FDY22XX(S) budget DSS signal generator from Ebay

[Warhawk]

Hi everyone,
I could not google any detailed information about FDY22xx (s) very low cost signal generator from ebay so I started a new topic. I don't want to mix it up with other DSS topics.

Q:
Is there please anyone who already have an experience with this gen.

http://www.ebay.de/itm/12MHz-Dual-Channel-DDS-Function-Signal-Generator-Sine-Square-Wave-Sweep-Counter-/121384829210?pt=Mess_Pr%C3%BCftechnik&hash=item1c4319811a

?

I am looking for a low cost signal gen as a general tool for my lab. Besides standard functions I would like to have a square signal with duty cycle control.

Google keywords:
FY2208S , FY2210S , FY2212S

[BobCollins]

I came here to ask the same question. The 12 MHz version (FY2212S) is available on US ebay for $87.40.

http://www.ebay.com/itm/12MHz-Dual-Channel-DDS-Function-Signal-Generator-Sine-Square-Wave-Sweep-Counter-/221506536174?pt=LH_DefaultDomain_0&hash=item3392d182ee

It looks too good to be true, but the Chinese have created some good stuff at some pretty incredible price points.

Any thoughts?

[godFather89]

Why can it generate faster sine wave than square wave? I see no samples/sec specified.

[BobCollins]

Why can it generate faster sine wave than square wave? I see no samples/sec specified.

The description is awful—I assume that it is a poor translation. However, I did see, for the 12 MHz generator, that sine was to 12 MHz and square was to 4 MHz.

Here is an equally-poorly written manual I found in another ebay listing: www.divshare.com/download/25797925-ce2

Edit: I'm not sure why EEVblog is unhappy with the URL. ;-)

There are many listings on the US ebay site for all the versions of the FY22xxS from 2 MHz to 12 MHz.

[godFather89]

I know the Fourier theory.
I'm not sure how this things work but I was thinking it's basically a DAC. So if it's 25 Msps, it should be able to output 12.5 MHz square or a 2.5 MHz sine (depending on the accuracy you wanted).
Do this things generate the square wave by generating multiple sine waves and combining them together?

[mikeselectricstuff]

My prediction :
It will sort-of work OK, but the user interface will be a complete dog, and the mains adapter will be a death-trap, covered in fake approvals marks.

[Mr Simpleton]

It is harder to generate a square wave vs sine wave. As you do need faster slewrate you see a square wave turn into the shape of a sine, when the output start to reach its bandwith limit.

I know the Fourier theory.
I'm not sure how this things work but I was thinking it's basically a DAC. So if it's 25 Msps, it should be able to output 12.5 MHz square or a 2.5 MHz sine (depending on the accuracy you wanted).
Do this things generate the square wave by generating multiple sine waves and combining them together?

[Rufus]

I'm not sure how this things work but I was thinking it's basically a DAC. So if it's 25 Msps, it should be able to output 12.5 MHz square

It would be able to but what would a 12.49MHz square wave look like?

[Warhawk]

I came here to ask the same question. The 12 MHz version (FY2212S) is available on US ebay for $87.40.

http://www.ebay.com/itm/12MHz-Dual-Channel-DDS-Function-Signal-Generator-Sine-Square-Wave-Sweep-Counter-/221506536174?pt=LH_DefaultDomain_0&hash=item3392d182ee

It looks too good to be true, but the Chinese have created some good stuff at some pretty incredible price points.

Any thoughts?

It is good that you could have found this thread. The price is unbeatable, especially considering fact that you usually pay same for a similar DIY housing only. 
The English description is terrible even for me, especially when they use sometimes FDY... and sometimes FY... for the product number.

Maybe we should buy one together and send it to Dave for a review 

[Fungus]

I know the Fourier theory.
I'm not sure how this things work but I was thinking it's basically a DAC. So if it's 25 Msps, it should be able to output 12.5 MHz square or a 2.5 MHz sine (depending on the accuracy you wanted).

Sure... so long as you assume the wires going to the output connector have zero indictance/capacitance and won't round off the nice square corner.

Me? I wouldn't put any money on that...

[godFather89]

It is harder to generate a square wave vs sine wave. As you do need faster slewrate you see a square wave turn into the shape of a sine, when the output start to reach its bandwith limit.

Got it!

[Mark_O]

It may not be terrible, but I'd steer clear of it, simply because better is available for not much more money.

By better, I mean (for example), square wave rise times better than the (poor) 100 ns this unit offers.  How they can claim:

"Square-wave frequency range:  0.01Hz ~ 4MHz (FY2205S, FY2208S, FY2210S, FY2212S)"

is difficult to comprehend.  Just so you don't have to do the math, a 4 MHz square wave has a period of 250 ns, or 125 ns high, then 125 ns low.  Rise time is 80% (10-90) of the total amplitude value, and the slew-rate on this is so bad that by the time you hit Vmax (125 ns), you've hit the downward part of the slope.  So such "square" waves can't possibly be square at all.

Other cheapy DDS generators can do 20 ns or better rise times, for less than 100 bucks.

[edavid]

Why can it generate faster sine wave than square wave? I see no samples/sec specified.

This is not an AWG... they use a DDS chip like the AD9830 to generate the sine wave outputs, but those chips don't have square and triangle wave outputs, so they have to do something else.  That something else is not as fast as the DDS.

[alex.forencich]

I know the Fourier theory.
I'm not sure how this things work but I was thinking it's basically a DAC. So if it's 25 Msps, it should be able to output 12.5 MHz square or a 2.5 MHz sine (depending on the accuracy you wanted).
Do this things generate the square wave by generating multiple sine waves and combining them together?

They most likely have a low pass filter at around 12.5 MHz to limit the aliasing you will get out of the DAC.  So if you try to output a 12.5 MHz square wave, you will get a 12.5 MHz sinewave.  This is generally how these devices are built.  And that's presuming you would even be able to get a decent signal out at 12.5 MHz, as they probably really cheaped out on the front end. 

[g7psz]

Well, I am new to this site and have ordered a FDY22010S from Fleabay.

I will see what t is like but bought it particularly as it has two outputs which could be combined (hopefully) to use as a 2-tone generator for testing linear amplifiers. I would have preferred one to a higher frequency but will at least be able to use it up to the 40m (7mhz) amateur band.

Steve, G7PSZ

[tggzzz]

Almost certainly the square wave is not generated directly by the DDS. The DDS output will be filtered to remove sampling harmonics, and the sine wave will be fed to a comparator to produce a square wave. Have a look at the sample circuits in the data sheet for the AD9851 to get an of what's happening.

[dom0]

I know the Fourier theory.
I'm not sure how this things work but I was thinking it's basically a DAC. So if it's 25 Msps, it should be able to output 12.5 MHz square or a 2.5 MHz sine (depending on the accuracy you wanted).
Do this things generate the square wave by generating multiple sine waves and combining them together?

They most likely have a low pass filter at around 12.5 MHz to limit the aliasing you will get out of the DAC.  So if you try to output a 12.5 MHz square wave, you will get a 12.5 MHz sinewave.  This is generally how these devices are built.  And that's presuming you would even be able to get a decent signal out at 12.5 MHz, as they probably really cheaped out on the front end.

By the way, the MSPS (update rate) of the DAC in a DDS and the master clock (e.g. 25 MHz) are usually not the same and depends on output frequency and LUT size.

Cutoff of the anti-aliasing filter and sinc-rolloff both play an important role near master clock / 2... also I doubt the output amplifier is designed for the plentiful bandwidth you need for a faithful reproduction of a square wave at 10 MHz or so (independent of the fact that a DDS alone cannot generate a square wave past, say, mclk / 20 or so. But if you add a fast comparator, or route the LSB of the LUT index word (!= phase accumulator, usually) outside (the latter will cause minor to major aberrations in duty cycle, depending on output frequency) the DDS to the output amplifier... that's how some integrated DDS do it.

[g7psz]

Well, I have bought one and it arrived quite quickly. a 10mhz model from Ebay

I immediately put it on a spectrum analyser (HP8590A) for a quick assessment.

At 9mhz sine, the spurs were about 40db down. The resolution bandwidth only goes to 3khz so the close-in noise sidebands were not possible to see.

The output level is adjustable in 0.1v steps; when set to 0, 9mhz output was about 50db down on the 10v setting.

On a frequency counter it appeared to be about 1 in 10-5 off frequency but this is probably adjustable inside the box.

Then I had a brainwave. I got a SW receiver in SSB mode (actually a Clansman 320 manpack), tuned in the 9mhz (it sounded quite clean) and then put the speaker handset near the mike of my Ipad with an audio spectrum analyser running (free Spectraview). I chose this as it has a linear frequency scale as opposed to a log one of most of the audio analysers. It seemed to have an effective resolution bandwidth of as little as several hz and the audio signal came out as a single pixel-width vertical line! The close-in noise 30db down was produced by the noise of the receiver. The receiver also produced an audio second harmonic!

So it must use proper DDS chips, not a fast pic which would have to "dither". I am quite impressed for a two channel generator for £43.

I opened up the case (4 screws) and it is ok inside.  Lots of hot glue to hold things down and loads of empty space to put other things inside. There does indeed seem to be trimmers for output level (or waveform) trimming and I think for frequency trimming of 125mhz ref xtal. All the electronics is on a single, well-made circuit board.

I think the device is ok and wish I had got the 25mhz version (but it costs about £100).

Steve

[JohnnyBerg]

Could you post some pictures?
Perhaps a mini teardown, or some pictures from the inside?

[g7psz]

I have got it, a 10mhz version.
I posted my opinion of it on another thread
I do find this particular site quite hard work, but the best way to find my opinion is to search
for my handle (g7psz, my amateur call sign).

Steve

[Zarnywhoop]

I have also risked buying one of these and it arrived today. The user interface is indeed pretty non-intuitive, and despite repeated attempts, I've been unable to download the manual from the link posted earlier in this thread. Has anyone managed to get hold of it? If so I'd be eternally grateful - well, for a short while at least - if someone could send me a copy :-)

[Zarnywhoop]

This damn thing is making me feel very stupid. I just cannot figure out how to use its alleged sweep function.      Has anyone here managed to? If so, can I have some pointers please? I have the 12MHz dual channel version.

 

[ion54]

Through trial and error I managed to find the way of sweeping a frequency range.
Hit "menu" and then go "parm" few times until you are able to adjust the frequency. Set the frequency to the desired start value.
Hit "menu" and you'll see on the second row the star followed by the option save and, usually "begin freq". At that point press the ADJ knob (encoder push contact)
Go to first step and select a second frequency (your desired end). Repeat the procedure and when presented with the save option rotate ADJ knob to the right until you see "*save end freq". Press ADJ knob to store.
Press function button "sweep". You should see "*lin-sweep:stop" If you want linear sweep press ADJ knob. If you want log sweep rotate knob (left) and then press the know to initiate. Display changes to "*log(or lin)-sweep:run
The signal is generated on CH1.
I'm still trying to figure out other functions
BTW my scope tells me the rise and falling time are in the range of 30 to 35 ns when generating 5MHz square wave in a 50 Ohm resistive load.

Last moment news: I've got the Chinglish manual in pdf format. About 250 KB only, so easy to email.

[g7psz]

As a further update, I tested the 10mhz output against a decent frequency counter.  The counter had  eight digit display and I set it to over count, so it was reading to 0.1hz. They were 40hz out but after three hours it they had drifted together so the error was only 9.5hz. Switching off the frequency counter for five min and then back on again and the error was the same after a two minute settling period, so I am sure it was the function generator, not the counter which was drifting.

Well, an error of 1 in 10-6 is ok. The function generator was positively warm after four hours so I am more certain it was the cause of the drift. I can put up with it, now I know the error.

Whether I can ever be bothered to put a tiny stabilised heater on the back of the the small surface-mount 125mhz xtal in the function generator ..............

Steve, G7PSZ

[ion54]

Here is the Chinglish manual in pdf.

[Neganur]

pictures or it didn't happen

(please)

[ion54]

Something else that was sent to me by the vendor is a small piece of software that will allow you to setup the generator much faster using the USB connection. As the file is above the size allowed by the forum I decided to upload it to my shared drive:
https://drive.google.com/file/d/0BwoyXU6sOGKLSTdEYi1nLXg4QTg/view?usp=sharing

The rar file contains both the manual and the application.

[jadew]

It's most likely using the AD9831 and AD9834 (for the 25 MHz version).

They're okish, if it's properly filtered but the amplitude will probably vary a lot based on the set frequency.

Anyone care to post some pictures of the board?

[Zarnywhoop]

Thanks for the sweep instructions, and for the manual. All working now... 

[pedorie]

Hi,
I also got one. Feeltech FY2200S (10MHz)
but i have a problem with the offset voltage. See measurements below. Can someone please check if they have the same offset issue?

Set      Measured
0.0       0.001
0.1       0.209
0.2       0.314
-0.1      0.045   (positive?  must be negative)
-0.2      -0.059

By the way, the signals are made with 8 datalines from a FPGA (altera max II) these digital lines are fed to a R2R network..

There are some adjustment potmeters inside. found the ofset one and the anplitude one. but there is one more..

Thanks,
peter.
 

[plesecq]

Thanks for the software. I received my FY2200SP today but the CDROM was infected by a virus.
Thanks
Paul

[pascal_sweden]

Isn't this the same signal generator as the other thread which I started in May?

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

Of course it's the 24MHz version in my thread, but should be within same series, not?