© COPYRIGHT NOTICE All Rights Reserved

AVO Mk4 tube tester repair, quick check up.
(Last updated: 09-Mär-2016 23:52 )


Make good note, this is not the complete procedure, I just add something here from time to time, when I find out something interesting. I think I said it before on this website, but unfortunately many Mk4 are sold with problems inside, many of which you can not fix. It gets very difficult when the seller compensated one problem with another. A well known "fake" repair, for instance when the tester in general indicates too low values, they change the mains calibration wrong on purpose. So when you set the mains calibration mark on the "~" symbol in fact you have set the mains voltage 10% too high. Of course, now those too low results are better. But also bad tubes become "good" again. That's not what you want.

It's amazing what an intelligent electronics repair guy can achieve by just trimming and adjusting things, knowing he will sell it anyway, and usually the new buyer is a newbee, who has no reference tubes anyway. So when his tubes produces 15% error, he can't verify it. Besides most people are happy when all their tubes measure very strong, so they praise such a tester that gives too high values. A crazy situation. It can take years before they find out the whole tester is a mess. So be prepared for anything whatsoever, ranging from a messed up tester with fraudulent "repairs" inside

The following is some general check ups, and not complete.

Good advise: Remove the mains plug when working inside as the switch is only one phase.

A meter within calibration is essential for the internal tester calibration, and for the mains calibration. If the meter is off calibration, not only will you read too much or too little, but you begin with a wrong mains calibration, and also the user's calibration can not have been done right. So a small error with the meter will give a big mess. The meter of Mk3, Mk4, and CT160 has a few parameters. 1) Full scale 30uA +/- 1.5uA. 2) 3250 Ohms impedance. 3) Linearity (unspecified, but yet important). Most problems come from the meter, and you can not repair a meter yourself. However when the meter is basically ok, it can useually be brought back in good condition. Ask for a quote.

  1. Panel meter test. Set the tester for a 6SN7, and see if it can display anode current as a basic function. So when it works, it's ok. Switch off the tester. Remove the tube, and set the tester to "leakage test". So the test that you need to do before you begin normally. Connect a precise resistor of 1 Meg Ohms from Anode Link1 to Anode Link2, and leave in both links. Now switch on, and do the mains calibration. Now do the leakage test, so the one before you start a tube test, and this means you are going to see a resistance of 1 Meg Ohms on the panel meter when the electrode switch is at A1 or at A2. If you see nothing you are doing the whole test wrong. If you get a result, tap the meter gently to see if that give changes. A small change is ok, like 4mA on the 100mA scale. Make the needle move, by switch on and off the tester, and you should repeatedly find the same value. If it is another value every time, you have friction on the panel meter. This means a specialists repair. It's hard to say what is acceptable, but I would say 4mA in the 100mA scale is acceptable, 2mA is very good. Only new made tout band replacement meter would have zero error. Don't open the meter yourself, as such a 30uA meter at only 3250 Ohms, is the most delicate meter that was ever made in history of panel meters, ever since. If bad, you can stop here. If good, connect a pot meter instead, and try to verify the EACH resistance value on in the scale. Like set the pot mete until the meter indicated 0.5 Megs, take out the pot meter, and measure it with a good ohms meter. Take into account the specified error or the ohms meter. Ohms meters are often not very accurate. They show many digits, but that is another thing.
  2. Needle shape. When a needle is bend, you can see this when the upper scale and lower scale are not at zero nicely together.
  3. Needle Paint. If you see little chips of the needle black paint missing, the needle is probably off balance.
  4. Static problems. The panel meter may have static charge inside. It may come from rubbing the meter glass, or sometimes just builds up by the needle movement. In stubborn cases the charge can stay on for days, and when you use it, new charge builds up. An indication for this is a sticky meter, which seems to stick at another position today is it did yesterday. Also the amount of stick is very variable. Anti static treatment of the whole meter needs to be done, including the glass and the needle. This is not job for DIY.
  5. Sticky meter. Unfortunately this happens more often now, as these meters just become too old, and magnets are extremely strong. Magnetic dust creeps through every little hole, and finds it way to the coil. Usually it becomes sticky at the left or right side. Reason can also be rust of the magnets. They are not treated for anti rust, and powdery layer adds to them, which is also magnetic. This powder creeps and clusters, preferably under the coil. So you won't even see it. You can not remove this with compressed air. It requires complete disassembly of the meter. In some cases a chip can be picked out magnetically, but where there is one, there are always more, and such a fix won't last long. Better is clean the whole meter.
  6. Meter Glass. Check if the meter glass is nicely glued on. Don't press on it!! As you can break it off that way.
  7. Meter coil deformation. This a bad problem, it will cause a sticky meter now or later. Also it gives scale unlinearity. This is caused by overload of the meter when it slaps into the right corner by mistake. The needle bends over to the right, and if too much, this even deforms the coil too. There is no way to bend the needle back simply. You do need to bend this back, but needle balance is so delicate, and this required desoldering of the spiral springs, and take out the needle out of the meter. You can't do this yourself. Smallest unbalance will easily give 10% error. Linearity was already checked with pot meter test (above). Here is as another test to discover a deformed coil. Set the tester to a certain value with a pot meter added to the Anode links.(like above). Set the meter to appr 25% full scale. You can read this from the 100mA Scale conveniently. Now place the tester on it right side, and change in the reading should be small. It can be more as you expect, but the smaller the better. Repeat for the left side, and for the back. All of this should ideally give the same value, such meters do exist. Though very likely you will see a difference. Excessive difference like 30% indicate an off-balance needle. And an off balance needed may indicate a deformed coil. Though it can happen, only the needle is off balance, and the coil is still good. Please note, these 30uA systems have by definition very thin spiral springs, with very little torque. So any off-balance of the meter will have a much larger effect as with a 1mA meter, which have 33x higher torque.
  8. If is passed until here, several things are good now. Including the internal mains calibration screw, and the high value series resistor to the panel meter, used for this test. Even so, in this case, consider NOT TOUCHING the internal mains calibration screw, as this is a nasty and difficult procedure.
  9. Panel meter test, another. You need to verify if the meter impedance is within 1.5%. You can SKIP this test when your ohms meter is not 0.5% or even much better. In any case, add the specified meter error to the error you measure. Here is what to do: Unplug the tester from the mains and remove one wire from the meter. Now measure the impedance of the meter. For this, take extreme care. You can ONLY DO THIS when you are sure of the following: When you connect the ohms meter, the needle should move to the left. Connect the ohms meter first with one end only, and the other end via two fingers. The tiny movement you see now, verifies the direction of the needle. If you do that wrong, the needle will sweep violently to the right, and the coil gets deformed when it hits the needle stop. Next what you need to take care of, some multi meters test an unknown resistor blindly with 1mA to see what happens, and if too low impedance, it goes too 0,1mA. Don't ask me where multi meter designers have their brains hidden, but believe me, I have this with very expensive meters too. Such a meter you can NOT use. So you need to check your multi meter first, using a second multi meter You set the first meter to ohms, and then measure a normal resistor of 3300 ohms. Put the second multi meter in series with the resistor, and measure the test current. This is the only safe way. You can NOT measure this meter with 1mA as this is an overload of factor 33x. That may or may nor damage the coil, but it is definitely too much current. So never connect a random ohms meter to this panel meter. Also not when it cost 2000$, or when it is a Fluke or Agilent. Impedance should be 3250 +/- 50 Ohms for a Mk3, Mk4, and CT160. Thus value may not be off tolerance, more than 50 ohms, because the meter resistance is part of the attenuator circuits when going from 1mA to 2.5mA, 10mA etc.
  10. Panel meter test, another. If the tester works, set the needle for 5mA on the 10mA scale. Verify if you 5mA on the 25mA Scale , and 100mA scale. Try it with some more values and all use all scales from 1....100mA. That is a tough test. It proves the attenuation circuits are ok, an your meter linearity is good.
  11. Panel meter test, another. You could put an external meter in series with the panel meter. However in this case, the whole tester won't work well, as the added meter has a certain impedance. To do this test in a better way, you have to use the Anode Link. The external meter must be averaging type RMS meter, not a true RMS meter. As the original meter is also averaging, and not true RMS. As the meter is pulsed operated with 50% duty cycle, a reading of 100mA on the internal meter, will correspond with 50mA on the external meter. This a bit nasty and difficult to understand. If you have huge errors, you probably do something wrong. Take care, on the anode link is high voltage.
  12. Panel meter test, another. With the power off, check if you can nicely set the panel meter to zero, and if the needle is not sticky at that position. When you tap the meter gently the needle with dance a little bit, but must come back to where it was. Some small changes are normal though. Test if the set screw of the meter works well. If not, probably somebody messed inside the meter.
  13. I prefer to take the panel meter completely out, and check if a non specialist opened the meter, without a special tool for the screws. You can easily see that.
  14. Roller switch. They run seldom smooth, that is kind of normal. On top of the switch is a bar which likes to crack. So inspect it carefully for repairs. Also check the screws of the roller switch. These screws are soft brass, so damaged screw slots indicate a repair.
  15. Mains calibration. Ideally you use a variac for this. At the internal fuses, behind the left panel, is written from what to what voltage they are intended. Check with the variac, if you can do the mains calibration as well at the lowest voltage as at the highest voltage. This should be possible. If not, you may have trouble with the mains internal calibration, and/or the panel meter magnet being weak.
  16. If you tester passed all tests so far, many basic things are probably ok.
  17. There is a protection capacitor for the meter, which is soldered inside the top right switch of the Mk4. The switch with "test" and "gas" written on it. It is a small size 10uF cap, and it is always dried up. As a standard procedure simply cut it out. The replacement cap is easier to solder it directly on the meter lugs, as that is the same connection anyway. Like this you can now use a 10uF small size foil cap. That will not dry out, and it has not leakage which gives a meter error.
  18. The original anti parallel protection diodes, don't be tempted to replace them, as you will have difficulties finding ones that give better protection at so low voltage
    only. That's why they're so big. Big size diodes have much lower forward voltage at very low current.
  19. Check if someone touched the factory calibration. (The set screws of the two dial plates). They only have to be adjusted if the potmeters behind the dial plates are replaced. Otherwise if you feel the "need" to change this setting, simply don't do it. The problem definitely has is another reason. Look for old dirt and dust still being there. If untouched, leave this as is. If you can see the potmeters were not replaced, but the set screws are scratched and not dusty any more, this indicates some trial and error by somebody, and you are only lucky when he set it back to where it was.
  20. As a standard procedure, take out, open, and clean the two back up potmeters as this will greatly increase the user's comfort. Replace them if they are extremely"touchy" still. They are uncritical wire wound potmeters.
  21. Take a small mirror, and look at the inside of the small panel where the calibration pot meter is on. If the original diodes are replaced, that is not necessarily bad, but it is does mean somebody has repaired the tester.
  22. Mk4: Check the optical condition od the two calibration pot meters, and the "link". If this looks like adjusted 100 times, that's not a very good sign. When the pot meters however were stuck, as you wanted to rotate them first time, that seems to indicate is much better. Then likely it was not sold for being impossible to calibrate.
  23. Check the tube sockets for oxide inside the pin holes, or rusty screws of the front panel. If so, the tester was likely stored humid, and that's a higher change in a rusty panel meter magnet at the inside.
  24. The sockets are connected to the deck with rivets. If some sockets like the Noval or Octal have screws, these are not original. This indicates intensive use.
  25. A clean valves book is worth minimum 60 Euro, specially recent editions. Is the original manual with it? How about the anode and grid lead?


© COPYRIGHT NOTICE All Rights Reserved