In late November last year, I reviewed the Rohde & Schwarz HMP4040.04 384W four-channel programmable power supply. This power supply was a beast of a supply with a hybrid linear-design and exceptional performance specifications which aligns well with its positioning as a performance-line supply. Despite being damaged in transit and eventually suffering a failure of its fourth channel, the unit has continued to function well. Quite recently, I dared to take it apart for a peek under the covers and attempted repair, and was impressed by the quality of components within.


However, recently, I have been involved with some further testing which resulted in a revisit to the RoadTest data for the reasons of comparison. Upon looking at the instrument performance data, I realised I had made a mistake which nobody seemed to have picked up upon. However, after discovering this, I could not rest comfortably unless I amended my errors.


The Errors

When revisiting the instrument performance data, I realised I had anomalies in the graphs for the current programming and read-back performance.


The originally graphed HMP4040.04 upper specification limits and lower specification limits are incorrect. According to the datasheet, the programming error should be 0.1% + 5mA and read-back error should be 0.1% + 2mA. However, due to a typo, I missed out a zero resulting in a tolerance of 1% + 5mA / 1% + 2mA instead. As a result, the graphs implied the HMP4040.04 had a wider tolerance than its datasheet specifications implied.


A second error was then discovered with the upper and lower tolerances for the readings from the Keithley Model 2110. Due to a programming error, the different %-reading error contributions for each of the current ranges were not taken into account for the 3A and 10A current ranges, thus implying the readings from the digital multimeter were more accurate than they actually were. The differing %-range contributions were correctly accounted for, however.


As a result of these two errors, the graphs incorrectly show a wider tolerance for the HMP4040's current programming and readback accuracy and also incorrectly imply that the Keithley Model 2110's readings were more accurate than they actually were. However, the absolute value of the programming and readback error were correct, with the miscalculations only affecting the lines which go around the actual value that indicate upper and lower limits respectively. I sincerely apologise that I did not catch this error earlier.


Updated Graphs

With some care, the graphs have been updated to reflect the correct error margins. I have also verified the other graphs and they do not contain errors to the best of my knowledge.

Current Programming Error


As the absolute values are unchanged due to the error, the absolute value of the programming error observed is impressive across all channels. Due to shunt heating in the multimeter, some error is observed around the 3A range change point which is not likely to be due to the HMP4040. In light of the correct (tighter) tolerances offered by the HMP4040, the Keithley Model 2110 does not provide enough accuracy to unambiguously guarantee the actual value of current is within specification when measuring in the 10A range, however, it strongly implies that it does.


Current Programming Error at Low Currents

Looking at small currents below 1A where the DMM is most adept, we can clearly see that the actual programming error is very close to zero and the margin of error sits entirely within the specification limits, exceeding the specifications.


Current Read-Back Error


Likewise, the current read-back error is actually tighter than the programming error. The absolute values show a good concordance with the correct specification limits, although the limitations of the accuracy of the multimeter really imply that it would be better to use the integrated metering as it is more accurate above 1.5A.


Current Read-Back Error at Low Currents


Looking at small currents below 1A where the DMM is most adept, we can clearly see that the actual value is very close to the expected value, but the margin of error sits wholely within the specification limits, clearly providing confidence that the supply meets and exceeds specifications.



In all, two errors resulted in the error margins for the current programming and read-back graphs to be incorrect, despite having the correct absolute values. This may have led to misinterpretations, thus I have re-processed the data with the correct error margin values and publish this follow-up. In all, this does not really change the conclusion in any significant way - instead, it shows that the Keithley Model 2110 perhaps doesn't have enough accuracy enough above 3A to be as confident of the results, but the values below 3A and the absolute error still strongly suggest that the supply not only meets the specifications, it beats them by a comfortable margin. It also now reveals that the read-back capabilities of the HMP4040 rival or even beat that of even a 5.5-digit benchtop meter, meaning external instrumentation may not be necessary at all.


I sincerely apologise to Rohde & Schwarz, element14 and all readers for this error on my behalf. I take full responsibility for what has been a rather careless mistake, and I thank you for your continued trust and support.



This is a part of the Rohde & Schwarz HMP4040.04 RoadTest - for the full review, see: R&S 4-Output Bench Power Supply, Prog (HMP4040.04) - Review