AVO Two Part Tester

[WaveyDipole]

I recently acquired an AVO two part tester from a member, sold as spares and repairs but complete and in a nice condition and already re-capped. This is the first time I have owned a valve tester so before attempting to use it, I set about downloading and studying all of the available material I could find online in addition to that already supplied on paper. I also downloaded three editions of the AVO Valve Data Manuals.

This instrument has a trip switch which presumably means it is the admiralty version? Initially there were problems zeroing it and a little investigation revealed some oily gunk on the top cutout switch contacts which has been cleaned. Also a joint on the MAV pot came loose and had to be re-soldered, so possibly there was a a dry joint.

With those two things done, the tester appears to be working - sort of. The SET ZERO control can now zero the movement instead of the needle swinging hard left on power on, There is also movement in the positive direction up to about 1 on the scale as the valve warms up, but then the cut-out activates, illuminating the red warning light.

I start with the unit powered off, setting the numbers set on the socket panel dials and the electrode voltages on the main panel. For the EBC33 there is no setting for the screen, so I set screen at minimum (60V). The short wire with the banana plug is connected to the GRID terminal and the croc clip attached to the top cap (the manual shows the top cap as the G1 electrode). MAV is set at 100 and SET ZERO to far right.

I see at least two adjustment screws on the cut-out assembly, but before I start tinkering on a trial-and-error basis, I just wanted to ask whether anyone has any documentation on this device and how to adjust it? I have so far failed to find anything online or in the materials I have downloaded so far. How does one re-set the cut-out? I see no control for this and the only way I have found so far is to gently push the switch carrying arm back into the 'on' position until it magnetically clicks against the core of the coil.

Incidentally, I am unsure whether this is a case of the tester working properly and the cut-out being too sensitive, or whether the cut-out is actually doing its job and protecting the meter due to some problem.

[Ian Blackbourn]

As I recall, the switch marked "mA/V and c.ins" will reset the cutout when moved to the right. I don't think it's mentioned in the instructions, leaving one to have a panic attack!

Ian

[WaveyDipole]

Ian, thanks. Will test that.

Forgot to mention that all voltages were checked and found to be Ok. Anode voltages were a bit high, but the manual does explain that this is to be expected, so as far as I can tell all voltages are as they should be.

[chriswood1900]

This thread may be useful:- https://www.vintage-radio.net/forum/...&highlight=avo
In Particular:-
The final problem hinged around the cut-out which is only on the navy versions, it was making poor contact and it did not seem to work properly, I could not find any setup details so I came up with my own. First cleaning the contacts with some thin card soaked in contact cleaner I then adjusted screw A so the contacts fully opened then adjusted screw B so that the spring had just enough tension to hold the contacts open the contacts normally being held closed by the residual magnetism until excess current is drawn. It is reset by a thin wire attached to the underside of the MA/V G.INS switch. If you see the last photo in the thread above I have marked the 2 screws.

[WaveyDipole]

Ian, your recollection is correct. Pushing the mA/V and C.INS switch does reset the cut-out.

Chris, I do see the wire you are referring to which serves as a push rod, pushing the swich contact back towards the coil.

I am starting to wonder whether the cut-out is, in fact, doing what it is supposed to. Is it normal for the tester to hum? The cut-out activates maybe a second or two after the valve warms up and the meter pointer starts moving in a positive direction, even if I try to back off the deflection. The meter has a background hum that intensifies to a loud hum just as the cut-out activates. I am therefore wonering whether something else is amiss.

I am testing an EF39. aV is set to 250V, hV to 6V and screen to 100V. Anodes is set to NORMAL. The top cap (g1) is connected to GRID. The dials are set as per the manual - 026 510 310. Not sure what else I would be missing?

[chriswood1900]

WaveyDipole
I can't remember the hum level but I would expect a moderate not loud hum when working correctly. Loud hum would indicate something drawing excess current or something loose. Your valve settings are OK but on a AVT the anode should be 100V and expect 2.2mA/V.
It should be noted to use the AVT 2 panel the data is missing from the later data books after about Issue 17.
Use this copy here:- https://frank.pocnet.net/instruments...AVO_VDM_17.pdf
the EF39 is on page 181.
I hope this helps.

[Chrispy57]

Hi Wavey - that change in hum level does sound suspicious, particularly as it seems to coincide with the cut-out tripping. It would be simple to monitor the mains current during that sequence, and if significant move on to monitoring HT and heater currents to narrow down the location of the increase. My non-admiralty 2 panel tester does not hum in that manner.

Cheers
Chris

[WaveyDipole]

Well it seems I have made two mistakes.

The first is looking at the wrong data. Chris (Wood), thanks for pointing out that I should be looking at the earlier versions of the manual. I thought those columns looked different in your printout!. The older version 15 and 17 do indeed have that extra column for the "AVT", whereas I was looking at the VCM Mk1-4 data in version 20.

I also tried the suggestion mentioned by Chrispy57, but what followed next changed everything.

My second mistake was downright stupid. It seems that the connecting plug between socket panel and meter will locate in two positions. Theoretically it shouldn't, but it does. I seemed odd that it felt a bit tight but as it went in almost all the way I thought no more of it. However, just a while ago I had the meter at the edge of the bench and with the socket facing me so that I could easily re-check all the voltages. I then plugged in the panel while the meter was still in that position with the socket facing me and it went in very easily and snugly which took me somewhat by surprise. This, then, must the the correct position. In my defence, there is no locating pin and when the meter is flat on the bench and the panel alongside it, connecting the plug is rather awkward at that angle as the cable catches the bench. I had inadvertently inserted it one pin counter-clockwise. I still feel like an idiot though!

With that sorted and the plug in the correct position, the meter no longer trips and has just a quiet hum which is more in keeping with expectations. However, it does not seem to be measuring much either. All voltages are still correct for heater and screen. For the anode voltages, D and REC are approximately at the stated voltage, the 125V to 250V are about 30% high (which is just about within limits), but the 60V and 100V settings are actually rather high 109Vac and 154Vac respectively. I Noted the figures using both a DMM and an AVO meter (the AVO closely matched the DMM readings), but hadn't actually done any calculations previously.

So its kind of back to square one, but some things learned along the way.

[ms660]

You're probably aware of this but for anyone who isn't the unloaded AC Voltage supply to the anode should be 1.4*The selected DC anode voltage, the unloaded AC voltage supply to the screen grid should be the same as the selected DC screen voltage, all within +- 5%

The Avo patent:

https://frank.pocnet.net/instruments.../AVOpatent.pdf

Lawrence.

[WaveyDipole]

Having checked the circuit diagram, I see that the anode voltage setting I was referring to as 60V is actually 80V, so 109V still falls within the 30% limit. This means that voltages on the secondary transformer are OK. That leaves just the 100V setting which turns out to be connected to a winding on the mains transformer. At 154V the 100V setting voltage does seem rather high even with 30% + mains variations. Still, I did check the mains voltage setting before I plugged it for the first time and it is set to 240V.

Originally Posted by ms660

You're probably aware of this but for anyone who isn't the unloaded AC Voltage supply to the anode should be 1.4*The selected DC anode voltage, the unloaded AC voltage supply to the screen grid should be the same as the selected DC screen voltage, all within +- 5%

The Avo patent:

https://frank.pocnet.net/instruments.../AVOpatent.pdf

Lawrence.

Thank you for the additional information. The "Second to fourth echelon work" document does state that the voltages may be 20% - 30% above nominal "to allow for transformer regulation" so I was aware that they would be higher. Your value of 1.4* would allow for them to be a little higher.

The good news is that I am finally getting somewhere, in part thanks to your comment about the screen grid. I had it set at 60V (the Valve Data Manual does not give a value so I just set it at the lowest) but after changing it to 100V to match the set anode voltage, I started to get some sensible readings. I also had to make a small adjustment to the mA/V + PERV pot as nothing happened until it was turned past MAV to 10. The wiper was slightly out of line. I slackened the grub screw, set the wiper dead centre of the MAV contact and re-tightened. Once adjusted, I could zero the meter with the knob on the ma/V setting. The meter read between 1.2 - 1.5 which seems a bit low. As per instructions I then set the ma/V knob to 2.2 which is the mA/V value stated for the EF39 in the Valve Data Manual, zeroed and pushed the ma/V switch to the left and got a deflection just into the green zone. I still get nothing on the 100 setting so will need to have a look at this another day.

I have downloaded the reference that you linked - not seen that one yet - for a bit of light reading tomorrow.

[WaveyDipole]

Lawrence, having read the patent document, the 1.4x factor makes perfect sense for rectification (as opposed to mere "regulation").

[ms660]

Originally Posted by WaveyDipole

Lawrence, having read the patent document, the 1.4x factor makes perfect sense for rectification (as opposed to mere "regulation").

The unloaded anode supply voltages and screen grid voltages @ +- 5% are shown in the attachments below.

Lawrence.

[Chrispy57]

Hi Wavey - sounds like you have made excellent progress, but one of your comments caught my eye - "I Noted the figures using both a DMM and an AVO meter (the AVO closely matched the DMM readings)" - did you try using an Avo Model 7, which has a much lower resistance than the later 20kOhms/Volt models, and is likely to be what was specified by Avo for making these measurements in that historical period?

Cheers
Chris

[WaveyDipole]

Chris, alas, the AVO7 I have is not ready yet to take measurements and is still at the project stage...
The measurements were taken with an AVO9 Universal. The good news on that front is that after a bit of a delay, I have just received the leads I needed that will enable me to make the precise 4-wire resistance measurements of the components immediately around the movement that are neccessary in order to progress that project.

I was however re-assured by the accuracy of the 50-something year old AVO 9!

[Chrispy57]

Hi Wavey - I'm not really questioning the accuracy of your Avo 9, which you say agrees with a DVM, but highlighting that an old Avo 7 would load the HT and give a lower reading - which may well be in spec, given that the measurement was intended to be taken with just such a meter back in the day.

My DVM, Avo 7, 8 and 9 are all calibrated but give different readings on an unregulated open circuit or very lightly loaded HT supply - you just have to know what to expect of each.

For instance, on their 100V DC range a DVM across the Avo terminals reads 50K on the model 7 and both the 8 and 9 read 2Meg, so a forty-fold difference.

Pending the refurb of your Avo 7, perhaps you could measure those voltages whilst clipping a parallel resistor across the DVM or Avo 9 terminals?

Cheers
Chris

[WaveyDipole]

Originally Posted by ms660

The unloaded anode supply voltages and screen grid voltages @ +- 5% are shown in the attachments below.

Lawrence.

Lawrence, thanks. Those tables are very helpful.

Originally Posted by Chrispy57

Hi Wavey - I'm not really questioning the accuracy of your Avo 9, which you say agrees with a DVM, but highlighting that an old Avo 7 would load the HT and give a lower reading - which may well be in spec, given that the measurement was intended to be taken with just such a meter back in the day.

Chris, I didn't take it that way and I appreciate the clarification. I was aware that an AVO can give a different reading to a DMM, but what I didn't appreciate was the difference in inherent resistance between an AVO 7 and an AVO8/9 so I do plan to re-test with the AVO 7 at some point.

[WaveyDipole]

I have another question I would like to ask about this meter. Before taking a reading, the meter is backed off to zero using the SET ZERO control. After taking the reading, should it fall back to zero?

I am finding that regardless of whether I am setting to the ma/V position and reading the gm, or setting to the rated ma/V and reading good/bad, the meter pointer does not drop back to its zero setting. It usually settles either well above or sometimes well below zero but never at zero. To take another reading, it has to be zeroed again. The same seems to be true for the slotted meter pointer adjustment when the meter is not enegized. I keep having to set it because it keeps shifting and not just slightly. Its hard to be exact, but the shift can be by as much as 10deg right or left. I have sometimes had to adjust it from one extreme of its travel to the other then back again because it will not stay put. Usually these things need adjusting only a fraction occasionally.

As a result, readings are erratic and tend to vary quite a bit.

[WaveyDipole]

Now that the AVO 7 is mostly fixed, I was able to check the anode voltages across the ranges. There was a slight difference between the AVO 9 and the AVO 7 readings but the difference was minimal. In some cases the reading was the same.

The 100V setting read 154.4 on the DMM, 151V on the AVO9 and 150V on the AVO7 so the difference between the readings from the DMM and the two AVO meters is more significant than the difference between the two AVO meters. The differences are still fairly small and whoichever reading you take, the anode voltage on this setting does remain a bit high. The mains transformer input is set up for 240V and the tags on the primary appear to be wired in the correct order (e.g. no issues such as 240 & 220 reversed or similar), so nothing is amiss here.

In my previous post I was also concerned about what seemed to be a rather quirky behaviour of the meter movement and suspiciously low readings, so I set up a test. The movement is rated at 600μA, so the PSU was set to 6Vdc and a 10kΩ resistor placed in series with the supply to provide a current of 600μ. I checked the current on the DMM - 596μA - close enough. The wire running to the movement coil was de-soldered from its tag and the meter physically zeroed via the slotted adjuster. The movement was then attached to the current source. The meter pointer landed almost exactly on 7, so 30% short of a full deflection. However, what concerned me further was that when the current was disconnected, the pointer would not settle back to zero but would always overshoot by approximately two minor marks. I repeated the test and it did the same every time. A gentle rocking of the panel might cause the pointer to settle back towards zero, or else it needed to be zeroed using the adjuster. Either way, it needed adjustment each time.

The result would seem to explain not only the suspiciously low readings, but also confirm the somewhat erratic behaviour of the movement. I will need a little help here to deal with this. I am assuming that the low reading must be due to loss of magnetisation? I have heard that adding neodymium magnets can help? If so where do they need to be placed? And what type (round, oblong etc) to use?

With regards to the overshoot, I am not sure quite what is going on here. Might be pointer be sticking slightly below the zero mark, or might this also be a symptom of reduced magnetisation? Or is there some other explanation? Some adjustment appears to be possible and judging from the appearance of the slot in the screw at the centre of the movement, has apparently been attempted in the past, but these things are a very delicate balance so I don't want to mess it up and make it worse.

I need a little help from the experts on this one.

[Chrispy57]

Hi Wavey - well done for getting the old Avo 7 back into operation and enabling a bit more progress on your valve tester problem. Sounds like you can be confident in the comparability of voltage measurements over the three meters from those tests.
What I was less confident about, if I understood your current testing sequence description correctly, was whether you measured the currents, 600uA and "7", concurrently?
In that situation can you make the needle move higher than 7 by increasing the PSU voltage a little, and what current does the DVM show for full scale, then does the meter stick at 7 on the way back down? This should help differentiate between sticking and weak magnetism.

If you have a weak magnet issue there are several posts on here addressing that issue for a range of Avo instruments that may help. Weak magnetism wouldn't cause the needle to settle back below zero, that sounds more like a bit of sticking in that area too, particularly if rocking cures it.

Cheers
Chris

I'm far from an expert on this tester, but my own unit did have a couple of places where the meter tended to stick, and I managed to improve the situation by following advice on here about using thin paper strips to dislodge "whiskers" in the movement and just exercising the meter.

[WaveyDipole]

Chris, thanks for your comment. Everything was still set up on the bench from last night so it was simple enough to repeat the experiment.

In answer to your first question, I had first set up the PSU to exactly 6V, confirming the voltage with the Fluke DMM and then reading the current through the 10kΩ resistor. It read just short of 600μA. I then attached the movement in place of the Fluke and read the dial. So these readings were NOT concurrent. The Fluke read 596μA . Then the movement, when connected in place of the Fluke read 7.

I repeated the experiment just now, but so that I could answer your other questions, this the Fluke was left in series so that the current could be monitored.

Adjusting the supply carefully upwards, the pointer continues to move smoothly up to 10. At 10 on the scale, the Fluke reads 940μA.

Adjusting the supply slowly downwards, the pointer moves smoothly past 7 and down to 1. There does seem to be a slight judder at around the 2 mark but it goes past and then gets stuck at about +3 minor marks on the scale. A light tap releases it and it drops to zero.

Evidently there is some sticking at the lower end of the scale. I guess that when the power is disconnected while on 7, there is enough momentum to keep the pointer moving and overshooting the zero mark, but then it gets stuck below zero instead.

So does this also answer the question regarding weak magnetism? It seems that almost a full milliamp is required for a full deflection.

I was not aware that the AVO movement was also susceptible to whiskers, so thanks for mentioning that. I found a handful of discussions on the subject, including a mention of whiskers and swarf as a possible cause of sticking, as well as exercising the meter using a signal generator. I have included some links below for reference:

https://www.vintage-radio.net/forum/...d.php?t=165476
https://www.vintage-radio.net/forum/...d.php?t=101476
https://www.vintage-radio.net/forum/...d.php?t=189317

I expect that I will have to remove the movement for a closer inspection, although I am not looking forward to that.