2 Panel AVO valve tester meter sensitivity

[Electronpusher0]

I have started the restoration of my 2 panel AVO valve tester and have been trying to find out what the basic sensitivity of the moving coil meter is.

Reference has been made in posts on this forum to 700uA sensitivity.

A search of available documents yields the following:
Y812 PART 2 – TESTERS, VALVE, AVO – TECHNICAL HANDBOOK - FAULT FINDING AND REPAIR DATA :
“M1 – 0.7mA f.s.d. – 5 ½ in. scale.”
Y814 – TESTERS, VALVE, AVO – TECHNICAL HANDBOOK – FIELD AND BASE REPAIRS – Section 6, repair of meter:
“The internal resistance of the meter is 27 ohm and the current at full scale deflection is 600uA”
TESTER, VALVE, AVO No 1 SECOND TO FOURTH ECHELON WORK Section 3, movement removal:
“The sensitivity of the movement should be 600uA and the resistance should be approximately 32 ohm.”

The following is extracted from - TESTER, VALVE, AVO No 1 SECOND TO FOURTH ECHELON WORK Section 10, check of meter sensitivity: (my italics)

“Unsolder, at the rectifier, the lead from the wiping contact of the SELECT ANODE switch to the rectifier. Take connections from the wiping contacts of the ANODE and SELECT ANODE switches to a series circuit comprising battery, variable resistor, and Avometer which has been standardized against a sub-standard meter. The battery and variable resistor should be chosen so that currents from 1 mA: to 100mA can be passed, and adjustment 10 the desired value effected with ease. Tum the SET MA/V control to the position marked 1, when F.S.D. on the meter of the valve tester should be obtained at 1 mA+/- 10%}. as indicated by the Avometer. If adjustment is necessary, the value of the 40 ohm (approx.) swamp resistor in series with the moving coil can be altered. As mentioned in para. 3, high sensitivity on the movement is often difficult to attain, and in such cases the value of the swamp resistor can be reduced to compensate for this. The highest possible value of resistor should, however, always be aimed at so that calibration is not unduly altered by variation in temperature. The swamp resistor is Wire-wound with Eureka wire on a fibre card and mounted above the SET MA/V control.
Turn the SET MA/V control to the position marked MA/V when F.S.D. should be obtained. at 10mA+5%. Adjustment, if necessary, can be effected by slight alteration of the value of the 9 ohm shunt resistor, which is wire-wound on fibre card mounted between the SET MA/V and SELECT ANODE.
Check the sensitivity at other Positions of the SET MA/V control between 10 and 1, when the F.S.D. in mA should agree with the setting, tolerance +/- 10%.
Turn the SET MA/V control to the position marked 100 when F.S.D. should be obtained at 100mA+5%. Adjustment, if necessary, can be effected by very slight alteration of the 1 ohm shunt resistor, which is wire-wound, next to the 9 ohm shunt.” (This is misprinted as 0.1 ohm)

I decided to go back to first principles and do some calculations on the meter and shunt circuit including the mA/V pot.
The schematic of this part of the tester is attached. It can be seen that the mA/V pot not only changes the value of the shunt but also changes the series resistance of the meter arm of the circuit.
I set up an excel spread sheet to calculate the current needed to be flowing in the meter / shunt circuit to achieve fsd on the meter at the various settings of mA/V. The meter sensitivity, meter resistance, swamp resistor and shunts are variables and can be changed to see the effects.
The excel sheet is attached. The values for RA (the resistance of the mA/V pot wiper to CW end) and RB (the resistance of the mA/V pot wiper to ACW end – the end of the wound section) are also variables, the values in the sheet as posted are given, to the nearest half ohm, to give the correct fsd at the intermediate settings of the mA/V pot. All variables are highlighted in yellow.

In practice it is best to unsolder the connections to the pot and measure RA and RB for each setting of the pot and enter them in the spreadsheet, my pot is actually 76 ohm instead of the rated 90 ohm, more on this later.

You can see from the spreadsheet, using the specified values for the pot, swamp and shunts, that if the meter fsd is 600uA then the error is 2% or less. If the meter fsd is 700uA the error is around 20%.

I therefore conclude that the meter basic sensitivity has to be 600uA.

As mentioned above, in my unit the mA/V pot is 76 ohm instead of 90 ohm. The swamp resistor is 27 ohm instead of 40. The meter has also lost sensitivity and was about 1mA fsd (I managed to make it worse by trying the trick of adding a small magnet to the side of the horseshoe magnet). When I plugged these values of pot and swamp into the spreadsheet I found that even if my meter was still at 600uA the accuracy would be around 11%.
In my next post to this thread I will describe how I used the spread sheet to optimise the shunt resistors and used an op-amp circuit to overcome the loss of meter sensitivity.

Peter

[Electronpusher0]

This is the second post, it describes the amp circuit I used to increase the sensitivity of the meter.

As described above, the main panel I bougt at RWB had two problems
1 the mA/V pot is 76 ohm not 90 ohm and
2 the meter sensitivity is much too low.

As received the meter sensitivity was 1mA fsd instead of 600uA, I actually made it worse buy trying to increase the sensitivy with additional magnets so it is now about 1.5mA fsd.

I decided to use an op-amp circuit based on those posted on this forum for the later Avo testers. The circuit is attached.
It uses a TL071 fet op-amp. R1 simulates the combination of the meter and swamp resistances. This value was optimised using the spread sheet posted in the previous post using the value of 76 ohm for the mA/V pot. Additional shunts were also needed to modify the values of the 9 ohm and 1 ohm shunts - this brought the 9 ohm down to 6.8 ohm and the 1 ohm down to 0.73 ohm. With these values the predicted accuracy is better than 5%.

I added back to back diodes on the input to protect the circuit and a 5.6V zener on the output to protect the meter, the normal output of the circuit at fsd is about 3.3V.

RV1 sets the sensitivity and RV2 is a zeroing pot.

R4 is added simply to damp the meter, it increases the current output needed from the op-amp for full scale to about 2.5mA.

For the power supply I used a Traco smpsu module that provides +/- 15V from a mains input and has 3000V input / output isolation. I basically used this because it was in the junk box. The isolation is needed because the meter in this tester floats at anode voltage.

I assembled it on a bit of vero board and mounted this on the pins of the Traco psu. The psu was glued to an aluminium bracket which was mounted on one the heater transformer bolts.

I calibrated it as described in the first post and tested it with an ECC82. it seems to work fine.
To confirm I will test the ECC82 at dc using bench power supplies to get a "standard" valve.

On with the rebuild of the valve base panel and rotary switch.

Peter

[Electronpusher0]

I have almost finished the valve base panel.
I clearly paid too much for the panel, It has obviously been wet at some stage and the "rings" surrounding the valve bases were all very rusty. The heater switch had failed and been bypassed, the roller switch segments were siezed and the wiring had been "got at" with 2 valve bases replaced with an aluminium panel bolted on carrying 2 alternative bases.

I decided a total strip and rebuild was the only way forward. I removed all the valve bases, switches etc and re-sprayed it.
I decided to fit new valve bases and fitted a sub chassis of 1.5mm black polystyrene on the underside. This carries the smaller valve bases.

The roller switches I stripped down, cleaned and rebuilt with vaseline to lubricate.

I rewired from scratch with copious ferrite beads. I have not fitted the capacitors.

I just need to fit the plug at the other end of the cable but this will have to wait for the wooden box to be finished so I know what length I need.

Peter

[IanG3XYV]

Excellent work Peter. I have a two panel AVO which basically works but one day will require some maintenance I suspect and what you have done is an excellent example of what I may have to do.

Ian

[Electronpusher0]

Thanks Ian, I resisted the urge to "dress" the wires neatly with tie-wraps and opted instead for the "birds nest" approach to minimise crosstalk.
I also took the opportunity to fill the engraved numbers on the roller switches with chinagraph wax pencil while I had the switch assembly apart.

If you decide to replace valve bases I have some B3G bases NOS which are on offer under the items offered section.

Peter