[SOLVED]HP/AGILENT/KEYSIGHT 34401A 2W Ohm Problem

[coldframe]

When I measure 2-wire ohms I get an offset in the reading.
 
For example.
I will get readings from 31 Ohms in 2-wire mode when measuring a 10 ohm resistor.
* In 4-wire ohms mode,  an accurate stable reading of 10.0 ohms.
* This is my 10 Ohm resistor Information  http://www.te.com/catalog/pn/en/1624323-4

This resistor measures a stable 10.0 ohms on a known good meter.
Any Idea?

Best regards.

[free_electron]

most likely the output protection cascade for the meter has taken a hit from overvoltage.

in 4 wire mode this cascade is bypassed and the current sources feed the 4terminal output directly.
the cascade is a chain of 6 or 7 transistors. most likely one of em is fried.

look in the schematic it is very easy to find the cascade

[coldframe]

most likely the output protection cascade for the meter has taken a hit from overvoltage.

in 4 wire mode this cascade is bypassed and the current sources feed the 4terminal output directly.
the cascade is a chain of 6 or 7 transistors. most likely one of em is fried.

look in the schematic it is very easy to find the cascade

free_electron..

I have often heard your name., I'm honored to meet you.

I'll get it checked soon.
Thank you for the tip.

*  most likely one of em is fried. , What does the "em" stand for?

[Fluxed Matter]

That's slang in the US for them.

[SeanB]

"em" means them. Replace all the transistors with new from the same batch and you will only have to test them to see which was faulty. They are cheap.

[coldframe]

That's slang in the US for them.

Ah.. I Got it. Thanks

[coldframe]

"em" means them. Replace all the transistors with new from the same batch and you will only have to test them to see which was faulty. They are cheap.

Okay, I'll do that.
SeanB Thanks for the help.

[free_electron]

yep , you got the correct block in the schematic

[Dr. Frank]

Most likely, the front/rear switch S1 on p.9-8 (pdf-page 150) of the service manual is faulty, i.e. either S1-A or S1-G is high ohm (20Ohm?). In 4 Ohm mode, any additional serial resistors in the current path are eliminated (that's the purpose of 4W ohm), so the reading is correct again.

You can make a very simple check, by switching from front to rear and using the rear jacks in 2W Ohm mode.

I predict, the reading will be okay.
In this case the switch needs to be replaced, or desoldered and cleaned internally - if you dare to make such a repair.

You can measure this contact resistance of 20 Ohm with an additional DVM, if you open the case, switch to front panel again and measure the resistance across the named switches. Normally, they have to be < 100mOhm.
Or perhaps a solder joint at the switch is faulty, that can be detected by measuring resistance from H or L input to the outputs of those switches, behind the switch, on the PCB, see schematics.

Similar errors already have been reported by several other users.
And my own 34401A shows the same behaviour, if I use the rear jacks.


Remark:
It is not true, that the high voltage protection transistor-bank Q203-Q210 is not active in 4W Ohm mode.
Wouldn't make sense, as the current source in 4W Ohm mode has to be protected the same way as in 2W mode
And the schematic clearly shows, that there is no switch to bypass it.

Conclusion: If 4WOhm is working correctly, the protection circuitry definitely is working properly.

Frank

[coldframe]

Most likely, the front/rear switch S1 on p.9-8 (pdf-page 150) of the service manual is faulty, i.e. either S1-A or S1-G is high ohm (20Ohm?). In 4 Ohm mode, any additional serial resistors in the current path are eliminated (that's the purpose of 4W ohm), so the reading is correct again.

You can make a very simple check, by switching from front to rear and using the rear jacks in 2W Ohm mode.

I predict, the reading will be okay.
In this case the switch needs to be replaced, or desoldered and cleaned internally - if you dare to make such a repair.

You can measure this contact resistance of 20 Ohm with an additional DVM, if you open the case, switch to front panel again and measure the resistance across the named switches. Normally, they have to be < 100mOhm.
Or perhaps a solder joint at the switch is faulty, that can be detected by measuring resistance from H or L input to the outputs of those switches, behind the switch, on the PCB, see schematics.

Similar errors already have been reported by several other users.
And my own 34401A shows the same behaviour, if I use the rear jacks.


Remark:
It is not true, that the high voltage protection transistor-bank Q203-Q210 is not active in 4W Ohm mode.
Wouldn't make sense, as the current source in 4W Ohm mode has to be protected the same way as in 2W mode
And the schematic clearly shows, that there is no switch to bypass it.

Conclusion: If 4WOhm is working correctly, the protection circuitry definitely is working properly.

Frank

Dr. Frank
Oh, I see. Let me check.
Thanks for the update.
I am so appreciative of your sharing your a work knowledge
Thanks for a detailed explanation.

P.S
Yes. 4WOhm is working correctly.
Vielen Dank   

[Dr. Frank]

Just checked my 24 years old instrument.
Simply shorted Hi/Lo on front and rear.
Initially measured 0.15 Ohm on the front.

Still has this 2W Ohm problem on the rear, when I actuated the REAR switch..
Reading is unstable, jumping between 1..40 Ohm.

Now, FRONT reading is unstable also... same error.. definitely switch.. oxydized contacts perhaps..

Next construction site for me, also...

Welcome in the club..

Frank

Update: Have opened my 34401A, and found root cause, as expected: Switch S1 , Hi contacts S1A are the "stinkers".
If I short those pins directly on the topside of the switch, namely pins #1-#2 for front or #2-#3 for rear, the readings become rock stable again at around 0.5 Ohm rear, 0.1 Ohm front. Otherwise several ten Ohms and jumping around.

Switch looks as if it can be opened, disassembled  and cleaned.

Follow-up soon.

[free_electron]

oops. brainfart on my part. DrFrank is right.

another thing to check while you got the machine open is the big 2512 resistors close to the big switch. i have seen those develop cracks on the electrodes. simply resolder them.

[Dr. Frank]

oops. brainfart on my part. DrFrank is right.

another thing to check while you got the machine open is the big 2512 resistors close to the big switch. i have seen those develop cracks on the electrodes. simply resolder them.

You mean those 2 x 6 x 24k ones?
You mean cracks in the solder joint?
Mine still look very well under magnification, but I'll check also.

Thanks!

[Dr. Frank]

Ok, here's an intermediate result and an obvious root cause found.

I desoldered the switch completely, (quite easily done) and measured all contacts.
Found 2 out of 8 switches which had higher resistance of several Ohms.

Then I carefully pulled out the lever inside a small box, because I ended up with 8 very tiny springs and 8 tiny wipers.

During pulling out and testing different wipers on the faulty contacts, I identified, that 2 of the wipers were the problem, not those 6 out of the 24 stator pins.

The stator pins were glued to the case, at its bottom and at the top  plate, with some sort of epoxy, which already showed signs of decomposition.
Inside the plastic case, small pieces of debris could be seen.
The debris originates quite obviously from the epoxy.
The surfaces of the stators seemed clean, as far I was able to look inside.
I could also see some abrasive dust (correct term?) originating from the lever.
So I also cleaned the inside with a cotton bud.

When I looked closer at the gold plated wipers, I noticed a faint film on the wipers, which can be carefully rubbed of with a dry, soft cotton bud.
This film probably caused the bad contact.
Otherwise, there were absolutely no signs of oxidation or abrasion of the gold surfaces.

My explanation for that film is evaporation of plasticiser from the case plastics, or from the epoxy.
A normal process for 25 years old stuff like that..

(I have seen Buzz Aldrins lunar space suit in the Smithsonian's Air & Space Museum, and all the plastics of that historic piece were in the process of disintegration, without any chance of holding that up, as it was mentioned there, I think)

Tomorrow morning, I will clean all wipers, and will try to reassemble those small puzzle pieces. Break-Before-Make wipers and Make-Before-Break wipers.... oh Dear!
Sometimes, small springs spring far away...

If I am successful, I will post a few pictures.

Good night - Frank

[coldframe]

Ok, here's an intermediate result and an obvious root cause found.

I desoldered the switch completely, (quite easily done) and measured all contacts.
Found 2 out of 8 switches which had higher resistance of several Ohms.

Then I carefully pulled out the lever inside a small box, because I ended up with 8 very tiny springs and 8 tiny wipers.

During pulling out and testing different wipers on the faulty contacts, I identified, that 2 of the wipers were the problem, not those 6 out of the 24 stator pins.

The stator pins were glued to the case, at its bottom and at the top  plate, with some sort of epoxy, which already showed signs of decomposition.
Inside the plastic case, small pieces of debris could be seen.
The debris originates quite obviously from the epoxy.
The surfaces of the stators seemed clean, as far I was able to look inside.
I could also see some abrasive dust (correct term?) originating from the lever.
So I also cleaned the inside with a cotton bud.

When I looked closer at the gold plated wipers, I noticed a faint film on the wipers, which can be carefully rubbed of with a dry, soft cotton bud.
This film probably caused the bad contact.
Otherwise, there were absolutely no signs of oxidation or abrasion of the gold surfaces.

My explanation for that film is evaporation of plasticiser from the case plastics, or from the epoxy.
A normal process for 25 years old stuff like that..

(I have seen Buzz Aldrins lunar space suit in the Smithsonian's Air & Space Museum, and all the plastics of that historic piece were in the process of disintegration, without any chance of holding that up, as it was mentioned there, I think)

Tomorrow morning, I will clean all wipers, and will try to reassemble those small puzzle pieces. Break-Before-Make wipers and Make-Before-Break wipers.... oh Dear!
Sometimes, small springs spring far away...

If I am successful, I will post a few pictures.

Good night - Frank

Problem SOLVED!!, All is back to its normal state

Be careful.  wipers!
They are not all the same, 2 wipers little bit difference 
Check out Attachment (Bottom 2  pieces)

Well, all right. I guess we both need some rest
I wish you all the best

Good night
Danke Dr. Frank

PS.
I checked  the PNP Transistors, P,N J-Fet one by one, There is nothing wrong
This is mechanical faults.

[coldframe]

oops. brainfart on my part. DrFrank is right.

another thing to check while you got the machine open is the big 2512 resistors close to the big switch. i have seen those develop cracks on the electrodes. simply resolder them.

Problem Solved! , All is back to its normal state

I'm very honored to meet you.
Thank you free_electron.

[Dr. Frank]

Well, I was also successful repairing my old 34401A.
And here are some more photos, as I assume, that this is a typical error of vintage 34401A instruments.
Many owners will have to do that easy repair.

The first photo shows, how to disassemble and reassemble the switch, taking care not to lose any of those tiny springs.
You can also see the epoxy bead, with signs of disintegration
All parts are displayed, in the box.
In the 3rd photo, the faint film is visible, contrasting the shining gold plated surface.
The surface has to be cleaned with cotton sticks and alcohol to avoid any scratches.
I also cleaned the interior of the switch assembly with alcohol and a thinned cotton stick

After reassembly, all contacts had less than 15 mOhm again. Before, 2 of them had up to 30 Ohm (unstable).

The best contacts now have 3 mOhm.

Afterwards, I had to recalibrate the instrument, especially the Zero calibration.

The polluted wipers had spoiled  a foregoing calibration totally.

Now, in 2W Ohm mode, Hi/Lo-inputs shorted, front and rear measurements show 3 mOhm maximum resistance; that is the variance of the switch contacts after actuating the front/rear switch.
 

[Kjelt]

Would an IPA bath not have taken care of it also without disassembling the switch?
Saves a lot of time but still nice to see the craftmanship at work.

[GreyWoolfe]

Congratulations on your repair.

[Dr. Frank]

Would an IPA bath not have taken care of it also without disassembling the switch?
Saves a lot of time but still nice to see the craftmanship at work.

IPA = iso propyl alcohol.

Well, I doubt that.

The film sticks very badly on the contacts, in a way that multiple actuation of the front/rear switch does not heal the problem at all.

The film can be mechanically rubbed off completely by a cotton swab already.

I could not determine, if the film dissolves in alcohol; I used that only to be sure to remove any grease like stuff, and to have a wet interface between the cotton and the surfaces, which would evaporate easily.

No, it's really safer to disassemble and clean mechanically.

Frank

[coldframe]

Would an IPA bath not have taken care of it also without disassembling the switch?
Saves a lot of time but still nice to see the craftmanship at work.

Hi, Kjelt
You right, this is time-consumer, but I enjoy this time of the year
in my case,This is mechanical faults
The best thing to do would be to disassemble/clean(polish)/assembly or replaced worn-out switch
without disassembling the switch?, I'm not sure it worked.
take the Instrument to the Local Service Center, Then you can save time and be more productive
I say this out of kindness
This is Not lawnmower, Do not add WD-40 Something

[coldframe]

Well, I was also successful repairing my old 34401A.
And here are some more photos, as I assume, that this is a typical error of vintage 34401A instruments.
Many owners will have to do that easy repair.

The first photo shows, how to disassemble and reassemble the switch, taking care not to lose any of those tiny springs.
You can also see the epoxy bead, with signs of disintegration
All parts are displayed, in the box.
In the 3rd photo, the faint film is visible, contrasting the shining gold plated surface.
The surface has to be cleaned with cotton sticks and alcohol to avoid any scratches.
I also cleaned the interior of the switch assembly with alcohol and a thinned cotton stick

After reassembly, all contacts had less than 15 mOhm again. Before, 2 of them had up to 30 Ohm (unstable).

The best contacts now have 3 mOhm.

Afterwards, I had to recalibrate the instrument, especially the Zero calibration.

The polluted wipers had spoiled  a foregoing calibration totally.

Now, in 2W Ohm mode, Hi/Lo-inputs shorted, front and rear measurements show 3 mOhm maximum resistance; that is the variance of the switch contacts after actuating the front/rear switch.

Congratulations!!
Did you Zero calibration?
according to Service manual (Page 77)  "4-wire short (copper) across"
What are you using to Zero calibration the short-bar?

Writing about a topic helps you integrate new knowledge with what you already
And this, obviously, is what Most 34401A users really want
You are proper person for the work ^^;
Dr. Frank, Post New Article Please 

Congratulations again

[robrenz]

Excellent place for Deoxit G100L a very good gold lube. I rub it in well and then wipe liquid off with a clean Kimwipe.

[coldframe]

Excellent place for Deoxit G100L a very good gold lube. I rub it in well and then wipe liquid off with a clean Kimwipe.

robrenz,
Thanks for valuable information

This is my solution ^^
I rub superfine fibres and wipe cotton

[Dr. Frank]

Congratulations!!
Did you Zero calibration?
according to Service manual (Page 77)  "4-wire short (copper) across"
What are you using to Zero calibration the short-bar?

Yep, I had to.
After repair, zero reading on the rear panel was 0.59 ohm, caused by the faulty switch and in turn wrong zero cal.

I soldered a real low ohm (<< 1 mOhm) and low thermal voltage (<< 1 µV) zero ohm bridge, see photo.

The bent copper wire on the left serves the same purpose, but for my HP3458A.

Frank

[coldframe]

a problem comes up

Hi,

It's all fine, except it's got a fault when measuring ohm with 2-wires again

For example.
I will get readings from 9.5 Ohms Value when Power On
But, Over time, Increase Ohm Value 10.7 Ohm
See attached.

*  When using 4-wires-measurement everything's alright.

Did anyone have that problem before?
Thank you

Coldframe

[coldframe]

Hi, Dr. Frank

How about you?
A problem has been encountered while measuring the Ohm
(increasing steadily Ohm Value)
Where do you think the main problems occur?

I'd be happy to hear what your thoughts are on this matter.

Regards, coldFrame

[robrenz]

a problem comes up

Hi,

It's all fine, except it's got a fault when measuring ohm with 2-wires again

For example.
I will get readings from 9.5 Ohms Value when Power On
But, Over time, Increase Ohm Value 10.7 Ohm
See attached.

*  When using 4-wires-measurement everything's alright.

Did anyone have that problem before?
Thank you

Coldframe

You may just be seeing the TC of the resistor as it warms up from the drive current of the meter.

[coldframe]

You may just be seeing the TC of the resistor as it warms up from the drive current of the meter.

Hi, robrenz
What does the "TC" stand for?
Could you be more specific?

Thanx!

[robrenz]

Temperature Coefficient.  In this case I implied TCR for Temperature Coefficient of resistance which is ppm change in resistance per degree change of resistor temperature.

[coldframe]

Temperature Coefficient.  In this case I implied TCR for Temperature Coefficient of resistance which is ppm change in resistance per degree change of resistor temperature.

Oh.. I see
Thank you for the explanation
So, This resistor measures a stable ohms on a known good meter.
* Temperature Coefficient (ppm/°C) ±5

[Dr. Frank]

a problem comes up

Hi,

It's all fine, except it's got a fault when measuring ohm with 2-wires again

For example.
I will get readings from 9.5 Ohms Value when Power On
But, Over time, Increase Ohm Value 10.7 Ohm
See attached.

*  When using 4-wires-measurement everything's alright.

Did anyone have that problem before?
Thank you

Coldframe

Don't understand your new problem, sorry. And my instrument is now rock stable in Ohm again, right from the beginning.


Zero Ohm reading is now ok, in 2W Ohm, in the beginning, and a short also reads << 1Ohm after warm-up / later?

Then you measure a 9..10 Ohm resistor, which reads always the same over time, when you use 4W Ohm?

And only if you measure the same resistor in 2W mode, then the reading will go up over time?

What kind of resistor do you measure?

Frank

[robrenz]

Here is a thought,  In 2 wire mode even if you zero out the lead resistance you are seeing the TCR of the resistor and the copper leads. Also since the connection points are carrying current and measuring voltage drop there will probably be a greater thermal emf generated. Reversing the connection and seeing the difference in readings will give a clue as to how much thermal emf is involved.

I am not saying that there isn't something wrong with the meter but you have to consider the above points.

[coldframe]

Don't understand your new problem, sorry. And my instrument is now rock stable in Ohm again, right from the beginning.


Zero Ohm reading is now ok, in 2W Ohm, in the beginning, and a short also reads << 1Ohm after warm-up / later?

Then you measure a 9..10 Ohm resistor, which reads always the same over time, when you use 4W Ohm?

And only if you measure the same resistor in 2W mode, then the reading will go up over time?

What kind of resistor do you measure?

Frank

Hi, Dr. Frank

Yes, When using 4-wires-measurement everything's alright.
Have you seen the Attatched Picture?
Only 2-Wires Over time, continuous increase

This is Resistor Specification.
http://www.te.com/catalog/pn/en/1624323-4

Do you have any good ideas?

Thank you
coldframe

[coldframe]

Here is a thought,  In 2 wire mode even if you zero out the lead resistance you are seeing the TCR of the resistor and the copper leads. Also since the connection points are carrying current and measuring voltage drop there will probably be a greater thermal emf generated. Reversing the connection and seeing the difference in readings will give a clue as to how much thermal emf is involved.

I am not saying that there isn't something wrong with the meter but you have to consider the above points.

Please understand,  I'm Not native English speaker

2-Wire
Attachment #1 Small Resistor Box
Attachment #2 Start
Attachment #3 10 Minute Later

4-Wire
Attachment #4 Stable

even now, on the increase....
what seems to be the trouble?   

Thanx

[Dr. Frank]

Well, please make a cross check over time, with a solid short in 2W mode.
Should read <0.2 Ohm or less, and be stable over time.
That's what I expect strongly, as your instrument should be ok again, after repair.

The resistor has got 5ppm/K, so no chance that the drift of 0.2Ohm in your picture comes from heating up.. that would require 4000°C.

You obviously simply wrapped the leads of the resistor around the posts- that makes no good contact at all.
This is the most probable root cause.
Please solder that resistor to the posts.

Perhaps the jacks themselves do not make contact very well.
That could be the reason, why the 4W measurement gives stable results... looks strange,, anyhow.

4W measurements give always smaller values than 2W measurements.. So test your instrument with a short first..

As you have a low Ohm resistor, you should use 4 jacks for the resistor box anyhow, i.e. 4 pole Kelvin connection.
All the contact resistances kill the accuracy of your resistor.

Frank

[coldframe]

Well, please make a cross check over time, with a solid short in 2W mode.
Should read <0.2 Ohm or less, and be stable over time.
That's what I expect strongly, as your instrument should be ok again, after repair.

The resistor has got 5ppm/K, so no chance that the drift of 0.2Ohm in your picture comes from heating up.. that would require 4000°C.

You obviously simply wrapped the leads of the resistor around the posts- that makes no good contact at all.
This is the most probable root cause.
Please solder that resistor to the posts.

Perhaps the jacks themselves do not make contact very well.
That could be the reason, why the 4W measurement gives stable results... looks strange,, anyhow.

4W measurements give always smaller values than 2W measurements.. So test your instrument with a short first..

As you have a low Ohm resistor, you should use 4 jacks for the resistor box anyhow, i.e. 4 pole Kelvin connection.
All the contact resistances kill the accuracy of your resistor.

Frank

All right, I will try immediately!
Thank you for the tip.

coldframe

[Dr. Frank]

Ups.. don't you own a simple, short lab cable with decent plugs, to short Hi and low in 2W mode???

Frank

[coldframe]

Ups.. don't you own a simple, short lab cable with decent plugs, to short Hi and low in 2W mode???

Frank

Dr. Frank
My apologies. I don't understand what you mean.

[Dr. Frank]

Ups.. don't you own a simple, short lab cable with decent plugs, to short Hi and low in 2W mode???

Frank

Dr. Frank
My apologies. I don't understand what you mean.

I asked you to make a test in 2W Ohm, shorting High and Low with a simple laboratory cable with good plugs.

What you are showing, looks not so good.

If you simply do the test, you can see, if the 2W mode is working correctly and if the zero is properly calobrated, after your repair.

Frank

[coldframe]

I asked you to make a test in 2W Ohm, shorting High and Low with a simple laboratory cable with good plugs.

What you are showing, looks not so good.

If you simply do the test, you can see, if the 2W mode is working correctly and if the zero is properly calobrated, after your repair.

Frank

I am sorry for the delay in my response.
Crash my Laptop, I have sent my Laptop be mended.

Would you Read The Personal Message?

Thanx

[Dr. Frank]

Hello Coldframe,

sorry, it took me a while to shoot "speaking" photos (I hope so).

Just repeat your measurements, like I show you in the pictures.
Use 6 digit, slow mode.

Pic 1: in 2W mode, short input with a decent (good) cable. Check, if reading is a few milli - Ohm.

(See my 34401A, it still has 22mOhm, although it has been calibrated after the repair.
The switch is clean now, but worn-out. That still leads to changing readings, when using front/rear switch,  but now several mOhm only, and that does not disturb the readings)

Pic 2: Use Null function to remove the offset.

Pic 3: The 10 Ohm resistor is now measured correctly in 2W mode

Pic 4: 4W mode does not need Null math. The reading is even smaller compared to 2W Ohm mode with Null.

Pic 5: The resistor is soldered, that is better than clamping. Small resistor values always need 4 pole connection.

Please tell me, what you have measured.
Good luck.

Frank

[coldframe]

Hello Coldframe,

Pic 1: in 2W mode, short input with a decent (good) cable. Check, if reading is a few milli - Ohm.

(See my 34401A, it still has 22mOhm, although it has been calibrated after the repair.
The switch is clean now, but worn-out. That still leads to changing readings, when using front/rear switch,  but now several mOhm only, and that does not disturb the readings)

Pic 2: Use Null function to remove the offset.

Pic 3: The 10 Ohm resistor is now measured correctly in 2W mode

Pic 4: 4W mode does not need Null math. The reading is even smaller compared to 2W Ohm mode with Null.

Pic 5: The resistor is soldered, that is better than clamping. Small resistor values always need 4 pole connection.

Please tell me, what you have measured.
Good luck.

Frank

Hi Dr Frank,

This is My Test Result, the Order are the same.

First of all,  press the button for the 2W Ohm and slow 6 Digit Mode Setting (A: MEAS MENU)

Pic 1: 0.09Ohm, Sometime 0.2Ohm... variable

Pic 6: When I Measure 2W Ohm Reading Value Increse  as time passes, This is My Problem
* This is Not Agilent, old HP
 
Thank you for your consideration.
Danke schön!

[Dr. Frank]

Ok, then your instrument still has a BIG problem with the front/rear switch.

You can test that, if you short the input in 2W mode, and engage (press) the front / rear switch.
After returning to "front", the reading will increase strongly, perhaps several ohm already, I assume.

Then try pressing this button a little bit, not completely, you will see the zero-reading change a lot.

I have the same problem with my cleaned switch, but not so strong.

If nothing helps, order a spare switch from agilent.

PS: The "old" HP 34401A are still good, but switches are always a problem.

The HP3458A has a similar switch, same supplier, same construction, it's about 13 years old, and  it shows also increased and unstable resistance on the 2W Ohm mode, but not yet so strong, also.

Frank

[coldframe]

Ok, then your instrument still has a BIG problem with the front/rear switch.

You can test that, if you short the input in 2W mode, and engage (press) the front rear switch.
After returning to "front", the reading will increase strongly, perhaps several oohm already, I assume.

Then try pressing this button a little be, not completely, you will see the zero-reading change a lot.

I have the same problem with my cleaned switch, but not so strong.

If nothing helps, order a spare switch from agilent.

PS: The "old" HP 34401A are still good, but switches are always a problem.

The HP3458A has a similar switch, same supplier, same construction, it's about 13 years old, and  it shows also increased and unstable resistance on the 2W Ohm mode, but not yet so strong, also.

Frank

Yeah, I guess so.
After change a part, I'll let you know the result

Thank you Sir.

P.S
I have not one but seven 34401A
This is My Fisrt, Oldest, Favorite, Still Good Condition Except 2W Ohm

[coldframe]

Problem has been solved

This is a complicated problem, Quick summary about it

First : Front/Rear Select Push Switch mechanical Falut (ie, oxydized contacts..)
Pic 1. New Part (C&K Components Silver Plate Push Switch)
Alternate Part# 51902 (F8UEE)
http://www.mouser.com/ProductDetail/CK-Components/51902-F8UEE/?qs=%2fha2pyFadujTidtN1Llg9KzbODfiX9eD%252bMABDo%2f5rXk%3d

Agilent Part# 3101-3162
http://www.home.agilent.com/myagilent/faces/partDetail.jspx?partNumber=3101-3162&imageStatus=YES&_afrLoop=688445535652000&_afrWindowMode=0&_afrWindowId=null#%40%3F_afrWindowId%3Dnull%26_afrLoop%3D688445535652000%26imageStatus%3DYES%26partNumber%3D3101-3162%26_afrWindowMode%3D0%26_adf.ctrl-state%3D1220ztkopw_13

If nothing helps, order a spare switch from agilent.
PS: The "old" HP 34401A are still good, but switches are always a problem.

Second : H/L Sense Seriese Resistor Fault
Pic 2. I can't see the problem with the naked eye
Pic 3. Desolder them and Check one by one(3 defect of the 12 resistors,  25 percent )
Pic 4. Soldered the circuit boards (72Kohm Seriese resistor temporarily)

Test : Function : 2W, Resolution : 6 Digit Slow Mode, Hi/Lo Shot and NULL function
Pic 5. Benchvue
Pic 6. Multimeter


CRCW251224K0JNEG   24K 1W 2512 Chip Resistor
http://www.mouser.com/Search/Refine.aspx?Keyword=CRCW251224K0JNEG

Agilent Part# 0699-3406
http://www.home.agilent.com/myagilent/faces/partDetail.jspx?partNumber=0699-3406&imageStatus=YES&_afrLoop=52685776462000&_afrWindowMode=0&_afrWindowId=2irg93od1_19#%40%3F_afrWindowId%3D2irg93od1_19%26_afrLoop%3D52685776462000%26imageStatus%3DYES%26partNumber%3D0699-3406%26_afrWindowMode%3D0%26_adf.ctrl-state%3D2irg93od1_31

another thing to check while you got the machine open is the big 2512 resistors close to the big switch. i have seen those develop cracks on the electrodes. simply resolder them.

After many complications the matter was settled But,
At the moment still has corroded solder joint problem but expected that

To Do.
Find Potential Problems in PCB(ie, Corroded solder joint, Cracked Joint, Pad Contamination..)
Ordering new parts
Calibration

Thank you again to Dr. Frank, free_electron for a work knowledge , and everyone who helped me out on this forum.

coldframe

[razberik]

Hello. I am writing to contribute to repair logs of 34401A.
I received cheap, broken 34401A with missing mains switch push rod, that plastic part holding front terminal in place and rear plastic frame.

1. problem - not booting
Multimeter was not starting properly. When switched on, it illuminated all display segments, then two beeps and then nothing. Shift was working.
Exact behaviour like this:

Turned out to be bad connection between front panel and mainboard. Here is what I found. Broken connector. Somebody torn off that plastic part holding ribbon cable in place. Temporary fixed with wires extension. Not very clean solution, but it works until I receive proper replacement.
After this repair, multimeter starts properly.

2. problem - measuring wrong numbers on some ranges
After ribbon cable fix, multimeter was showing something, but nonsense numbers. When switched to 100V, 1000V range, everything works. All DC amperes OK. 4W Ohms OK. Random errors from selftests.
Then I realized that 2W sometimes work innacurately, sometimes dont work at all.
So I started to trace the signal path when some little voltage applied to input. And then it suddenly started to work.
Turn out to be bad connection on R104 2512 13K resistor. When I was touching that place, results were changing randomly. Resolved by applying flux and reflowing solder joints on R104.

Therefore I am confirming this:

another thing to check while you got the machine open is the big 2512 resistors close to the big switch. i have seen those develop cracks on the electrodes. simply resolder them.

and indirectly this (from this thread):

I then started to follow the sense tracks and, hello 24K resistors... why when I press you the sense line goes from open circuit back to good?
So I reflowed all those six 24K bastards. Actually it looks like three of them had a bad joint (they were invisible to me, I could spot nothing with my naked eyes).

After repair, multimeter passes all tests and seems to be working properly. Compared to another two 34401As.
I hope it might help somebody.

Now I should sketch up missing plastic parts for 3D printer. We will see...

[bitseeker]

Nice hunting to a successful repair. Having 3D models for replacement plastic parts would be great as more and more of them become unavailable from Keysight's parts inventory.

[coldframe]

hello razberik,

I congratulate you on your success 
I'm sure great help to someone.
coldframe.

[HighVoltage]

Nice repair and good price!
I had a broken 2512 resistor before as well and it is an easy fix.
This seems to happen more often than expected.

[bitseeker]

Yeah, makes you wonder what some of these instruments go through during their lifetime.