AVO CT160. Protecting the meter.

[glowinganode]

Martin, thanks for all the info.
The effect of diodes whilst testing other valves is much smaller, and insignificant in the set~ / heater cont / leakage positions as such large current pulses are not present.
You say that the 6080 is not the best tube to carry out these tests, but it is the effect of such tests (high current, low anode voltage) that I wanted to investigate.
It's interesting that the manual mentions problems testing this valve.
Rob.

[Dekatron]

If you have read my post in this forum about the inner workings of the AVO Mk IV, which I and my friend Euan MacKenzie have been working on for a long time, there is a short passage about the calibration being different for some tubes.

It seems that some tubes need a larger value for the calibration value which AVO states should be 0.52 or 0.525, the true mathematical value should be 0.5554 to give the correct readings. If I correctly understand what Euan MacKenzie has found so far it seems that this higher factor is necessary for valves like the one you are testing, high current valves with a low Vg. We will write more on this later when we have doen some more testing.

Until then I think that it is best to follow the advice from AVO that these tubes might give you the wrong reading. The CT160 should be better as measuring these tubes than the Mk IV and earlier testers as it is somewhat better built for that purpose but it might still struggle with them.

Have you tried to see if the valve is oscillating when you measure it?

[glowinganode]

Martin, I did read your other post briefly and shall read it again when I have time to absorb all the information it contains.
Just to clarify, my tester is working as it should, gives accurate readings and is not prone to oscillations.
I just wanted to bring to the attention of other members the fact that high peak voltages occur across the movement, and that clamping with diodes will affect the calibration.
All of my findings are from practical experiments to which I have tried to apply some theory. I'm sure I've only scratched the surface, but I feel my findings are valid and relevant.
Rob.

[Dekatron]

Rob,thanks for explaining that your tester is working as it should, that clears things for that part.

My friend Euan did some tests in his home yesterday on his own CT160 and I am still waiting for an answers from him if I can publish those tests here. Euan has explained to me that he has been using diodes across the meter movement in his CT160 since he first became aware of the necessity to do that to protect the meter (my guess is that it was back in the 90s as he is referring to an article by Don Beswick from 1993) and that he is getting the correct readings compared to DC measurements with the same valves. I asked him to sum his experiences up in a short text that I will publish here as soon as possible. Euan and I never really discussed his CT160 since I did not have one to make comparisions too, we were only discussing our Mk IVs and VCM163s.

I will also try to get my hands on a C/VT160 as soon as possible so I can reverse engineer it and also make comparisons between an original meter movement and an OP-AMP driven meter as I am guessing there is something quite different between those two solutions. I am guessing that the OP-AMP solution is suffering more from diodes across the input terminals than an original meter movement will do - but that is just my guess. There can of course be other stuff that affects the readings. It would be very good to look at an oscilloscope trace with and without the OP-AMP circuit in place, maybe just by unplugging the OP-AMP from the circuit when the test is done, that will surely show if the signal is affected by the OP-AMP (with the diodes across its inputs).

In general the C/VT160 is an AVO Mk III with one of the backing of potentiometers replaced with a switch with resistors and the other backing off potentiometer left in the circuit. I wish I did not sell mine so fast when I got hold of my VCM163 but I did not have the economy to keep them both.

I'll write more here in a day or two when I hear from Euan again!

[Dekatron]

Euan MacKenzie gave me permission to write about his tests with his AVO CT160, which has been fitted with silicon diodes across the meter movement giving the correct readings since several years with this modification done to the circuit.

---------- From Euan MacKenzie 11th of July 2010, email sent to Martin Forsberg ----------

I have no experience of using op-amp meter amplifiers, so I can't pontificate on them; however for the record, this afternoon I opened up my CT160, disconnected the microammmeter from rest of the CT160 circuit; instead connected directly to it and measured its FSD and resistance.

The results are as follows:-

Real current for FSD indication on CT160, ie '100µA' on the 'gas' scale was 32.0µA; as measured with two DMMs in series. Voltage across CT160 at FSD 101.2mV, across 3.000k ( two 0.1% RC55Y resistors) in series 96.3mV; hence CT160 meter resistance is 3,153 Ohms at 13°C.

Placing two back to back 1N4004 diodes across the CT160 meter at FSD made absolutely no detectable difference. However placing two back to back Germanium diodes (OA91s) did drop the indicated current to '78µA' on the 'gas' scale; just as Don Beswick said it would.

----------

Euan is referring to a written paper by Don Beswick which I have included one section from as text below and also as a complete PDF-file.

---------- Section from "Mistakes in AVO VDM" by Don Beswick February 1993 ----------

One worthwhile modification AVO introduced on the Mk IV is to fit a silicon diode, OA210 or equivalent, across the meter terminals, in the forward ie conducting direction, which limits the voltage drop across the meter to about half a volt. It might be thought that germanium diodes, such as the OA85, would give better protection because of their lower voltage drop; however it is too low and introduces non linearity in the meter reading at the top end of the scale, eg a meter reading of 90mA drops to about 80mA, which is unacceptable.

The AVO valve characteristic meter is an item of professional equipment; it has been described as the Rolls-Royce of valve testers and like a Rolls-Royce it needs to be driven with due care and attention.

Don Beswick, February 1993

----------

Euan also continues to write about the similarities between the AVO Mk III and the CT160:

---------- From Euan MacKenzie 11th of July 2010, email sent to Martin Forsberg ----------

I don't buy the "it won't work in a bridge" argument either; if you pause to think about it, the CT160 is a Mk III with the coarse backing off pot swapped for a switch with resistors corresponding to 10mA steps, and the fine anode current control (0 - 10mA) is the same as the fine backing off pot; the only minor difference being it has a 10mA range instead of about 7.5mA. All this for reasons best known to the military committee responsible for its (peculiar) specification...

----------

I have no reason to doubt Euan on this as he knows his way around his CT160, and then that rules out the balanced bridge part as AVO and others have described the CT160 as having on some cases. I just bought the bridge explanation as the truth myself since I have read it in so many places, so there I was evidently wrong! I will check some more to see what AVO says about the construction, maybe it is one more of their famous typos to describe it as a "bridge" construction. I just got hold of a handbook for my AVO B150 Mk3 measuring bridge and that is a full of errors as anything else that AVO have published, so it might even be so with the V/CT160.

Euan also told me about the document "A P 2537 2 1 65 Guidance" addition to the CT160 manual by the Military, which I have included one section from regarding low Ra valves, and the rest as a PDF-file.

---------- A P 2537 2 1 65 Guidance, slightly edited to fit text format ----------

7. In practice, the transformer, cut-out coil and metering resistances add up to some few hundred ohms resistance (375 ? on Va = 400V). This order of source resistance is negligible for most valves but a few types are coming into service that have very low a.c. impedances (Ra). Examples are:-

CV2984 = CV5008 = civilian Type 6080
CV2975 = civilian Type EL84
CV4079 = civilian Type A2293
CV5077 = civilian Type PL81

These low Ra valves are not tested to static test conditions but under dynamic conditions because the source impedance of the power supply is no longer negligible. As a result the mutual conductance figures may be low by a factor of one half to one quarter of the nominal value. This effect is a design limitation of the CT160 that cannot be eliminated by modification of the instrument. It applies to any valve that has an a.c. impedance lower than about 400 ohms.

----------

I am myself in the process of making new reverse engineered schematics for the AVO Mk III and the AVO Mk IV. I will try to make them more or less identical in the way they look so it is much more easy to spot the differences and the likenesses in most places. I will post them here on this forum when I am finished.

If you have any questions regarding the tests that Eaun MacKenzie did you can send me a PM or an email and I will forward them to him.

[Karsten]

When attempting to protect the original 30 uA/3250 Ohm meter in the CT160 with two antiparallel clamping diodes, their clamping voltage must be higher than the voltage peaks across the movement coil, otherwise incorrect readings will occur. Positive and negative voltage peaks increase with plate current and become maximal at full meter deflection and the rated 100 mA current maximum.

The 1 mA/V mark on the meter scale is at 75 % of full deflection, so if the peaks at full deflection stay below the clamping voltage this will result in some additional headroom at the 75 % / 1 mA/V point.

The movement measures the arithmetic mean of the current passing through the meter coil. This arithmetic mean is NOT influenced by the capacitance of a parallel damping capacitor. The capacitor however very effectively lowers the voltage peaks across and thus the current peaks through the meter coil. Therefore an interdependence exists between clamping voltage and minimum capacitance.
I have measured the positive and negative voltage peaks at full meter deflection with an oscilloscope for various plate currents and capacitances of 0, 10, 20, 47 and 100 uF and plotted the values in the attached Excel chart.

Key results are (for 100 mA plate current):
** 600 mV Silicon clamping diodes require >10 uF. 20 uF will suffice.
** 200 mV Schottky clamping diodes require >47 uF. 100 uF will suffice.

I use 1N5711 Schottky diodes and a 100 uF 6.3 V bipolar SMD ceramic capacitor. Even with this rather large damping capacitance the needle moves swiftly but with greatly reduced acceleration. For example the needle does not buzz anymore when driven against the left or right stops.

I hope this will contribute positively to the discussion about meter protection measures.

[Dekatron]

Thank you for a new look at this problem and for such a thorough check!

[Karsten]

In the attached .pdf I have recorded some oscillograms showing the principal waveform of the voltage which drives the meter. The pictures also demonstrate the influence of a damping capacitor. An example measurement shows how standard silicon clamping diodes together with 10 uF capacitance cause faulty meter readings at high anode currents.
The CT 160 used has the solid state rectifier modifications as proposed by Martin Forsberg. All measurements were taken with a 6L6GC valve load.

The oscilloscope is a Fluke 97 Scope Meter without a computer link. Therefore I had to photograph its screen. Even though I did my best with the equipment available pictures are of minor quality. I apologize for that.

[Karsten]

In my last post I forgot to mention that all oscillograms have been recorded with the gm scale in the "set zero" position. I have added this information plus the values of the various peak voltages to the .pdf.

[Karsten]

Here are the details of the tested and recommended meter protection circuit consisting of two Schottky diodes and a tiny biplolar SMD capacitor with rather large capacitance 100 uF.

Features are:
** No rattle of meter needle against end stops
** Low mechanical forces on needle due to high damping
** Smooth but swift needle motion for long meter life
** Lowest possible clamping level for best mechanical protection
** Thoroughly tested. No influence on meter display even under extreme conditions: max. Anode current (100 mA) plus full meter deflection
** Moderate cost: 30 Eurocents per diode plus 1 Euro for SMD capacitor

Be aware however, that this circuit has only been tested with a "siliconized" CT 160 with additional Anode voltage rectifier. Before you build the protection circuit into a different AVO valve tester make sure that the circuit does not upset the meter: fit only the 100 uF cap first. Then use a suitable valve and adjust Vg for max. tester Anode current and full meter deflection. Connect clamping diodes and observe meter. Meter should not move. If it does, try additional 47 uF (just as tiny as the 100 uF, so no space problem should occur).

Good luck!

Karsten

[Karsten]

Addendum to my post above:

When testing whether the meter is influenced by the clamping diodes, meter circuit must be in its most sensitive position, e.g. "set zero" on the CT160.

[Radio Wrangler]

With the opamp meter approach, it is possible to use a string of a few Si diodes in series to protect the opamp circuitry, and have the clamping effect at a significantly higher voltage than would introduce any detectable non-linearity within the metering range. The opamp gain and the series resistor for the meter could be arranged so that given the power supply voltage for the opamp, the meter current is limited before the meter is put at risk. A little filtering can also limit the pointer speed as well.

A 21st century valve tester design could be rather interesting. An SMPS to do 1.4 to 100V heaters with a hundred watts or a few amps as limits would not be too difficult. Similarly supplies for anode and screen 50-500V. Realistic heater-cathode leakage tests with appropriate voltages would be easy.

Lack of high current switches may force a patch-lead or bank of relays approach to pin/electrode combination switching.

There's no need to have a huge table of sockets all at once. A simple FR4 square piece with a socket and up to a dozen flying patch leads could clip on and the leads be plugged into appropriate ones of a group of 4mm sockets. Add a couple of sizes of top cap clips and voila!

Do swept DC tests for a curve tracer function
Do pulsed DC tests to keep heat down with higher power devices
Plot Gm versus anode current.

Use a 1/4 VGA LCD and make a home for an Arduino or raspberry.

Have a compendium of valve data and tests on a flash stick.

Lots of evil new technology, but in a good cause.

It wouldn't sell enough to interest any manufacturer, but it could be fun as a hobby project.


David

[Dekatron]

David,

One very nice 21st century tester designed by Helmut Weigl, is found here: http://www.roehrentest.de/ (you can translate the german language with Google Translate) the software is now available in English too, it currently has a database of 14564 valves!

Another very cheap and simple one in development by Ronald Dekker, is found here: http://www.dos4ever.com/tubetester/tubetester2.html There have been two previous versions which you could look at since they contain a lot of the development information and steps to reach the thrid version. The current version, V3, produces very nice measurements but there is no information on when it will be finished. It is very interresting to follow the development process though!

[Karsten]

Attached is a summary of my contributions so far with some additional comments which also may shed some more light on what the anode current backing off circuit really effects. They are not aimed at transforming the vintage CT 160 to a 21st century tester but rather at the desires of those forum members who strive to preserve the irreplaceable meter in their unit for future generations.

Regards, Karsten

[Karsten]

There has been concern from a forum member about the suitability of the 1N5711 clamping diode as it is only rated for 15 mA DC and the peak currents in the two extreme cases where the gm scale is in "set zero" and
A. Anode current = max. = 100 mA and backing off controls at zero or
B. No valve and backing off controls at max. (90 mA + 10 mA)
are considerably higher than 15 mA.

I have measured and roughly calculated the current load limit integrals for both cases by graphically evaluating the current oscillograms.
Some data sheets state that the 1N5711 can withstand a 2 A 10 us current pulse which gives a current load limit integral of I^2*t = 40 A^2*us.
Both measured load limit integrals are well below this limit.
This is certailnly only a very rough estimation as a load integral specified for 10 us cannot simply be converted to a meaningful value in the 10 ms time range. However, the measured load integrals are so much smaller than the stated data sheet value that they justify the suitability of this diode. Also, the diodes have withstood both test situations.

I have added two oscillograms for the two extreme cases to my little study.

[mark-jacobs]

I would like to thank Karsten for all his work and information on protecting the AVO CT160 meter. I had previously started out with a 10uF PP capacitor + diodes, which was an improvement. I then tried a 50uF NP electrolytic, but still had issues with needle overshoot and there seemed to be a slow decay to the final reading. Might have been due to the dielectric properties of the electrolytic. I ended up using AVX Multilayer Ceramic Capacitors (MLCC) ST203C107MAJ05 - Leaded 100uF 25VDC along with the 1N5711 Diodes and the CT160 meter is well dampened, yet responsive with minimal overshoot at both ends of the travel. I have several meters for the CT160 and they seem to vary on their dampening characteristics. The first meter that I installed this in was the least dampened, and the needle would rattle when testing power tubes before using the 100uF capacitor. I have not seen any issues with the combination of a 100uF MLCC capacitor and 1N5711 diodes. Of note, the 1N5711 diodes measured a static forward voltage of 0.356V on my Fluke DVMs, higher then most of most of my other schottky diodes with the same DVMs. So I would caution using any schottky diode that you may have in the tool box.

[Karsten]

The 1N5711 has a rather "soft" Uf(If) forward characteristic starting with approx. 200 Ohm dynamic resistance which grows further with increasing forward current. Therefore the forward voltage measured for instance by a multimeter depends considerably on the current in the measuring circuit. This current varies with the type of instrument.
The easiest method to check whether the clamping diode influences the CT160 meter reading is:

• use a valve suitable for 100 mA anode current
• adjust grid voltage and backing off controls to 100 mA
• observe meter needle with and without clamping diodes. If needle moves: increase damping capacitor (100 uF should be enough) or use diodes with higher forward threshold.

I will post more on meter protection soon.

[Karsten]

Martin Forsberg ("dekatron") has pointed me to the risk of excessively high anode current in the case of a sudden inter electrode short where the overload relay will be too slow for reliable protection. Therefore the meter protection circuit should be able to handle such an event.
However, the voltage across R10 caused by the high peak anode current in the case of an anode-cathode short will drive a current through the small signal Schottky diodes which could possibly fuse them or the voltage might even damage the meter. I have therefore added a pair of standard Silicon rectifier diodes to cope with such a rare event. I have also measured the overall efficiency of the protection circuit in terms of remaining meter overload. See the attached .pdf for details.

[woodchips]

Hi, wasn't quite certain into which thread to put this, please move as required.

I have come across some sensitive meters used in safety ohmmeters. Made by Ernest Turner, they only seem to measure 40 ohms but have a marked sensitivity of 8uA. I have had a quick play with one and get 25uA from the centre mark to the edge of the mirror.

They are taut band type, and I can't see any parallel resistors across the coil. Wire diameter doesn't look gossamer either.

Any use as a replacement? These ohmmeters are 2 a penny a the moment, no one wants them. Does anyone want to experiment?

Bob

[Dekatron]

Bob,

There are a few reason which I think will make these meters hard to use in an AVO.

Firstly; it looks like the needle is very short on these meters, can you give an overall size of the meters. Lengthening a needle requires rebalancing of the movement which is not easy.

Secondly; the fact that the resistance is very low would mean that a different type of temperature compensating swamp resistor arrangement would have to be used than the original one in the AVO meters.

Thirdly; changing a centre movement to a full swing movement is not so simple as the meter has been balanced for this kind of centre use, it would probably have to be rebalanced if it would be changed.

They might be worth buying anyway, you never know when a good meter is needed!