Review: Hantek DDS 3X25. Anyone own one?

[Mechatrommer]

back to hardware review. this thing is noisier than i thought, about 0.1Vpp noise, in peak detect mode.

[clonecrp]

Mechatrommer:

Thanks for clarification on a number of 3X25 points. Found your thread Thursday ... enlightning to say the least. GREAT Job !

Just received a 3X25 Friday but I haven't hooked-up yet. Will you be changing your software to accomodate other 3X25 functions?

Thanks!

Doug

[Mechatrommer]

...Will you be changing your software to accomodate other 3X25 functions?

such as?... give me good reason, then i will welcome to the club.
ps: i got suggestions earlier, but i will find a time for it.

[Bored@Work]

...Will you be changing your software to accomodate other 3X25 functions?

such as?... give me good reason, then i will welcome to the club.
ps: i got suggestions earlier, but i will find a time for it.

You could release the source code if you don't want to work on it. Or you could sit on it and continue gloating how great you are.

[Mechatrommer]

...Will you be changing your software to accomodate other 3X25 functions?

such as?... give me good reason, then i will welcome to the club.
ps: i got suggestions earlier, but i will find a time for it.

You could release the source code if you don't want to work on it. Or you could sit on it and continue gloating how great you are.

ask for it, i'll release it, including all the classes and the modules necessary. i can see vb6 is not everybody's interest nowadays, and here is aint a code monkey party. and many people lose track already, all they need is just the API functions in the dll (programmers guide in the installation cd). i can release full bloated version of it with all the features, if i really want to gloat, but i'm just a normal person, i need to find time for it. my primary intention is to demonstrate what bundled software cannot provide, just the basic stuff. but if i'm asked to do something bundled software already can do, or has workaround to do, then there's not much point imho. and i do things mainly influenced/based on interest, if you can persuade me to do something interesting, then i can put any busy'ness aside and jump in into the crazyness. if all you need is the source code, then ask again... explicitly, its no secret, if it is, you wont see a single thing from me, and i dont want to gloat and show on how good i code things.

[clonecrp]

Mechatrommer:

I for one appreciate your work. You have provided a great deal of information/methods with regard to the 3x25 which is a new type of device for me. Keep up the good work!

Other contributions on the subject matter are appreciated too.

I am curious however, if you have given any thought to the digital input/output part of the generator ?

Thanks!

Doug

[Mechatrommer]

I am curious however, if you have given any thought to the digital input/output part of the generator ?

i'd loved to, but i think:
1) i need to study more on that
2) i dont think i have necessarry equipment to test that

yes its pretty sad that for now i only know how to use 3 IO on the hantek (signal, synch, freq counter), there is more value that it can provide, such as those digital io, cascaded unit? trig in etc. but so far i have no idea, sorry. and for programmers outside there, below i provide the API list that dll's provide. and all you need is the bundled demo program to get started...

Code: [Select]

Declare Function DDSSearch Lib "DDS3X25Dll.dll" () As Integer
Declare Function DDSCheck Lib "DDS3X25Dll.dll" (ByVal index As Integer) As Integer
Declare Function DDSSetFrequency Lib "DDS3X25Dll.dll" (ByVal index As Integer, ByVal freqHantek As Double, wavePointNum As Long, wavePeriodNum As Long) As Boolean
Declare Function DDSDownload Lib "DDS3X25Dll.dll" (ByVal index As Integer, ByRef buf As Integer, ByVal number As Integer) As Boolean
Declare Function DDSResetCounter Lib "DDS3X25Dll.dll" (ByVal index As Integer) As Boolean
Declare Function DDSGetMeasure Lib "DDS3X25Dll.dll" (ByVal index As Integer, ByVal bFreq As Integer, ByRef Value As Double) As Boolean
Declare Function DDSSetTrigger Lib "DDS3X25Dll.dll" (ByVal index As Integer, ByVal bInter As Boolean, ByVal bEdge As Boolean) As Boolean
Declare Function DDSGetDigitalIn Lib "DDS3X25Dll.dll" (ByVal index As Integer, ByRef Value As Integer) As Boolean
Declare Function DDSSetDigitalOut Lib "DDS3X25Dll.dll" (ByVal index As Integer, ByVal Value As Integer) As Boolean
Declare Function DDSSetDIOMode Lib "DDS3X25Dll.dll" (ByVal index As Integer, ByVal mode As Boolean) As Boolean
Declare Function DDSSetPowerOnOutput Lib "DDS3X25Dll.dll" (ByVal index As Integer, ByVal bOpen As Boolean) As Boolean
Declare Function DDSSetSingleWave Lib "DDS3X25Dll.dll" (ByVal index As Integer, ByVal bSingle As Boolean) As Boolean
12 API calls only. i wish i have more.

[clonecrp]

Mechatrommer:

No programmer here so this stuff is Greek to me...

I have a question though. I read your program directions panel for "Angle Phase Calibration"

Is it correct to say that the calibration process works for:

1. Sign
2. Square
3. Triangle
4. Sawtooth

collectively  or is a cal.csv needed for each type of wave?

But I think you mean:

1. Sign => auto
2. Square => Auto => Reconstruct
3. Triangle => Auto => Reconstruct
4. Sawtooth => Auto => Reconstruct
 
Where the resultant cal.csv is calibration file for all Sign, Square, Triangle and Sawtooth then cal.csv moved to application directory.

Is this correct  ?

Thanks in advance !

Doug

[Mechatrommer]

no. you choose one signal, then it will create one single calib.csv only, and then you'll use it for any other type of signal later on (by loading it from app path/calib.csv). i suggest you use sine during calibration since its mostly used, i only provide flexibility, say if you want to work mostly on square, then you may want to calibrate using square signal. but if you look carefully, the method is not so perfect, esp at high frequency. if you want different signal with different calibration file, then you should save each files yourself. to load it, rename to calib.csv and put in application folder, open the application and it will be loaded.

about the auto/manual button. AUTO will calibrate your device at predefined frequency set in auto.csv in app folder. MANUAL will calibrate your device based on frequency (min, max, step) that you put in application textboxes (the same textboxes to be used as sweep function).

reconstruct button... say if you did manual calibration at frequency 100-1000Hz, and later on decided to calibrate at 1000-1,000,000 Hz, the older (100-1000Hz) calib.csv will be overwritten, but not the raw files (many files you can check in My Hantek Files\Calibration). now you got new calib.csv for 1000-1MHz only, to combine all the 100-1000 and 1000-1MHz raw files into one single calib.csv, then you should press reconstruct button, when its finish, you'll got a calib.csv for 100-1MHz.

sorry its not very clear, i was lazy to type a dedicated help file for that particular function. there only a few of us here, so i guess they can read it here as well, so i guess my part of explaining it is done in this post

[Mechatrommer]

here is the latest version with the source code (if anyone interested, vb6 required to compile). just minor changes:
1) calibration should a bit faster and different calibration method for low frequency.
2) 0 button to stop 3x25 from sending any signal.
3) signal boost in synch option (should be better (max amplitude) for unstable (non synched) high frequency signal >75MHz)
4) you can place your mouse on top of the "synch option"'s checkboxes to see popup text on "pro and con" of the option

[clonecrp]

Nice .... ... I need to play with this some more !

Doug

[clonecrp]

Mechatrommer:

Is atattched more or less what you mean ... 

Doug ...

[Mechatrommer]

Mechatrommer:
Is atattched more or less what you mean ...  
Doug ...

exactly what we need. great job! i've changed the doc a bit, made a help button in the little software to automatically open it. i think i got last minute change during the prior version (calibration process only doing sine iirc), so i fixed and checked before attaching it (latest version) here. you may want to overwrite your existing *.exe (and add the readme.docx made by Doug) in the zip file. i also included the source that i've change. Cheers

[clonecrp]

Mechatrommer:

A couple of corrections that might be of interest ....

Doug

[seattle]

Hi, has anyone looked to see if the amplitude adjustment on the 3x25 is done via an analog gain stage, or if it's done via limiting steps in the output DAC? It seems like the noise floor comes up quite a bit as the output amplitude is reduced. On a 2Vpp sine at 500KHz, the noise floor down quite a ways in the audio region (at the limits of the 10-bit scope ADC), but as the 3x25 output is reduced to 40 mVpp, the noise appears to come up quite a bit (though I'm at the limits of the scope there it looks like).

Overall, however, 3x25 is a great deal and I'm really glad I found this thread. PS. Has anyone seen some USB flakiness? It can be fixed by unplugging and replugging. Just wondering if it's my machine or others have seen this.

[Mechatrommer]

Hi, has anyone looked to see if the amplitude adjustment on the 3x25 is done via an analog gain stage, or if it's done via limiting steps in the output DAC?

i think someone who are familiar will be able to tell from my teardown pictures earlier (relinked below). from my limited knowledge, i think its not by analog gain stage, its all comes from the 12bit pins of DAC. and as no info on the DAC model, last time i search to no success. hint... 200MHz clock 12 bit DAC, the closest i can find is ISL5861 210MSPS DAC Chip but looking at the pictures, i dont think it is.

Has anyone seen some USB flakiness? It can be fixed by unplugging and replugging. Just wondering if it's my machine or others have seen this.

yes its with my machine too. it hanged from reporting any value to pc software. un/re-plugging solve the issue. but its with my homemade software (api calls) i dont know if its happened with the bundled software, never experience that. so it could be the machine, or it could be the software.

thanx for reviewing and joining in seattle. i believe you have more capable machine of testing the noise floor than me. the spectrum on SA 3.4vpp.JPG looks sexy, do u mind to tell what machine it is?

[seattle]

it hanged from reporting any value to pc software. un/re-plugging solve the issue. but its with my homemade software (api calls) i dont know if its happened with the bundled software, never experience that. so it could be the machine, or it could be the software.

Yes, it does happen a fair bit with the bundled SW on my machine. So your SW is likely at the mercy of their libs in terms of stability

thanx for reviewing and joining in seattle. i believe you have more capable machine of testing the noise floor than me. the spectrum on SA 3.4vpp.JPG looks sexy, do u mind to tell what machine it is?

It's a chinese USB scope called QA100, dual 100Msps 10-bit adc with 32 channel logic analyzer and super deep memory.

[Mechatrommer]

it seems somebody got mixed up between RIGOL DG2041A and HANTEK DDS 3x25...
RIGOL DDS-3X25 video signal generator

[torch]

it seems somebody got mixed up between RIGOL DG2041A and HANTEK DDS 3x25...

I don't know much about these things, but neither of the two standalone units pictured made me think "cost-effective, USB function/arbitrary waveform signal generator (card) (virtual)." 

[torch]

So my DDS3x25 arrived today, and I fired it up to put it through it's paces.

First thing I found is that the manual is wrong -- at least for Vista 64. The drivers are buried in an archive on the CD and must be extracted manually -- the install program copies over the 32 bit drivers, not the 64 bit ones.

With that sorted, I fired up their program and made sure things were talking. Then I shut it down and fired up Mechatrommer's version. Might as well try out the auto configuration routine, right? So I dug out another bnc cable, two terminators and another USB cable, hooked it all up and let it go. (The little graphic was great -- makes the connections idiot proof -- I need things like that ;-) )

Once it was done, I started playing a bit. I noticed that the rise time on the trigger output was much quicker than the rise time on the square wave output when I cranked up the Rigol to 2ns. It also had a tiny ring that the output didn't have. That seemed odd, so I swapped terminators -- no difference. Then I swapped cables and the signal shapes swapped. On closer examination, the cable that came with the DDS3x25 was an RG58/U. The one I had in the drawer is an RG58/UA. I've never had either a signal generator before, nor a scope fast enough to see this level of detail, so maybe you all know about these differences already, but for any other newbies shopping for cables, the RG58/UA seems to be the better choice here.

Ok, I'm an idiot. It's not the cables, it's the function generator.

[Mechatrommer]

a picture will worth 1K of word. the trigger output (or should it be input? shouldnt it?) is something never been reviewed. and pls dont blame anybody (incl me) if you found out this unit is not for you. you've made the decision and all the good and bad that we have found already stated in this thread. and praise to master saturation for bringing us in here will be looking forward for your point of view (review) Cheers

[Lunat1c]

I just stumbled upon this thread while searching for a signal generator to start up my lab and I must say that this couldn't have happened at a better time!

You guys are awesome! Working together and with the help of Mecha you made this little baby worth owning Thanks a lot for your efforts!

I have gone through the thread and just to make sure, in a nutshell we could say that the upgrades that Mechatrommer has performed are through the software that one uses to interface with the signal generator, right?

From what I've been reading on DDSs, this piece of hardware seems to be the best bang for the buck AWG. The Instek SFG1003 provides signals up to 3MHz while this one has more features and has a higher frequency range for the same price. Please correct if I'm wrong as I'm a beginner when it comes to this stuff and wouldn't wanna be disappointed if i purchase this

Thanks

[Mechatrommer]

1 thing to add. if you rely on sweep function, then dds is not much of a wonder. there will be signal distortion when you try to change the signal (shape/amplitude OR frequency OR both, during transition only). i'm not sure whether it is a DDS weakness, or a "Cheapo" DDS weakness.

[Lunat1c]

oh. If the problem is during transition only I can't see it affecting me that much at this point. At least I hope not

One question, what's the maximum frequency of all waveforms with the latest modifications you have posted?

[Mechatrommer]

One question, what's the maximum frequency of all waveforms with the latest modifications you have posted?

if you are talking about all waveforms, the device's DAC sampling rate/clock need to be observed closely = 200MSa/s. lets take the worst case waveform ie square (max rise time), 2 adjacent samples from lo to hi will need 5ns to rise (ideally), so about how long it has to stay at that hi level (pulse width) will depend on your definition of what good a square signal is, i'm not sure the real figure/standard, but for me, maybe like... 1 portion rise, 8 portion hi, 1 portion fall and another 10 portion (fall,lo,rise). so altogether 20 points needed to generate 1 cycle of acceptable square (my standard). so calculate it 200MSAPS / 20SA = 10MHz. if you need higher quality square, you need to lower that figure. higher and higher frequency signal, they will all converge to resemble sine signal, ie poor square, poor triangle, poor arbitrary etc. But as for sine, it is usable up to 100MHz (2 points per cycle), but no accurate noise level has been figured at that highest frequency. the most accurate so far i think, is presented by seattle above, with his 10bits scope.