HP 3456A help please

[GhettoTronix]

I have a couple questions. This is a new machine for me and a big step up. I have gone thru the manual and I have done some searching without luck.

I think my "B" trimpot is toast. Poor little guy has been adjusted too many times in the last 30 years. I can't figure out where it is in the schematic to determine its value

I don't want to tear it down until I am ready for the repair. I have the same scanned manual that is on the HP website. Can one of you experts recommend a suitable replacement trimpot? Perhaps point it out in the schematic? I imagine I want to spend a couple extra bucks for something good.

Next up is the "MEAN" function. How do I set it up? The manual doesn't explain it well enough. It tells you what it does, but not really how to set it up for use.  I tried pushing buttons to figure it out but it only throws an error at me.

Thanks in advance for your expertise

[Dr. Frank]

Next up is the "MEAN" function. How do I set it up? The manual doesn't explain it well enough. It tells you what it does, but not really how to set it up for use.  I tried pushing buttons to figure it out but it only throws an error at me.

Thanks in advance for your expertise

Obviously, you have the wrong manual, service only?

Download the operating manual, "introductory users guide", available on the agilent site.

That's very well described, page 19-20 "How do I use a Math Feature?".

You have to engage the STAT function, and after at least 3 readings, you can read the math registers MEAN, VAR, COUNT, UPPER, LOWER, Z.

The error applies beacuse those registers have not been filled, because STAT was not started.

Frank

[Dr. Frank]

I have a couple questions. This is a new machine for me and a big step up. I have gone thru the manual and I have done some searching without luck.

I think my "B" trimpot is toast. Poor little guy has been adjusted too many times in the last 30 years. I can't figure out where it is in the schematic to determine its value

I don't want to tear it down until I am ready for the repair. I have the same scanned manual that is on the HP website. Can one of you experts recommend a suitable replacement trimpot? Perhaps point it out in the schematic? I imagine I want to spend a couple extra bucks for something good.

Do you have the correct manual, the compilation, including service manual 03456-90006 from the agilent site?

In that part of the document, pdf page 155, Section V - Adjustments, all the different pots are described, so you can look in the schematics, where to find them.
Pot A is coarse adjustment of 10V DC, B is fine 10V DC. How to adjust is described in paragraph 5-14 on pdf page 156.

That should be clear, in advance.. Why do you think, Pot "B" is not ok?

The 10V is directly calibrated on the reference schematic, A20 / 5A or 5B, depending on which reference is built into your instrument. "A" - coarse is a rotary switch, "B" fine sits direcly beneath it, that's a trim pot.

Btw.: That's pdf page 276 or 278.

PS: Why do you want to trim pot "B"?
Do you have something less uncertain as a calibration standard, than this box?


Frank

[GhettoTronix]

Ahh that makes sense on the service manual. I thought the 400 page PDF is all there was and it seemed to be lacking on information. I will look for more.

The B trimpot does not act right when I tweak it. Seems like it is worn out. I bought one of those reference modules off ebay that uses the AD584LH chip (5ppm). I know this machine has not been NIST traceable for at least 5 years, probably longer, and it was 10 counts high compared to my reference that is claimed to be checked against a NIST traceable 3458A. Probably nothing to worry about but I want it to be spot on to my best known reference even if it is slightly off. I am a hard core perfectionist and the little details matter to me. Finding the sweet spot in the trimpot is a PAIN and it seems to move by a turn or two when I make an adjustment. I have it dialed in best I can for now but I don't trust it. I do plan to spend more money on a better reference down the road when I can afford it but this is what I have for now.

I am keeping record of time, temp, measurements, and tweaks. I know my reference pretty well and I know exactly how much it drifts with temperature down to tenths of a degree in the range of my lab, around 21-26*C. I can measure the difference from the heat of my hands by handling it and always let it settle to room temp. I can get a couple degrees higher by using the heat radiated from the 3456A but I am not keeping record of those measurements. It is winter and cool right now so I will get a few more degrees out of it once it warms up outside. Working on the time part now to see how well it stays in tune in the coming months.

When I do buy a good reference I want to replace the trimpot as well so my confidence in the machine is upped. This trimpot just doesn't seem right and it's not like I never turned one before. This is the first time I have ever suspected a problem from a trimpot. I might be wrong but it seems to be the case.

[sync]

The B trimpot does not act right when I tweak it. Seems like it is worn out. I bought one of those reference modules off ebay that uses the AD584LH chip (5ppm). I know this machine has not been NIST traceable for at least 5 years, probably longer, and it was 10 counts high compared to my reference that is claimed to be checked against a NIST traceable 3458A.

I would trust the HP 3456A over the Ebay reference. And IFIAK these references are calibrated against a 34401A. Which can off by 10 counts at 6.5 digits and nonetheless well within spec.

btw:
AD584LH  5 ppm/K (and that's only for the IC itself)
HP3456A  2 ppm/K + 2 counts/K 0.02 counts/K (10V range)

[Dr. Frank]

....I bought one of those reference modules off ebay that uses the AD584LH chip (5ppm). I know this machine has not been NIST traceable for at least 5 years, probably longer, and it was 10 counts high compared to my reference that is claimed to be checked against a NIST traceable 3458A. Probably nothing to worry about but I want it to be spot on to my best known reference even if it is slightly off.

I am a hard core perfectionist and the little details matter to me.

If you are a perfectionist, I recommend you to first learn about the basics of metrology.

Obviously, you confuse temperature coefficient of the AD584L (i.e. 5ppm/K) with the uncertainty - measured in a different unit, namely ppm - of the assembly from ebay, and the uncertainty / stability of your 3456A, which is 8ppm/24h or 23ppm/90days .

If the room temperature, where you now operate your AD548L, deviates only 2°C from the calibration temperature, the output will deviate already by 10ppm, that's what you observe already.

To calibrate the 3456A, you need a reference, which is uncertain to 1ppm relative to NIST, and has a T.C. of < 0.1ppm/K.

I have seen descriptions of those chinese 4-channel references based on an AD584L, but the uncertainty (again: in ppm)  is to my knowledge not defined at all.
They show photos with a 34401A only, and that instrument - IF freshly calibrated - is uncertain to 20ppm only.

Even if they would calibrate it against a 3458A, the AD548L is way too unstable (over temperature and time)  to make a useful reference for calibration of 6.5 digits DMMs.

A reference for the 10V range alone also is not sufficient to calibrate that instrument, for DCV only you need those 5 Cardinal Points: 1000V, 100V, 10V, 1V, 0.1V.

Better trust the 3456A as is.

Frank