Fluke 8600A

[Zenith]

This was bought from one of the usual traders at a swapmeet for a tenner. He deals in small bags of components for £1 each, but occasionally has TE. He said it worked and he'd used it as his bench DMM for years.

The bottom segment of each 7 segment display didn't light. Opening it and wiggling the strange flexible PCB which connect the front panel and the motherboard solved that. A dry joint, a cracked track? I didn't investigate further. After that, it worked but wandered a bit. I thought it was probably damp from being unused for a time and left it on. When I came back an hour later it was showing strange results and the relay was chattering. Time to check the power supply voltages. The test points for + & -15 V described in the manual weren't there and I didn't have the energy to find where the voltages could be found. I later found the pins on the can enclosed op amps where they appear. I changed the reservoir caps and the tants across power lines.

It now worked and the power lines were at the right voltages with very little ripple. It seemed in close agreement with a newish Brymen DMM. Then it started doing strange things. The display would go blank when the auto button was pressed and the meter had to be restarted to get it back. Then it started giving a fleeting flash of the display on power up and then the display was unlit. The + and - 15 V lines were at the correct levels with no ripple but the +5 V line was at 0.9 V. Powering it up with a DSO on the 5 V line, showed the level at 5V for a short time after power up and then collapsing.

I thought it couldn't possibly be the three pin 5 V regulator. Those things never go wrong. It was cold to the touch, so it didn't appear to be an overload causing the 5 V line to be so low. I desoldered the regulator and checked it with a bench PSU and it seemed OK, but I didn't test it with a load, which is important. It was supposed to be an LM340T-5 but this was a 7805. I decided to try the DMM with a bench PSU supplying the 5 V. It didn't work, either with the 5 V supplied before or after the DMM was powered up. It did show the 5 V line was drawing just under 300 mA. I had an LM340T-5 in the collection and soldered it in.

Success! It ran for hours agreeing very closely with the Brymen and didn't act up when the auto button was pressed. There was very little ripple on the 5 V line. I touched the heatsink of the 5 V regulator and it was uncomfortably hot. An IR thermometer showed the nut and bolt to be at 50C. This was with the case off. The case is plastic and unventilated. The temperature would be higher in normal operation. The unregulated supply to the regulator is about 11 V measured lazily with a meter. That's not an unreasonable drop with adequate heatsinking, but the heatsink is quite small and there isn't much room for a bigger one. Even with a bigger heatsink, its effectiveness would be limited by the small, unventilated case. It's bound to shorten the life of electrolytics. No wonder the original regulator eventually failed. Dubious design decisions there. They may have used that transformer voltage for some option, such as a rechargeable battery. I haven't looked into the options they offered. I don't dig the flex circuit/PCB either.

I removed and re-glued the metal strip showing the functions of the terminals and buttons. It had become detached at one end and was interfering with the operation of the power button.

It agrees pretty closely with the Brymen on the DC ranges and resistance, which is remarkable since it's around 50 years old and I assume it has never been calibrated - there are no cal stickers. I'm wondering whether I should tweak the calibration or leave well enough alone. I should also clean the switch contacts. A bigger heatsink would be good. Maybe tasks for a rainy day.

My impression is that it's a very nice basic 4½ digit bench meter with 1 GOhm input on the 200mV and 2V ranges. No GPIB or anything like that. The front panel is in a slighty tatty state, but the handle is there and works and it doesn't seem knocked about. I like LED displays. I prefer the push buttons to the inconveniently small piano key switches on the 8800A. The documentation is in most ways very good, but in some ways is awful. A diagram showing where the test points are would have been useful and being directed to test points that weren't there was infuriating. The schematics aren't the best.

The meter measuring the open circuit voltage of a tired 9V battery.

DSCN4329.JPG

Close up of the 5V reg with tiny heatsink. Note the radial electrolytics bodged in because I didn't have the right axial ones. I might at least have found some sleeving. The + and - 15 V regulators have no heatsink. I assume they are only supplying a few mA and the voltage drop across them is 8 V, so they don't need them.

DSCN4328.JPG

Top view of the motherboard showing the flex circuit/PCB at the top right.

DSCN4324.JPG

[tggzzz]

That's generally similar to my 8000, same:

• switches; I'll bet it also routes naked mains to the front panel switch
• smoothing capacitors
• flex PCB
• EDIT: forgot to note the same novel attitude to safety w.r.t. that IEC socket

but the solderless breadboard is missing (with trimpots and daughterboard in that location), and it looks like there is LSI rather than special ICs in the 8000.

[Zenith]

It does indeed route mains to the front panel switch. They could have used a plastic shaft to activate the mains switch on this, but there is space left for various options such as a data output card and they may not have had room. The mechanical link to the mains switch on the 8800A is a strange thing, necessitated by the piano keys, which may have been cool around 1980, but are rather a pain, especially since the keys are so small.

The controller is a single 40 pin DIL chip designated C2506. I can't find too much about it, other than there was a CMOS version, later replaced by a PMOS version. The 8800A documentation mentions a jumper to be added to switch between the two versions, but I haven't been able to find it on my 8800A. It supplies control signals and presents data and reads input from the front panel. I'd guess it was a special version of a µprocessor that was around in the 70s, maybe Z80. I suppose it might have been dedicated logic. It is known to fail and I wonder if it could be replaced by an Arduino or similar.

I'm sure the 8600 was a refined version of the 8000 and the others, such as the 8800 were further refinements.

[MED6753]

Here's some internal views of my 8600A. Those Philips crapacitors in the PSU had to go. I went exclusively radial caps since axials are getting rare and expensive. Replaced some of the tants too. I haven't had any issues with the regulators. I also have a parts unit just in case some unobtainium part craps out. I found the 8600A to be very accurate once it has warmed up for about an hour.

[AVGresponding]

This has me wondering whether I should open one of my 8800A's to see if it also has dubious main power switching. I do know all my other Fluke mains powered gear is better wrt separation.

[Zenith]

The 8800A has piano key switches and the mains switch is at the rear and in a compartment. It's operated by a rod, which is moved by a lever arrangement from the piano key. The lever can come loose and is a bit fiddly to reinstall, but generally causes no problems.

[MED6753]

The 8800A has piano key switches and the mains switch is at the rear and in a compartment. It's operated by a rod, which is moved by a lever arrangement from the piano key. The lever can come loose and is a bit fiddly to reinstall, but generally causes no problems.

I have had that whole assembly jam up which was quite a PITA to disassemble and fix.

[Zenith]

Those piano key switches were a 60s and 70s thing. I find the keys on the 8800A a PITA altogether. They are very small and I find it easiest to use them with a pencil with an eraser on the end. This is not ergonomic design. As plastic ages it becomes brittle. They might at least have made the on/off switch a straight push button and avoided that peculiar lever arrangement.

I'm impressed with the way those 8000 series meters are usually pretty much within spec after so many years. They sold well, and I doubt anyone thought they would still be being used 15 years after they sold, leave alone 50 or so years later.

[AVGresponding]

When they wanted to make them well, they did. My 8300A's both are within a few LSDs of my 8840A, which agrees with my Keithley 2000, 2015THD, and late Agilent 34401A. Not to mention that all my Fluke handhelds agree in a similar way, from the 8060A to the 289.
The 8000A really does seem to be an aberration

[tggzzz]

Zenith's second picture of his 8600 shows the same mains socket insulation made of perforated paper. So clearly it was an aberrant line rather than a specific meter

[Zenith]

Med's first picture of his 8600 shows an block of what looks like silicone sealant, between the questionable IEC socket and the mains transformer. I suppose it was put there for reinforcement.

There's no doubt that a few iffy design decisions were taken with these things.

[MED6753]

You guys are forgetting that up until very recently it was common practice on nearly all equipment to have mains exposed internally and it was up to you to be aware and be safe. And if you weren't then you shouldn't be in there.

[Zenith]

450V - 0 - 450V transformers with connections made to bare tags on top - no cover, no attempt at safety insulation. Live chassis TVs and radios........

Some people don't like the mains wiring going to the front panel switch on the 8000 and 8600, and mains voltages close to sensitive metering circuitry - although that doesn't seem to cause problems.

I find the IEC socket flimsy and I feel uneasy about the way it flexes when the plug is inserted. It seems a surprisingly makeshift thing.

[tggzzz]

I find the IEC socket flimsy and I feel uneasy about the way it flexes when the plug is inserted. It seems a surprisingly makeshift thing.

The metal prongs are pretty thick and heavy in the direction in which the plug is inserted. I'd check whether it is the PCB flexing!

Side to side the prongs are easier to bend, but that shouldn't arise in practice.

I'll take more pictures when I've returned to home.

[Cubdriver]

Zenith's second picture of his 8600 shows the same mains socket insulation made of perforated paper. So clearly it was an aberrant line rather than a specific meter

That's fish paper - it's a vulcanized fiber material that's lightweight, tough and tolerant of temperature extremes, and has been used as insulation since the late 1800s. There's nothing wrong with it in that application; in fact that's one that it is ideal for.

https://www.americanmicroinc.com/electr ... ish-paper/

-Pat

[Zenith]

Zenith's second picture of his 8600 shows the same mains socket insulation made of perforated paper. So clearly it was an aberrant line rather than a specific meter

That's fish paper - it's a vulcanized fiber material that's lightweight, tough and tolerant of temperature extremes, and has been used as insulation since the late 1800s. There's nothing wrong with it in that application; in fact that's one that it is ideal for.

https://www.americanmicroinc.com/electr ... ish-paper/

-Pat

Yes, well, and a good engineer can do for a shilling what any damned fool can do for ten quid, and views of electrical safety have changed over the years and so on. 50 years on they wouldn't be expected to meet modern safety standards.

Nonetheless, I don't like it. It feels crap when you push in the IEC plug and feel it flex. Someone chose to supplement the highly specified fish paper in Med's example with a dose of what looks like silicone sealant. Similar misgivings, or an owner previous had thrown away an apparently useless bit of paper? The earth pin is longest and would be expected to contact the body of the transformer first.

If anyone had been electrocuted by these devices, Fluke would have been sued out of existence. They weren't. Since the case is all plastic, it's hard to imagine that happening, unless you decided to pay with it in the bath.

I really don't dig that LM340-T5 ( 7805 whatever ) with its miniature heatsink, particularly the heatsink. The datasheets claim they indestructible. Hmm, fairly indestructible. It seems a bit naff round the edges.