Products > Test Equipment
Rigol DP832 - Firmware list and bugs
skander36:
--- Quote from: garrettm on March 17, 2020, 11:54:08 pm ---
I decided to do the print format edit first. Now everything is lined up and prints out values in volts and amps. So no mucking about with annyoing scientific notation. I also reduced the settling time to 2 seconds for volts and 3 seconds for current. This is in anticipation of the larger cal point arrays, which will make the calibration take longer.
Let me know if you have any problems with new changes and thanks for using the script!
UPDATE: There was a minor typo in TelnetTest that I fixed. It would cause the DMM to report an error because of an unfinished measuremnt. Accidentally called readDMM() instead of dmm.read() inside the for loop. Whoops.
--- End quote ---
Now it look very good, thank you :
ch1 DAC-V calibration
step 0, cal point: 0.0v, meas val: -0.2345v
step 1, cal point: 0.01v, meas val: -0.2349v
step 2, cal point: 0.03v, meas val: -0.2352v
step 3, cal point: 0.1v, meas val: -0.2554v
step 4, cal point: 0.2v, meas val: -0.1541v
step 5, cal point: 0.7v, meas val: 0.3462v
step 6, cal point: 1.0v, meas val: 0.6448v
step 7, cal point: 1.2v, meas val: 0.8427v
step 8, cal point: 1.7v, meas val: 1.3375v
About extended calibration points . I've done it (using your script) but I don't see any difference. My problems on CH (0-40mV) is still present and can be solved only by manual calibration and other channels are still about the same than before .
1mV output > 1.91mv
2mV output > 2.94mV
3mV output > 3.99mV
I have done better results before . Looks like DP832A cannot benefit from more granular calibration points . Or maybe values must follow a specific algorithm ?
I have attached script used and output .
Maybe you or other can repeat the test . Maybe I was doing something wrong .
garrettm:
@skander36 I've improved the readback accuracy and got ch3 spot on, 1mA and 1mV resolution essentially at all programming values. Ch 1 and 2, however, are a different story as you've seen. I'm still playing around with different cal points for ch 1 and 2 to see if I can get closer to 0mV when programmed to 0 and minimize the "off" state leakage current. Right now leakage is 10mA when off and some 200uA when on and at 0mV. A rather ironic result.
On a side note, I wish Rigol would have put relays in these units to do auto tracking (parallel, series connection internally) and disconnect the output transistor completely. My old GW Instek often times feels superior because of these simple additions. Though I much prefer the screen of the DP832 and its ethernet connectivity.
skander36:
Yeah you cannot have it all .
What I was want to say is that I cannot improve DP832 over it's limits, so extended range of points seem that doesn't bring more precision .
Your script is working well now and is very usefull for calibrate DP832 into factory specifications .
The others scripts also do the same thing but with more hassle to setup, unless someone make a binary executable for them .
JDubU:
I've been testing various values and quantities of cal points and it seems like specific values at specific places in the range have more effect than just the quantity of points. I wish I knew what interpolation algorithm Rigol is using for this and what the DAC hardware resolution is. I did manage to get all of the channels on my DP832 to be accurate to the millivolt for almost the entire voltage range, but there are some rare small spots where the output voltage goes off by as much as 3mV. For example, if I set the voltage to something like 1.423V and then scroll the first digit to the left of the decimal from 1 to 32, the last three digits might be xx.423 for every single volt change except at 19 volts where it will be 19.420. This occurs even when the cal points were only integer volts between 10V an 32V. It's weird!
I also discovered that the DAC voltage cal points for my DP832 need to be different between channel one and channel two in order to get millivolt accuracy close to zero on both. They just do not calibrate the same way with the same cal points. The list of cal points that, so far, have produced the best results for my DP832 is attached.
garrettm:
@JDubU thanks! I managed to improve the readback accuracy with using more points and similarly got channel 3 spot on with extra cal points. I also noticed that custom cal points are needed for each channel, as, like you say, they each respond differently during calibration.
I'll give your point list a try tomorrow when I have some free time.
As for the DAC, I would hope Rigol used a 16-bit DAC for ch 1 and 2: 0 to 32 volts at 1mV resolution is 32,001 discrete values--which is about 15-bits. Better to have more bits than not enough. Though they may have cheaped out and used a 14-bit DAC.
As you say, it would be great to get the algorithm they used. It wouldn't be too hard to use Matlab to help optimize the cal point list to better linearize the output. Now that I think about that... I might just set the PSU into cal mode and step through at 10mV increments (using the same "cal step" with just different set values) to see what the raw output looks like. Might go with 1 second settling time, as this will take about 2 hours to collect at 100nplc.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version