Hp 438A & Hp 8481A problems.....

[SV1EOR]

I have bought my Hp438A from eBay 2 months ago and an Hp8481A sensor too with the cable.

I have a problem with that meter it didn’t recognize that it had a sensor in the cable so after some measures I

Understand it had to be changed....

I bought an A4 input Amplifier PCB from eBay too and after a long service it is with the new A4 PCB in.

But a new problem that I can't understand came up.

The sensor starts with an unstable measurement and then it goes out saying please zero in input A & input B.....

Has anyone meet such an attitude.....and solved it...??

I haven't got another sensor to use.....!!!!

https://youtu.be/T3MNp_7ASgs

[ManateeMafia]

This may not be of any help but I have repaired a 438A. The inner board, connected with the ribbon cable, has a few wet tantalums. They were leaking and caused some damage.

[MarkL]

It would help if you could tell us what you've already checked besides replacing the A4 board.  There's a bunch of writing on your schematic; I assume you've looked at a number of things.

Without knowing what you've already done, here are my initial thoughts:

Have you tried the B input?  Have you checked that the reference output is ok with a scope?  Can you zero the sensor if it's not connected to the reference output?  How about measuring a different signal source?

Have you checked all power supply rails for in-spec voltage and ripple?

Does wiggling the cable at the connectors have any effect?

There's an extensive set of troubleshooting steps in the service manual.  It's section 8 and it's 130 pages long.  Have you attempted any of those?

[SV1EOR]

Have you tried the B input?

Input A & input B behave the same

Have you checked that the reference output is ok with a scope? 

The reference output is ok with scope and counter...

Can you zero the sensor if it's not connected to the reference output?

No it can't zero connected or disconnected

How about measuring a different signal source?

It gives me the message can't zero and stops

Have you checked all power supply rails for in-spec voltage and ripple?

Voltages yes.... ripples.....no

Does wiggling the cable at the connectors have any effect?

no it hasn’t any effect

There's an extensive set of troubleshooting steps in the service manual. 

Its section 8 and its 130 pages long.  Have you attempted any of those?

I have checked with service sheet 6 on service manual (Hp Part No 00483-90039) input amplifier circuit troubleshooting:

A) Power supply checks ..................page 8-87 (ok)
B) 220Hz buffer tests......................page 8-90 (ok)
C) 220Hz multivibrator Test .............page 8-90 (ok)
D) Channel selection circuitry Test.....page 8-91 (ok)
E) Decoupled Power Supply Test .......page 8-92 (ok)

F) Amplifier Test ............................page 8-89

     On A4TP3 I have: no Ac/-0.91Vdc/220Hz

The service manual gives open cable between A4J8 and J2 but I don't have indication error 31!!!

I have opened the Hp8481A because the screen of power meter says it can't zero

to see as mentioned at page 16 service manual (Hp Part no 08481-90049) the white-black wire and the white brown wire

But I can't understand if the problem is on power meter or in the power sensor.....

I don't have the Range calibrator Hp 11683A so ......I'm confused.....

[SV1EOR]

This may not be of any help but I have repaired a 438A. The inner board, connected with the ribbon cable, has a few wet tantalums. They were leaking and caused some damage.

The Wet tantalums are in good condition as I can see I don’t think that there is the problem ...

[SV1EOR]

After some new reading and testing.....

I have found the extended test error messages

And more I have check the input 220 Hz gate drive in the sensor in order to see if the sensor is off

or the power meter as the sensor manual says and I found input in the sensor:

White-black (POS-PIK:-0.02V/NEG-PK:-10.69V/220.2Hz)

White-Broun (POS-PK 0.00V/NEG-PK:-10.81V/220.2Hz)

The manual says-9V is it ok?

(It's not but there isn't any regulator ...and 220Hz multivibrator  & 220Hz Buffers works...)

[MarkL]

After some new reading and testing.....

I have found the extended test error messages

And more I have check the input 220 Hz gate drive in the sensor in order to see if the sensor is off

or the power meter as the sensor manual says and I found input in the sensor:

White-black (POS-PIK:-0.02V/NEG-PK:-10.69V/220.2Hz)

White-Broun (POS-PK 0.00V/NEG-PK:-10.81V/550.2Hz)

The manual says-9V is it ok?

(It's not but there isn't any regulator ...and 220Hz multivibrator  & 220Hz Buffers works...)

Are you sure the White/Brown is 550.2Hz?  It should be the same frequency as White/Black, except 180 degrees out of phase.

I have a 438A here.  I will check voltage levels in a bit to compare with yours.  I will do it from the meter side, however, since opening the sensor voids the calibration and I don't want to do that to any of my sensors.

The manual does not say which of White/Black and White/Brown wires is connected to cable pins "G" and "H".  Since you have your sensor open, can you post a pinout of colors vs. connector pin letters?

[SV1EOR]

White-Broun (POS-PK 0.00V/NEG-PK:-10.81V/220.2Hz) it’s my fault.....sorry....   (I HAVE CORECT IT IN MY POSTING)

WHITE-BLACK   IN  H (WAITING SIDE)   AND PIN 1 IN THE A4J8 & A4J9 SOCKET
WHITE-BROWN  IS G  (WAITING SIDE)   AND PIN 3 IN THE A4J8 & A4J9 SOCKET

[MarkL]

...
F) Amplifier Test ............................page 8-89

     On A4TP3 I have: no Ac/-0.91Vdc/220Hz

The service manual gives open cable between A4J8 and J2 but I don't have indication error 31!!!

Ok, let's see what we can do without a range calibrator.

First, connect the front panel reference output to a scope and make sure you have a good signal.  Into 50 ohms you should see 0.632Vpp @ 50MHz.  We will use this as our test signal.

Now plug the sensor into the reference output.   You should see on A4TP3 a DC signal of -0.9V +/- 0.2V, and an AC signal of 134mV pk-pk.  At least that's what I get with my 8481A sensor.  It's a very noisy signal at this point, so use BW limit and/or averaging, and reject HF triggering on the scope.

If you turn the reference source off you should see the AC signal go to 0V, but it will still be at -0.9V DC.

Regarding the measurements on the White/Black and White/Brown, my reading of the schematic says you should see the SAME square wave but shifted 180 degrees from one to the other.  In other words, the service manual says this:

• At the black/white wire: -0.05 ± 0.05 Vdc (top of square wave).
• At the brown/white wire: -9Vdc (bottom of square wave).

But I think it really means that you should see the same magnitude square wave on both.  And I wouldn't worry if it's not exactly -9V.  It may be less because it just needs to be negative enough to fully turn off the JFETs in the sensor's U1.  Mine goes to -10.48V on both pins H and G.

[SV1EOR]

First, connect the front panel reference output to a scope and make sure you have a good signal. 

Into 50 ohms you should see 0.632Vpp @ 50MHz.  We will use this as our test signal.

Ok in the first & second pic You can see the ref out connected to 50 ohm input of the scope & the counter

[SV1EOR]

Now plug the sensor into the reference output. You should see on A4TP3 a DC signal of -0.9V +/- 0.2V,

You can see A4TP3 Dc signal

and an AC signal of 134mV pk-pk.  At least that's what I get with my 8481A sensor. 

It's a very noisy signal at this point, so use BW limit and/or averaging, and reject HF triggering on the scope.

with the ref output on at A4TP3 I can meter 0.190 Vac p-p

[SV1EOR]

If you turn the reference source off you should see the AC signal go to 0V, but it will still be at -0.9V DC.

and with ref output off at A4TP3 I can meter 0.187 Vac p-p  and not 0Vac

when I tryed to zero A chanel look at the video......

https://youtu.be/yBBPmQk0_CI

But at tfe end it can't zero!!!

[SV1EOR]

Regarding the measurements on the White/Black and White/Brown,

my reading of the schematic says you should see the SAME square wave but shifted 180 degrees

from one to the other. 

I think that the one of the two pins has biger magnitude and not the same because the one JFET in the sensor's U1  is not working well!!!

[SV1EOR]

In other words, the service manual says this:

But I think it really means that you should see the same magnitude square wave on both. 

And I wouldn't worry if it's not exactly -9V. 

It may be less because it just needs to be negative enough to fully turn off the JFETs in the sensor's U1. 

Mine goes to -10.48V on both pins H and G.


Ok it has to be bigger than -9.0V in order to turn off the JFETs in U1.

But if the one JFET is not working well it takes more current and gives lower voltage in the buffers of 220Hz multivibrator,

 and when Hp 438A trays to zero it can't take the lower voltage prices as it says in "Extended Test Messages"
 

[MarkL]

Ok, from your video it appears that the meter is trying to zero.  That would tend to say the zeroing DAC and related circuits are operable.

If you wanted to further confirm that, you can measure the voltage between connector pin L (Sensor Auto Zero) and pin E (Sensor Ground Reference) while it is trying to zero.  The voltage can vary between +15mV and -15mV (nominal).  Mine varies from -12mV to +17mV.  Yours will probably settle at one extreme or the other since it is unable to zero.

If you have access to a GPIB interface, you can send the meter "ZDA0EN" and "ZDA255EN" commands to manually force the zeroing DAC to the two extremes when in test mode (as described on page 8-53).  You should be able to set roughly +/- 15mV.


On your scope photo of the two 220Hz chopper waveforms, it looks like they have roughly the right levels.

What am I looking at in the second photo (PICT0092_mini.JPG)?  Please always label or describe your waveform photos!

If you think one of the chopper signals is being loaded down by a bad JFET gate, you can unplug the sensor while you're watching the two chopper signals from the meter side.  The waveforms should be unchanged since the JFET gates should not be loading either signal.  If one or both does change significantly you are likely right you have a bad JFET in sensor U1.  If the JFET gate is shorted, you should see a drop created by the voltage divider formed by one of the output resistors on U5 of 215R and 1k on the input of U1.  On the -10V swing it would be around 10*215/(1k+215) = 1.77V, but I didn't see that in your previous measurement posts.

Or... If all of this seems to be ok, I think at this point I would proceed with the sensor troubleshooting steps starting on page 41 of the Agilent Sensor Operating and Service Manual (pub #08481-90173).

If the thermocouple has the right resistance (200 Ohms +/- 10 Ohms), you could try shorting U1 between pins 1 and 9 and try zeroing the meter.  Or you could also apply a short at the location where the "gold wires" were connected.  Both of these short the input.  Use your judgment on mechanically what's possible - I'm at a disadvantage here since I've never had one of these sensors open.  Perhaps someone who has taken one apart can offer further advice.

If you can get a short in place, the meter should be successful in zeroing.  If it still won't zero, we need to keep digging.

[SV1EOR]

If you wanted to further confirm that, you can measure the voltage between connector pin L (Sensor Auto Zero) and pin E (Sensor Ground Reference) while it is trying to zero.  The voltage can vary between +15mV and -15mV (nominal).  Mine varies from -12mV to +17mV.  Yours will probably settle at one extreme or the other since it is unable to zero.

The voltage was varying between +20mV and -11mV but I'm not sure I have meter it correctly with a dvm

I haven’t got a GPIB interface to send the commands…

If you think one of the chopper signals is being loaded down by a bad JFET gate, you can unplug the sensor while you're watching the two chopper signals from the meter side.  The waveforms should be unchanged since the JFET gates should not be loading either signal.  If one or both does change significantly you are likely right you have a bad JFET in sensor U1.  If the JFET gate is shorted, you should see a drop created by the voltage divider formed by one of the output resistors on U5 of 215R and 1k on the input of U1.  On the -10V swing it would be around 10*215/(1k+215) = 1.77V, but I didn't see that in your previous measurement posts.

I don’t see any change in the chopper signals when I disconnected the sensor so it isn’t loading the Buffer no one of the JFET's gate.

Or... If all of this seems to be ok, I think at this point I would proceed with the sensor troubleshooting steps starting on page 41 of the Agilent Sensor Operating and Service Manual (pub #08481-90173).

If the thermocouple has the right resistance (200 Ohms +/- 10 Ohms),

I can’t meter a resistant on the gold wires it looks like an open circuit.

You could try shorting U1 between pins 1 and 9 and try zeroing the meter.  Or you could also apply a short at the location where the "gold wires" were connected.  Both of these short the input.  Use your judgment on mechanically what's possible
If you can get a short in place, the meter should be successful in zeroing.  If it still won't zero, we need to keep digging.

I have short at the location where the "gold wires" were connected pict0092_mini.jpg and started zero A but it didn’t zero error 01.

After that I have meter the voltage at A2Q1 and it was -0.0683 Vdc tha means the first amplifier is working well (service manual page 42 #08481-90173)

[MarkL]

...
I can’t meter a resistant on the gold wires it looks like an open circuit.

This says the thermocouple in the A1 bulkhead is open.  This is the #1 failure mode for these sensors and is caused by too much input power.  The A1 assembly is unfortunately not repairable.

The A1 assembly 08481-60042 and alternate 08481-60009 are both obsolete.  I don't think I've ever seen them on ebay or elsewhere.  And even if you could find one and replace it, the sensor would require recalibration which would be additional cost.

At this point, I would start looking for a replacement sensor.  If you're looking on ebay, make sure you get one that is "Used" since according to ebay's definition of that category it's supposed to be working.  That will give you a chance to receive it, test it, and return it if it's not working.  You can sometimes find a bargain from someone selling one that is "Used" but untested.

I don't know what power levels you're looking to measure, but if you can find a good deal on a diode based sensor such as an 8481D, you could buy that instead along with some quality attenuators.

And one other note when shopping for sensors: Make sure it has the calibration chart on the sensor.  If you don't see it in a photo, ask.  For some reason, many sensors I've seen have the chart ripped off or defaced and this will require you to get the sensor calibrated to regenerate the chart.

You could try shorting U1 between pins 1 and 9 and try zeroing the meter.  Or you could also apply a short at the location where the "gold wires" were connected.  Both of these short the input.  Use your judgment on mechanically what's possible
If you can get a short in place, the meter should be successful in zeroing.  If it still won't zero, we need to keep digging.

I have short at the location where the "gold wires" were connected and started zero A but it didn’t zero error 01.

I would expect the meter to still be able to zero.  So either there is still a problem, or it may be the long leads you are using are either introducing too much noise, or too much offset from thermal effects.

The range for zeroing at the input is extremely small.  With the thermocouple intact, +/-15mV from the meter through A2R1 404k, and then through the thermocouple 200R + U1R1 15R results in +/-8uV.  With a test short the range is even smaller.

At this point you've already determined the sensor is unquestionably dead.  I would find a replacement sensor before concluding anything about the meter.

If you still want to see the meter zero, you could try holding a 200R SMD resistor across the input pads with a toothpick.  There's not much you can do to make the sensor any worse at this point.

[SV1EOR]

  There's not much you can do to make the sensor any worse at this point.