Rigol's New Function Generator and DMM: DG800Pro / 900Pro / DM858

[josfemova]

If recording the screen was not in consideration for me, I think I would be inclined to go with the Siglent
https://siglentna.com/operating-tip/sdg-remote-scpi-command-to-request-a-screen-image/

Ok that makes my decision harder now haha. I still think that being able to show/control the scope via the web interface is better constantly fetching the scope screen without showing the buttons I'm touching, but I guess for my use case not showing the buttons is not that much of an issue

[Njk]

I wish they to stop designing new things and to dedicate the rest of their life to fixing the existing designs. By now, the nomenclature is very wide so it's a big opportunity. And that'll be a real mission for Rigol

[csuhi17]

I have found two bugs so far.
Unfortunately, I don't use it much, the FY6800 is often enough for me.

One is when the dolphinfish attack, the device becomes unusable. 

I don't know what's causing it, but I can't change the setting of channel 1, only channel 2, but it has no output, it responds to the buttons. This is the second time it has come up.
Third, only now at the counter.

The other is that at and around 1Hz square wave there is a small sine wave, the location and frequency of which depends on the frequency.
and the duty must be between 20-95%.
It can also be found for the pulse, but only if the trailing is 3.0ns.
It doesn't do it at 3.1ns.

Due to the lack of memory, it is not visible, but I measure it as a 312.5Mhz with the DHO924. I don't have a faster scope.

[ZhuraYuk]

So, does anyone figured out the way to unlock second channel and full bandwidth? At leas not going to DG900 pro. I assume the way is similar to DHO800/1000 using key file and python script to generate license codes?

[DaneLaw]

Nope, not yet' I still just trying to get my unit just to play along...
The firmware on my unit is one sick puppy..
its all native with the FW/SW it arrived with  (UI 09-12 DG821PRO) , its literally getting more & more weird each time I turn it on.
I haven't tried to hack it or mess with its software, this is all on Rigol's cape.
Couldn't even activate the output or the counter and the "back to reset" didn't do jack.
Tried  some of the settings, but it clean the devicemem if you adjust the int/ext clock, and the infamous dolphinfish will enter..
Doesnt seem to be any newer fw, than V00.01.00-2024-01-12   https://www.rigol.com/products/DG_detail/DG800Pro
- that do match the fw my unit arrived with 00.01 tough dated on the device as UI/FW 09-12-23.. anybody installed this Rigol site fw?

[gbix]

Nope, not yet' I still just trying to get my unit just to play along...
The firmware on my unit is one sick puppy..
its all native with the FW/SW it arrived with  (UI 09-12 DG821PRO) , its literally getting more & more weird each time I turn it on.
I haven't tried to hack it or mess with its software, this is all on Rigol's cape.
Couldn't even activate the output or the counter and the "back to reset" didn't do jack.
Tried  some of the settings, but it clean the devicemem if you adjust the int/ext clock, and the infamous dolphinfish will enter..
Doesnt seem to be any newer fw, than V00.01.00-2024-01-12   https://www.rigol.com/products/DG_detail/DG800Pro
- that do match the fw my unit arrived with 00.01 tough dated on the device as UI/FW 09-12-23.. anybody installed this Rigol site fw?

new fw is same as fisrt, but have readable description in english

for hack device you send me pm

[ZhuraYuk]

Right now there are huge discounts on Aliexpress, you can buy Rigol DG821 Pro for 316USD with VAT with promocode! https://www.aliexpress.com/item/1005006674035954.html
Apply one of the codes at the end before paying,  I tried 50 USD one and it worked, the list of codes is here: https://www.reddit.com/r/ChineseWatches/comments/1d370w2/comment/l6743vg/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_button

[zrq]

gbix did the upgrade from DG811 Pro to DG 922 Pro for me, apparently works fine except the amplitude calibration (and probably worse, phase calibration???). You need to send him the setup.stp in the /rigol/data/ folder and the CPU serial you get from cat /proc/cpuinfo. You can fetch and replace that file by adb through network just like the DHO series.
Note that the encryption algorithm is completely different from the DHO series. There is no xxtea but algorithm that I could not recognize.
The calibration is significantly off after modification, probably gbix know how to correct it? I'm also looking at the cal_xxx.hex files.

DophinFishSE have all the symbols and debug info not stripped, so it can help greatly in the reverse engineering. But somehow GHIDRA get choked by the file for some reason (figured out, it's the the thumb instructions), I have to use IDA.

[zrq]

Interesting, it appears the calibration data of DG800 pro is relatively easy to understand, I started to get some sense of it. So there are three ranges for CAL_AFG_HIFREQ_FLAT...

Code: [Select]

with open(os.path.join(fn0, 'cal_hfflat1.hex'), 'rb') as fh:
    a=np.fromfile(fh, dtype='<f8', sep='')
plt.plot(a)
with open(os.path.join(fn0, 'cal_hfflat2.hex'), 'rb') as fh:
    a=np.fromfile(fh, dtype='<f8', sep='')
plt.plot(a)
plt.vlines([27, 50, 77], ymin=0.99, ymax=1.01, colors='r')
plt.xlabel('Index')
plt.ylabel('cal_hfflat*.hex')
plt.figure()


It would be very nice if someone own a DG900pro that can share the calibration file (which can be fetched by adb -s you.rip.add.res:55555 pull /rigol/data/cal_hfflat2.hex) .

[zrq]

Thanks to a user sending me the calibration from a 912 pro, these is how the files compare. One would guess...


unfortunately, it seems an accurate DIY amplitude calibration is not as simple as I thought, apparently there is some reflection artifact due to imperfect source impedance/matching. I connected the generator to either the hacked DHO1074 scope or a spectrum analyzer, and also with a minicircuits 20dB attenuator to improve SWR of the scope or SA's port. I always see these kind of oscillations, with the beat frequency dependent on the cable length.

[ZhuraYuk]

So correct me if I wrong, we have two files cal_hfflat1.hex and cal_hfflat2.hex they are calibration tables for each channel. They contain 78 values, each valued is dB calibration offset. But for what frequencies? Also what about gbix statement that there is 71 points for 200kHz+ range and 78 points for range below 200kHz?
I suspect that at factory calibration was done by some undocumented SCPI command automatically. One might search for them in firmware.

[zrq]

So I managed to calibrate it to my scope by hacking the cal_hfflat?.hex. Perfect upgrade from DG821 Pro to DG922 Pro seems possible given one have an amplitude calibration standard. The 20 dB attenuator is actually causing more reflection than a direct connection to the scope set to 50 Ohm via a short BNC cable. Looks pretty flat against my scope's frequency response after adjustment, the jump at 20 MHz is because I decided to keep the original calibration data intact.






DG821 pro seems to be only factory calibrated to 20 MHz  , and DG 912 pro only to 150 MHz. The frequency points for cal_hfflat?.hex is found in CApiCalibration::servFlatCaliGear (warning: shit code ahead) and is given here

Code: [Select]

200000000000000 cal_HF_FLAT_offset = 27 (not used for rng2)
175000000000000 cal_HF_FLAT_offset = 26 (not used for rng2)
150000000000000 cal_HF_FLAT_offset = 25 (not used for rng2)
120000000000000 cal_HF_FLAT_offset = 24 (not used for rng2)
100000000000000 cal_HF_FLAT_offset = 23
90000000000000 cal_HF_FLAT_offset = 22
80000000000000 cal_HF_FLAT_offset = 21
70000000000000 cal_HF_FLAT_offset = 20
60000000000000 cal_HF_FLAT_offset = 19
50000000000000 cal_HF_FLAT_offset = 18
40000000000000 cal_HF_FLAT_offset = 17
30000000000000 cal_HF_FLAT_offset = 16????, not used, bug?
20000000000000 cal_HF_FLAT_offset = 15
10000000000000 cal_HF_FLAT_offset = 14
8000000000000 cal_HF_FLAT_offset = 13
6000000000000 cal_HF_FLAT_offset = 12
4000000000000 cal_HF_FLAT_offset = 11
2000000000000 cal_HF_FLAT_offset = 10
1000000000000 cal_HF_FLAT_offset = 9
900000000000 cal_HF_FLAT_offset = 8
800000000000 cal_HF_FLAT_offset = 7
700000000000 cal_HF_FLAT_offset = 6
600000000000 cal_HF_FLAT_offset = 5
500000000000 cal_HF_FLAT_offset = 4
400000000000 cal_HF_FLAT_offset = 3
300000000000 cal_HF_FLAT_offset = 2 (200k to 300k)
200000000000 cal_LF_FLAT_offset = 14
100000000000 cal_LF_FLAT_offset = 13
90000000000 cal_LF_FLAT_offset = 12
80000000000 cal_LF_FLAT_offset = 11
70000000000 cal_LF_FLAT_offset = 10
60000000000 cal_LF_FLAT_offset = 9
50000000000 cal_LF_FLAT_offset = 8
30000000000 cal_LF_FLAT_offset = 7
10000000000 cal_LF_FLAT_offset = 6
9000000000 cal_LF_FLAT_offset = 5
7000000000 cal_LF_FLAT_offset = 4
5000000000 cal_LF_FLAT_offset = 3
3000000000 cal_LF_FLAT_offset = 2 (100 to 300)
1000000000
There are 3 ranges that have different flatness calibration value, the Vpp used in factory calibration are 1: 12649111e-7 V, 2: 50000000e-7 V, 3: 6324556e-7 V . The 3 ranges corresponds to the three sections in cal_hfflat .

[ZhuraYuk]

So what about simple instruction about how to do the calibration with oscilloscope? The most unclear part is how to edit the calibration file based on oscilloscope readings.

[Construct]

DophinFishSE have all the symbols and debug info not stripped, so it can help greatly in the reverse engineering.

Leaving symbols and debug info in place is quite the gift to the reverse engineer.

Thanks for sharing your progress.

[zrq]

DG821pro's THD measured with my sound card (has a negligible THD floor) @ 1 kHz, 1 Vpp, it meets well with the 0.1% spec and is kind of among the average of cheap AWGs. Somehow the channel 2 on my DG821pro (hacked, of course) shows a consistently lower distortion.

[ZhuraYuk]

So I measured my hacked DG821 Pro and here is the result,  1V PtP  hiZ output in to 100k Ohm load, the result is identical for both channels. Far worse than people are measuring for SG2042x 0.06% vs 0.005%

https://www.eevblog.com/forum/testgear/the-siglent-sdg2042x-thread/msg3606484/#msg3606484

[NE666]

Would I be correct in saying that the DG800/900 series lack the ability to create/load/store user-generated waveforms (other than user-configurable sequencing of built-in waveforms)?

I don't see any reference to such a feature in the user documentation.

[ZhuraYuk]

Would I be correct in saying that the DG800/900 series lack the ability to create/load/store user-generated waveforms (other than user-configurable sequencing of built-in waveforms)?

I don't see any reference to such a feature in the user documentation.

I tried to load stored waveform but there is no button on screen for that. There is physical button named Store/Recall, but actually there is no way to recall or load anything. So I guess they left that for future software updates. Previous DG900 generation had this feature.

[csuhi17]

In a previous comment, it is stated that the software is still under development, so this function is not available.
And in another comment, if I remember correctly, a method was described, but it is inconvenient and a very fiddly method.

It is impossible to know when it will be available.
Rigol updates are also slow, nothing new half a year.
Maybe years...

[NE666]

In a previous comment, it is stated that the software is still under development, so this function is not available.

That's regrettable. I'd hoped, against prior experience and knowledge of the brand, that this might have been addressed by now. And knowing Rigol as we do, if it's not been addressed already, it will more than likely be years out, or never.

What a wasted opportunity.

[zrq]

Spectrum of DG900 pro's noise output, looks like sampled at 156 MHz, slower than I would expect the hardware is capable of.

[Njk]

Seems odd. Spec data for noise waveform:
Noise (-3 dB): Typical (0 dBm), >250 MHz bandwidth