Re-calibrating a AVO Model 8 Mk5

[Specmaster]

Hi George,

Thanks for your observations and great help and advice. Let me flesh out my results a bit more.

I soldered two pin headers into the holes on the switch PCB next to the test point for the meter (marked on the drawing as 100mV test point) and then connected a 6.5 digit meter (manual switcher so no auto ranging errors) and I applied 3v from a bench power supply to the 3Vdc position. I also connected a very high Gohm impedance DMM (Fluke 8842A) to prevent any impact on the OPV of the Avo, The fluke verified the 3V input and the Avo displayed 3V FSD, and the DMM on the test point displayed 100mA. That gave me confidence in the meter, and also confirmed my earlier reports that the 3V range did read correctly.

Page 10 of the service manual also states that if the resistors R16 to R24 are low, the instrument will read high, but if the R's are high, the instrument will read low and I must have had a brain fart or something that day and took it the other way round 

So I removed the switch PCB to gain access to the bottom switch contacts and the ribbon cable terminations which were slightly oxidised. Cleaned then and reflowed the terminations and then checked the readings, via the switch terminations and ribbon cable back to the various range drop resistors, and they are low as follows:-

R16      58K      58.2K
R17     140K    139K
R18     400K    398K
R19     600K    565K
R20     800K    795K
R21     4M       3.62M
R22     6M       5.1M
R23     8M       6.4M
R24     40M     41M

I can only assume that when I first read the values of the resistors on the resistor PCB, the leads were a bit contaminated but now they have been cleaned up and hence the lower readings, which is now entirely in keeping with page 10 of the manual, i.e. resistors low results in the instrument reading higher., which is perfectly logical.

Now as these values are specifically made up against large orders, have effectively 3 options open to me:-

1- Source a donor meter and hope that the resistors are OK.
2- Try to make up the equivalent values if space permits, from available values.
3- Unhook 1 leg of each resistor and connect a series resistor to restore them to near enough their originals. There is just enough space to allow a maximum of 1 series resistor per value. The original resistors may well continue to drift slightly as they age even further, and then again maybe they won't. As long as the meter provides a pretty good indication of the voltage it will be fine. Where accuracy is required, I use DMM's.   

[George Edmonds]

Hi

Now having read your last posting there are more things that I do not understand, for instance in sixty years I have never cleaned some resistors and their mounting PCB and had the resistance go down, up yes, as the unwanted effectively in parallel resistors resulting from the contamination are cleaned away.

The resistor value table that you provided also does not make sense, for instance R22 (600VDC) which should be 6M you say is 5.1M and R23 (1000VDC) which should be 8M you say is 6.4M, in my experience this is simply too large a change to be correct AND its simply in the wrong direction.  Have you lifted one end of each resistor and then independently measured it?  It may take some time for the high value resistors to fully recover from the thermal effects being lifted.

I have taken my AVO 8 Mk5 which is of 1973 vintage, with still has intact factory seals, and measured with a Keithley 2000 the resistance between the plus and minus meter terminals on the DC ranges.  Again these are just quick measurements and subject to some small Keithley 2000 lack of nulling errors.

I got the following resistances:-

3VDC      60.2K.  (58K of R16 plus 2K of movement resistance, within my measuring error)
10VDC   200K.
30VDC   604K.
100VDC   2.04M. (again within the 2% error I am seeing at full scale)
300VDC   6.3M.
600VDC   11.97M.
1000VDC   19.7M.
3000VDC   60.9M.

This is entirely consistent with the observed full scale DC Voltage errors of my AVO 8 Mk5.  For example on the 100VDC range the terminal to terminal resistance should be 2.00M, it is actually 2.04M which is about 2% high which will give a 2% low full scale error reading, which it does.

Just a thought, has your AVO 8 Mk5 ever had a leaking D cell in it?  The gold “long lasting” ones that were sold as none leaking actually leak very badly with time, when they do you get an Alkaline corrosive and conductive film covering deposited which is almost impossible to remove unless done chemically, washing with IPA will not remove it.  Is it just a coincidence that the most effected resistors are located very close the D cell holder, any sign of corrosion on the D cell contacts? .

I still have a scrap AVO 8 Mk5, maybe time to pull the series resistor board and check all of the resistors.

I still keep a number of moving coil meters just to prevent having to do constant mental maths on measured voltages when repairing late valve and early transistor kit.  My go to meter is a Simpson 260, not an AVO.

Keep up the fight.  Keep in mind in my experience, on their own resistors never go down in resistance with age, only up.

George G6HIG Dover in the UK.

[Specmaster]

Hi George

After I posted last time, in response to your post I began to doubt my findings, so I stripped out resistor board and checked the readings again, I still got the same sort of results. So I lifted one end of each and low and behold, they were spot on. I then removed the AC transformer and popped the board into a ultrasonic cleaner and gave it a couple of goes. Once it had all dried out nicely, I resoldered the lifted ends back on the board and checked values, again spot on. Then I reattached the transformer and ribbon cables, checked values again, and they were back to the original values

Prior to the above I had stripped the meter down and disconnected everything as in the manual and removed the top switch board. Ribbon cables all look to be in good condition, the top switch board seems to be OK, contacts look nice and clean, no pitting etc and good contact is made with the moving contacts.  Bottom contacts also inspected and the ribbon cable connections checked and all are good, a real head scratcher this one.

Today, following your post, I rechecked the total resistance per range via the plus and minus terminals on the DC ranges and using my Brymen BM867 (goes upto 50M) so I had to use a cheap manual switcher to read the 3,000V reading as it goes upto just over 60M. Any way I obtained the following readings.

DC Volts   Resistance.
3             60.3K
10           201.5K
30           666K
100         2.06M
300         5.03M
600         7.78M
1000       10.8M
3000       58.8M

I did notice that the readings were creeping upwards if I left the meters connected, just like a capacitor being slowly charged?

I then got out my other 2 Model 8's, not to sure which marks they are but both have the 3,000 range as opposed to more normal 2,500v One is the normal domestic specification and the other is a NATO issue one and weighs a bleeding ton. These are the reading obtained from them in the same fashion as the Mk5.

Normal one
3           30.1K
10         200.8K
30         603K
100       2M
300       6M
600       11.99M
1000     19.88M
3000     40.5M

NATO one
3           60.3K
10         200.3K
30         603K
100       2M
300       5.9M
600  Not supported on this meter
1000     19.4M
3000     39.98M

Neither of these two meters exhibit the slow upwards creepage that the Mk5 does and the capacitors on the Mk5 are minimal values and should not be in circuit as the AC side is switched out of circuit? Could the ribbon cable be acting as a capacitor I wonder?


To answer your other question about has a D cell leaked at some point, I think the answer to that is yes some evidence of slight corrosion on the plus terminal but zero sign of anything inside the meter or the battery compartment itself.

[George Edmonds]

Hi

Following your latest posting I am now almost 100% convinced that you have a chemically contaminated AVO 8 Mk5.  The only other possibility is that excessive applied voltage has caused some localised tracking damage.  It is of course virtually impossible to be absolutely certain without seeing it and doing some high voltage insulation tests on it.

The final clincher is when you said “the readings were creeping upwards if the meter was left connected, just like a capacitor being slowly charged”.  There are no capacitors in a AVO 8 Mk5 large enough to show this behaviour and as you correctly say they are on the AC side and not in circuit when making DC measurements.

The inter track capacitance of the ribbon cables will be in the order of pF’s and can be totally discounted but their insulation resistance cannot.  If you can see the above creeping effect using a DMM with an applied voltage of around one volt it must be serious.  If push comes to shove all of the printed ribbon cables can be replaced with simple point to point wiring.

My best guess at this time is that the 1.5V D cell has possibly caused this problem.  If a NiCad had been used in place of a normal dry cell these can produce a highly conductive/corrosive out gassing vapour when they reach the end of their life.  The damage/contamination caused by this vapour can be extremely difficult to see visibly.  My suspicion is that you are not seeing a capacitor charging but an electrochemical reaction taking place, don’t forget that with your AVO 8 Mk3 set for over 30VDC full scale the resistances involved are over 1M Ohm so a very little contamination can/will have large consequences.

Had I have been working on the AVO myself and once seen the series resistors changed downward in value I would have first done a Litmus paper test for the presence of an acidic or alkali contamination and then with an AVO BM403 started doing a though series of insulation tests on it with the moving coil meter isolated and finally with the printed ribbon cables removed so as to isolate various sections of the meter.  The AVO BM403 can do insulation tests at three different voltages up to 1KV.

There may well be other small problems with your AVO 8 Mk5, BUT, the main problem as shown by your resistance measurements must be addressed and resolved first. It may be worth considering that the “contamination” has been absorbed and cannot be totally by the PCB material of the series resistor deck and the range switches.

George G6HIG in Dover UK.

[Specmaster]

Hi

Yes, I'm inclined to agree with your thoughts as I've exhausted all the other possibilities.

As an educated guess, where do you think the problem is, the PCB board the resistors are located on, the ribbon cables or the switch board?

I suspect the reason why the resistance measurements are working OK is that the range selector and most of the control is done with the right switch and the DC volts is on the left switch, meaning the battery voltage, although it maybe 15V on the higher range, but that the signal path of that is away from the contaminated area?

What would you recommend cleaning everything with now to try to effect a cure?

[Specmaster]

I could remove certain resistors totally from the board, say R21 through to R25, leaving just 3,10,30 and 100V connected and see if there is any différances

[Specmaster]

OK then, I removed completely from the board, all the resistors from 100v tap upwards, but there was still a high resistance connection on the 300V and above positions. This did not rule out the ribbon cable, so I traced the connections in the ribbon and isolated the 300 and above connections to the resistor board and the readings went open. This proves then your theory that there had a chemical contamination. Now can I scrub the board with some solution or other, or do I need another donor board?

The readings for 3, 10, 30, and 100V ranges now reads as

3V     66K
10V   207K
30V   635K
100V 2M

[George Edmonds]

Hi

My own gut feeling would be to remove the shunt and series resistor boards, this will effectively cut the problem in half, then do insulation tests on the range switches and printed ribbon cables.  With the switches and ribbon cables first do insulation tests between adjacent switch contacts.

At this point it would also be a simple matter to measure the overall resistance of the series resistance chain between the cold end of R16 and the hot end of R23 and check to overall resistance for stability.  Given that R25 and R24 effectively have their hot ends floating I would not include them in any initial tests so as to allow your better quality DMM to be used.  The total resistance you should be seeing is 19,998M which if you add the 2K moving coil meter resistance will give you 20M which is exactly what a an AVO 8 Mk5 should be on the 1KVDC range.  Not the 10.8M you reported when you made the terminal to terminal resistance measurements.

At this point I saw your more recent posting, fully agree that a spare board is probably the only way forward as I doubt if the contamination can ever be fully removed from the original board.

I have removed the series resistance chain board from the scrap AVO 8 Mk5 that I have, washed it quickly with IPA.  It now needs some time to recover fully, but my Keithley 2000 at turn on reads 20.043M from the cold end of R16 to the hot end of R24.  Most importantly the resistance is slowly drifting slightly upwards, not downwards, which is what I would expect from a newly washed in IPA board with almost forty year old resistors.  At these resistance values even a fingerprint can produce serious errors.

What I don’t wish you to do is to start off with yet another unknown board.  I will come back to you later on Sunday with what I find.

George G6HIG in Dover UK.

[Specmaster]

Hi,

That sounds like it could be good news, I'll check on the total chain value tomorrow and post the details here, but it will only be a summation of the total values as I have removed some of them and I have put the meter away for the night now.

Edit:

I just had to measure the entire chain from end to end, thats R16 cold end to the hot end of R24 and it is 20.076M, plus the 2k would 20.078M, just 0.37% error which would be a good result?
 

[George Edmonds]

Hi

Overnight test, R16 to R24 is 20.06M

George G6HIG

[Specmaster]

Hi,

There is something very odd going on here, when I measured the chain last night, I measured from R16 to R21 and then measured R21, R22 and R23 and added them onto the R16 to R21 because I had removed those last 3 resistors from the board. This morning I was unable to find any evidence that there was a problem with the board, no matter what or where I probed, I could not any get any sign of leakage, zero.

So I resoldered those resistors in place and measured it again and from the cold end of R16 to the hot end of R23 (8M) I get 20.17M. Add on the 2K meter circuit, totals 20.19M near enough to be within the quoted tolerance of 1% 

The ribbon cable is just about shot to pieces now with the desoldering and resoldering taking its toll on it, so I have ordered 1 metre of 28awg multicolour ribbon cable to replace the plastic rubbish, and wire this point to point, doubling up on the same colour where the blue plastic ribbon has a couple of thicker tracks.

This cable will hopefully be with me on Wednesday, as should my insulation tester, so I'll know more then.

I'm wondering if the old ribbon cable was to blame all along and if so, the damage only seemed to manifest itself when the meter was all reassembled at which time the ribbon cable is pinched quite tightly between the switch board and the meter reverse and cutout switches. 

[2N3055]

Just a note, old phenolic paper boards can absorb acids/bases/salts and they then become very humidity dependent...
I'm not saying that is problem, but something to keep in mind. Also on one device I found a flat cable that was basically strands of wire caught between two parts of PAPER adhesive tape.. Battery leak made it a defacto resistor pack...

[Specmaster]

Yeh, thanks, that is the problem with those old type of boards, but I would have thought that if the paper had absorbed something that the board itself might have started to delaminate, warp or bubble but it shows none of these signs. It has been a complete mystery from the start. I got from a retired electrical engineer who said it had in a drawer for years and that was all he could tell me about other than it worked last time he used it, but that might not have been on DCV. He said that he was only selling it because his Son was joking with him saying that when the time came, he would not have a clue about what the various bits of his tools were or how they worked etc. So that prompted him to dispose of them now while they may of some value to someone.

The larger of the two ribbon cables is all connected to the range switches is mounted in such a fashion that it makes it impossible to really inspect the contacts 100% unless the complete ribbon is removed which would completely destroy the cable anyway. Maybe when the new ribbon cable arrives, I might discover the reason for the strange happenings. This is my third Avo 8 meter, the other 2 are properly built to a specification rather than a price.

The mark 5 meter onwards was not Avo's finest period of history coming along as it did, after Thorn acquired them and set about as most companies do, milking the maximum profit from the products

[Specmaster]

Update time, as you will have learnt from previous posts, I am going to rip out this god awful plastic ribbon cable and instead wiring it with some flexible multicoloured ribbon cable which will be more suitable I feel as it will not tend to crack and will generally make it far easier to remove boards for inspection etc should it ever become necessary in the future.

To make this process easy, it is advisable to trace out just where the connections on the existing ribbon actually go to reduce the chances of cross wiring as not all the terminations are clearly identified on the various switches and boards.

Here are some preparations I have I done so far towards this goal. Later I will remove the top switch board to gain access to the bottom layer of switches and the old ribbon, which is not numbered or identified as to what each terminal is doing and I will post these pics, both before I remove the cable and after for clarity to aid the next person who decides that they want to this.

[George Edmonds]

Hi

The complete layout for the main printed ribbon cable is in the AVO 8 Mk5 manual.

George G6HIG in Dover

[Specmaster]

Hi

This is absolutely correct but only the switch positions are marked (only on the layout drawing, and nothing for the resistor board end) the actual printed ribbon does not carry any printing on that I can see. I aim to take a detailed photo of the chassis, showing all the switch terminals and then label them and post that as belt and braces for anyone to use if they so wish.

I just know that I'm going to have to destroy that printed ribbon when it comes to getting off the switch terminals as my desoldering gun cannot handle the spade terminals on those switches.

[Specmaster]

Eureka, I'm pretty sure I've found the smoking gun on this son of a bitch, conductive gunk, hidden and trapped between the range switches and the ghastly printed ribbon cable, the latter was totally fucked in the process of unsoldering it from the switches, just I predicted, so it will be point to point wiring now on this meter.

Now to bathe and scrub these switches to banish this gunk once and for all, fingers crossed the rest works OK after this.

[Specmaster]

Update on Avo time. Switch contact rings popped into ultrasonic cleaner on a long cycle and then left to air dry, switch mounting wells also given a thorough cleaning with IPA and air dried. Tested contact ring and  there is no sign of any residue of the conductive gunk and no matter where I pop a can find nothing conductive that isn't meant to be.

Bottom section of switches reassembled and now labelled ready for the new ribbon cable to be delivered and installed.

[Specmaster]

Gah  this is evil, did a quick check this morning and there are still a couple of contact positions that actually have a 30M connection between them 

Back into the ultrasonic cleaner they go again. If the meter thinks its going to beat me, it'd better think again, I'm going to whip its candy arse into line yet 

[George Edmonds]

Hi

I finally took my scrap AVO 8 Mk5 apart, the already damaged printed ribbon cable was completely destroyed in the process, as expected.  The switches were very clean and showed no sign of any contamination.

I cannot see how the contamination got into yours unless it was done during manufacture.

George G6HIG in Dover UK 

[Specmaster]

Yes, with care you can remove the printed ribbon from the resistor board a number of times, but the switches, no, that is certainly not possible. A piece of poor design IMO. The manual makes me laugh, it very clearly states that soldering the ribbon, to use only just enough heat and minimal solder to make the connection. On the switch there was a massive amount of solder. This meter still had the makers seals over the screws so if the gunk was the residue flux from that solder, then yes it was done at manufacture and has taken this long to become conductive. On the other hand, if a D cell had leaked, then the switch with the problem is directly below the D cells location, but no sign of leakage inside the case.

We will never know just how it came to be there, but it is a real PITA.

[Specmaster]

@George Edmonds

Have you stripped your scrap AVO 8 Mk5 right down to the contact rings that the printed ribbon connected to?  If yes, could you measure the resistance between the contacts on the DC range switch for me? Thats the switch ring on the right when the meter is on its face with the switches nearest you, starting position 1 as on the pic below (switch indicates AC) and progress round in (from the back clockwise), 10A, 1A etc and see what resistance figures you get on your contacts?

I think my contacts have been buggered internally and are tracking in the moulding and the resistance readings I have are dramatically affected by temperature, the higher the temp, the lower the resistance.

Contacts     Value
1 & 2          30M
2 & 3          44M
3 & 4          O/L
4 & 5          41M
5 & 6          O/L
6 & 7          O/L
7 & 8          37M
8 & 9          24M
9 & 10        23M
10 & 11      34M
11 & 12      O/L
12 & 13      50M
13 & 14      O/L
14 & 15      44M
15 & 16      50M
16 & 17      O/L
17 & 18      O/L
18 & 1        O/L

Considering that contact 11 is 3V, 12 is 10V, 13 is 30V, 14 is 100V, 15 is 300V, 16 is 600V and 17 is 1KV then those resistances on mine are not very helpful.

[George Edmonds]

Hi

Will get back to you as soon as I am able to make the measurements

George G6HIG in Dover UK.

[Specmaster]

Measured the same contact ring this morning and the results are as follows:-

Yesterday                Today
Contacts     Value
1 & 2          30m       55M
2 & 3          44M       111M
3 & 4          O/L        142M
4 & 5          41M       62M
5 & 6          O/L        500M
6 & 7          O/L        500M
7 & 8          37M       55M
8 & 9          24M       45M
9 & 10        23M       38M
10 & 11      34M       125M
11 & 12      O/L        250M
12 & 13      50M       500M
13 & 14      O/L        125M
14 & 15      44M       77M
15 & 16      50M       111M
16 & 17      O/L        67M
17 & 18      O/L        200M
18 & 1        O/L        200M

Both sets of readings were taken using a Brymen BM867s, yesterdays were taken in straight ohms range, which tops out at approx 51M. Today I noticed many more readings were O/L therefore in excess of 51M so I switched into nanosiemens and measured again and then converted the readings into Mohms.

Now this has set me thinking, does this mean that these contact rings have become sort of porous, taking in moisture when the humidity is a certain point, therefore meaning that almost any meter reading taken on this AVO will never be trustworthy. Or is this something that I may have caused, unwittingly by using a switch cleaning lubricant, in this case Servisol Super 10, which I have used for years with no problems. Has this one occasion been one where the cleaner may have reacted with excess flux/and/or the printed ribbon and made the cleaner conductive even  In which case I have fffed up my own meter?

In view of the above results, I think I now need to repeat the same test on the AC switch contacts with the meter set nS mode. I do that and post the results shortly.

[Specmaster]

Here are the measurements for the AC switch ring

Contacts     Value
1 & 2          125M
2 & 3          1250M
3 & 4          52M
4 & 5          S/C
5 & 6          62M
6 & 7          270M
7 & 8          500M
8 & 9          1250M
9 & 10        200M
10 & 11      200M
11 & 12      1250M
12 & 13      530M
13 & 14      263M
14 & 15      500M
15 & 16      750M
16 & 17      200M
17 & 18      17M
18 & 1        66M