HP8903A First impressions

[27AUDIO]

Hi everyone, I recently bought a used HP8903A. It seems to work well, only when it is turned on, with nothing connected , the right display shows a slightly high value, or maybe it is normal. should it not be around 0.0070 / 0.0090mV about??

I then noticed that on the back there are 2 BNCs jumpered, if I remove this jumper the DX display indicates 0.0064mV, but it no longer measures anything.
Do all HP8903As have this jumper on the back?

attach the Pics

Thanks
S.

 

[PRR]

I can believe that a wide-open BNC jack will pick-up 34mV of hum in a room, especially with all the power-points behind it.

Build a shorted BNC plug, or short a BNC probe/cable (you will need something), and try again.

The back jacks have hand-label "filter". I assume that if you do not use/need an external/special filter, they *must* be jumpered to send the signal through.

 

[27AUDIO]

Hi PRR ,thanks for your reply,.. there is a short plug BNC in there! With 1kHz @ 1V (80khz filter on) THD is 0,0022% is this good ? Changing the PS filter capacitors can better the performance or I need to calibrate the instrument?

S.

 

[1audio]

Looks like an external filter loop. It may be switched with the internal filters. .0022% is what I would expect.

 

[warrjon]

Opt 001 are rear terminals I have not see one with this conifuration usually in that position are the High and Low outputs.

I would NOT re-cap it, i had a very old 8903A in my work station and we would NEVER open them unless they failed. The caps in these things are extremely high quality.

There should be a pullout card under the front IIRC 41.0 SPLC resets the instrument.

Then connect a cable between output high and input high switch both float switches to gnd, switch filter off and and set a frequency to say 1KHz and amplitude to 1V the levt display will read the voltage or what ever you switch for the measurement.

 

[as_audio]

Somebody made these additional BNC sockets for a special purpose. The label reads "filter throughput" in german. You see no indication of factory work. "Philips Krefeld" owners label ..

 

[27AUDIO]

I looked at where the two coaxial cables connected to the BNC . go to the RED board, Input Amplifier board.
The previous owner made the holes for the two BNC himself!



A pcb track has been cut and soldered the two cables! (in the schematic at pin 14 of the U10D)ruined printed circuit!


There are 2 keys that remain pressed, have lost their spring!



Maybe I haven't make a good deal to buy this instrument!

 

[1audio]

Have you looked into where in the circuit that link was added? it can be really useful to add a low pass filter in the input circuit for testing digital stuff. The HF in switching amps etc. can really prevent accurate readings of distortion and noise. The analog filters are too far into the circuit and the autoranging circuits will be confused. They box predates most digital audio stuff.
I have seen far worse butchery of PCB's from factory service shops. I have added and modified internal circuitry on similar instruments to keep them useful. You can do some careful opamp upgrades and some other optimizations to improve the performance. There should be some discussions sharing suggestions in the web universe.

 

[as_audio]

This mod was done probably in order to use weighting filters not included
in the 8903a. I think you should not worry too much about this.

The push buttons can be repaired or replaced, you may find knowledgeable
people here. If you are in Germany we can compare the characteristics of
your instrument with mine. How much did you have to pay ?

 

[27AUDIO]

Have you looked into where in the circuit that link was added? it can be really useful to add a low pass filter in the input circuit for testing digital stuff. The HF in switching amps etc. can really prevent accurate readings of distortion and noise. The analog filters are too far into the circuit and the autoranging circuits will be confused. They box predates most digital audio stuff.
I have seen far worse butchery of PCB's from factory service shops. I have added and modified internal circuitry on similar instruments to keep them useful. You can do some careful opamp upgrades and some other optimizations to improve the performance. There should be some discussions sharing suggestions in the web universe.

Thanks Martin , I want to restore the instrument, and now I will remove this external filter mod. anyway thanks for the valuable advice!.

as_audio:

I live in Italy , I bought it in Germany for 370 €

I was thinking of replacing only the power supply capacitors, since the display brightness is poor (compared to my HP8903E)

 

[tomchr]

Replacing the supply caps is not likely to restore display brightness.

Step 1: Get the service manual.
Step 2: Measure the ripple voltage and supply voltage on the three supplies in the instrument. If they're within spec, don't "fix" anything.

Before you remove the filter mod, I'd look at what it does. It may be useful. If you decide to remove the filter, you'll find the schematics in the service manual useful in bringing the instrument back to its original state. The BNCs for the filter mod were clearly added. They have no HP labels and aren't even aligned vertically.

This instrument is NOT an OPT001 version. It clearly has the front panel connectors.

+1 for building a shorting plug. Or use a 50 Ω terminator when measuring the noise floor. If you perform a loopback test (1 kHz sine, 1 V RMS), you should see around 0.00x % THD+N (80 kHz BW). I seem to recall mine getting down to 0.002 % (best case, 1 kHz).

Tom

 

[as_audio]

I bought it in Germany for 370 €

This price is not cheap but fair enough I think.

If you perform a loopback test (1 kHz sine, 1 V RMS), you should see around 0.00x % THD+N (80 kHz BW). I seem to recall mine getting down to 0.002 % (best case, 1 kHz).
Tom

My instrument shows 0.0035% in this case.

 

[27AUDIO]

Ok, thanks all, I will check the PSupply ripple!.

Any advice for a better display brightness?
Thanks
S.

 

[as_audio]

PS re post 1 etc the difference in "self noise" is not large here between open and shorted input : 14 and 10 μvolts.
Always observe with a scope connected to the "Monitor" output to be sure what is going on.

 

[27AUDIO]

suggestions for better the display brightness ?


Thanks
S.

 

[PRR]

Replacing the supply caps is not likely to restore display brightness.....

It "might".

If it is supposed to be a cap-input DC supply, and that cap is quite dead, the average current may be 2/3rd, or even less if LED voltage is a large part of design DC voltage.

And they will flicker.

I had a microwave suffer a lightning near-hit. After that the display was not as bright and if you waved your hand/head it "strobed" (flicker). My guess was that half the diodes went poof and it was eating half-wave instead of full-wave; but maybe the cap went poof. (This was a VFD so probably did not share a power supply with anything else in the box.)

 

[1audio]

Its quite unlikely since the audio channels would have severe hum as well.

The LED displays are old and early generation so not as bright as newer LED's. You may be able to change the current setting resistors. However the red filter in front may be quite dirty and scratched. Quickest and cheapest fix- turn down the room lights. . .

I recently upgraded the lighting in my shop and went too far. Even new monitors are washed out by the light BUT I can see everything inside electronics (and every bit of dirt and crud in the room).