DIY Calibrating a HP-3478A

[bingo600]

Guyzz

I have a HP 3478A that's 80ppm off (high) when feeding it 10v from my Geller reference.
I trust my 10v reference, as it is spot on. Whwn measured with my Keithley 2015 (Cal'ed May-2011)

I would like to calibrate "just" the 10v range , is that possible.

Any hints would be nice ....

I have the Servicemanual , and have read the Calibration section.

But i can't see if it's possible to just calibrate the 10v range , or if i have to make all calibrations.

I don't have equipment to calibrate AC or DC above 10v.

TIA

/Bingo

[staxquad]

Guyzz

I have a HP 3478A that's 80ppm off (high) when feeding it 10v from my Geller reference.
I trust my 10v reference, as it is spot on. Whwn measured with my Keithley 2015 (Cal'ed May-2011)

I would like to calibrate "just" the 10v range , is that possible.

Any hints would be nice ....

I have the Servicemanual , and have read the Calibration section.

But i can't see if it's possible to just calibrate the 10v range , or if i have to make all calibrations.

I don't have equipment to calibrate AC or DC above 10v.

TIA

/Bingo

disclaimer: I have never calibrated an HP 3478A, but need to investigate, then calibrate my meter with the standards I can muster up, so I am interested in the process.

after doing the performance tests 4-21 (to log results for a plan of attack)

First of all, the HP 3478A has 5 ranges for DC voltage, 30mV, 300mV, 3V, 30V, 300V

0 and max for each of those ranges, so your 10V can't apply (except for the 30V range, 1/3 full scale)
 
from reading parts of the manual, you can calibrate any one of those listed ranges individually, and you have to calibrate DC voltage before DC current, then AC voltage, the machine calculates AC current from the DC voltage, DC current and AC voltage values.  Ohms is calibrated either 2 wire or 4 wire on the 30ohm, 300ohm, 3Kohm, 30Kohm, 300Kohm, 3Mohm, 30Mohm ranges (entering zero and max of range for each range pass, full scale or 1/3 scale possible, so using either 30ohm, 300ohm, 3Kohm, 30Kohm, 300Kohm, 3Mohm, 30Mohm standard resistors or 10ohm, 100ohm, 1Kohm, 10Kohm, 100Kohm, 1Mohm, 10Mohm standard resistors).

calibrating DC voltage

you would do section 4-60 from a to b

c. Set the 3478A to the single trigger mode by pressing the SGL/TRIG button.  Take the 3478A out of Autorange by pressing the downrange button.

c-1 set the 3478A to the required range, either one of 30mV, 300mV, 3V, 30V, 300V.

d. Install a short across the 3478A's HI and LO INPUT terminals (this sets the low value of the range)

e. Press the blue shift button and then the LOCAL (CAL) button.

f. Press the SGL/TRIG button once.  When the display shown CAL FINISHED, remove the short form the input. (zero for the range entered)

g.  Connect the DC Volts Standard output terminals to the 3478A INPUT terminals as shown in Figure 4-2.  Set the DC Volts Standard for the same output as selected on the HP 3478A (30mV, 300mV, 3V, 30V, 300V).

h. Press the blue shift button and then the LOCAL (CAL) button

i. If the actual output value of the DC Volts Standard is known, use the 3478A uprange or downrange buttons to set the 3478A display to the same value.  Then press the SGL/TRIG button.  If the actual output is not known or if the Standard has a high enough accuracy, just press the SGL/TRIG button. (max of the range has been entered).


i-1 (I assume CAL FINISHED is shown on the display)

repeat from c-1 for all required ranges

in red is my interpretation, go ahead, take my head off

it's my interpretation of the text I read

please feel free to correct my interpretation

[cellularmitosis]

bingo, did this work for you?  I need to do the same with my SVR-T and 3478A.

[bingo600]

I sent mine off to a Cal Lab

/Bingo

[alexwhittemore]

There are more than a few of us interested in calibrating the 3478a, I suppose it's become pretty popular on the forum, and it's still pretty cheap on ebay

Given the popularity, it seems likely one or a few of us will come up with a calibrator design that's applicable, I plan to try, though it'll likely be multiple weeks before I have time to get anywhere.

As for the ORIGNAL question, yes, you can calibrate any individual range and not the others, and you could conceivably do it at 10v, but as stated, you'd be at 1/3 of full scale on the 30v range, so it'd be a bad idea. You really want good .03/.3/3/30/300 references to do a good job.

[agaelema]

Today I found the Technical Manual, it details the calibration process. This manual has many useful information for meter maintenance.

https://www.manualslib.com/manual/1016815/Hp-3478a.html?page=45

[alexwhittemore]

Today I found the Technical Manual, it details the calibration process. This manual has many useful information for meter maintenance.

https://www.manualslib.com/manual/1016815/Hp-3478a.html?page=45

Interesting - it says "or in some cases, 1/3 full scale input" but the procedure sure looks like you can take your pick of full scale or 1/3 full scale however on each range you please. So maybe a super high quality 10V reference and accurate dividers are all you need, rather than a fancy arbitrary reference generator.

Of course, full scale should give you 1/3 the absolute error...

[Tj138waterboy]

I have recently acquired a 3478a multimeter and while going through calibration and verification testing I discovered that my unit has a 2.00000 volt calibration. Not sure if this is normal but im also having issues zeroing in on a 2.5v reference. Everything is dead nuts on using 1 and 10v scale using a 1 volt and 3 volt reference. My 2.5 volt reference reads correct on 10 volt scale but is consistently .0125 volts off on the 1 volt scale.

[ArthurDent]

I bought a half a dozen HP3478A multimeters with various problems locally for $20 each and the seller threw in the remaining one with a dead display for free. I used a number of parts from the donor one to repair some of the others. Some of the problems were dead batteries or a defective power switch. It was some work but I ended up with 6 good ones. I then had to dig into how to calibrate them and that turned out to be pretty straight forward.

If you don't have any error messages at turn-on you can just calibrate the range of the function that you know is off. If you had to replace the battery then probably all the backed up cal data is scrambled and you'll have to do at least one range for each function to get rid of the error message.

I have a stable D.C. standard that goes up to 300 volts so I used my HP3457A, for the indication as to what the output actually was. in parallel with the HP3478A under test and calibrated all the D.C. voltage ranges. I have an L&N 15A shunt that I could use with a bench supply and the HP3457A for D.C. amps then just a variable A.C. isolated supply to do the A.C. ranges that I wasn't that interested in and didn't need to be that close.

For resistance I again used the HP3457A and either some standard resistors or an ESI decade box and set the value I needed as read on the HP3457A then used that value to calibrate the range on the HP3478A. I wasn't quite as precise on the extremely high resistance ranges.

Using a higher accuracy known good meter as a transfer standard seemed to work pretty well for my needs although it isn't the best way to do it. It meant that all the standards I used didn't have to be that accurate as long as I could use the HP3457A to set the values.

[Tj138waterboy]

Just went through voltage only cal 2 more times and still same situation. I believe I read that the resistance cal may be the cause of this but it seems odd that 1 volt and 3 volt are exact matching on both 1 and 10 volt range but 2.5 is wrong on the 1 volt scale. Any chance if anyone else has applied a 2 volt reference and see that the 3478a has a calibration menu as well or do I have a one of a kind.

[Miti]

I would like to calibrate "just" the 10v range , is that possible.

Any hints would be nice ....
/Bingo

If you have a GPIB interface, you can adjust the 10V range instead of calibrating it. Take a look here:
https://www.eevblog.com/forum/repair/hp-3478a-how-to-readwrite-cal-sram/75/#lastPost

[Mp3]

Apologies for bumping an old thread, but it's already been revived a few times.

I have a bench power supply that only puts out 5v through 60v. I see you can do 1/3 range, but is there any way to use 5v or 6v for a voltage reference?

I just want to get it close enough to use at my home workbench where I really don't NEED more than one or two decimal places, so i don't have to keep seeing the "uncalibrated" warning blinking.

Edit: I've misquoted my power supply's capabilities, it seems that i'd be able to do 3v and 30v but nothing higher and nothing lower. It seems now I'm not sure how I'd manage calibrating AC, as I only have a handful of AC 12v wall warts, and can't trust the accuracy of those.
https://www.parts-express.com/7a-dc-60v-adjustable-step-down-regulator-nc-power-supply-current-voltage-meter-buck-module--320-615

[Tj138waterboy]

The first question I would ask is if your meter has uncalibrated message when you turn it on. If no uncalibrated or error message pops up when you power it on, does the voltage reading you get not look right? And lastly yes you can self calibrate with resistor divider or just put a few 2 aa batteries in series to get the 3V but in doing so without a known precision voltage I wouldnt suggest doing so. If you think one voltage range is not in tolerance, manually switch it to next higher range to cross check such as if you measure 3 volts on the 1 volt range you will get something like +2.99899, if you switch it to the 10 volt range ideally you will get +02.9990 and again on 100 volt range it should read + 003.000. They all shoul match really closely.

[Mp3]

The first question I would ask is if your meter has uncalibrated message when you turn it on. If no uncalibrated or error message pops up when you power it on, does the voltage reading you get not look right? And lastly yes you can self calibrate with resistor divider or just put a few 2 aa batteries in series to get the 3V but in doing so without a known precision voltage I wouldnt suggest doing so. If you think one voltage range is not in tolerance, manually switch it to next higher range to cross check such as if you measure 3 volts on the 1 volt range you will get something like +2.99899, if you switch it to the 10 volt range ideally you will get +02.9990 and again on 100 volt range it should read + 003.000. They all shoul match really closely.

Thank you for your reply, it does say uncalibrated when i power it on, and it blinks CAL. And thank you very much, i have other trustworthy meters though only trustworthy to 1 decimal place really, besides that i have been getting quotes from cal labs so if i am not able to get it "close enough" I will send it off. Thanks very much for your help!

[Tj138waterboy]

If im not mistaken mine would still give semi accurate voltage reading after performing the short calibration or 0 volt. You will have to use a shorting plug and manually change each voltage range. Pay close attention to what range you are working on though. But if you have a decent meter you trust, you can use the value it gives to calibrate with. Luckily I have 2 meters so I just made them agree with each other.

[NoisyBoy]

I hate to say this, but when you don't have the requisite stable calibration sources, along with another meter that you only trust up to 1 decimal place, you are not going to get a two digit decimal place calibration out of what you have. 

And before you send it off to a cal lab, find out why it lost calibration in the first place and fix it.  Otherwise you are just wasting money.

You will need to shop around for a cal lab, as many labs will probably charge you what the meter is worth to have it calibrated.