Review: Hantek DDS 3X25. Anyone own one?

[Mechatrommer]

if you have the data in text form (csv,xls,txt) i hope you can attach here rf-loop, or bigger picture of the frequency plot.

[rf-loop]

if you have the data in text form (csv,xls,txt) i hope you can attach here rf-loop, or bigger picture of the frequency plot.

Sorry I have not data string about measurement. I have take this picture with this resolution and I do not want do this work all agen, it is not as fifteen minute work. This picture have done with J.Miles HP7470A (HP-GL/2) plotter emulator (http://www.thegleam.com/) and nice Prologix USB-GPIB  (HP-IB)  adapter..

I have not programming manual for this Rohde&Schwarz spectrum analyzer so I can not send command what return trace data string via GP-IB.

[saturation]

Thanks rfloop for the good bode plot.  If you could, a real good help would be a harmonic content of its sine wave output; what's the SNR?

At worst,  could you just check 20 Hz, 2000 Hz, and 20,000 Hz for the audio spectrum to cover bass, mid and high frequencies to see what harmonics appear and what amplitude it is?

If your devices can sweep to the high MHz and just get a cumulative harmonic content that would be perfect!

if you have the data in text form (csv,xls,txt) i hope you can attach here rf-loop, or bigger picture of the frequency plot.

Sorry I have not data string about measurement. I have take this picture with this resolution and I do not want do this work all agen, it is not as fifteen minute work. This picture have done with J.Miles HP7470A (HP-GL/2) plotter emulator (http://www.thegleam.com/) and nice Prologix USB-GPIB  (HP-IB)  adapter..

I have not programming manual for this Rohde&Schwarz spectrum analyzer so I can not send command what return trace data string via GP-IB.

[kx5]

Many thanks to Mechatrommer and the rest of the guys for your support
It is a great update comparing it to the crappy - dysfunctional original software.

[Mechatrommer]

I have only experience with a Hantek 8060 five in one, Rigol function generator, Rigol 1052SE digital oscilloscope (similar the one you have pictured), and a single channel analog oscilloscope. 
Lawsen

i've been following rcgroup's thread on hantek 8060. the generators output spec is identical to this dds3x25 but i'm not sure if they use same circuit. but i already have dso,dmm,fg so i dont need 8060 anymore, but i believe its good for money. just if i dont have the rigol+uni-t+hantek 3x25, i would get this 8060. i'll get extra... an isolated (battery powered and handheld) dso/fg.

[marmad]

I've been reading with interest this thread because I've been thinking about buying either the Hantek DDS-3X25 or the Velleman PCGU 1000 - they are currently the best bang-for-the-buck AWGs available - and they cost close to the same amount from my location (~€102 vs ~€126).

The Velleman seems more stable, has less distortion, better resolution, galvanic isolation of the USB, and a bigger waveform sample depth than the Hantek (8KSa vs 4KSa) - but it's lacking the external trigger in, frequency counter, or ARB Express file compatibility - and it's waveform editing software is primitive.  The Hantek has a much higher frequency range, the aforementioned I/O and file compatibility, and it can work independently of the PC (which might be useful) - but it appears more buggy, and it seems to have a rather crappy set of programming functions and procedures.   It's important to me that I'll be able to write functional software for which ever I buy - I need to create Bode Plotter software with it and the DS1052E.

I guess I'm asking any owners of the DDS-3X25 in general what their current feelings are about the Hantek - and Mechatrommer in particular about the pains of programming it - e.g. it doesn't appear to even have functions for turning the DDS generator on and off - or for elegantly handling sweeps.

BTW, if any current owners of the Hantek have not downloaded Tektronix's free waveform editing software (http://www.tek.com/products/signal-generator/arbexpress.html) - you should get it immediately.  It's MUCH better than what Hantek provides you - and because Hantek uses Tektronix's ARB format - you can easily create waveforms in it (keeping the sample depth below 4096, Vpp below 3.5, and export as CSV) and read them directly into Hantek's software (and AWG).  It also comes with a great library of predefined waveforms as well.

Edit: The link to the actual download page of Tektronix's software: http://www2.tek.com/cmswpt/swdetails.lotr?ct=SW&cs=sap&ci=17700&lc=EN

[Mechatrommer]

I guess I'm asking any owners of the DDS-3X25 in general what their current feelings are about the Hantek - and Mechatrommer in particular about the pains of programming it - e.g. it doesn't appear to even have functions for turning the DDS generator on and off - or for elegantly handling sweeps.

the pain? the pain is in learning programming language and how much complexity of your software, nothing to do with the device. i programmed in VB for this particular application. turning the DDS generator on and off? with little hack in function calls and parameters, you can set DDS to off (no signal), its in my so called "Open Source" link earlier (Halt Button). elegantly handling sweeps? this is yet to be studied. my current method is not elegant due to no available device api function that do this specifically/directly. or maybe thats the limitation in software, hardware, fpga or the dds awg design that we cannot do much about it.

[marmad]

the pain is in learning programming language and how much complexity of your software, nothing to do with the device.

@Mechatrommer - well, learning a language is certainly the bulk of the work... but certain devices can be a headache when trying to get them to perform certain functions.  I program a lot of devices - many of them of my own design - and even some of them can be a pain in the ass. 

my current method is not elegant due to no available device api function that do this specifically/directly. or maybe thats the limitation in software, hardware, fpga or the dds awg design that we cannot do much about it.

Yes, I've studied the API calls - though I don't have a DDS-3X25 to try them on - and it seems that there's no way to know exactly where the FPGA is in the sample memory - which could lead to glitches when changing frequency.  The Velleman AWG has the procedure: SetSweep(freq1, freq2, ampl, offset, time, form) to avoid this problem.  But the English in the Hantek manual is terrible, so it's sometimes not easy to know exactly how a procedure works without a device to test it on.  But I'm studying your source code now for more info. 

Selamat malam  -  Mark

[saturation]

Some side comments.

The Hantek is disappearing, so get it while you can.  Last I checked gone are vendors from EU, America, and what was 2-3 pages of product listing is down to just 3 sellers, all now in China.

Since Velleman has more software tools listed, and they have been demonstrably good in responding to their product issues such as posted on eevblog,  your chances are good of getting tech support from them as you write your software; support is unlikely from Hantek, given too, the device is probably off production.

https://www.eevblog.com/forum/index.php?topic=2885.0

I've had trouble with ArbExpress; for a time Tek changed the arb file format many times but they didn't change the file extension.  So what version works and is compatible is unknown.  Its easier to use the CSV format.

The Velleman is 2 MHz, the Hantek is at 20 MHz.  See rf-loop post on the 'flatness' of the spectrum, suggests there are minimal sine harmonics to 3 MHz to be on par with the Velleman for say, audio work, to compare with Velleman, were a tests has it it clean to ~ -90dB.

http://forum.velleman.eu/viewtopic.php?f=11&t=2041

https://www.eevblog.com/forum/index.php?topic=2170.msg50564#msg50564

I've been reading with interest this thread because I've been thinking about buying either the Hantek DDS-3X25 or the Velleman PCGU 1000 - they are currently the best bang-for-the-buck AWGs available - and they cost close to the same amount from my location (~€102 vs ~€126).

The Velleman seems more stable, has less distortion, better resolution, galvanic isolation of the USB, and a bigger waveform sample depth than the Hantek (8KSa vs 4KSa) - but it's lacking the external trigger in, frequency counter, or ARB Express file compatibility - and it's waveform editing software is primitive.  The Hantek has a much higher frequency range, the aforementioned I/O and file compatibility, and it can work independently of the PC (which might be useful) - but it appears more buggy, and it seems to have a rather crappy set of programming functions and procedures.   It's important to me that I'll be able to write functional software for which ever I buy - I need to create Bode Plotter software with it and the DS1052E.

I guess I'm asking any owners of the DDS-3X25 in general what their current feelings are about the Hantek - and Mechatrommer in particular about the pains of programming it - e.g. it doesn't appear to even have functions for turning the DDS generator on and off - or for elegantly handling sweeps.

BTW, if any current owners of the Hantek have not downloaded Tektronix's free waveform editing software (http://www.tek.com/products/signal-generator/arbexpress.html) - you should get it immediately.  It's MUCH better than what Hantek provides you - and because Hantek uses Tektronix's ARB format - you can easily create waveforms in it (keeping the sample depth below 4096 - and the Vpp below 3.5) and read them directly into Hantek's software (and AWG).  It also comes with a great library of predefined waveforms as well.

Edit: The link to the actual download page of Tektronix's software: http://www2.tek.com/cmswpt/swdetails.lotr?ct=SW&cs=sap&ci=17700&lc=EN

[Mechatrommer]

but certain devices can be a headache when trying to get them to perform certain functions.  I program a lot of devices - many of them of my own design - and even some of them can be a pain in the ass. 

from my experience working with 3x25, its painless. every functions do what they suppose to do. its just tinkering on how to make the dds "OFF" in software, you need to find out the indirect way (several api calls), because there is no direct way. and there's no SetSweep function like you said, it will be alot more painless if it is.

about the code. maybe you need some effort digging the functions call in my code. the simpler explanation is to send "zero valued" (volt) data to DDSDownload and then call the DDSSetSingleWave (not compulsory). if you want something about pain. DDSSetSingleWave is the most painfull function to call, i dont know what it does, its just stop the device from outputting data (which i was expecting it to output a "one time/once" data from the name.

and my "not elegant" method in sweeping is by repeatedly calling DDSSetFrequency and DDSDownload, that also can be a pain, depending on your definition of "pain".

[marmad]

Thanks so much for your feedback, saturation!

The Hantek is disappearing, so get it while you can.  Last I checked gone are vendors from EU, America, and what was 2-3 pages of product listing is down to just 3 sellers, all now in China.

Yes, it appears it is... although I can still find one or two here in Europe.

Last I checked, ArbExpress 2.5 doesn't work; the reason is for a time Tek changed the arb file format many times but they didn't change the file extension.  So what format it really is remains unknown, and whether it will work on 2.6 ArbExpress is unknown too.  Its easier to use the CSV format.

Yes, I meant using the CSV format from ArbExpress (will edit the above message).

For practical electronic use, the Hantek is far more capable, but it has its blemishes, fixing it is what hacking is all about 

Your message has finally tipped me towards the Hantek -  I do love to hack around with things    Now I'm just waiting to hear from a vendor here in the EU if they will ship it excl. VAT - which would make it just a few bucks more than the Chinese dealers on eBay.

@Mechatrommer

DDSSetSingleWave is the most painfull function to call, i dont know what it does, its just stop the device from outputting data (which i was expecting it to output a "one time/once" data from the name.

Yes, it certainly seems from the manual that it should be used for single-shot waveforms.

repeatedly calling DDSSetFrequency and DDSDownload,

I guess there's not much more to work with than those two functions for wave generation.  One question, can you poll DDSCheck to see if the device has finished the last command?  I've been trying to make a more understandable list of the functions (see attached .docx).  Feel free to check it and correct me if I've got something wrong.

In any case, as you might have read above, I've decided to buy one myself, so I can help you with the software experimentation on it.  Maybe we can reverse engineer it 

[Mechatrommer]

your doc looks like description from manual, cant go wrong i think. there's no polling, you wait until the api function returns and thats it. completed or not, is not accesible from your software, thats the mercy of the driver and hardware/firmware implementation. different seller will try to add different flavour in their ads, you probably looking at the same item in your example above. i also cannot download ArbExpress, dont know why the site keep hanging. i'm looking forward to listen to your idea/comment when you get your 3x25. cheers.

[marmad]

completed or not, is not accesible from your software, thats the mercy of the driver and hardware/firmware implementation.

That's too bad - but I'll just have to make do with what's available.

i also cannot download ArbExpress,

I pm'ed you some other links to try - maybe it works from there.

i'm looking forward to listen to your idea/comment when you get your 3x25.

I'll let you know when I've got it warmed up 

[saturation]

I hope you get one marmad, and share your new developments with us.  If you are in school and have access to a good spectrum analyzer, I'd be curious to see how clean the sine waves are to 2 MHz, compared to the Velleman.  See the link I posted earlier about the tests on the Velleman forum.

[marmad]

@saturation

Sorry, I'm not in school - and I don't have access to any better analog testing equipment than you - having just replaced my 30 year old Tektronix analog scope (which finally stopped working) with the DS1052E.  But I was thinking of cracking it open and having a look inside for possible mods.

And I'm more than happy to share any software and source I write for the device.  Hopefully I can contribute to Mecha's work.

Mark

[Mechatrommer]

i'm working on a new UI for it. but its been halted for sometime now. i only finished the GUI, no code yet. thats going to be a hell for me

[marmad]

The Velleman is 2 MHz, the Hantek is at 20 MHz.  See rf-loop post on the 'flatness' of the spectrum, suggests there are minimal sine harmonics to 3 MHz to be on par with the Velleman for say, audio work, to compare with Velleman, were a tests has it it clean to ~ -90dB.

@saturation

Just a note: even though I'm tipping towards buying the Hantek (for the raw speed and wealth of other hardware that they jammed into their device), the comparison above is not quite fair.  From what I've read, Velleman is being conservative with their specs, applying them equally to all waveform outputs, including arbitrary - while it appears Hantek is being slightly optimistic in theirs.  The Velleman does clean simple waves (like the sine) to 5MHz and can do perfect sine sweeps up to 25MHz (half the 50MHz clock speed) from their delivered software.  So from what I gather, the Hantek has perhaps a 4-fold advantage in real speed comparisons with clean sine waves (5MHz to 20MHz), but less with complex waveforms since it has half the sample size (4KSa vs 8KSa in the Velleman).  The Velleman also has a switchable low-pass filter on the output which the Hantek is lacking - but I'm hoping to rectify that.

BTW, not trying to dis the DDS-3X25 - I'm on the path to buying it myself - but I just thought the Velleman deserved a more accurate comparison 

[saturation]

Hello marmad,

Yes, no worries, a critical eye always helps to identify more things to fix or exploit.  I was using their advertised frequency capability to compare both.  I agree with your comments.  Clearly both have pros and cons the more you dig into it.   For example, you could play with the sweep capacity to output long durations of a single frequencies far above 2 MHz to overcome its limit.

More info on the forums:

http://forum.velleman.eu/viewtopic.php?f=11&t=3267

http://forum.velleman.eu/viewtopic.php?f=11&t=6079

The Velleman is 2 MHz, the Hantek is at 20 MHz.  See rf-loop post on the 'flatness' of the spectrum, suggests there are minimal sine harmonics to 3 MHz to be on par with the Velleman for say, audio work, to compare with Velleman, were a tests has it it clean to ~ -90dB.

@saturation

Just a note: even though I'm tipping towards buying the Hantek (for the raw speed and wealth of other hardware that they jammed into their device), the comparison above is not quite fair.  From what I've read, Velleman is being conservative with their specs, applying them equally to all waveform outputs, including arbitrary - while it appears Hantek is being slightly optimistic in theirs.  The Velleman does clean simple waves (like the sine) to 5MHz and can do perfect sine sweeps up to 25MHz (half the 50MHz clock speed) from their delivered software.  So from what I gather, the Hantek has perhaps a 4-fold advantage in real speed comparisons with clean sine waves (5MHz to 20MHz), but less with complex waveforms since it has half the sample size (4KSa vs 8KSa in the Velleman).  The Velleman also has a switchable low-pass filter on the output which the Hantek is lacking - but I'm hoping to rectify that.

BTW, not trying to dis the DDS-3X25 - I'm on the path to buying it myself - but I just thought the Velleman deserved a more accurate comparison 

[Mechatrommer]

...but less with complex waveforms since it has half the sample size (4KSa vs 8KSa in the Velleman)...

i still dont have a need for complex arbitrary wave, sine is still the mostly i used. and i still dont have a clear clue on what application of deep memory for complex arbitrary is really essential, except maybe in a very specific (but seldom) app like medical device that read heartbeat or simulating sensor output etc. but i guess that will be seldom for me. anyway, the IO output (12bit) the 3x25 provided, i'm thinking a way if we can output digital data like out of it. my current imagination is an arbitrary (analog) wave editor and at the same time we can see the equivalent digital data, or vice versa, we specify the digital data (serial or parallel) and we can see the equivalent analog wave. well, thats something for me to play later.

[marmad]

the IO output (12bit) the 3x25 provided, i'm thinking a way if we can output digital data like out of it.

You can use Digital Output pin 11 (in generator mode it would reflect the the MSB of the DAC) as the Sync out.  If you generate full-scale simple waves (sine, square, triangle or saw) DO11 will be the same as a sync pulse out - although it is still just another unbuffered pin of the FPGA - although it might not suffer crosstalk problems with the Output since it's laid out farther away on the PCB.

Sample generator output image - 10 bits in image, but it's the same with 12 bits.

[alm]

What's the point of buying an arbitrary waveform generator for just the sines? A dedicated function generator or even frequency generator will have a cleaner output, better amplitude control and proper controls.

[Mechatrommer]

What's the point of buying an arbitrary waveform generator for just the sines? A dedicated function generator or even frequency generator will have a cleaner output, better amplitude control and proper controls.

i bought a sine generator up to 100MHz for $100++, and am not an audio people, and am not a true professional engineer. + other bonuses mentioned. - standalone"liness", lcd screen and low noise.

[marmad]

What's the point of buying an arbitrary waveform generator for just the sines? A dedicated function generator or even frequency generator will have a cleaner output, better amplitude control and proper controls.

True, but I'm not sure where you get a 20MHz function generator (with the added counter and digital IO stuff this thing has) for ~$150... and it's the proper controls that cost much of the money.  This has 6 digital inputs... for a quick and dirty FG with set frequencies, it would be easy to rig some controls to the front of the unit and write some simple control software - for example:
1 bit for push-button function select: alternate between sine, square, triangle, ramp, noise, am/fm, exponent, etc.
3 bits for range: 1Hz, 10Hz, 100Hz, 1kHz, 10kHz, 100kHz, 1MHz, 10MHz
2 bits for multiplier: x1, x2, x5 ,x7.5

I guess my point, as saturation pointed out, is that the DDS-3X25 is a cheap hacker's delight 

[marmad]

Just bought the DDS-3X25... now I just have to wait on the long trek of China Post 

[saturation]

Great! Welcome to the club.  Please give us feedback also on the buying process and what the package was like when you finally received it.  Enjoy.

Just bought the DDS-3X25... now I just have to wait on the long trek of China Post