Sniffing the Rigol's internal I2C bus

[ted572]

Mr Ted572, Thank you for your great great fix ...  And Thanks to all the people who worked on this project.
I am sure it took many hard hours of work!!!

Good job, and congratulations.  All that I did was put together the information I found here.  That is all.  We are all grateful to the people that did all the hard work developing the these upgrades.

[bandgap]

Well that is great news, congratulations.  I will be going ahead also of course, but I have to wait a few days due to a family medical issue.  In the meantime it would be nice to see your before/after spreadsheet data.  It is encouraging to hear that you were able to successfully complete the process.  Thank you for describing your method and offering to share your data with us.

Ted,

As promised, here's a plot of my pre-cal / post-cal (for Channel 1, Channel 2 is similar) after I hacked my DG4102 to a DG4202. As you can see, it looks DANG good after the cal.

Here's a little more description of the steps I did (I didn't do things in the cleanest way, but it gave good results, so...)

1) Set cal item to AC amplitude (all measurements here are at 1kHz if I recall correctly.)
2) Go through and record each voltage value (RMS) using DM3068 and DS2202 ("DS2302"). I set each value in the DG using the DM3068 measurement.
3) Fit a line through DS2302 vs DM3068 measurements (needed to do this so that I could correct my scope measurements since it had a small offset)
4) Save and set cal item to LF flat.
5) Set each point value to the measured value on the DM3068.
6) Save and set cal item to HF flat. (DG will automatically set output to 50-ohm for this cal item.)
7) Connect a 50-ohm through terminator to the scope.
Use the corrected RMS value from the DS2302 (using the relationship in step 3) and calculate dBm. dBm=10*LOG10(Vrms^2/(50*0.001))  Record this value into the DG for each point.
9) Save and enjoy flat response from DG.

The only problem I have is that I have quite a bit of drop off when I get to low frequencies, say at 1Hz and below. (Is this normal? - Probably not. I'm redoing my channel 2 to see if this problem exists on channel 2.)

EDIT: Added channel 2 plot. Again, below 10 Hz, I'm seeing rapid fall-off (maybe something wrong with my cal method.)  At 1 Hz, I've got around -4 dB (the plots start at 10 Hz, so this isn't shown.)   10 Hz and above is very flat, though. 

Thanks,
Clayton

[Pinkus]

Maybe I missed this information in this huge thread:
At a DS2072 (not DS2072A), improved to 200/300 Mhz: can the CAN protocoll analyzer be enabled in any way or not?

[Marc M.]

...see the values that an Owner with a DG4162 can help as a base line values...

I took the time to record the cal values in my DG4162 for the community effort and stuck them in an Excel 97 xls file.  I recorded the values for CH1 first then started to do CH2 and realized they were all identical.  I then did some spot checking of random values and couldn't find any differences between the 2 channels so I didn't spend the time to finish going thru all of CH2. 

I also had a problem reading all the values for LOAD except the first (1-1).  I kept getting a dialog box stating that I needed to remove the 50ohm load.  Both channels were set internally to HI-Z and there were no loads connected to the generator.  I was unable to select any other item under load apart from 1-1. 

I also encountered odd behavior under HFLAT.  When I initially started recording values, the units were in dBm.  At some point I accidentally exited the cal procedure.  When I re-entered it I got a warning dialog that the units were in mVrms .  I power cycled the generator and re-entered the HFLAT cal table and it was back to dBm.  I finished recording the dBm values, I jumped out and back in to trigger the mVrms values to record those.  It turns out that they are identical to the dBm values   ?? ?? ??

I hope this helps.  If there is anything else I can do to help, feel free to ask.

Marc -

[ted572]

Well that is great news, congratulations.  I will be going ahead also of course, but I have to wait a few days due to a family medical issue.  In the meantime it would be nice to see your before/after spreadsheet data.  It is encouraging to hear that you were able to successfully complete the process.  Thank you for describing your method and offering to share your data with us.

Ted,

As promised, here's a plot of my pre-cal / post-cal (for Channel 1, Channel 2 is similar) after I hacked my DG4102 to a DG4202. As you can see, it looks DANG good after the cal.

Here's a little more description of the steps I did (I didn't do things in the cleanest way, but it gave good results, so...)

1) Set cal item to AC amplitude (all measurements here are at 1kHz if I recall correctly.)
2) Go through and record each voltage value (RMS) using DM3068 and DS2202 ("DS2302"). I set each value in the DG using the DM3068 measurement.
3) Fit a line through DS2302 vs DM3068 measurements (needed to do this so that I could correct my scope measurements since it had a small offset)
4) Save and set cal item to LF flat.
5) Set each point value to the measured value on the DM3068.
6) Save and set cal item to HF flat. (DG will automatically set output to 50-ohm for this cal item.)
7) Connect a 50-ohm through terminator to the scope.
Use the corrected RMS value from the DS2302 (using the relationship in step 3) and calculate dBm. dBm=10*LOG10(Vrms^2/(50*0.001))  Record this value into the DG for each point.
9) Save and enjoy flat response from DG.

The only problem I have is that I have quite a bit of drop off when I get to low frequencies, say at 1Hz and below. (Is this normal? - Probably not. I'm redoing my channel 2 to see if this problem exists on channel 2.)

EDIT: Added channel 2 plot. Again, below 10 Hz, I'm seeing rapid fall-off (maybe something wrong with my cal method.)  At 1 Hz, I've got around -4 dB (the plots start at 10 Hz, so this isn't shown.)   10 Hz and above is very flat, though. 

Thanks,
Clayton

Thank you very much Clayton.  This is great.

Was it necessary to go through Menu 1 (AC) and 2( LFLAT)? Could you have avoided some of the LF issues (10 Hz and below) by leaving these as they were?

I was assuming before that it may be possible to skip these and just deal with Menu item 3 (HFLAT) (?).

   Thanks for all the info and your help, Ted

[ted572]

...see the values that an Owner with a DG4162 can help as a base line values...

I took the time to record the cal values in my DG4162 for the community effort and stuck them in an Excel 97 xls file.  I recorded the values for CH1 first then started to do CH2 and realized they were all identical.  I then did some spot checking of random values and couldn't find any differences between the 2 channels so I didn't spend the time to finish going thru all of CH2. 

I also had a problem reading all the values for LOAD except the first (1-1).  I kept getting a dialog box stating that I needed to remove the 50ohm load.  Both channels were set internally to HI-Z and there were no loads connected to the generator.  I was unable to select any other item under load apart from 1-1. 

I also encountered odd behavior under HFLAT.  When I initially started recording values, the units were in dBm.  At some point I accidentally exited the cal procedure.  When I re-entered it I got a warning dialog that the units were in mVrms .  I power cycled the generator and re-entered the HFLAT cal table and it was back to dBm.  I finished recording the dBm values, I jumped out and back in to trigger the mVrms values to record those.  It turns out that they are identical to the dBm values   ?? ?? ??

I hope this helps.  If there is anything else I can do to help, feel free to ask.

Marc -

Thank you Marc:  This excel listing is great and will serve as a nice sanity check when we input our data. If something looks off from what you have, then it will at least put up a red flag to reconsider what we are doing.  And your note about HFLAT switching from dBm to Vrms will certainly save us a potential WTF, or worst yet a heart attack.  Hi

I asked Clayton (above) if it was necessary to go thu AC and LFLAT settings. You of course thought it was required also. Is it clear that this is required once you get started? I originally assumed it was required, then later that maybe it could be avoided (?), with just possibly HFLAT having to be done.  Thanks for your assistance, Ted

[bandgap]

In the meantime it would be nice to see your before/after spreadsheet data.  It is encouraging to hear that you were able to successfully complete the process.  Thank you for describing your method and offering to share your data with us.

I'll clean up my spreadsheet sometime and upload it in the next day or so. You CAN calibrate the HF flat curve without having going through the AC amplitude and LF flat curve. You might want to start with HF flat and see if that flattens everything up on the high end while keeping your low end in tact. I may even restore my preset values and do the same.

I re-calibrated my LF flat curve very carefully and also did a calibration of the DC offset (very easy to do that one.) However, this didn't fix my drop-off below 10 Hz. The fact that I have the same thing for both channel 1 and 2 suggests that it wasn't just a silly error of entering a number incorrectly.

-Clayton
 

[ted572]

In the meantime it would be nice to see your before/after spreadsheet data.  It is encouraging to hear that you were able to successfully complete the process.  Thank you for describing your method and offering to share your data with us.

I'll clean up my spreadsheet sometime and upload it in the next day or so. You CAN calibrate the HF flat curve without having going through the AC amplitude and LF flat curve. You might want to start with HF flat and see if that flattens everything up on the high end while keeping your low end in tact. I may even restore my preset values and do the same.

I re-calibrated my LF flat curve very carefully and also did a calibration of the DC offset (very easy to do that one.) However, this didn't fix my drop-off below 10 Hz. The fact that I have the same thing for both channel 1 and 2 suggests that it wasn't just a silly error of entering a number incorrectly.

-Clayton

Clayton:  How are you measuring below 10 Hz?  Isn't that a little tricky, as in, don't you have to use DC coupling in your O'Scope to see it.  And I don't think most DMM's will do very well at this low of a frequency.  Excuse me, I'm sure you know all this, but just in case you forgot it.  Ted

[bandgap]

Clayton:  How are you measuring below 10 Hz?  Isn't that a little tricky, as in, don't you have to use DC coupling in your O'Scope to see it.  And I don't think most DMM's will do very well at this low of a frequency.  Excuse me, I'm sure you know all this, but just in case you forgot it.  Ted

Yeah, the DMM certainly can't do it very well. I was using the scope to measure it when I was evaluating the before/after. I don't believe any of the calibration points are at frequencies that low, though.

-Clayton

[Radardude]

1ns TB, 200M BW Limit, and correct DS2302 Model type *yeah* ;-)

Orange noticed that the trigger is off by 3divs (lags behind) if u enable the 2nd channel - same happening, but only in 1ns mode.
maybe that is the reason why there is no official DS2302 version - anyhow i can live with that small limitation as it vanishes above 1ns TB

here is the version that will take care of model type string

http://www.filedropper.com/ds2302rilol

the "recalc" of the string is triggered by option un/install - so flush keys, and reapply them and it will become active.

did a selfcal on top of that - everything good.

Thanks to everyone who contributed to making the DS2000 series an awesome scope. I was looking at the Hantek until I came across this message board. I'm about to retire from my ATC electronic technician job. So I was in need of a oscope. After reading about DS2072, I purchase one from Tequipment. As we all know that the higher the bandwidth the higher the noise level. So here are some photos confirming the bandwidth increase using cybernet modified firmware ds2302rilol. First photo is using the latest factory firmware and all option installed. The rest of the photos are using cybernet modified firmware. Notice that the DS2202 has 840 uv of noise and the DS2302 has 1.16 mv of noise.

[ted572]

In the meantime it would be nice to see your before/after spreadsheet data.  It is encouraging to hear that you were able to successfully complete the process.  Thank you for describing your method and offering to share your data with us.

I'll clean up my spreadsheet sometime and upload it in the next day or so. You CAN calibrate the HF flat curve without having going through the AC amplitude and LF flat curve. You might want to start with HF flat and see if that flattens everything up on the high end while keeping your low end in tact. I may even restore my preset values and do the same.

I re-calibrated my LF flat curve very carefully and also did a calibration of the DC offset (very easy to do that one.) However, this didn't fix my drop-off below 10 Hz. The fact that I have the same thing for both channel 1 and 2 suggests that it wasn't just a silly error of entering a number incorrectly.

-Clayton

Clayton:  I'm looking forward to seeing your spreadsheets when they are done.  Thank you for offering to share them.

I'll do the HFLAT cal. and post the results.  Thank you for suggesting going ahead without doing AC and LFLAT.

Do you think there is anything to be gained by doing the AC and LFLAT calibrations, or do you think we can skip them without loosing something?  Or is it too soon to know?

Ted

[bandgap]

I'll do the HFLAT cal. and post the results.  Thank you for suggesting going ahead without doing AC and LFLAT.

Do you think there is anything to be gained by doing the AC and LFLAT calibrations, or do you think we can skip them without loosing something?  Or is it too soon to know?

Ted

Well based on your results, we should know! I may also reset mine to preset values and just do the HFLAT calibration to see how things turn out.

-Clayton

[thetooth]

As promised here is GEL file for DS2000A: http://thetooth.name/dev/DS2000(DSP)update%20(2).rar

Version is 00.02.01.00.03 November 5th, 2013, i have not tested it but hopefully someone here will find it interesting.

[neslekkim]

So DS2000 and DS2000A series are the same then, almost?, signature in the top are similar/same? DS2202
00.02.01.00.03


from the file:
DS2072A
DS2102A
DS2202A
DS2302A
MSO2072A
MSO2102A
MSO2202A
MSO2302A
MSO2072
MSO2102
MSO2202
MSO2302

USB
RS232
SPI
I2C
CAN
56M/28M
100M BandWidth
200M BandWidth
300M BandWidth

[Pinkus]

Well, the firmware also shows these strings:
CAN
LIN
USB

So, how can somebody enable CAN / LIN / USB protocol analyzer at a DS2072??

[Mark_O]

Well, the firmware also shows these strings:
CAN
LIN
USB

So, how can somebody enable CAN / LIN / USB protocol analyzer at a DS2072??

The USB may be only for trigger selection, not protocol decoding.

As for LIN, that's not available on either the 4000 or 6000-series, so I doubt that's available in any way.

[AndersAnd]

So DS2000 and DS2000A series are the same then, almost?, signature in the top are similar/same? DS2202
00.02.01.00.03...

The .GEL firmware name is also exactly the same for both the DS2000 and DS2000A series: DS2000Update.GEL
So maybe the DS2000 series can also be updated with the DS2000A firmware.

from the file:
DS2072A
DS2102A
DS2202A
DS2302A
MSO2072A
MSO2102A
MSO2202A
MSO2302A
MSO2072
MSO2102
MSO2202
MSO2302
...

From the .GEL file it looks like Rigol has plans to also release a MSO2000/MS2000A series with built-in logic analyzer.
So far they only make MSO4000 and DS1000D series with built-in logic analyzers.

[Tom01]

So maybe the DS2000 series can also be updated with the DS2000A firmware.

Is anyone going to risk the upgrade firmware on your DS2000 from DS2000A?

[AndersAnd]

So maybe the DS2000 series can also be updated with the DS2000A firmware.

Is anyone going to risk the upgrade firmware on your DS2000 from DS2000A?

Probably not before the DS2000A firmware has been hacked.

[marmad]

Is anyone going to risk the upgrade firmware on your DS2000 from DS2000A?

Since even their dealers/service personnel might mistakenly use the wrong FW file at some point, I would hope Rigol has given some thought to handling this possibility gracefully:
i.e. A-version FW won't load - OR - A-version FW is compatible - OR - A-version FW is not compatible but an older non-A version can just be boot-loaded over it...
- but perhaps I'm giving them too much credit 

I'd give it a try - but am in the middle of a job and can't run even the small risk of down time on the DSO right now. Perhaps in another week or so.

Meanwhile, I'll email Drieg and see if he knows the answer.

[alank2]

I'd give it a try - but am in the middle of a job and can't run even the small risk of down time on the DSO right now. Perhaps in another week or so.

You are brave!!!  Given their power on / help button method which seems to invoke some sort of bootloader that does very little (no display, etc.), probably you can *try* it and it will either (1) load the fw or (2) not load it, and then possibly B (1) boot it or (2) not boot it.  If it boots it may be buggy or it may be ok.

With any luck, hopefully no matter what the outcome, you will be able to return to ds2000 firmware using the help button method...

[alank2]

I'm going to dig around to see if I can find secret key sequences, like the menu7-menu6-menu7 type thing.  Does anyone know how the control panel connects to the blackfin?

[AndersAnd]

Since even their dealers/service personnel might mistakenly use the wrong FW file at some point...

They already have mixed up the firmwares, when they sent the latest DS2000 frimware to "thetooth" when he asked for the latest firmware for DS2000A: https://www.eevblog.com/forum/testgear/sniffing-the-rigol's-internal-i2c-bus/msg341839/#msg341839

[g***!]

Is anyone going to risk the upgrade firmware on your DS2000 from DS2000A?

Someone had to try running the firmware  Runs fine on my DS2202, but no CAN decoding showing in the options list unfortunately.

[Tom01]

Someone had to try running the firmware  Runs fine on my DS2202, but no CAN decoding showing in the options list unfortunately.

So D2000A is different from hardware D2000, right? Someone wrote that it was just a different firmware. Hmmm ....