Hantek HDG2002B AWG: 5Mhz or 100MHz? Let's see!

[smgvbest]

anyone knows a way to comunicate this instrument with a c or python or another interface with a ubuntu 14.04 computer over usb?

Python
https://pyvisa.readthedocs.org/en/stable/
Perl
http://www.labmeasurement.de/
C and other languages
sreach for NI-VISA


that should get you goiing

[robert_]

Finally got around to finish my HDG2k "improvement" project:
Now with a safe PSU, isolation between output common and ground/USB and less loud, more effective cooling.
If there is any sort of interest, i could be motivated to publish the psu design files (not that great, as it was built around components i had around) and some infos on the isolation mod.

[lukier]

Finally got around to finish my HDG2k "improvement" project:
Now with a safe PSU, isolation between output common and ground/USB and less loud, more effective cooling.
If there is any sort of interest, i could be motivated to publish the psu design files (not that great, as it was built around components i had around) and some infos on the isolation mod.

Nice job. Sure, let us know the details.

Why this big toroidal transformer? Is that a linear supply for the analog part and the rest (black brick & misc) is a switcher for the digital? What kind of improvement did you get by using linear supply for the floating side?

[robert_]

Why the transformer: I cant get bothered to design a 5-output SMPS for a one-off project. It takes umm, a little while to get things like transfromer design, EMC etc sorted. I wanted something that works properly on the first PCB revision...
After all, i need floating +5V and +-18V for the output/analog side, and +5V (grounded) for the controller/usb side, plus 12V fan supply.
Using a transfromer to supply +-18/+12V via linear reg, +5vf via a non-isolated switcher and the other +5V via an isolated dc/dc module is an easy way to get it done.

Advantage: Standard PSU is unsafe POS (insuficient creepage distances within the transformer, if it breaks down instant line voltage at BNC connectors without tripping a fuse) that also happens to have 4.7nF between line and output common, which gives you nice ~ 115V AC at the BNC "ground" as the psu output is NOT grounded even though it really needs to be. Can deliver enough current to be felt and to destroy components.
The output "ground" is floating somewhere but as there is no isolation between it and the USB connector, connecting it to the PC now gives you a nice ground loop to somewhere else. Better than the previous situation but still not good as it will inject all sorts of noise, plus relying on the USB cable shield as "safety" grounding isnt exactly good.

Now i have a less dangerous PSU, safely grounded controller/USB side AND floating (but without significant leakage current!) output side to eliminate ground loops in my setup.

[pascal_sweden]

The Hantek HDG2000B series already had a 16-bit D/A converter, which Siglent now recently started using in the new SDG2000X series.

How does the Hantek HDG2000B series compare to the new Siglent SDG2000X series?
Anybody out here who has both units in his lab, for taking up a practical comparison?

The new Siglent SDG2000X series can handle a really low rise time. What is the lowest possible rise time on the Hantek HDG2000B series?

[Howardlong]

The new Siglent SDG2000X series can handle a really low rise time. What is the lowest possible rise time on the Hantek HDG2000B series?

I measure it at about 11ns on a square wave on an Agilent MSO7104B 1GHz scope. Amplitude doesn't have a big effect.

[pascal_sweden]

Good question actually:

When measuring the rise time on a signal, the main focus is the output of the signal generator, but if the scope is not meeting specs, it could contribute to a lower measured rise time, isn't it?

How to exclude the impact of improper performance of the scope from the actual measurement of the rise time with regards to the signal generator output?

[Howardlong]

Good question actually:

When measuring the rise time on a signal, the main focus is the output of the signal generator, but if the scope is not meeting specs, it could contribute to a lower measured rise time, isn't it?

How to exclude the impact of improper performance of the scope from the actual measurement of the rise time with regards to the signal generator output?

The rise time of that scope has been measured at about 330ps in real time. The system rise time is the square root of the sum of the squares of the individual rise times. The effect of the 330ps scope rise time is therefore negligible on the DUT rise time when the system rise time is 11ns.

[Mark_O]

Good question actually:

Silly question, actually.

How to exclude the impact of improper performance of the scope from the actual measurement of the rise time with regards to the signal generator output?

If the scope is performing improperly, you CAN"T exclude the impact on the measurements.     You switch to a working scope.

~~

Howard answered a completely different question, which is, how do you compensate for limitations in measurement in one instrument, when attempting to evaluate another instrument?  With the answer being that when there is more than an order of magnitude difference in the contribution from the two, the impact on the net result will be minimal.

[Berni]

Okay here are some performance tests of my Hantek using a 4GHz scope.

Small square waves get a 6.8 ns rise time on mine



Large square waves are a bit faster at 6ns but look at all that ringing



Then the akilies heel of DDS signal generators, an oddball non integer output frequency(4.95MHz in this case). Yeah tons of jitter due to being at only 250MS/s fixed.



Here is some more analysis on that. As you can see its actually outputting two frequencies that are close, having them average out to 4.95MHz in the long term



And here is a 19MHz Sine wave spectrum, there is jitter in there too as you can see by the two modulated side band tones on each side of the main sinewave.

[Howardlong]

Age is creeping up on me, I did some tests before here: https://www.eevblog.com/forum/testgear/hantek-hdg2002b-awg-5mhz-or-100mhz-let's-see!/msg759319/#msg759319

I didn't get any ringing. The firmware used is from the link mentioned here: https://www.eevblog.com/forum/testgear/hantek-hdg2002b-awg-5mhz-or-100mhz-let's-see!/msg753380/#msg753380

All below with 50 ohm termination, scope is Agilent/Keysight MSO7104B 1GHz BW scope at 4GSa/s. 10MHz square wave with different amplitudes to check output buffer slew rate. 8 averages taken.

100mVp-p (50mVp-p with 50 ohm termination)


500mVp-p (250mVp-p with 50 ohm termination)


1Vp-p (500mVp-p with 50 ohm termination)


10Vp-p (5Vp-p with 50 ohm termination)

[Berni]

Age is creeping up on me, I did some tests before here: https://www.eevblog.com/forum/testgear/hantek-hdg2002b-awg-5mhz-or-100mhz-let's-see!/msg759319/#msg759319

I didn't get any ringing. The firmware used is from the link mentioned here: https://www.eevblog.com/forum/testgear/hantek-hdg2002b-awg-5mhz-or-100mhz-let's-see!/msg753380/#msg753380

Hmm interesting. For you there seams to be a tad of undershoot before it settles in. Might be that there are some hardware changes going on. Would also explain why mine has almost twice as fast of a rise time. I am guessing there is a different value cap somewhere in the AFE.

[Howardlong]

Age is creeping up on me, I did some tests before here: https://www.eevblog.com/forum/testgear/hantek-hdg2002b-awg-5mhz-or-100mhz-let's-see!/msg759319/#msg759319

I didn't get any ringing. The firmware used is from the link mentioned here: https://www.eevblog.com/forum/testgear/hantek-hdg2002b-awg-5mhz-or-100mhz-let's-see!/msg753380/#msg753380

Hmm interesting. For you there seams to be a tad of undershoot before it settles in. Might be that there are some hardware changes going on. Would also explain why mine has almost twice as fast of a rise time. I am guessing there is a different value cap somewhere in the AFE.

For the avoidance of doubt, probing was with a direct BNC-BNC 50 ohm patch lead and using the scope's internal 50 ohm termination.

[Berni]

For the avoidance of doubt, probing was with a direct BNC-BNC 50 ohm patch lead and using the scope's internal 50 ohm termination.

Same here. 1m of BNC to BNC coax cable going in to a 50 Ohm terminated scope input from Ch1 of the signal gen.

So i guess these Hanteks might not be the most consistent. But the jitter tests should be valid for all of them since they don't have a fractional clock generator for its DAC

[jashugan]

Hello,
I made a complete calibration (channel 1 and 2), but I got at least two problems:
1) On channel 1, the calibration has been performed and gets good values up to 35 MHz, after which the waveform reduces its value. Very good DC values has been obtained. I can't reach nowhere near the 100 MHz, because after let's say 60 MHz the waveform is almost flat.
2) On channel 2, I have the same frequency response of channel 1, but I don't have good DC values. I can't calibrate them, and more, if I output let's say 5 V DC, the AWG outputs -5 V more or less.

Anyway, I modified the original script to work with Tektronix MDO3014 I have. I think it should work for all MDO3000 and MDO4000 series.

[Berni]

In the script i modified for my Agilent scope i made it auto range after each step so that i could accurately capture the waveforms when they get tiny. Worked pretty well in the end for me all the way to max freqency.

[jashugan]

Also this script auto ranges amplitude and frequency. It seems as if my Hantek can't output high values at high frequency. Could it be possible? I see that the amplitude begins to reduce at 35 MHz. Berni, which version do you have? I have PCB 1004, with the latest firmware available.

[Howardlong]

Also this script auto ranges amplitude and frequency. It seems as if my Hantek can't output high values at high frequency. Could it be possible? I see that the amplitude begins to reduce at 35 MHz. Berni, which version do you have? I have PCB 1004, with the latest firmware available.

You may notice that the three range relay steps kick in at different p-p voltages dependent on the frequency during the cal, I assume that the output buffers start slew rate limiting.

[50ohm]

Hi!

Just bought this (well actually it's HDG2032B) and I'm wondering is this "normal" or serious failure. When running self test it starts, generates few relay clicks and then reboots whole unit. No results shown. Software version is 1.00.3.02 (150325.0). Otherwise the unit seems to work quite normally (it's able to generate different waveforms), however I have not figured out how to generate own waveforms, tried with PC with no luck.. Also, it would be really nice to have the display settings permanent. Every time it's rebooted the screensaver is on and sceen is too bright.

I have to decide now to keep it or return it. So I'd really like to know if the self test issue is normal feature of the software or not..

[Freedom]

Hi,
I had the same problem with self test until I upgraded SW to 150807.0. Now I see after a self test the test result.
I didn't find a way to keep my display brightness setting

[50ohm]

Does anyone know what type of signal is suitable as external reference clock? It looks that Hantek HDG2032 does not lock properly to standard sinewave reference from rubidium master clock. If the output signal is measured with oscilloscope, it looks just fine. The frequency and phase are both locked to external reference just fine - according to oscilloscope...

But... when measuring the signal with spectrum analyzer there's huge difference between internal and external reference clocks. With exterenal clock there's wideband noise covering all over the band!

It looks that the standard 10 MHz reference signal is not enough for Hantek. It has output of +10 dBm. I was able to amplify this to about +20 dBm, which gave much better results with Hantek. At most of time the Hantek output was stable, but still there was moments with noisy output. So it might be that +20 dBm is just on the edge of working reference level.

But: +20 dB is very high level for reference signal! Really; it means 100 mW transmitter power, same as in WLAN! It's really very HUGE signal as clock reference. Too much. Something must be wrong with Hantek's reference input circuit.

The internal reference seems to be off about -200 Hz, when the unit is warmed up more than 30 minutes (normal warming period for all lab instruments). This was verified with 10 MHz output setting and from reference output jack at the backside. Both shows exactly same result with 10 MHz setting - like they should do.

With external reference we can have stable and correct frequency at the output - for the short moments when it's working without noise. At most of the time threre's wideband noise but even when it occurs the fundamental output frequency was correct according to specrtum analyzer's counter.

Here's my tests with 10 MHz/1V output setting. I also did some wideband noise/spurious measurements with 15 MHz setting. It seems that the output blocks are not properly shielded because the 10 MHz reference leakage is clearly visible at generator output. Also some other spurious frequencies and raising noise level around fundamental was detected. The 10 MHz leakage is clearly visible and verified with spectrum analyzer's counter and it's same as Hantek's erratic reference frequency. Also it's harmonic was visible on 20 MHz.

As for comparison I took a same measurement from Marconi 2022C RF generator ouput with 15 MHz / 0 dB settings with external reference. This Marconi is not even very high quality RF generator yet - but like we can see the Hantek's output is not very clean.

So there's some modification work to be done to get the external reference working and to shield the output blocks so that there's not so much spurious.

[markone]

I'm not following this precious 3ad for several months, now i'm the hurry to make some RF work and i have a straight question in order to decide what to do with this instrument (ie sell it for a siglent/rigol one) :

is the fault of uncalinbrated amplitude frequency sweep function solved by the last FW update and/or FW hack ?

Right now i'm away from home and i'm not able to test it with the last firmware revision.

Several months ago i put it straight in the box as soon as i discovered that freq. sweep function does not (did not, at least) take in account calibration data, ie if you sweep from 1 to 50 Mhz, the amplitude of the signal will decrease A LOT with frequency rising.

Any help will be appreciated !

[50ohm]

I'm not following this precious 3ad for several months, now i'm the hurry to make some RF work and i have a straight question in order to decide what to do with this instrument (ie sell it for a siglent/rigol one)

Here's 100 kHz...30 MHz frequency response plots from unmodified Hantek 2032B with 1V output setting. Firmware version is 1.00.3.03.05 (150807.0). Calibration data is also untouched, original data from factory.

Results are measured with spectrum analyzer using trace max hold function.

First attached image includes three traces:

Green one is what you asked about: it's frequency response of automatic sweep from 100 kHz to 30 MHz from 2032B.

In the blue trace there's manual sweep with 50 kHz steps from 2032B. If we compare these two traces, we can notice that calibration data is not applied with automatic frequency sweep. Total amplitude error between 1 MHz and 30 MHz with frequency sweep was measured with delta marker and it was 2,09 dB.

Please note the notch at 24,6 MHz on manual sweep trace. At this point, relay clicked inside Hantek. Maybe some kind of output filter was disabled and/or the output impedance was changed because the level increased about 0,3 dB there. This relay is not clicking in automatic frequency sweep mode, even with very slow sweeps like 120 seconds.

In the red trace there's output from Marconi 2022C RF generator with -1 dBm setting and manual frequency sweep with 50 kHz steps. Those little 0,1...0,2 dB notches in this (and other) traces are spectrum analyzer calibration errors (this graph has 1 dB vertical scale while it's normally 10 dB so the spectum analyzer's frequency vs. level calibration table errors become also visible in this measurement setup).

If you plan to do any RF work, please consider to buy RF generator instead of any function generator. At least with Hantek, possibly with other ones also you will face some serious issues if you try to do RF work with it:

- General spectral purity problems and huge reference/spurious frequency leakage (please see my previous post)

- Insufficient frequency area for any kind of modern RF tuning. Suitable only for IF blocks testing.

- There's no dBm setting for the output - and the voltage setting does not correspond to any real world dBm in the "50 ohm. world" anyway! It shows correct voltage levels only against oscilloscope's 1 megaohm input, when spectrum analyzer and all other RF instruments have 50 ohm. impedance.

- You cannot generate any faint signals. With minimun setting the output is 2 mV. This is shown as -56 dBm at spectrum analyzer. This is too loud signal for any kind of receiver tuning etc. You will need external step attenuator to generate needed levels. With RF generator, you can go beyoud -100...-130 dBm depending on model. For example, most receiver tuning/testing tasks require usually less than -100 dBm signals.

- There's huge leakage from generator output also when the output is turned off! Output relay does not cut RF signals. Please see the second measurement, it's time domain level analysis when the 30 MHz output is first on with 1V level, then output turned off. The output relay has only ~56 dB attenuation at 30 MHz, so in the RF world it doesn't cut the signal at all... If you want the generator to be silent you will also need to set the amplitude to the minimun value or turn of the whole generator. This leakage is frequency related; with 1 MHz the relay had 82 dB attenuation.

- Output connectors are BNC type, when all RF instruments have N type. If you use it with different kind of RF stuff you will need adapters.

I hope this helps.

[markone]

-snip
I hope this helps.

Sure it helps, many thanks for your post.

I agree with all your statements, for serious purpose we should start from something like Rigol DSG815.

For HF receivers sensitivity test i built a 60dB fixed attenuator to put in line with other fixed attenuators, nothing critical, just to get some rough estimations.

During this day i flashed the hantek with last FW (150807) and i get the very same result of the past, that is identical with yours.

I addition, now AM modulation function with internal mod. source is no more working properly, see the attached image 

The funny thing is that when i enable the function it works properly for a couple of seconds, then starts with the strange amplitude envelope that you can see in the scope screen shot.

I fear that i have to reset / re-flash the instrument, maybe losing cal data.

Current instrument SW / HW info :

SW 1.00.3.03.05 (150807.0)
Kernel : 3.2.35
FPGA 21
KEYBOARD 3

PCB 1004

Any help will be appreciated.

[Berni]

Yeah the firmware is horrible. Most modulation features will do something weird if you set them to just the right values, sometimes the waveform preview on the main screen glitches and shows markers that cut in to the text above it or just plain shows something that's not right. Even the output enable button does not work properly. I had a case where i somehow got the unit to have the output enable button lit up signaling the output is enabled but there was no output, pressing the enable button again the button stayed on but there was waveform output after that. Some settings also crash the unit and make it reboot, if you have selected to remeber the last setting it will reload them and get stuck in an infinite reboot loop until you factory reset it. It just goes on and on.

Its a real shame since the hardware in there is pretty nice but the software ruins it.