Currently, the unit’s setpoint accuracy isn’t all that great (readback resolution is limited as well, but shows the offset): Regardless of the range it has an offset of about +20 mA. Sure, for a unit capable of 60 A that seems reasonable, and well within its specs of ±0.1% ±75 mA. But since I’ll use it in the low range (6A) most of the time and likely often even well below 1A that doesn’t seem too great, does it?
Calibration/adjustment procedure
A word of caution: The voltages found inside the unit can cause damage, severe injury or death. If you are not authorized or qualified to work on such devices, don’t do it. I don’t take any responsibility for actions you take or the results of these actions. You do everything at your own risk. Be very careful and stay safe. Any information is provided “as is”.
So it’s time for an adjustment. The first thing I wanted to do is the “post repair calibration”, which is checking that the offset voltage produced by U5 is within ±0.5mV (procedure in service manual). Mine was just below +0.5mV and was easily brought down to <<0.1mV by adjusting the only trim pot of this unit: R155.
Next comes the actual calibration/adjustment procedure. Unfortunately, HP doesn’t let you adjust the ranges individually, no, you are to calibrate/adjust all ranges in a certain, strictly specified manner. Now, consider that you usually have at least one low point and one high point per range and don’t forget about the transient and readback circuits. Then you’ll see that the calibration procedure isn’t completed within 5 minutes.
HP didn’t bother to include a calibration menu in the user interface – to the best of my knowledge you have to use GBIP for that. This, by itself, wouldn’t be a huge deal. Requiring a 5V/60A/300W low noise power supply, however, might be. And measuring such large currents precisely is a huge challenge in itself. I, for one, neither own a 0.01 Ω @ 100 A, 0.04% @ 100 W current shunt nor a 0.1 Ω @ 15 A, 0.04% @ 25 W current shunt HP lists under “Equipment required”. But this isn’t all. you would also need 15V/10.9A, 10.9V/15A and 60V/6A power supplies to calibrate the voltage and resistance ranges. Except for the low current range I don’t own a single power supply capable of any of these settings (with only one channel used).
Makeshift calibration setup
So first let’s have a look at what I got, and then how I will make it work to achieve a calibration that I consider good enough for my simple purposes.
Power supplies that were available to me for the calibration:
- Power supply 1: 0-20V/0-10A
- Power supply 2: 0-15V/0-7A
- Power supply 3: 2x 0-30V/0-5A/80W; 0-5.5V/0-5A
DMMs:
- Agilent 34401A (3A max.; 0.12% of value + 0.02% of range)
- Handheld DMMs (10A; 6,000 to 60,000 counts; about 0.3% of value + some digits, or worse)
With this equipment it’s obvious that my calibration won’t be perfect. But just better than now, especially for the low current range, is everything I’m looking for.
Luckily, HP allows to adjust the unit at much lower currents than full scale. Even the exemplary calibration program described in the user manual specifies a current for the high calibration point of about 50 A, matching the capabilities of HP 6032A auto-ranging power supply quite nicely. But as I found out, the 6060B doesn’t complain if the high calibration point is reduced even further, in my case to 32A. (At one point I also tried 25A without any problems.) An equivalent approach works for voltage and resistance adjustments too. (Be aware that the approach has obvious and maybe even less obvious limitations.)
The plan was to pair the power supplies/channels with a DMM for the current measurement. The power supplies/channels are then connected in parallel or in series to match the requirements as good as possible, but always trying to use as few power supplies as necessary for a given setting. The total current would then be determined manually by adding up the DMM readings for the supplies connected in parallel. The voltage measurements are much simpler and can be directly performed with the Agilent 34401A. It surely will make a difference though, whether you take the voltage measurement at the front or the rear terminals.
But in order to get a feeling for the accuracy of the handheld DMMs used, I compared them with each other (up to 10A) and with the Agilent 34401A (up to 3A) as well. My higher-resolution DMMs agree with each other very closely. At 5A the indicated values of the 6000 counts meters are 1 or 2 mA off compared with the higher resolution meters – still good enough considering.
I ported the BASIC script shown in the operating manual to Python (jupyter notebook, using pyvisa for the SCPI communication). This simplifies the procedure significantly. It’s far from fully automated though, mainly because the setup changes very frequently and measurement have to be taken and entered manually. As described in the operating manual, the exemplary calibration procedure takes some shortcuts – doing a more thorough readback adjustment might improve the readback accuracy. In case someone would like to use the script as a starting point, you can find it on GitHub: https://github.com/sbstnh/HP-Agilent-keysight-6060B-calibration. (Some improvements were made after I used it for the calibration, untested, so bugs might have been introduced doing that.)
Current ranges calibration
- Low and high point of the low current range calibration/transient (0.3A, 5.1A), low point of the high current range calibration (3A):
- Power supply 2, 60,000 counts meter (600 mA/6A range).
- High point of the high current ranges/transient (reduced to 32A; this is also the maximum my test leads are specified for), power supplies parallel:
- Power supply 1 (10A), 50,000 counts meter (10A range)
- Power supply 2 (7A), 60,000 counts meter (10A range)
- Power supply 3 (3x 5A), 6,000 counts meters (6A range)
Voltage range calibration
- All calibration points, channels in series:
- Power supply 3 (2x 30V, reduced to 2.677A; maximum of this power supply)
- Agilent 34401A (voltage)
Resistance ranges calibration
- Low resistance range/transient, power supplies parallel:
- Power supply 1 (15V, 10A), 50,000 counts meter (10A range)
- Power supply 2 (15V, 0.9A), 60,000 counts meter (6A range)
- Agilent 34401A (voltage)
- Low point of mid resistance range, power supplies parallel:
- Power supply 1 (10.9V, 10A), 50,000 counts meter (10A range)
- Power supply 3 (10.9V, 5A), 60,000 counts meter (6A range)
- Agilent 34401A (voltage)
- High point of mid resistance range/transient, power supplies parallel:
- Power supply 1 (10.9V, 10A), 50,000 counts meter (10A range)
- Agilent 34401A (voltage)
- High resistance range/transient (channels in series):
- Power supply 3 (2x 30V, reduced to 2.677A), 60,000 counts meter (6A range)
- Agilent 34401A (voltage)
Results
The adjustment was successful to the extent that could be expected: The large +20mA offset is gone, with the actual current in the low range tested to be within <2mA of the setpoint, and in most cases significantly below that.
The current readback accuracy for low currents, however, isn’t all that great – by today’s standards 12-bit readback resolution with only one current measurement range simply doesn’t cut it. (60A divided by less then 4096 usable steps would likely result in >15 mA resolution.)
Conclusion
Performing a proper calibration of the HP 6060B requires a lot of high current precision equipment most home/hobby electronic labs including mine don’t offer. This complicates the procedure significantly. That being said, I’m pretty happy with the results of this approach.
In the future DC loads will be discussed in way more detail, at least that is my plan.