2048 byte limitation?  Would it be possible to simulate a signal such as the example shown below:

[HoracioDos]

Also, one of the previous posts mentioned a USB isolator (http://www.ebay.com/itm/NEW-USB-Isolator-Magnetic-Coupling-Isolation-ADUM4160-USB-to-USB-/141998972441?hash=item210fcc5e19:g:aVAAAOSwEa5XPrRb) as well.  Anyone tried this?  Would be simple way to fully isolate the device, albeit between the USB isolator and PSU isolator would almost double the unit's price.

Also found this one with a ADUM3160, It's quite small and cheaper but not full speed.
http://www.ebay.com/itm/1500V-USB-to-USB-Isolator-Board-Protection-Isolation-ADUM4160-ADUM3160-Module-/201771509748?hash=item2efa84e7f4:g:N8YAAOSwUKxYa8xz

[sata]

Guys, using a 220v/220v isolation transformer would fix the grounding issue? It would have the added bonus of being a floating generator.

[Kleinstein]

The isolation "problem" is due to the capacitor needed to ensure EMC. Like with many other switched mode supplies that don't use PE, they have two small class Y caps (e.g. 5 nF range) from both of the mains lines to the output ground.  These are more or less needed to prevent RF interference from the switcher. Just an isolation transformer would solve the problem, but is rather heavy. In this case the fully linear supply would be the more logical choice - if could be an old heavy wall wart and linear regulator in the case.
The original generator should be already floating, except the USB and the RF caps. The transformer will not isolate the USB of cause.

I think for the USB, there is an easier solution than doing it at USB level - AFAIK there is video showing the USB isolation inside together with a linear supply mod. AFAIR the USB goes through a UART to USB bridge anyway and isolation could thus be done at UART speed.

[richie]

Hi guys, just received my new FY3200 24MHz FG today and I've noticed that the frequency is off by approximately 5%.  So, when the FG is set to 1k the hardware counter on my DSO reads 952Hz and at 10k it reads 9.52k.  My DMM has a basic counter and it gives the same reading as the DSO.  There's no offset; function=sine.  Have I done something wrong?    Is this a common problem?

it seems that my unit has the same problem. I think that this is a weird software problem.

I got the thing from banggood and after struggling around with them 5 weeks to get appropriate support they told me an undocumented calibration function that resets the unit. Press the left-arrow and the param key together, switch the unit on, wait 2 secs. Then you come into a calibration menu. Don't change this value (write it down), press the blue turn knob. You can also monitor the Ch1 Signal to set a exact value. If you power cycle the unit, the produced frequency is right.

BUT my unit disadjust and produces the wrong frequency when I use the sweep-Function in a particular way ! In detail, I set the start frequency to 0Hz and store it with the blue knob, then I set the stop frequency to some value (1kHz for example), store it, set the time (3secs), store, let the sweep run for some cycles, press the Ch1-Button to switch to normal mode, switch the unit off, switch it on again: frequency is off 15% or like.

My unit has software version 3.5.

Can you confirm that your unit has exactly the same problem ?

[claytonedgeuk]

it seems that my unit has the same problem. I think that this is a weird software problem.

I got the thing from banggood and after struggling around with them 5 weeks to get appropriate support they told me an undocumented calibration function that resets the unit. Press the left-arrow and the param key together, switch the unit on, wait 2 secs. Then you come into a calibration menu. Don't change this value (write it down), press the blue turn knob. You can also monitor the Ch1 Signal to set a exact value. If you power cycle the unit, the produced frequency is right.

BUT my unit disadjust and produces the wrong frequency when I use the sweep-Function in a particular way ! In detail, I set the start frequency to 0Hz and store it with the blue knob, then I set the stop frequency to some value (1kHz for example), store it, set the time (3secs), store, let the sweep run for some cycles, press the Ch1-Button to switch to normal mode, switch the unit off, switch it on again: frequency is off 15% or like.

My unit has software version 3.5.

Can you confirm that your unit has exactly the same problem ?

Thanks for posting that.  My offset was more like 25% (125Hz reading was actually more like 100Hz) - but now aligns with the DSO and DMM.  Many thanks!

[richie]

Thanks for posting that.  My offset was more like 25% (125Hz reading was actually more like 100Hz) - but now aligns with the DSO and DMM.  Many thanks!

Does it deadjust again if you use the sweep function (like I described) ?

Mine does and this makes the reliability of the unit more than questionable.

[claytonedgeuk]

Thanks for posting that.  My offset was more like 25% (125Hz reading was actually more like 100Hz) - but now aligns with the DSO and DMM.  Many thanks!

Does it deadjust again if you use the sweep function (like I described) ?

Mine does and this makes the reliability of the unit more than questionable.

Hey RIchie...well I followed your test...and yes, does the same thing.  Even easier...it looks like it's tied to when you save the end freq parameter.
Looks like some overflow error somewhere...but can do this.
1.  Set frequency to 60Hz, 3.00v, Sine
2.  Press 'sys' button once (to select "save p_on freq")
3.  Press blue button to save it
4.  Press 'parm' to go back to main menu
5.  DMM reads 60Hz
6.  Power cycle reads 60Hz
7.  Press 'sweep' three times to select "Sweep end freq". 
8.  Set to "1000.00Hz"
9.  Press blue button to save
10.Power cycle
11.DMM now reads 52Hz
12.Power cycle and press "Parm" and left arrow together for two seconds while starting up.
13.Press blue button to save
14.Press "Parm" button
15.DMM now reads 60Hz

Seems to hold this setting as long as you don't save your end sweep frequency.    I think it's reliable enough....but a little on the buggy side.  For your info, if you change the end frequency value, you change the error offset in the process!

[hellzakk]

Hi all, yesterday i have replaced the original psu from my f3224s with one of the "Hiland DIY USB Boost Single Turn Dual Power Module Linear Regulators Multiple Output Power Kit" ( Input voltage: 5-24V DC, Output voltage: + 12V, -12V, + 5V, -5V , +3.3V, Output current: 300mA (per channel) ) like the one in this video

The unit can power on and the buzzer do the beep, but no interface on the screen and no response from any input.
After this i have reinstalled the original psu and measured the +-12V output pins with a result of about +-16V...
Any hint about this?

UPDATE: powered with 20V 1A and it's working, need to check with oscilloscope the waveforms (and study the theory behind )

[Tom620]

Today I received my FY3224S unit from bang good:

It's hw version 3.6 and the software on the cd is v3.0.

A sine wave of 1kHz shows on my Rigol DS1054Z as 1.0001kHz. Pretty spot on.

Edit: it's hw version 3.6

[innkeeper]

My version of using transformer isolation to reduce residual voltage and leakage current.
initialy i had 46vac to ground and 30 ua leakage current. now 0

The isolation transformer i used is a 50va (yes it's overkill) triad N-68X
you can source it from most of the usual places, mouser, newark, digikey and so on for about 17 usd.
I got one off ebay pretty cheap as a pull. It is "electrostaticly shielded" and i attached the shield to the ground of an EIC inlet filter.
i drilled 2 holes and bolted down the transformer using 10-32 nylon pan head screws and nylon bolt so no exposure of the transformer outside the case.
nibbling out the power opening I bolted in a filtered EIC plug i had in my parts drawer. delta electronics 06GEEG3E It can be had for 6 bucks at most places.
just for good measure i also added a rf choke to the dc lines. It is probably not needed but i had it on hand and threw it on.

here is some pics on how it turned out..the layoyut worked out very well, and very neat and it was a pretty quick install.
i was a bit worried about noise from the power supply having moved it near the pc board, but so far the output looks clean post install, well as clean as when i started anyway.

note: the center 'black' wire you see on the picture on the secondary is NOT a centertap, that is the electrostatic shield connector, it is usual it is white but i had to bodge on a wire as it was chopped short when i received it.

[bitseeker]

That turned out nicely, inkeeper.

[kpbosma]

I have just tried the 1 khz 2khz check on my FY3224S using my DS1054Z scope on CH1 and Ch2.  Observing the display in XY there is a slow drift in the lissajou display.  A complete cycle of the the difference takes just under 3 minutes (about 2 secs under).  Also the scope frequency meter (on CH1) seems to cycle between 999.999 hz to a minimum of 999.997 hz. Just to be sure I checked with both sources set to 1khz but one channel set to a 45 degree phase difference to make it easier to watch.  There was no noticeable drift under this circumstance although the frequency counter on the scope CH1 still showed a drift as mentioned earlier.

Interesting.

John

[mcmusic]

Hi guys,

for the workbench use of the FeelTech FY3224S I made a card for quickly looking up the available preset waveforms without looking into the manual. The "x" and "-" reflect that one can change the frequency of all PREx and ARBx waveforms, but not the duty cycle.

I have also made space for the arbitrary waveforms and the 20 memory spaces - just fill out by hand what you have stored under which ARB or Memory position, for example

Select.  NAME       f                duty   Voltage and offset
ADDR4  Square  1.024 kHz   30%   5V+2.5

I found it interesting to see that the 2 TTL connectors in the back put out not only the same frequency as the corresponding channel in the front, but they do have a variable duty cycle - an adjustment which seemingly did make no sense in case of a selected sine wave on CH1 or CH2, but the duty cycle entered here (as well as in all other waveforms) appears on TTL A (or B for CH2). So these TTL connectors are just great if you need a signal for any digital input. You do not have to pay attention to the amplitude and offset value if you adjust for a square wave on CH1/2, the TTL outputs deliver the adjusted frequency and duty cycle always with 0-3.3V, enough for both 3.3V and 5V designs.

I made the x-dimension of the card just wide enough that it can fit between the 2 front legs of the case (print it about 10cm wide). With adhesive film and 2 cardboard stripes I made a simplistic mount: I can slide the folded card (printed on stiff paper)  into this mount for fast storage and retrieval.

Maybe that is of help for some of you.

Greetings from Germany,
mcmusic

[pascal_sweden]

Very cool mod! German engineering and inventivity!

FeelTech ... Das Signal Generator

[bitseeker]

Great idea. Thank for sharing!

[HoracioDos]

Another successful modification. My unit is 25 Mhz with firmware 3.6V

a) C14 socket with ground
b) Fuse socket
c) Earth / floating ground switch

Thanks for this thread

[bitseeker]

Nice work, Horacio. Looks like it was made that way.

[derosoft]

HI!  I just got my brand new FY3224S signal generator - Can anyone explain to me the value or the use of a DC wave...?
Thank you!

Derosoft

[Savoyarde]

You can change the Offset and get different DC Values.

[mikeekim]

Hi all, i'm new in that forum and i need your help to solve a issue if is possible.

Recently i acquired a Fy3224s, but the signal generator have a issue.

Cannot adjust the signal generator in frequency, the screen show a frequency but the signal is more high.
Ever is 0.3x factor over that the screen show in hertzs.
For example, if i want to use 1000khz, then i need to put 700khz and then the signal is 1000khz in any signal.
If i use 2Mhz then i need to put 1.4Mhz.
And ever is the same, someone know how i can solve this problem?
See the images plz.
Thx for your help.