another 34401A Resistance Problem looking for advice

[smgvbest]

I've tried reading through posts here and am a bit lost on this so I hope you'll excuse me while I explain what behavior I am seeing and ask for help?

My 34401A passes all self tests is the first thing.
I've done a zero calibration on the unit which does complete.   I use a Fluke Shorting Block to do this.

When I switch into 2W resistance I notice a high initial reading when shorted in the 0.500Ohm to 2 ohm range that gradually decreases to around 10mOhm range after about 10-60mins depends on how high it starts at.   it drops off fast then as it gets closer to zero.   Actuating the Front/Rear switch seems to have effect on this. same behavior in 2W mode is on the rear as well but not as promounced.  ie.  it seems to start in the <50mOhm range and goes down and settles within 5minutes.

4W mode does not seem to be effected.  readings front and rear are in the <1mOhm range
All reading at in 6.5 digit, Fast mode.

I'm thinking the Front/Rear Switch is contaminated but the 4W behaving differently puts doubt on it.  I would expect 4W to behave the same if it was the F/R switch.

If I null out the value and read a 50Ohm 0.1% resistor I have it reads 49.997.x on 2W and 49.999.x in 4W
so if nulled it reads very accurately  (I do have a Rigol DM3068 which reads the same resistor at 49.987.x in 4W mode, just to see what another meter shows)

I checked the F/R switch and while most terminals show 10mOhm a few are much higher.
5-6      0.456Ohm
17-18   0.374 ohm
22-23   0.423Ohm

Trying to understand the test points and these reading to me say I should see this on the rear inputs not the front
unless I'm reading things wrong.

[deadlylover]

I think that's normal behaviour, it happens on my 34401As as well and I get the same jumpiness when switching from front to rear. The spec sheet does require you to use null on 2W mode for it to meet spec.

I'd imagine it's nothing to be worried about. 

[Dr. Frank]

When I switch into 2W resistance I notice a high initial reading when shorted in the 0.500Ohm to 2 ohm range that gradually decreases to around 10mOhm range after about 10-60mins depends on how high it starts at.   it drops off fast then as it gets closer to zero.   Actuating the Front/Rear switch seems to have effect on this. same behavior in 2W mode is on the rear as well but not as promounced.  ie.  it seems to start in the <50mOhm range and goes down and settles within 5minutes.

4W mode does not seem to be effected.  readings front and rear are in the <1mOhm range


I'm thinking the Front/Rear Switch is contaminated but the 4W behaving differently puts doubt on it.  I would expect 4W to behave the same if it was the F/R switch.

If I null out the value and read a 50Ohm 0.1% resistor I have it reads 49.997.x on 2W and 49.999.x in 4W
so if nulled it reads very accurately  (I do have a Rigol DM3068 which reads the same resistor at 49.987.x in 4W mode, just to see what another meter shows)

I checked the F/R switch and while most terminals show 10mOhm a few are much higher.
5-6      0.456Ohm
17-18   0.374 ohm
22-23   0.423Ohm

Trying to understand the test points and these reading to me say I should see this on the rear inputs not the front
unless I'm reading things wrong.

Sandra,
a contaminated switch very well explains the malfunction in 2W Ohm. It is usually contaminated if the instrument is older than about 10 years.
My 34401A showed exactly this behavior, and even my 3458A from 2001 has such a switch, which already shows an offset in 2W mode.

4W Ohm mode explicitly removes any residual resistances of the Input High and Low paths, as it measures the resistance (or the created voltage) directly at the DUT.
Therefore, the observed difference between 2W and 4W mode is correct, more than that, it has to be that way.


2W measurements includes all resistances of the external and internal cabling and components, up to the internal sense points, which you may identify in the schematics, pages 9-8 and 9-9.
The switch pins you mentioned are all NOT involved. Anyhow, these too high resistance values indicate a contamination of the whole switch.

Instead, pin 21-20 is the front Input Low path, pin 19-20 rear Input Low, to AGND.
Pin 3-2 and 1-2 are the respective ones for Front / Rear - Input High.
This latter path can be traced to the point between K101, pin 3 and L101, which is the High reference point, as there the reference current is injected, and the High voltage is tapped off, towards the input multiplexer U101-B.

So only pin 21-20, pin 3-2 and K101 are inside the instrument , and also the involved PCB traces and solder joints, which may effect your 2W measurement, for the FRONT setting.

Using your Rigol, please measure these components, or the complete paths from the Input jacksto the reference points (AGND, DCV_Low), if they show high resistances and/or changes after actuation of front/rear switch.

A fraction of this error is removed during instrument zero calibration, so reading of the 34401A and of the Rigol may be different.

Frank

[deadlylover]

I stand corrected, that'll explain why all 3 of my >15y old 34401A show similar behaviour. 

Thanks Dr. Frank, I will have a fiddle with my units later. As a small gift, here's a fun fact, the R6581 uses relays instead of a switch for front/rear. I'll see if I can measure any fluctuations when switching.

edit:
After a few minutes of switch actuations, it seems the 34401A's have settled from ~50-100mOhm to <10mOhm shorted on 2W. I don't think mine were contaminated, they just needed a little exercise. I was also using the Fluke short.

0.5-2 ohms sounds like you might need to open it up and clean the switches. 

Dr. Frank, in regards to the R6581;
On 2W shorted the residual reading is 144mOhm with Ohm-COMP, and negative(!) 340mOhms on normal. What a silly goose. 
As expected with relays though, it only changes <0.1mOhm when switching front to rear.....I'm not looking forward to replacing them when they wear out.

[HighVoltage]

On one of my older 34401A meters it helped to completely clean the switch.
But another one showed some more corrosion and I had to replace the switch.
With a new switch, everything was just perfect again.

[smgvbest]

@Dr. Frank
Thank you I will check that those points out and see what I get.
I do have one question,  the schematic state that S1 is shown in the rear position so doesn't that make Pins 1-2 the front and 3-2 the rear?

If it is the switch what is the best process to clean it?
 

[Dr. Frank]

@Dr. Frank
Thank you I will check that those points out and see what I get.
I do have one question,  the schematic state that S1 is shown in the rear position so doesn't that make Pins 1-2 the front and 3-2 the rear?

If it is the switch what is the best process to clean it?
 

You are correct about the pin assignment.

You'll find the repair thread here:
https://www.eevblog.com/forum/testgear/hpagilentkeysight-34401a-2w-problem/

Frank

[smgvbest]

I checked the resistance values at the 3 main points
S1A1-S1A2  reads <10 milliohms
K101-P3 <40 milliohms

I did note that repeated actuation of S1 has lowered the reading to <1 milliohm after about 60seconds of on time
switching from Front to Read and back it jumps randomly anywhere from 30 milliohm to .6 milliohm but settles quicker now

I'm not seeing a direct read that says the switch is bad but repeated actuation's seems to have helped it.
So I'm inclined to go with the switch being the main issue.

I've ordered one to have and in case I fubar the one in the meter when I try to dissemble it and clean it.

Will post an update when I have any news

Thanks

[smgvbest]

4w mode does not care about DMM internal resistance and test lead resistance, otherwise why we use it?

Not sure why you're saying this.
4W is not in question here.  2W is.
Can you explain further your comment

[kwass]

I found the same problem on my old 34401A.  The fix was to clean the banana jacks using DeOxit soaked into a Q-tip.  It's amazing how much dirt can get in there.

[smgvbest]

4 wire mode uses one pair of probes to source current, and another pair to measure voltage. The first pair sources current, not voltage, so serial resistance does not matter. Sensing pair has very high impedance, so probe resistance does not matter either. This is called Kelvim connection.

In 2 wire mode, tehe same probe set sources current and measures voltage, so the dmm sees voltage drop of both test leads and dut, hence the probe and switch's resistance will be mesaured along with dut.

@blueskull
I understand how 4W works and why and when I would use it but I'm not sure why it's relevant to my 2W issue.   
I want 2W working because it's broken on my meter

[smgvbest]

I found the same problem on my old 34401A.  The fix was to clean the banana jacks using DeOxit soaked into a Q-tip.  It's amazing how much dirt can get in there.

Thank you.   I will give that a try as well.

[smgvbest]

I just replaced SW1 in my meter (cheaper than ordering contact cleaner, plus brand new switch) and it's behavior is much better now in 2W.   Switching between F/R and back comes back down to the calibrated value in <1 second.   So after that I continued the performance checks and its passing all just within specs except for AC Current.   I'm using a DMMCheck plus from voltagestandard.com with a 1mA reference current in AC.   the Rigol 3068 shows 1.0002 as the current but the 34401A is showing <500uA and it jumps wildly from 0-500uA.   Looking at the schematic it looks like probably block 4 is the problem.  Is there any common issue's in the AC block to look at first or anywhere for this issue?

TIA

[deadlylover]

The 1A ACI range on the 34401a is only specced down to 1% (10mA), you will have to activate the hidden 10mA range by sending command "DIAG:POKE 25,0,1" which will let you measure down to 100uA.

More info here.

[Dr. Frank]

I just replaced SW1 in my meter (cheaper than ordering contact cleaner, plus brand new switch) and it's behavior is much better now in 2W.   Switching between F/R and back comes back down to the calibrated value in <1 second.   So after that I continued the performance checks and its passing all just within specs except for AC Current.   I'm using a DMMCheck plus from voltagestandard.com with a 1mA reference current in AC.   the Rigol 3068 shows 1.0002 as the current but the 34401A is showing <500uA and it jumps wildly from 0-500uA.   Looking at the schematic it looks like probably block 4 is the problem.  Is there any common issue's in the AC block to look at first or anywhere for this issue?

TIA

First question:
Did you order an original replacement switch from Keysight, or a similar one from a distributor?
If you got no original, does this include gold plated contacts and also the correct make / break before break / make scheme?

2nd question:

You're trying to measure 1mA AC on the 1A AC range, correct?

That is no regular verification point, as it's only 1/1000 of this AC range.
All AC ranges are specified down to 1% of F.S. only, between 1% and  5% of F.S., the accuracy already degrades.

Maybe you know, that the 34401As AC ranges are clipped by firmware for even lower readings, see here: http://www.gellerlabs.com/34401A%20AC%20zero.htm
Therefore, the behavior you see, is correct, and your instrument is very probably fine.
I also assume, that 34401A already measured ACV correctly, so there's no reason to assume a failure in the AC block.


If you have GPIB or RS232 control, then you may activate permanently the hidden 10mA AC range, realized by the 5 Ohm shunt.

Please send 'DIAG:POKE 25,0,1' to the 34401A, and then you may measure 1mA AC w/o problems in this 'new' range.

This trick is published here: http://www.keysight.com/main/editorial.jspx?cc=DE&lc=ger&ckey=1000001267:epsg:faq&nid=-536902435.536880933.02&id=1000001267:epsg:faq

Frank

[smgvbest]

Hi
I think I have typed this reply now 4 times, walked off to do something and forgot to hit post   
So I'm going to do a much quicker reply and hit post this time

I did not get the switch from keysight.  it is correct make/break but it is not gold plated.  It is however working great at this point.
I've also taking the switch that was in it and cleaned it up with contact cleaner and could try reinstalling it at some point.
I would order from keysight but anything less than 50 has a flat 20 shipping charge and there's just something about paying 20 to ship a 15 dollar part (or whatever keysights price was)

I did read that but thought that applied to the mV range not mA range so my bad there.
I have a USB->GPIB adapter coming and will do the thing to enable a  the 10ma range.

[Dr. Frank]

Hi Sandra,
If that switch has the same break/ make coding, then it's fine.. I'm asking, because my 3458a also starts having this problem.. And Keysight wants about 60 bucks for a spare part.
So, where did you find that switch?
 Frank