I've injured my Keithley 175 DMM

[mole42uk]

I'm rather sad to report that I've damaged my Keithley 175 DMM by testing a voltage that I thought was 270V only to find that it was nearly 2kV.

The meter passes the self-tests but when switched to measure DC voltage and the input shorted the display shows a cycle of several different voltages and sometimes the over-range indication. Selecting the Ω range gives similar nonsense.

Visual inspection of the PCB does not reveal any damaged component or fuse.

Has anyone experience of this DMM and can suggest a possible line of diagnosis? I do have the official service manual.

[Freya]

I had one a while back that had a similar fate when I inadvertently brushed the anode of a line output valve, it took out the input protection FETS, a few of the Multiplexers and the micro, yours may have done more damage though with that many kV
It was a few years back now, an oscilloscope and the manual was great for the diagnosis and repair.

[Chris55000]

Hi!

OH DEAR!

I've downloaded the 175 Manual and I'm giving it a careful perusal with the intention of posting detailed fault–finding/isolation tips!

Note – the switching FETs in the input ends of meters can be a bit fussy for their pinch–off voltages, and the zero–gate bias drain current (Idss) and it may be necessary to select replacements if you can't get the types specified in the parts lists!

The gate–leakage current Igss is important as well, as if this is too high you'll have bother with erratic zeroing and unstable readings on the lower voltage ranges, and also problems with "Ohms" not returning to a true 'overflow' indication on open–circuit on high ohms or autorange.

Your description of the fault after the damage occurred suggests floating or damaged input returns at the input to the A/D converter, and that's the point to work back from. I recommend, rather than trying to work from all the drawings in one go, get yourself a cheap block of squared or graph paper, and draw out the components for one range at a time, omitting any superfluous switching relays, CMOS gates, etc., except those directly connected to the range your dealing with!

Alternatively you can achieve the same result on Keithley's OEM diagrams with a pack of multi–colour highlighter pens!

(Obviously use a PDF manual to work from as you can scribble on, cross bits out and highlight things as often as you need to, since fresh reprints are always on hand from your friendly printer!)

Generally speaking, begin with the lowest DC range first, concentrate on that first, as you need to prove no significant damage has occurred in the a/d converter and it's control circuits from the '6805 micro.

I don't think, from the sounds of it, that the '6805 has been damaged – the symptoms quoted suggest the damage occurred farther back nearest the I/P sockets!

Let us know how you're getting on, I'm keeping an eye on this thread in case you get mega–stuck!

Chris Williams

[PETERg0rsq]

The quick and dirty method i use to check a digital multimeter to isolate the fault to analogue or digital circuits is to apply a DC voltage straight into the input of the Analogue to digital converter circuits, and see if the display corresponds

In this case it would be to the source/drain of JFET Q110 of the multiplexor.

[mole42uk]

The meter passes it’s own self-test and allows the various test functions to happen so I’m fairly confident that the processor and memory works.

Checking the power supply reveals an oddity. The 5V is spot on, but the negative rail around U117 is pulsing at about 2 c/s and doesn’t seem to produce the required -7V. There’s an inverter at U101 which has no output although the +5V supply is good, it’s negative rail is tied to the -7V which might be stopping the inverter. I suspect there is something dragging the -7V into overload so that the regulator is shutting down.

Further checks will reveal more, I’m sure.

[Chris55000]

Hi!

Do you have a bench power supply with digital voltage & current metering?

I've found in all these sorts of faults such a power source connected temporarily to the failing rail (in this case -7V) is by far the most straightforward means of isolating the problem, either to the -7V source or the circuits fed from it - U117, etc.

Alternatively, use a PP3 battery, a series 470 ohm resistor and a 6V8 shunt zener - a 500mW one will do - (with due consideration to polarity of course!) to supply the -7V, if this temp. supply won't hold up then you need to check the devices on this rail!

Chris Williams

[Chris55000]

PS!

If anyone's following this thread and can't read the horrible scans of the circuits in the original 175 Manual this revised one, called "Rev C", has much better diagrams in it:-

https://www.industrialcalibration.co...n%20Manual.pdf

I've printed 'em out for study to assist further if need be!

I think the differences will be only insignificant, mainly in the mains connector, backlighting & software rather than a major physical - difference - the basic topology between the drawings looks identical to me!

Chris Williams

[dave cox]

Ahh, a readable schematic, thanks Chris

My thoroughly unreliable crystal ball says U109. I've replace a few analogue switches in the past AND it's on the -7V rail


The JFETS near the input as well but its not clear if they can pull down the -7V

dc

[mole42uk]

I downloaded five different manuals for the 175 before I got to the "Rev C" version! It's the only one with completely readable circuit diagrams. It's strange but the downloadable one from Tek, the current owners, doesn't have the circuits!

[mole42uk]

Removing the -V supply to U103 settles the instrument down, the display reads 00.00 on any voltage range but it does respond to the range pushbuttons so I think the µP is working.

U109 may be the culprit but I've got no further than clipping the +V leg to that which made no difference to the -7V. I have ordered a couple of TLC271 chips to replace U103 - not only is that the regulator for the -7V but it's also in close proximity to the probe connections on the front panel.

[Chris55000]

Hi!

Originally Posted by mole42uk

I downloaded five different manuals for the 175 before I got to the "Rev C" version! It's the only one with completely readable circuit diagrams. It's strange but the downloadable one from Tek, the current owners, doesn't have the circuits!

I think Tek are purposely "improving" all their archived manual store by carefully deleting all the drawings, layout diagrams and schematics before re–uploading them – they were one of the first exponents of the " we no longer provide provide schematics in our service manuals" policy and of course, all the other OEMs followed like sheep & lemmings!

It's just as well we've got Bama, K04BB, Kevin Chant and docs.devs, etc., to step in for us!

Good luck with the Keithley 175 repair Richard – it looks you've certainly found at least part of the damage!

Chris Williams

[mole42uk]

An update - Here's a picture of two naughty components, U103 (the TLC271) which provides the -7V rail, and U109 (the 4053) which, as predicted by Dave, is mis-behaving, oscillating at about 2 c/s.

I've replaced the 271 which has restored order and calm to the power supplies, but I'm still waiting for the 4053 to be delivered.

Just a thought, when it's working again I'd like to check that the calibration isn't too far away. Are thiose £5 voltage references that can be bought from ebay accurate at all?

Further updates to follow.

[The Philpott]

The AD584L derived Vref circuit boards you refer to are quite popular, (mine is 2.5v,5v,7.5v and 10v) and judging by at least one youtube review the accuracy is satisfactory for 'our' purposes. One supplier abroad is infamous for providing calibration certificates that are simply photocopies that always have the same numbers on them, but it seems this is just laziness/cost-cutting and the product itself is OK. The supply voltage should ideally be around 15volts which is a little inconvenient.

The other (cheap) Vref i use came from Pimoroni, is the size of a phone SIM card, and gives you 2.048v or 4.096v. Simply needs wiring to a switch and a PP3 and putting in a project box.

Dave

[Andrewausfa]

Mark did a review of some units here

https://www.markhennessy.co.uk/ad584_references/

[chriswood1900]

Richard
Give me a call and I can sort you out a reference.
Chris

[mole42uk]

Thanks Chris! Let me get it working properly first!

Mark’s review of the little AD584 references is comprehensive and surprising. A pre-aged chip, who’d have thought that?

[mole42uk]

Now, an update but it's not good news at all.

I replaced U109 which made not a jot of difference. I tried a small voltage at the FET end of R114 which gave no indication of its' presence on the display.

I haven't done anything else yet but I have obtained a logic probe with which to work through the troubleshooting modes as referenced in Section 5 of the Instruction Manual.

[PETERg0rsq]

Originally Posted by mole42uk

I tried a small voltage at the FET end of R114 which gave no indication of its' presence on the display.

That wont work!!!

You are on the wrong side of the multiplexor!

Q110 to Q113 switch in signals in a controld sequence, such as signal, 0V, reference.

The microprocessor controls the A-D converter to look at and process these levels to get the correct reading. The signal at R114 will be a multiplexed signal of four components, sequenced for the A-D conversion.

As I said in an earlier post you need to connect your test voltage to the signal side of Q110 at R107.

Also check the logic to the four gates is correct before looking at the A-D and processor circuits.

U106 and U109 control the FET's and act as single pole double throw (SPDT) analogue switches.

[mole42uk]

Ah, so that shows that I shouldn’t try testing faulty equipment after a long day at work. Looking at the circuit diagram again I see what you mean.

[mole42uk]

I now see that I mis-understood your earlier post and injected a test voltage at the wrong side of Q110.