Rigol DP832 - Firmware list and bugs

[lunxg]

So these days I have to review why all my design prototypes fail. Finally I found  all of my prototypes using DP832 with Oscilloscope and soldering iron  to debug and work made all transistors (8050 in this case) fail.
It spend a couple of days to figure out why and I lost those day (and prototype broads) for my work. Well done. Rigol.

[mcinque]

So these days I have to review why all my design prototypes fail.

The DP832 is indeed not a state of the art PSU... but saying that by using that unit every transistors based circuit will fail, probably it's a little bit excessive: we all have seen how much current we're dealing with this issue...

Just my opinion.

[bapou]

just to put the DP832 current in relation. My Selektra soldering station (which is great in any other aspect)
has ~120VAC with ~100uA AC measured from the soldering tip to ground, likely induced by the toroidal transformer and I still didn't notice destroyed parts after using it for years (although after noticing it know, I don't feel very comfortable when soldering Mosfets)
Photos (text is in german):
http://www.mikrocontroller.net/topic/361296#4054746

[thn788]

I've checked our 4 DP832As and 2 DP832s: all show similar measurements as posted here by others already.

We use these Rigol PSUs on a daily basis for more than one year, now (replacing Agilent E3649A PSUs) to power our own custom designs, as well as off-the-shelf eval-boards; sometimes powering the devices "only" through their on-board regulators, sometimes powering chips directly from the Rigols, e.g. in cases where on-board regulators of the designs are not yet working properly or are not present at all in the first place. This also includes powering 32-bit SoCs or PMICs directly from the Rigol, sometimes.

So far not a single device (of the probably hundreds or thousands we have used) has ever suffered any damage from the Rigol PSUs; no matter if the devices are only connected to the PSU alone or also to Ethernet-Switches, USB-, RS232-, CAN- or other ports of different PCs at the same time or whether somebody uses oscilloscopes or logic analyzers from various different brands on "life" devices.

[mcinque]

This confirm my idea. This "issue" is not an issue. And I bet that similar readings can be found on other brands too. I will check on my Aim-TTi.

Probably you should check your designs, lunxg. They are not failing because of the DP832. This model has its bugs, but it's not a circuit killer.

[lunxg]

This confirm my idea. This "issue" is not an issue. And I bet that similar readings can be found on other brands too. I will check on my Aim-TTi.

Probably you should check your designs, lunxg. They are not failing because of the DP832. This model has its bugs, but it's not a circuit killer.

Can you explain more about you idea? Better with examples showing other supplies did this reading too.

I have compared the DC output rail to the ground with DP832, HP6632A, HP6642A and KIKUSUI PMC70-1,Xantrex PS,even cheap KORAD PS, ONLY DP832 showing this AC reading to the ground.

Some of my non-grounded SMPSs do showing high AC value to the ground. BUT MOST of my CHEAP SMPSs with ground connection are absent with high AC voltage to the output.

[mcinque]

Can you explain more about you idea?

As I stated in a previous post, I'm confident that those small currents are not an issue for a circuit, and this is confirmed also by the last two posts.

And I guess that even if we were dealing with higher current values, we should consider that probably your circuits are not connected to mains ground, so no AC current can flow thru your components. Am I wrong?

Better with examples showing other supplies did this reading too.

I can test only a PL303 by AimTTi, I will do the measurements tomorrow.

[Macbeth]

And I guess that even if we were dealing with higher current values, we should consider that probably your circuits are not connected to mains ground, so no AC current can flow thru your components. Am I wrong?

I was wondering the same thing. I mean most circuits are isolated from mains ground or else the grounds are tied. I have a railway track not far from my house with 25,000 VAC on its overhead lines. I daresay if I extended my multimeter leads between any piece of test equipment and that line I will read some kind of voltage. So what, it's not a part of my circuit anyway. Same as the mains ground isn't unless I deliberately connect it to my negative rail in which case there is zero potential difference. 

[H.O]

The possible issue is that when the circuit is connected to the powersupply (which is turned off) and you go in with your soldering iron (which has a soldering tip that is typically connected to mains earth) there's a path for the AC current from the supply output back to mains earth thru the soldering iron.

EDIT: Or you connect your oscilloscope ground clip (which is also connected to main earth).

Personally I've been using my DP832 for 18 months and haven't had any circuit fail on me.

[lunxg]

First, I have to make it clear again:

All my circuits fail because I used DP832 with my oscilloscope and  soldering iron.

As I stated in a previous post, I'm confident that those small currents are not an issue for a circuit, and this is confirmed also by the last two posts.

Refer to any specification of semi-conductor, exceeding its maximum voltage will have a chance to damage the component. Regardless how small the current it is.

And I guess that even if we were dealing with higher current values, we should consider that probably your circuits are not connected to mains ground, so no AC current can flow thru your components. Am I wrong?

I think most of the user will use DP832 with oscilloscope(and the clip is connected to main's ground). Who will expect they can't make floating measurement with isolated power supply?
On the other hand, human body is ground referenced object. So you will create a return path for AC current when you are touching the circuit which connected to DP832. Of course, human body is a huge value resistor and dividing most of the AC voltage. making the chance of damage less.


P.S. Yes, the most proper way to do iron work is to disconnect all the power wire first, the on/off channel function of DP832 really make me skip this step.

[mcinque]

hmm... you're damn right. Just connecting any DSO ground clip you're making voltage flow thru the components.

On the DP832 I measured these values between the ouput (+ or -) and mains ground:

• CH1 20.460VAC 58uA
• CH2 21.876VAC 61uA
• CH3 21.868VAC 61uA

I will do some experiments to see if I can kill intentionally any component with that output.

While on my PL303 by Aim-TTi these ones:

• CH1 22mVAC 214uA

And these are the scope screens while the PSU is ON and OFF (even after shorting the output to discharge the filter cap, I can read some AC voltage on any output  ).

[diyaudio]

I've checked our 4 DP832As and 2 DP832s: all show similar measurements as posted here by others already.

We use these Rigol PSUs on a daily basis for more than one year, now (replacing Agilent E3649A PSUs) to power our own custom designs, as well as off-the-shelf eval-boards; sometimes powering the devices "only" through their on-board regulators, sometimes powering chips directly from the Rigols, e.g. in cases where on-board regulators of the designs are not yet working properly or are not present at all in the first place. This also includes powering 32-bit SoCs or PMICs directly from the Rigol, sometimes.

So far not a single device (of the probably hundreds or thousands we have used) has ever suffered any damage from the Rigol PSUs; no matter if the devices are only connected to the PSU alone or also to Ethernet-Switches, USB-, RS232-, CAN- or other ports of different PCs at the same time or whether somebody uses oscilloscopes or logic analyzers from various different brands on "life" devices.

This a great news, I just ordered a DP832 which cost me alot and reading the last few posts made me rather nervous.
 
 

[H.O]

hmm... you're damn right. Just connecting any DSO ground clip you're making voltage flow thru the components.

Well, no, voltage doesn't flow thru anything but never mind....

On my DP832 Ch2 and Ch3 are the worst ones with 19VAC 50Hz open loop voltage and a "short circuit" current of 13uA measured with a FLUKE189.

If I measure the AC voltage while "loading" the supply with the impedence of my body by placing my left hand index finger on one probe and my right hand index finger on the other probe the 19VAC drops to 5VAC. Measuring the resistance from finger to finger shows around 2M which suggest that the "supply" is pretty weak.

[Howardlong]

Microchip have an interesting application note here http://ww1.microchip.com/downloads/en/AppNotes/00521c.pdf where they advocate current limiting 110VAC to 35uA by using a 5Mohm current limiting resistor straight onto a device pin, using the clamping diodes to limit voltage.

[Guni]

I've just found this by case. Small unboxing of DP832 on chines website. Nothing special except... full color display on non-A model. It looks like some early firmware. 
Color display is not a big deal but I'm curious if you know something about this ?



Source:
http://www.hkepc.com/forum/viewthread.php?tid=1953803

[sdufour]

Hi,

I have tried the keygen from multiple source (as the riglol website is now down) and got the same key with every one of them, using the built in private key.

Using Mxxx code is says that the licence is invalid but with Fxxx code the "installation fails".

I think it is an issue with the version 1.11, should i upgrade to 1.13 first?

Digital Version : 00.01.11.00.00
Analog Version : 02.02.02.02.02.02
boot version : 01.09


Edit: I have absolutely no voltage (or current when shorting) between the channels and ground. I may have a different board version than yours... maybe it would be interesting to compare the differences.

I had that issue at work where the ground in the power bar plug was sometime unconnected. The EMI filter capacitor in the input of the oscilloscope were acting as a voltage divider and I got 40V ac between the oscilloscope ground and my soldering iron. We lost a week trying to figure out what it was, blowing multiples mosfet in the process, and now the power bar is in the gargbage...

Edit 2: Think I found my issue...

Also note there is a 12 hour (run-time) lock out period if an invalid code is entered too many times. Also, no dashes. Your code doesn't look long enough either, mine on the DP832 were 28 characters long.

[lunxg]

Two weeks before I have sent email and my video to Rigol and vendor about the AC voltage problem but still seeking for reply. I was expected they can act faster since all I use is their native Chinese language to state the problem.

[McBryce]

I prefer the monochrome look actually. The colours are a bit too "Hallo Kitty" for my liking.

McBryce.

[Guni]

My Chinese is terrible but I think they are saying that they send a command over the LXI interface... Maybe the MON command? They are definitely saying you should be able to replicate it on other units.

Using Google translator, I didn't found anything about hacking. But it is probably field for some experiments.
On this video Rigol demonstrates A-version, and screen composition is not triangle layout. Now there is a update (01.09) for A version with "classic", non-A layout.

I prefer the monochrome look actually. The colours are a bit too "Hallo Kitty" for my liking.

McBryce.

Depends on color combination. Sometimes colors is just more readable the mono

[McBryce]

Yes, I find the green version best / most readable.

Btw. I've ordered a 3.8mm drill (not a drill size one tends to have lying about) to attempt the positive post repair. As my DP832 has already been hacked / opened etc. I don't have the option of sending it back. I'll report back on how it goes. Worst case (if it goes pear-shaped) I can replace the posts completely.

McBryce.

[bson]

I ported over LaurentR's script to Python and thought I'd share it here.

I refactored it a bit, so there's an LxiInstr base class, a KeysightTrueVolDmm class for measurements, and a Dp832PowerSupply to talk to the PSU.  I wanted to move the DMM support into a separate file to make it easier to use it for other purposes, such as data logging, or to implement the same interface for other DMM's.  (I kept LaurentR's DM3068 switch to Agilent mode, but not sure if it works.  It would be better to add a native implementation for it.)  calib.py is a generic wrapper pretty much, and all of the calibration guts are in Dp832PowerSupply.  They're DP832 specific anyway, and not easily reusable.

I tested this on OS X 10.10, NI-VISA, PyVISA, LAN devices, 34465A DMM.  Python2.7 is required.
But it should work fine with GPIB, USB also.  (<= famous last words!)

The code is somewhat short on comments...  Anyway, it's a first stab at creating a generic Python library for VISA instrument support.

[Stupid Beard]

The code is somewhat short on comments...  Anyway, it's a first stab at creating a generic Python library for VISA instrument support.

Why don't you just use https://github.com/python-ivi/python-ivi which already supports the DP832, as well as a bunch of other DMMs, scopes, PSUs, etc?

[bson]

Oh cool, had no idea such a thing existed!

Edit: it doesn't seem to know how to calibrate or do much of anything else device specific...

[Macbeth]

I ported over LaurentR's script to Python and thought I'd share it here.

Thanks for this. I have a Rigol DM3058E but never got around to using LaurentR's code because I don't have Matlab. I don't believe any Rigol owner has tested it yet?

Anyway, I'm using Windows 7 and have never used Python before, but what the hell - lets see what this snakey thing is. It wasn't as simple as just download python and run, there are a few dependencies as well:

Python 2.77
Microsoft Visual C++ Compiler for Python 2.7
pip 6.11

then use pip to download and install

PyVISA 1.6.3
NumPy 1.9.2 (go and make a cup of tea while waiting for this one to compile/install!)

Results in next post...

[LaurentR]

I ported over LaurentR's script to Python and thought I'd share it here.

I refactored it a bit, so there's an LxiInstr base class, a KeysightTrueVolDmm class for measurements, and a Dp832PowerSupply to talk to the PSU.  I wanted to move the DMM support into a separate file to make it easier to use it for other purposes, such as data logging, or to implement the same interface for other DMM's.  (I kept LaurentR's DM3068 switch to Agilent mode, but not sure if it works.  It would be better to add a native implementation for it.)  calib.py is a generic wrapper pretty much, and all of the calibration guts are in Dp832PowerSupply.  They're DP832 specific anyway, and not easily reusable.

I tested this on OS X 10.10, NI-VISA, PyVISA, LAN devices, 34465A DMM.  Python2.7 is required.
But it should work fine with GPIB, USB also.  (<= famous last words!)

The code is somewhat short on comments...  Anyway, it's a first stab at creating a generic Python library for VISA instrument support.

Got it to work on Win64 and it's great, thanks. I used Matlab out of convenience (and easy drawing/debug), but was always wanting to give pyvisa a try (I am not a Python expert though). Thanks for doing the work.

Unfortunately, as pointed out:

Anyway, I'm using Windows 7 and have never used Python before, but what the hell - lets see what this snakey thing is. It wasn't as simple as just download python and run, there are a few dependencies as well:
...

The Windows install is not straightforward for us GUI people, but it works 

The only issue I had with your code was that early on, you check for the presence of ":" in the device Visa addresses, but Visa supports aliases , which typically don't have ":" in the name. For instance, my 2 devices are aliased (rather imaginatively) "DP832" and "34461A".