HP6632A Correcting LCD Readout?

[SirAlucard]

So I've had this HP6632A Power supply for a little while now, and while the programmed voltage is pretty accurately being read on my 3478A, the reading on the actual unit itself is off. For instance if I set VSET 1V and hit enter my VSET will actually end up showing up as 1.001, but the reading output will show .998v, but the actual measured voltage on my DMM is 1.00000v. I'll throw out a little chart of voltage set, voltage it think I set, PSU reading and DMM reading.

Desired Voltage |	VSET Voltage |	Front Panel Measurement |	DMM Measurement
.5v	.498v	.496v	.49735v
1v	1.001	.998	1.00000
5v	5.002	5.000	5.0020
10v	9.999v	10.001v	9.9995v
12v	11.999v	11.996v	11.9984v

This gets ever so annoying if I try 5mV increments, this PSU is suppose to do 5mV increments, and for the most part it does, however there are a few instances that it doesn't quite get it. For instance if I set my voltage to 7v my Vset will be at 6.998 the front panel will read out 6.996v, and the DMM will say 6.9978, but if I put in 6.995v nothing will change. I'll then go to 6.99, and it'll go down to 6.992 on the Vset, the front reading will be 6.990, and my DMM will give me 6.9925. I'll go 5mV further to 6.985, and weirdly my VSET will read out as 6.987v, my front panel readout will be 6.990, and my DMM will read 6.9872.

So I guess what I'm asking if there was a way to fix these readings so what I'm pressing in as my requested voltage is what is being read back correctly. Or if this is a quirk of these power supplies? Secondly if I can somehow fix my front panel reading to be more accurate, I went through the process to calibrate the unit to try to see if that would help, it didn't much. If anyone has come across this before or can give more insight it'd be appreciated. Also the insides look great, no caps are bulging but I've thought about replacing them in case? Dunno let me know.

[Dr. Frank]

Your 6632A is doing fine, according to its specifications, there's nothing to be repaired, and nothing to be fixed or made more 'accurate'.

Explanation: These PSUs have a 12 bit DAC for setting the output voltage, at 20V max. output, that is about 5mV resolution of the output voltage, exactly the granularity you already recognized. The accuracy of this ouput is specified as 10mV, so you're already below the accuracy specification.
See specification parameters 'programming resolution' and 'programming accuracy', or alike.

The actual output is read back by a separate A/D converter, for the A model, originally it was also a 12bit one with 5mV resolution.

The later 6632B has even a 16bit ADC, the 'DC measurement accuracy' is 3mV and its resolution is a bit more than 1mV (I think).

The discrepancy simply arises from the different DAC vs. ADC used, i.e. at which point the respective bits flip over.

In the end, it's just a precise PSU (within the specification limits and DAC / ADC limitations), but not a kind of voltage standard, so please do not exaggerate the comparison with your 5 1/2 digit DMM.
Instead round the digits appropriately, i.e. to about 5 mV.
Frank

[SirAlucard]

Thank you for your response. I was hoping to hear that nothing was wrong with it. However I just didn't know what to expect, as I expect it to agree 100%, but that's my own Volt nut thing. I've been very happy with this PSU and I just wasn't sure about that aspect of it. So thank you for clearing up my confusion.

[Dr. Frank]

Thank you for your response. I was hoping to hear that nothing was wrong with it. However I just didn't know what to expect, as I expect it to agree 100%, but that's my own Volt nut thing. I've been very happy with this PSU and I just wasn't sure about that aspect of it. So thank you for clearing up my confusion.

You're welcome.. Your confusion was also mine, when I used my 6632B initially.. therefore I could also very well comprehend your 'volt-nuts thing' 

Frank