AVO 2 panel valve tester Meter and other Oddities

[Matthew kane]

Have 3 of these to restore and 2 of which so far have some underlying problems. 3rd unit is a Patern version so I'll leave that for now. Lots of inconsistencies across both units.

Unit 1, Outer 1.5-1.8k set zero wirewound resistor card is damaged. SWAMP measures 16.9 ohms. Meter actually measures 20.3 ohms. Have not measured the FSD. Will leave this out of the way now due to fact the set zero wirewound is cactus. mA/V pot measures 30 ohms.

Unit 2, SWAMP measures 23 ohms. Meter measures 27.1-27.4 ohms, SWAMP measures approx 23 ohms. I did a quick test with FSD on the meter and it registers around 980 uA FSD from 0 to approx 10 on the scale. When adjusted for 600-700 uA the pointer is about 2/3 on the scale of true FSD. mA/V pot measures 76.5 ohms. 9 ohm shunt resistor measures 8.9 ohms and 1 ohm 100mA shunt resistor measures approx 1.2 ohm.

Based off the EMER manual and some old snippets of testing the meter it is specified with some inconsistencies with right values for the meter FSD. One manual suggests its a 600 uA between 30-32 ohms resistance and another manual suggests 700 uA 27 ohm resistance. The mA/V pot should be 90 ohm and 40 ohms approx for the SWAMP.

Now the question is, why are there so many inconsistencies with both testers? Could it be because the parts were hand wound and hand crafted that they had to compensate for where omissions are made and/or old age of the tester or did the engineers at AVO use different lots of stuff here and there based off manufactured batches?

In one of the pdf snippets it says to refer to a factory manual where a battery and variable resistor circuit used to measure the ideal optimum FSD of the meter, anybody have this actual circuit? I've been fiddling around with the magnetic shunt and south and north pole set zero adjustments on one of the meters with little success.

Any tips/helpful information pointed my way is greatly appreciated.

[Craig Sawyers]

I don't know whether the original 1936 patent gives any further insights

Craig

[Electronpusher0]

I did a lot of work on the meter sensitivity and the effects of the variations of the components around the meter circuit. Se here
https://www.vintage-radio.net/forum/...d.php?t=162366

Attached to that thread is an Excel spread sheet which allows you to plug in the actual values you have for the swamp and pot and see what the effect on accuracy would be.
On one of my units I ended up optimising the values of the shunt for the (low) value of my pot using the the spread sheet.

The op amp circuit I mention in that thread was not successful, it worked at dc by the AVO uses half wave ac so the peak values were too large.

Peter

[Dekatron]

I've only owned one of these for a short while, it was shipped to me and arrived in a broken state so I had to return it to the seller, so I haven't had more than a quick look inside one of these and only had a quick look at the circuit - however I have more experience on later models.

I have several different circuit diagrams and texts that can be found on this forum and elsewhere on the Internet and they speak of 600uA, 660uA (Patent GB480752 on page 9 - they also specify the Eureka wire used to get the shunt resistances in this patent), 700uA and 750uA movements and they all have the series resistor as (approximately -adjusted to get correct readings) 40 Ohm and the meter internal resistance at some 23 or 32 Ohm (the lower value could be a misprint or AVO might have changed the resistance of the meters to compensate for different FSD).

Since some documents mentions that you should adjust the sensitivity by adjusting the resistance of the 40 Ohm series resistor AVO therefore want a total voltage drop across the meter and the series resistor to match a certain value so that the meter indicates correctly in the circuit with the shunt resistances. Tracing the circuit and making a simplified circuit will probably help you to calculate the values needed for the circuit to work properly. Don't forget that we are probably speaking about half wave rectified voltages/currents here so on the 100mA range the circuit will see a mean value of 50mA due to the valve itself working as a rectifier - I might be wrong here but AVO used this design in later models and I believe they did here too.

[Dekatron]

If you measure the three resistances in the mA/V control as well as the 40 Ohm series resistance and the internal resistance of the meter it should be possible to calculate what AVO used for FSD current on your meter when it was new - unless someone has tampered with any of these resistances. The three resistances in the mA/V control should be close to 90, 9 & 1 Ohm.

[Electronpusher0]

Originally Posted by Dekatron

If you measure the three resistances in the mA/V control as well as the 40 Ohm series resistance and the internal resistance of the meter it should be possible to calculate what AVO used for FSD current on your meter when it was new - unless someone has tampered with any of these resistances. The three resistances in the mA/V control should be close to 90, 9 & 1 Ohm.

If you go to the thread linked in post 2 I have posted an excel spread sheet that performs the calculations

Peter

[Dekatron]

Yes, I saw the Excel-sheet but what I was thinking about was to use the resistances to calculate the FSD that is needed, the FSD is a factor you enter in your spreadsheet.

On a few occasions I have seen people measure the resistances in-circuit, so they were measured with all of the other resistances in parallel which resulted in incorrect values - I don't think this is the case here from the values in the first post, but it might be, so therefore I asked for a re-measurement of them to calculate the necessary FSD so it could be compared to the current FSD of the meter.

In some occasions the moving coil and series resistor and the shunt resistors have been overheated resulting in resistance changes which might also explain the differing results.

[Electronpusher0]

Originally Posted by Dekatron

Yes, I saw the Excel-sheet but what I was thinking about was to use the resistances to calculate the FSD that is needed, the FSD is a factor you enter in your spreadsheet.

Agreed, but if you plug in the resistor values you have you can then try different values for fsd and see what one gives the minimum error.

If you like, measure the resistance of the shunts, the swamp, the meter and the mA/V pot and I will do the calcs.
I measured the values of the pot ends to wiper for each mA/V setting but if you just measure across the ends I should be able to work it out.
Of course all the components will have to be measured out of circuit and you will need to subtract the meter lead resistance or use 4 wire measurement for the low value shunts.

As I set out in the thread I linked to I am convinced that 600uA is the correct fsd for the meter.

There have been some references to 700uA, my theory is that the same movement was used in both the AVO 7 and the 2 panel valve tester.
In the AVO 7 the movement is indeed 700uA BUT OVER A GREATER ANGLE OF MOVEMENT.
I have tested an AVO 7 movement over the more restricted angle range of the AVO valve tester and guess what, it becomes 600uA fsd over the shorter range.
The early AVO 7 movements with the horseshoe magnet are identical to the 2 panel AVO valve tester, the scale mounting strips and pointer length are different.


Peter

[Electronpusher0]

Going back to the original post where the OP states that the meter fsds are too high this reflects my experience with both my 2 panel AVO testers. The old horseshoe magnet seems to lose magnetism and is very sensitive to being touched with magnetic screwdrivers, just using a magnetic screwdriver to remove the meter caused it to become even more insensitive. I ended up buying a non-magnet screwdriver.

One unit I managed to calibrate by reducing the swamp resistor, this gave acceptable results.
The other way one was far to high an fsd to use and set me off building a re-magnetiser, reported here:

https://www.vintage-radio.net/forum/...d.php?t=165159

In post 48 of that thread I reported that I could only increase the sensitivity to 700uA and gave up. Instead I dumped the horseshoe magnet and made a new one using a Neodymium magnet and iron pole pieces. That was very sucessfull and acheived a sensitivity of 320uA which required a shunt resistor to get it to 600uA:

https://www.vintage-radio.net/forum/...2&postcount=48

Peter

[Matthew kane]

Replied to your PM Peter.

On the other hand to summise what I've replied to Peter's PM. Due to the unbalanced and "contaminated" hairsprings on both units, I ventured into new territory and carefully disassembled both meters to clean and rebalance and align the hairsprings.

The only problem is, it seems the first unit (meter already had issues) the needle is basically free flowing. No magnetic field in the center between both magnet poles to dampen the movement when current flows through. Fortunately the coil resistance is still measurable and not open circuit at an approx 26 ohms. The horseshoe magnet still has magnetisim, but there is no reaction to the center iron cylinder piece between the needle coil when current flows through.

Second tester meter, same thing with this one, disassembled and cleaned. Made sure no contact between each turn of the hairsprings. Decided to do a proper FSD current reading and this meter measured 1.5mA, a far cry from the AVO spec-ed 600-700uA range. FSD resistance is around 76 ohms. Tested via my Agilent DMM in series with a 10k resistor and my HP DC bench supply. The needle is very loose with almost no damping.

Is it possible to replace both meters altogether with a 3.5 digit range 200mV DPM?

[Electronpusher0]

Hi Matthew, I'll reply to your PM but to answer your question regarding replacing the meters with a 3.5 digit range DPM.
Unfortunately this is unlikely to work without additional circuitry. The problem is that the AVO tester applies ac to the valve and relies on the valve half wave rectifying the ac.
Hence the current through the meter is pulses of half wave rectified ac.
For an average fsd of 600uA the peak value of the current will be 1.887mA (Average is 0.318 of the peak)
The analogue meter smooths the pulses (somewhat, the meter still has visible vibration) but if you applied this to a DPM you will get a widely varied reading depending on exactly when the DPM samples the input.
To use a DPM you would need a circuit to turn the pulses to dc but with a fast enough response that the meter is not over damped.

Having said that it may well be worth persuing, these meters all seem to have lost sensitivity over the years.
Also using a DPM gives an opportunity to extend the gm range beyond the 10 that the AVO can indicate with the analogue meter.

Peter

[Matthew kane]

I recall somebody did something similar in the avo mk4 thread (forgot link now).

Wonder if I can apply that to this.

[Electronpusher0]

Originally Posted by Matthew kane

I recall somebody did something similar in the avo mk4 thread (forgot link now).

Wonder if I can apply that to this.

Someone did use an op-amp circuit to overcome a weak magnet hence an insensitive meter, this amplified the half wave pulses and the analogue meter provided the averaging. A dpm is a different challenge for the reasons given above.

However I have sketched out a circuit based on a Sallen-Key lowpass filter, I will breadboard it and post the results.

Peter

[Electronpusher0]

OK, the low pass filter idea was not right.

What is needed is a sample and hold with decay to grab the peak value.

Attached is the schematic of such a circuit using a TL072 or similar dual op-amp.
The DPM is a cheap readily available dpm module.
The output is 2V dc for a gm indication of 20, this can be attenuated with a divider to give the voltage to suit the dpm.
The pot is used to calibrate the output.
The supply is + / - 9v, I chose this because 9V is also the voltage required for the DPM supply, but anything between 5 and 15V can be used. The supply will have to be isolated as the circuit floats at Anode voltage.

To use this in the AVO 2 panel valve tester the meter and swamp resistor are replaced with a 68 ohm resistor, the voltage derived across this resistor becomes the input to the sample and hold as shown on the schematic

I have breadboarded the circuit and tested it using a simulation of the input waveform derived from a low voltage transformer with diode and attenuator to give the required 264mV peak half wave rectified waveform which is what would be expected from the AVO at a gm reading of 20.

The circuit works well with a stable display as can be seen from the photos.
The 2.2uF capacitor and 470K resistor set the decay time, reducing the value of these increases the response but also increases the meter jitter which shows as the least signicicant digit "hunting" by one or two counts.

I now need a meter panel from a 2 panel AVO valve tester, one with a broken meter, to test the circuit fully. I am loth to play with my calibrated unit.

Anyway I offer it as an idea.

Peter