Products > Test Equipment
Old Fluke Multimeters
<< < (110/192) > >>
med6753:

--- Quote from: med6753 on September 21, 2018, 02:22:25 pm ---
--- Quote from: GregDunn on September 21, 2018, 03:48:08 am ---Well, Flukers (is that too crude?   >:D ) it looks like I have an 8800A/AF on the way.  Guaranteed working, pictures supplied of the unit allegedly powered up and working in multiple modes.  Fingers crossed, it's supposed to arrive tomorrow.

Now, business: this, and my 8600A, are likely in ticking time bomb mode as far as electrolytic caps are concerned.  I've read almost all of this thread, and it seems maybe the aluminum poly caps are the best replacements?  How about for replacing the tantalums?  I'm putting together a BOM to get my meters updated so (without starting any arguments) what's the consensus?

--- End quote ---

My 8600A just had a major crap out due to tantalums. See here...
https://www.eevblog.com/forum/testgear/test-equipment-anonymous-(tea)-group-therapy-thread/msg1839170/#msg1839170

--- End quote ---

Revisiting this issue with the 8600A. It appeared at first that the drifting problem on the 20VDC range had been fixed. It wasn't. After the DMM was shut off for several hours on initial power up it was grossly inaccurate. For example: against a 9.99691V reference it would sometimes read as low 9.945V. Over the course of a half hour or so the voltage measurement would slowly increase to 9.997V and then remain stable. I finally traced this back to the DC input divider board and believe it or not it was one of the reed relays. When the relay was “cold” it had high contact resistance and as it warmed up the contact resistance decreased. And I proved it with freeze spray and a heat gun. Freeze spray would lower the measured voltage to as low as 9.7V and then apply heat and right back up to normal.

The 8600A has a total of 7 of these reed relays. 3 on the DC divider board, 1 on the Ohms board, and 3 on the AC board. I decided to change the reed relays on the DC divider and Ohms boards but for now leave the AC board alone since I don't have appropriate standards to perform a calibration.  All of the reed relays are 5 volt coil and SPST NO type and they have a finite life span which after 40-45 years has been greatly exceeded. Obviously getting exact replacements is near impossible so I had to find acceptable substitutes. One of the items that needed to be checked was coil resistance. Turns out 1 relay had a coil resistance of 125 ohms and the other 3 had a coil resistance of 600 ohms (Didn't check the relays on the AC board).

Here's the DC Divider board. The reed relay on the left is 125 ohm coil, the 2 on the right are 600 ohm coil. The defective relay (K2) is in the lower right.



The Ohm's board. The single reed relay is 600 ohms coil.



Here's what I found for substitutes from Mouser. The bag on the left is a 125 ohm relay. Mouser P/N 934-HE3351A0500. The bag on the right are 500 ohm relays. Mouser P/N 934-HE3321A0400. I could not find 600 ohm relays that were reasonably priced so I decided to go with 500 ohm. The difference in drive current is about 2ma greater and I figured it would not be an issue (Proven to be true).



Here is the DC Divider board with the new relays installed. They are a smaller form factor so they are mounted sideways against the board and held in place with 3M automotive grade emblem tape. Then leads soldered to the pins.



The Ohms board. The relay mounted the same as the DC Divider board.



The results. No more drift. Accurate results right at power up. I also did a complete recalibration of the DC volts and Ohms. If you have a vintage Fluke with these reed relays I recommend changing them out before they cause problems. My 8800A has at least 4 of these reed relays and it's on my project list to replace them. 

GregDunn:
Interesting!  I haven't seen this sort of misbehavior on my 8600A or 8800A; they seem to be pretty accurate within a minute or less of power-up. 

How much change in adjustment was required?  And what are you using for the DCV and resistance standards?  I don't have access to anything accurate enough to re-calibrate either of my meters to that degree of precision.
med6753:

--- Quote from: GregDunn on October 13, 2018, 06:13:26 pm ---Interesting!  I haven't seen this sort of misbehavior on my 8600A or 8800A; they seem to be pretty accurate within a minute or less of power-up. 

How much change in adjustment was required?  And what are you using for the DCV and resistance standards?  I don't have access to anything accurate enough to re-calibrate either of my meters to that degree of precision.

--- End quote ---

Valid questions. How much change was required? Not much actually. Just some tweaks. The 8600A was pretty much spot on before it started acting up.

I have an assortment of references that, while not traceable, do a credible job. First, I have two AD-584M references. One as primary and one as backup. They are accurate and stable enough to check up to 5.5 digit DMM's. But they only cover 2.5VDC to 10.0VDC. I have 2 home built references for 100.0mV, 1.000V, 10.00V and 190.0mV, 1.900V, and 19.00V that are really only accurate enough to do 3.5 digit DMM's but if I monitor their output with the known accurate 8800A I can adjust the 8600A pretty much dead nuts.

I have an inverter card that generates 190.0VDC but it isn't very stable but if I monitor it with a known good DMM it at least get's me in the ballpark.

For resistance I built up a decade box of 0.1% resistors from 1.0 ohm thru 10.0MEG. Not perfect but accurate enough for me.

 
bd139:
I actually had one of the original relays' coils go open circuit on my 8600A. Very very annoying that was. Did the same sort of replacement.
GregDunn:

--- Quote from: med6753 on October 13, 2018, 08:01:07 pm ---I have an assortment of references that, while not traceable, do a credible job. First, I have two AD-584M references. One as primary and one as backup. They are accurate and stable enough to check up to 5.5 digit DMM's. But they only cover 2.5VDC to 10.0VDC. I have 2 home built references for 100.0mV, 1.000V, 10.00V and 190.0mV, 1.900V, and 19.00V that are really only accurate enough to do 3.5 digit DMM's but if I monitor their output with the known accurate 8800A I can adjust the 8600A pretty much dead nuts.

I have an inverter card that generates 190.0VDC but it isn't very stable but if I monitor it with a known good DMM it at least get's me in the ballpark.

For resistance I built up a decade box of 0.1% resistors from 1.0 ohm thru 10.0MEG. Not perfect but accurate enough for me.

--- End quote ---

I have three inexpensive but calibrated voltage references (2 AD584 and 1 LM399) which I definitely trust in the range of 1-10VDC and am also preparing a resistance standard which should be good enough for the meters I currently own - that's why I was interested in comparing your methods to my approach.

I am almost finished repairing an HP6920B which I hope will be good and stable enough to at least let me sync my meters as you described.  Checking it against my Keithley and 2-Fluke bench meters it seems to be consistent if not 5-digit accurate up to well over 100V (that mechanical dial is probably the least accurate part of the unit...) and 1A. I expect I'll do a standard deviation calculation on these three meters vs. the HP and see how they line up over a wide voltage range - that should be good enough for the work I do here.  I rarely need anything better than 0.1V accuracy under any circumstances, but it's nice to have better for drift and other relative measurements.

So perhaps I should stock up on relays in my next parts order just to be prepared for the inevitable.  Have you looked at the 8800A relays yet?  The BOM indicates that the two units use the same Fluke part numbers for the relays though perhaps not the same number of them.
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod