Review: Hantek DDS 3X25. Anyone own one?

[saturation]

I concur with mecha, but to simplify it, for square waves, for a nice clean sharp edge,  I've got at least 10 MHz, and with sinusoids at least to 70 MHz, and useful up to its limit of 100 MHz.  This is the limit of what I can see myself, using my scope, not that its truly the limit of the 3x25.  I start to see edge roll off at about 20 MHz, where the harmonics are over 100 MHz, beyond my unmodded Rigol 1052e capabilities.

Compare to what you get with other function generators of 20 MHz, such as the Rigol 1022 at $400+, and you see you get a good bang from the lowly 3x25.

The cleanest waveforms for all types are in definitely in the 10 MHz and lower region.

So it gives good high frequency for general purpose use well matched to a home lab scope.

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?

[torch]

OK, pictures. Channel one is OUTPUT, Channel 2 is SYNC OUT.


5MHz


10MHz


25MHz

[saturation]

Thanks torch, looks pretty good given the rise time is 4.7ns on the sync signal.  I recall there is a filter on the V-out so it causes more rise time delays but your unit looks like its performing as it should. 

OK, pictures. Channel one is OUTPUT, Channel 2 is SYNC OUT.


5MHz


10MHz


25MHz

[Mechatrommer]

confirmed mine is the same as torch's. maybe for my definition a good square will be up to 2MHz or at most 3MHz. but thats subjective and depend on what you are doing. so far, i still use the sine signal the most, so no issue at higher freq.

[torch]

and pls dont blame anybody (incl me) if you found out this unit is not for you.

Absolutely not! Please don't misunderstand, I'm not saying anything of the sort! Thanks for all your work on this.

I'm new to all this and so I was not sure if this was a normal waveform distortion or not. I thought it might be abnormal because the SYNC OUT waveform was obviously much more square at higher frequencies than the OUTPUT waveform. I thought it should be the other way, if anything. But if this is normal and this unit does the job for so many who know more than I, then I'm happy with the deal -- as you pointed out, this is far cheaper than any comparable product.

BTW: I should mention that the sinusoidal waveform is very clean all the way up to 100mhz, and at that speed, the SYNC OUT is also sinusoidal.

[Mechatrommer]

BTW: I should mention that the sinusoidal waveform is very clean all the way up to 100mhz, and at that speed, the SYNC OUT is also sinusoidal.

its not the 100MHz SYNCH OUT is sinosuidal, it is still squared. its just the oscilloscope BW/sample rate limitation showing it as sinosuidal.

[alm]

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

The amplitude range is much more limited compared to most stand-alone function generators like the SFG-1003. Depending on how and why you work, the lack of front panel may also be an annoyance. I find that I like computer control for complicated stuff like arbitrary functions (which I use fairly rarely), but prefer real knobs and displays for simple stuff like generating a sine and adjusting frequency until I get the result I'm looking for.

its not the 100MHz SYNCH OUT is sinosuidal, it is still squared. its just the oscilloscope BW/sample rate limitation showing it as sinosuidal.

Is it just BW limitations in the scope, or also the generator? What's the rise time spec for the SYNC out?

[saturation]

If the current trend to iPad popularity continues, more widespread touch screens enabled laptops will be commonplace and be a better emulator for hardware knobs.  Although you can only move one knob at a time, and there is no physical knob sensation or feedback, its far faster than the PC keyboard or mouse to actuate controls.

[Mechatrommer]

Is it just BW limitations in the scope, or also the generator? What's the rise time spec for the SYNC out?

err, i was just assuming. since the synch out comes directly from spartan pin, so probably its digital instead of sine. gotta look the very detail of spartan datasheet for this.

[Lunat1c]

just to confirm that I'm not misinterpreting the spec sheet of this device... is the maximum amplitude really +-3.5V?

[saturation]

Yes, its built around LV CMOS.

just to confirm that I'm not misinterpreting the spec sheet of this device... is the maximum amplitude really +-3.5V?

[Mechatrommer]

is the maximum amplitude really +-3.5V?

not anymore, if you want to put some effort on it... preliminary: https://www.eevblog.com/forum/index.php?topic=3182.msg47122#msg47122

[saturation]

Wow, that's awesome.  Not sure why I never saw that reply, fantastic work mecha!  Will review it through.

is the maximum amplitude really +-3.5V?

not anymore, if you want to put some effort on it... preliminary: https://www.eevblog.com/forum/index.php?topic=3182.msg47122#msg47122

[Mechatrommer]

just by coincident, reply #161 and the link are about the same time posted this early morning.

[nubelube]

Is anyone else having problems with the SOFTWARE having a resource leak?  Specifically, it allocates handles and then doesn't release them.  Eventually (after an hour or so) the software crashes because it has exhausted the resources.  Using "Process Explorer" I have confirmed this on at least two computers (one XP and one Windows 7).  I contacted Hantek a few months ago but got no response.

My apologies if any of this is redundant -- I did a quick search but didn't find anything related.  This thread is impossibly long, but it appears to still be active (and bad software still seems on topic with whether to buy the device), so I decided to post in here.

p.s.  It has been a few months since I looked into it, but I know that the file handle issue is a problem with the software.  The DDS-3x25 is perfectly stable when controlled through LabView.  In fact, I made a half-assed attempt to write a LabView implementation of the native software, but gave up after a few hours of trying to figure out LabView hooks into native drivers.  The API is actually pretty damn simple, but I want the pretty pictures; my current workaround is to just start the software and then suspend it whenever I'm not actively making changes.

(Edited to remove a paragraph about bad download on Hantek website -- I emailed them simultaneously with my post and they fixed it.)

[Mechatrommer]

i dont care much about resource leak. terminate the program will terminate the leaks too, i believe its automatically handled in windows. i got software 3.2.1.7 here that i think downloaded from hantek web decompressed correctly, so i'm not sure with your.

while playing with opamp this few days, i got repeatable hanging in the software when i try to change the wave the 2nd time. connecting and reconnecting usb solve this. but then the hang repeat again. i think this is because i'm feeding the buffer opamp thats the 3x25 feeding into, with noisy power supply. my suspect is when noise got into the unit, maybe the fpga will stop responding to usb.

while still in it. from my observation in previous post that the shape of synch signal will be affected by the where the output signal/phase is, there's some sort of cross talk between this 2 lines, maybe this way the noise will be able to sneek in directly to fpga. i think so.

[saturation]

I've not yet experienced any lockups or resource issues, but even so, once the Hantek receives its marching orders you can actually set it so it retains the commands and you can log off the client.

I haven't used mecha's software as extensively but its great, and any issues I find, just turning the software off then back on solves it.

Its not elegant, but it happens so infrequently, its the price to pay for a $160 100 MHz FG.

BTW, the thread is long because its has a lot of technical detail.  Its probably the most detailed dissection of the 3x25 there online and it was made to give the reader faith that this cheapie can do the job.

[Mechatrommer]

hi again. i've made profiling (flatness test) on hantek dds3x25 signal output. its "not so good" method measured with rigol ds1052e (100mhz hacked), open end (unterminated) measurement direct BNC connection from DUT/FG to DSO, max amplitude feed of ±3.5V sine wave from "virtual" pc controller panel. with 50 ohm terminated signal, please expect ~half the value. please note the sine signal will become distorted at 50MHz < freq < 100MHz, rendering the result is less reliable, esp somewhere 70MHz <  freq < 100MHz. if you know what i'm talking about. so, for what its worth, the picture and the xls file...

[saturation]

Thanks mecha for the data. If you're looking to do a VAC vs Hz graph is better to do it terminated with 50 ohms.  What you are looking for is a transfer of power to the load, and even if the amplitude is reduced, the power is constant.  So the roll off will begin much later; from your graph is looks like if you halve the amplitude, the roll off will begin at 40 MHz.

What you'd like in a signal source is as much flatness throughout the spectrum rather than maximum amplitude.  Without termination, at 1.5 Vpp you could get that from 0-100 MHz, which is fairly good.  That means all the harmonics generated are also flat, at least to the limits of full frequency response of the source and the capacity of your test gear to see it.  After all, the purpose of a FG is not to provide power output, but signal fidelity and wide bandwidth for testing.

For example, in the old analog days video signals are all under 2 Vp.

The limitation of the Hantek is if you were to use it as a test input or clock for 5V logic.  But unterminated, you could easily drive it to 20 MHz without the front end op amp you designed and posted on another thread.  

[Mechatrommer]

If you're looking to do a VAC vs Hz graph is better to do it terminated with 50 ohms.  What you are looking for is a transfer of power to the load, and even if the amplitude is reduced, the power is constant.  So the roll off will begin much later; from your graph is looks like if you halve the amplitude, the roll off will begin at 40 MHz.

a quick report on terminated reading... i told you so, you dont believe me! just halve the volt will ya? and how do we measure power anyway? i dont have gossen metrawatt?! as i said, its not so good, not really accurate, just a rough estimate. cheers!

i try to give raw data here. so later viewer can make the voltage division by themself, be it if they want to terminate it properly with 50ohm, 10ohm, 1Mohm whatever.

and even if the amplitude is reduced, the power is constant

sorry i'm not following this... how come the power is constant if the voltage is reduced?
if V is halved, so do I. so P should be quarter, not constant?

P = (I/2)^2 . R
P = I^2.R / 4

P is the function of I, and I is the function of V. how come P is constant if V is changed?

[saturation]

Hi mecha, I haven't had a chance to detail a reply but you are right and I am wrong.  I've made a correction to my post, hope it hasn't created too much confusion for you, my apologies!

[Mechatrommer]

you are busy man eh? thats alright i have something coming kinda busy but boring whatever! i really should have not reply this!

[Lawsen]

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

[rf-loop]

Some small measurement. DDS3x25 is original, not modified. (some day I will modify for cleaning signal... there is lot of noise, harmonics and also 200MHz harmonics. No, I really do not need myself this kind of toys but... still these may be very useful for some peoples. This is not bad if compare price and some special things. But software is really poor. Example freq adjustment... what hell they think. maybe this peole who write this soft have never use in real world any signal generator for any work. Hope hantek do small work for make this SW usability better. Also they may littlebit think signal cleaning. But it is not bad if look just maximum medium voltage level harmonics and non harmonics level. later if I have time I may add some test pics about signal quality. (harmonics level)
Well... frequency jitter is very strange... in some special cases.

But noise in this curve is not related DDS signal quality, it is just interfere from spectrum analyzer sweep and RBW and stepping DDS3x25). Stepping DDS with keyboard give lot of noise as DDS change frequency.

Spectrum input is of course 50ohm (it is extremely good "high-end" spectrum analyzer made by R&S)
I have adjust so that 500kHz (3x25 is nearly flat from zero to 3MHz) is exactly 0dBm on the spectrum display. (spectrum is adjusted so that one division is 2dBm (dB related to 1mW) and  freq scale is log. (spectrum own noise level is minimum around < -145dBm )



Do not think R&S RBW... selection is 2dB/div... so...

[Mechatrommer]

Some small measurement. DDS3x25 is original, not modified. (some day I will modify for cleaning signal...I have adjust so that 500kHz (3x25 is nearly flat from zero to 3MHz) is exactly 0dBm on the spectrum display...

very thank you rf-loop! i'm not good at power db, but comparing to my result, i think its closer. yours -3db@23MHz, -6db@36MHz. mine 20log(2.46/3.36)=-2.7db@23MHz, 20log(1.84/3.36)=-5.2db@36MHz. both plot showing flatness to 3MHz. your measurement should be more accurate since you use proper tool.

there is lot of noise, harmonics and also 200MHz harmonics... But software is really poor... peole who write this soft have never use in real world any signal generator for any work...Also they may littlebit think signal cleaning...lot of noise as DDS change frequency

it pretty much covered in earlier posts, except the noise thing, since now we have more advance measurement tool. thanks to you!

frequency jitter is very strange...Well... frequency jitter is very strange... in some special cases....

i think you are refering to "dajones glitch" and "leap year effect". page 2 and page 5

No, I really do not need myself this kind of toys but...

now thats an insult!

still these may be very useful for some peoples. This is not bad if compare price and some special things.

now its more relieving. its usefull to me no matter how noisy you say it is. i'm working on something right now yes the output has not been modified/filtered whatsoever. what you get is a raw output from an amplifier which get the signal from dac which get the signal from fpga, the simplest fpga based dds signal generator you can build i think.

later if I have time I may add some test pics about signal quality. (harmonics level)

i'll be looking toward this with interest.