Fluke 8600A used buy- what things to check for?

[nightfire]

Just got a used 8600A (standard version without batteries according to picture from the underside) from ebay for a decent proce- and awaiting delivery.

As I want to prepare for that new member of my workbench, I want to ask if there are things to check for that might be known with this model.
Pictures on ebay showed the instrument in good cosmetic condition, clean from the outside, and being able to measure reasonable values. Pics showed measurements of a battery and mains voltage (230V here in germany).

My ideas are to open the case, clean dust, check for issues with the caps used, maybe clean up some minor issues with IPA and/or a glassfiber pen.

What about the contacts of the buttons? Are there any experiences/recommendations how to check or clean them?
Regarding the internal PSU: Would it make sense to check for stability of output voltage?

For a "poor-mans-calibration check" i just ordered one of those el-cheapo AD584KH modules, from the pictures it seems to be a relabelled KKMoon model.
(I mean, for the things i will do with them, the absolute accuracy is not my primary concern, as long as all my meters (that 8600A,  Agilent U1272A and a LabjackU6) all agree to the values...)

As the Fluke 8600A will be primarily used for measuring DC Voltages of batteries and small DC Power supplies/USB Connections, this should be fine and hopefully sufficient for my usage scenario.
Comments?

[bd139]

Main failure points are dead Reed relays, dodgy switches, idiot previous owners (front end protection isn’t great), power supply problems, display board damage and rework (front diffuser plastic falls to bits) and if it’s a battery option one, leaked batteries. Check that the fuses are intact and not bypassed first.

Switches usually clean up with standard no residue contact cleaner.

If you got a good one they’re nice meters though and certainly fine for what you need to do.

[med6753]

As BD mentioned. My experience is shorted capacitors in the PSU. If the range switches are dirty be careful with the contact cleaner. It's a very high impedance area and too much cleaner with permanently affect it and you'll have erratic readings. 

[BobRyan]

My 8600a developed an issue with auto ranging where it would constantly switch ranges but would read accurately when ranged manually. I sprayed some deoxit d5 onto the switches and cleaned the board with IPA which resolved the issue.

It's been a good meter and I love the 7 seg display but the lack of support for shielded banana plugs is limiting.

[nightfire]

I made the decision for my lab (that i am currently building resp. getting equipment together) not to use unshielded 4mm plugs. Simple reason: I also work on mains equipment, and when in a hurry, leads can be exchanged or interchanged and therefore you have some real potential for acccidents.
But there are test leads with retractable spring-loaded shields, i got myself some Stäubli XZG425/SIL leads i really like, and I was able to get them new for below 10 €/1 Meter length(per piece)- and they are rated for 32A/2.5 mm²)

[Paceguy]

You made the wise decision to go with the line operated unit instead of the battery one. I have both line and battery, but way prefer the line one. These were my go to bench meters before I got a 8840A. I like the auto range function and never can really say that I had problems with mine except for dirty switches. If you intend to remove the circuit boards, be careful not to break those small red pin connectors.

[nightfire]

Update: Unit arrived today, and it looks quite nice. That is, apart from a broken handle due to rough handling and rudimentary transport protection... 

Also my generic AD584KH voltage reference arrived (late version, with switch at the side and brown stickers on the sides of that clear plexiglass...)
Sadly, there is no sticker at the bottom with calibration data like in some earlier versions.

My Agilent U127A (also an ebay find) could basically agree to the values presented within its specs, and the Fluke 8600A also measured very reasonable values. Those were all a few mV higher than of my Agilent, but after probably 30+ years, that is still impressive that at the 10V measurement, we are talking about a difference of about 6 mV between the Agilent and the Fluke here...
On all values presented from the AD584KH it measures a bit higher than rated  (at 2.5 V nearly spot-on, at 5V 3mV higher, 6mV with 7.5V, and at 10V about 10mV)

From the inside, it is quite neat and clean, no dust buildup, some light oxidation on the springs of the front switches, but no cap spill or similar stuff.
Transformer is a bit crusty, probably molten glue from the insulating tape (that is really crispy by now...)

Pictures will follow, hope I can take them later.
Also tests for  the mA range have to follow.

[Hexley]

Congrats on the nice 8600A. You had asked about points of failure, so here are a few  problems that surfaced in my 8600A:

1. The automatic range selection goes through a 7488 PROM, U9, that does the relay logic. This chip failed completely in my case, so the meter was stuck on low range. Rather than try to source and program that obsolete part, I made up an emulator using a PIC16F628A to replace it. That has worked fine for the last 6 years.

2. The 5-volt regulator, U102, an LM340T5, had inched up in voltage over the years, and was at 5.48V when I measured in 2018. The part had a 1974 date code, so it was 44 years old. This was a bit above the expected 5.25 volt upper limit, so I replaced it even though there were no symptoms of bad behavior yet. But it seemed like an unnecessary risk to the TTL logic to leave it in.

3. A few years ago, the meter started reading 4% off on the lowest DC volt scale. Turned out to be a fault in the CMOS switch U17, a CD4016 -- that part was injecting excessive leakage current into the 100 mV reference node, causing it to rise slightly. I swapped in a CD4066 and that fixed it.

It is a nice meter, and still my go-to DMM when I want to measure something Right Now, since it has no boot up delay.

[nightfire]

Just read the datasheet on the AD584KH in question, and then after fully charging that battery, i ran some measurements. From what I see, both my Agilent U1272 and the Fluke 8600A as oldtimer are well within the tolerance of that reference. So without knowing what drift the reference has, I can only assume everythings running smoothly...

nominal Voltage	Tolerance +/- mV	lower limit	upper limit	Agilent	Fluke
mV		AD584KH	AD584KH	U1272A	8600A
2500	3,5	2496,5	2503,5	2499,1	2500
5000	6	4994	5006	5003	5007
7500	8	7492	7508	7499	7504
10000	10	9990	10010	10004	100010

[nightfire]

As promised, some photos- I also took the liberty of using some equally old glass for doing this and mounted an old Micro-Nikkor 55/3.5 on my trusty Fuji camera ;-)
Judging from the datestamp on some IC, this Multimeter has been build around 1976...
PCB Revision H is stated on some place on the PCB...

Anyone has seen those fuses before? The winding style seems to be... somewhat tricky...
Caps seem to be clean and good, depicted here is some kind of valve? on C103 (yellow one)
The trimmer left (in focus) is C30, right would be C24 according to schematics...

[coromonadalix]

For the fuses  they are slow blow fuses ... Kinda "timed rupture"   

For a start the obvious capacitors who needs to be changed, tantalums and chemicals.

Carefully avoid to touch any  "screw" post  "photo DSCF7636.jpg",  normally they are frequency compensations if i recall

Seems pretty clean.   

[nightfire]

Just checked the measurements against my other Multimeter (u1272A) with the exception of AC due to lack of devices that will output ACV- and I did not want to plug it directly into mains...

I was impressed- most ranges spot on after 40+ years- even Diode check of a red LED with about 1.7V only differed 20 mV to my Agilent DMM...
also on a 100 Ohm resistor we were talking about 50 mOhms difference- seems pretty good to me ;-)

As the board is pretty clean and the caps show no signs of spilling, I would use it as it is now and some time later, where I have natural lights in the evening, maybe replace the caps after some preparation to get the right ones- and if I would have to solder with lead...

[med6753]

I recommend changing out the capacitors in the PSU. Those blue Philips capacitors are known for going short.

[bd139]

Nice and tidy. I'll second med's recommendation to replace those caps. They do indeed go short.  BC/Vishay sell suitable replacements.

[nightfire]

I need some help clarifying the notation in the manual:
For C101/C102 there is noted: "Cap, elect, 220uF, +50/-10%, 40V  and: Mfg Part Nr. ET221X040A01
C103: "Cap, elect, 2000uF +100/-10%, 15V and MFG Part Nr. B4101-2200/15

Does this really mean that I can use a cap with 50% more capacity for C101, but not much lower?

As I understand, I could also use a cap with higher Voltage than specified? - I have noted that sometimes an exact 40V is hard to get, depending on supplier...

Does ESR plays an important role in this scenario? Or can I get any good quality manufacturer that matches the above specs?

[bdunham7]

I need some help clarifying the notation in the manual:
For C101/C102 there is noted: "Cap, elect, 220uF, +50/-10%, 40V  and: Mfg Part Nr. ET221X040A01
C103: "Cap, elect, 2000uF +100/-10%, 15V and MFG Part Nr. B4101-2200/15

Does this really mean that I can use a cap with 50% more capacity for C101, but not much lower?

As I understand, I could also use a cap with higher Voltage than specified? - I have noted that sometimes an exact 40V is hard to get, depending on supplier...

Does ESR plays an important role in this scenario? Or can I get any good quality manufacturer that matches the above specs?

Those specs are simply those of  the part that they originally used.  You can use a 330uF +/- 20% 63V for the first and a 2700uF +/- 20% 25V or 35V for the second.  The trickiest part is measuring the dimensions, including lead spacing, and finding one that fits neatly.  There's no need to 'match' the voltage spec and on older stuff you often go to a higher voltage just because that is the only way to get close to the original dimensions.

ESR will be a non-issue.  The caps are huge in proportion to the power supply current, deliberately in order to reduce ripple.