Avo EA113 temperature-related zero drift

[Phil G4SPZ]

I repaired an EA113 a while ago. One of the components I replaced was C34, a 10uF tantalum bead capacitor near the input to the DC amplifier, because this component seemed to be very sensitive to temperature changes.

Over time, the temperature-related zero drift has returned. The drift is downwards (i.e. below zero) when it's cold, and upwards when warm. I am guessing that any DC leakage through C34 will upset the zeroing. Does this capacitor need to have a particular specification, or leakage or temperature performance? Annoyingly, the service manual doesn't mention zero drift as a possible fault, although it does describe C34 as "solid tantalum".

I've not come across this problem mentioned in any previous threads about the EA113. Can anyone advise, please?

[karesz*]

Hi Phil,
select an film-i.e. polypropilen- SMD condenser please, I think it hase enough place if you will leaded it and so do solder into, and it will the leakage zeroed...
rgds, Karl

[AndyGilham]

The original design used a 10uF 10V Tantalum from STC. Its referenced as a Tag 10/10. Thats all the info i have.

[Phil G4SPZ]

Thanks Andy. I assume that just means 10uF at 10 volts. My problem is that I'm struggling to understand fully the operation of the DC amplifier and I'm not actually convinced that leakage through C34 could cause temperature-related zero drift - despite my observation of the symptoms and the apparent initial cure by replacing C34.

As I understand it so far, VT1 and VT4 form the chopper modulator and demodulator respectively, and VT2 and VT3 form the fixed gain AC amplifier. I think that R11 and C5 (1uF +/-20% at 100V) form the low pass filter, and therefore the circuit around C5, VT6 and VT5 is more likely to be the cause of the temperature-related zero drift. But I'm probably going to have to try the old focused heating/cooling trick to try to pinpoint the component responsible.

[pmmunro]

Phil.

Do you have a copy of the service manual for the EA113?

PMM

[Phil G4SPZ]

Yes, Peter, thank you. It came with the meter and is called "Servicing Information", but apart from the useful exploded diagrams, and the circuit diagrams (which contain errors) and the components list, the manual I've got isn't really much help. Zero drift isn't mentioned in the fault finding guide. There is also no description of the circuit operation. If you know of any better documentation, I'd be very grateful!

[pmmunro]

Phil,

I've found a copy of the manual which has pages numbered 31 & 32 which have brief circuit descriptions which may be missing from your version.

I've also attached a copy of the DC amplifier circuit diagram in the hope that someone with experience of these may have some suggestions. The one component with known ageing characteristics, other than C32, is C2. It's value is not critical but leakage could be. From the selection procedure given in the manual, the JFET VT1 is obviously critical, so with the possible difficulty of getting a batch of the from which to select one with the required characteristics, I would hope it's not causing trouble.

PMM

[ronbryan]

The DC amplifier (VT5, VT6, VT7 and VT8) can be considered as an integrator, with input from demodulator transistor VT4, input resistor R11 and integrating capacitor C5. If VT4 saturation voltage changes with time, then there will be a drift in the zero setting. To eliminate the DC amplifier/integrator from the investigation, you could try temporarily shorting VT4 collector to emitter, and see if the zero setting drifts under those 'test' conditions.

Ron

[Phil G4SPZ]

Thank you both for the helpful information. Peter, I hadn't got those pages, and they have been very useful in enabling me to understand the circuit better.

Ron, many thanks, I'll definitely try that test, although the zero drift is not related to time but to ambient temperature. I can actually cause the zero to drift upwards by holding my hands round the outside of the case; after half a minute or so, the zero creeps slowly upwards. Take my hands away, wait two minutes and the zero has drifted back down again. In cold weather, the pointer goes down below zero, but in warm weather the pointer sits well above zero. At constant temperature, the zero is perfectly steady.

I perhaps didn't mention that the meter reads absolutely spot-on accurate on all ranges. The drifting DC zero makes it tricky to use the left-hand zero setting, but when set to centre zero using switch 'C' all ranges reads correctly, as do the ohms ranges.

One thing I have noticed is that the pointer appears very heavily-damped and slow-moving and takes several seconds to reach its final position. Setting FSD ohms and the centre zero is also very sluggish. Are these instruments all like this? It's almost as if C5 is too high in value, or possibly leaky.

Looking at the circuit and trying to understand how the bias stabilisation works, the AC amplifier VT2 and VT3 appear straightforward, and C3 prevents any drift getting to the DC amplifier anyway. With no input, and the chopper running, am I right in thinking that the voltage at VT4's emitter will be around 0.2 volts above zero? I can see that if this drifts, the output at VT8 will drift also.

I have two other electronic multimeters by other manufacturers, and they both have a front-panel mounted 'set DC zero' control! So easy...

[Phil G4SPZ]

Correction - VT4's emitter will be around 0.2 volts negative with respect to 0 volts...

[ronbryan]

I think that chopper transistors are used in 'inverted' mode to achieve a very low saturation voltage. In addition, some transistor chopper designs I have seen, use high frequency transistors such as the 2N918 or MPS918 to minimise chopping spikes being carried through by inter-electrode capacitance.

Here is a link to a recent EDN article which states that V(CE)sat is in the low mV (edit: actually it says microvolts) when inverted mode is used. I have always assumed that this is a positive voltage.

https://www.edn.com/electronics-blog...mode-switches-

Ron

[barrymagrec]

Originally Posted by Phil G4SPZ

One thing I have noticed is that the pointer appears very heavily-damped and slow-moving and takes several seconds to reach its final position. Setting FSD ohms and the centre zero is also very sluggish. Are these instruments all like this? It's almost as if C5 is too high in value, or possibly leaky.

I used one of these when it was new, it always seemed a bit sluggish to me but at the time it saved carrying around two separate meters.

Somewhere I have a part broken unit I just kept because the movement seemed to good to junk, if anyone needs any parts I`ll look it out.

[pmmunro]

Barry,

I have an EA113 with burnt switch contacts on the PCB, so I would be interested in acquiring a replacement board if there's one available.

If this a part you have, please send me a private message letting me know what you would like for it.

PMM

[Phil G4SPZ]

Originally Posted by ronbryan

I think that chopper transistors are used in 'inverted' mode to achieve a very low saturation voltage... V(CE)sat is in the microvolts [region] when inverted mode is used. I have always assumed that this is a positive voltage.

Thanks Ron, it makes more sense now. I was trying to puzzle out how on earth VT4 could behave as a shunt switch as drawn. It's made more complex by the dual voltage rails.

Here's my stab at what is happening with no DC input. The voltage at the junction of C3/VT4 emitter is zero (or just a few microvolts above zero, due to the clamping action of VT4) and hence the voltage at VT6's base is also zero. However, VT6 is conducting due to its emitter being returned to the -2.8 volt rail via part of RV4 and R14. VT5 is also conducting, to a degree set by the position of the 'set zero' pot RV4. VT7 and VT8 also conduct, to an extent such that the net potential difference across the meter, i.e. between the junction of D4 cathode/R17/R16 and RV1/R20, is zero.

With a positive-going DC input present, VT6 conducts harder and the current through VT5 falls, causing VT7 to turn on harder. The current flowing through emitter-follower VT8 falls and the potential at its emitter and at the cathode of D4 moves more negative, creating current flow through the meter and moving the pointer clockwise. The same mechanism applies with a negative-going DC input.

I think this means that with no input, any drift of zero caused by temperature change could be caused by VT5, 6, 7 or 8 or any of their associated components, but not VT4?

I also think I can see a feedback mechanism whereby any drift in the output sends an opposing potential back to the input, via the gain-setting resistors R8, R9 and R10, which would act to correct the drift.

My head is now spinning after studying this circuit for the past two evenings. I'm going to give myself the weekend off, then tackle the meter on Monday, armed with a soldering iron and some freezer spray!

[Phil G4SPZ]

Sorted! Peter, you were spot on. The faulty component was indeed C2, a 64uF 4 volt electrolytic, but in this case rather than going leaky it had lost its capacity virtually completely, and when I snipped it out it measured just 270pF! Although the components list states that C2 is 64uF, the component originally fitted was a 47uF and I used a Philips electrolytic of the same value to replace it. Rather than dismantling the instrument, I used the time-honoured trick of snipping out the old component and soldering the new one to the wire stubs on the component side of the board.

I will try to write this up as an extension to my original Success Story because it has been an interesting, albeit frustrating, repair.

Thanks to everyone for their interest and help.

Phil

[Phil G4SPZ]

Bill AC/HL has kindly re-opened the original Success Story, and I have brought it up to date.

[pmmunro]

Phil,

I'm really pleased for you. You now have a very useful meter and it's good to know it's with someone who can appreciate it for its technical qualities.

I think you have done the right thing soldering the new capacitor to the old leads rather than risking any damage to the PCB tracks. Some authorities advised this as the correct method of repair when PCBs first became common.

PMM

[Phil G4SPZ]

Peter, thank you very much for your kind comments and also for the helpful advice you provided. I believe the repair technique was adopted by many TV repairers as the quickest and most cost-effective way of replacing individual components where visual appearance was of no consequence. One precaution I do take when working this way is to clamp a self-gripping pair of forceps or tweezers around the wire 'stub' to act as a heatsink and prevent the heat travelling down the wire and melting the solder on the print side.

In normal circumstances where the PCB is accessible from both sides, I would do a 'proper job' including removing the old solder from the print land with a sucker and desoldering braid, but in this case the print side was completely inaccessible without major dismantling. There's an element of laziness in there too! A future owner will be able to see where the repair was done, but as this is a piece of operational test gear and not a valuable or museum-grade vintage radio, I don't see anything wrong with such an approach.

I must say that the more I use the EA117, the more I like it and the less I am using my traditional Avometers. It takes the place of a traditional multimeter, plus an AC/DC millivoltmeter, and its five ohms ranges covering from a couple of ohms to over 50 Megohms are more than a match for most resistors I'm likely to encounter. I used it just today to check the movement of another multimeter with a 9.5uA FSD. Quite a piece of kit, and no wonder they are sought-after.

Thanks again,

[mole42uk]

Just checking one of mine today I found that when set to DC the meter needle swings left, below zero, and the meter doesn’t respond to DC voltages. The AC ranges and the battery test work perfectly well.

I haven’t started serious testing yet, it was too hot in the shed, but I suspect that it will be necessary to delve into the DC amplifier so I thank you for the circuit description.

[Phil G4SPZ]

Hi Richard,

It was bloomin' hot in my shed today, too. When I went in this evening for the first time since yesterday, I realised that I had inadvertently left the EA113 switched on and set to the 1 Megohm range. Happily, this unscheduled 24-hour soak test with the movement sitting at FSD has had no ill-effects, despite the ambient temperature hitting close to 30 degrees, and the zero has not drifted one iota, so I'm confident it is now well and truly 'fixed'.

If you don't mind my making a couple of suggestions, the first thing I'd check is the action of the set zero preset RV4. If you can get the pointer to zero, even if it subsequently drifts, it suggests that VT5 to VT8 inclusive are working OK. Do this check with the meter set to the 100 volt range or lower. Does the centre-zero function work properly? If not, it suggests that the chopper isn't running.

The next step is to check that the multivibrator VT9/VT10 is running, as without the chopper operating there will be no pulses coming from C3 even with a DC voltage input. The easiest way to check the multivibrator's operation is to use a signal tracer (or a scope, or even a pair of high-resistance headphones) between 0 volts and the cathode of D3, when you should hear or see the square wave at about 600Hz with an amplitude of around 5 volts peak-peak.

Hopefully these will localise the fault and point to the next steps. For what it's worth, I measured the DC voltage at VT3's collector at +1.84 volts. VT8's emitter was sitting at +0.5 volts.

Whether by accident or design, it is fortunate that the majority of the components on the component side of the DC amplifier PCB are accessible for fault-finding without having to remove the range board. The AC amplifier section is totally covered by the range board, but fortunately yours seems to be working OK!

Good luck, and please let us know how you get on.