A word on Tube Testers

[EL34]

I have added a note to my web site Tubes page regarding using home tube testers to check the New Sensor pre tested tubes I sell.

I have run into a couple customers that are taking the tubes that have been tested nine ways to Sunday by New Sensor and coming up with some goofy results based on their home machines.

I don't accept tube returns from guys testing New Sensor tubes on home tube testers, no matter how great you think your machine is.
Use your tube tester to weed out flea market tubes and find flat out dead garage sale tubes.

Your ears are the only home tube tester your can trust.

Here's what Allan Otto at New sensor has to say about the testing procedures of the tubes I buy from them and home tube testers.
Otto is a highly respected tech guy. I have had quite a few tech dealings with him to know that this guys knows what he is talking about.

Emission testers such as Sencore Mighty Mite and B&K Dynajet types do not give an accurate indication of the actual operation of a vacuum tube. They only check cathode emission. These testers also check for interelectrode leakage (in which the tester criteria can far exceed typical application and will test many good tubes as defective). In the emission test a voltage is place between the cathode and grid through a load resistance and meter with the tube being tested as a diode. Many times this is an AC voltage and can be as low as 40 volts (depending on the tester design). This tells you nothing about the transconductance (measured as mutual conductance on Hickok testers), which is the tube's ability to amplify. We use two testers to test small signal tubes.
 
 The main tester is the George Kaye Small Signal Tube Checker. This tester places the tube in an actual small signal amplifier circuit to test for gain, voltage output, and distortion. Both sides of a twin triode can be compared for balance. This is especially useful when selecting tubes for phase inverter service. This tester also has a small speaker built in to check for noise and microphonics. If there is anything wrong with a small signal tube, this tester will quickly and thoroughly analyze the tube's condition. Vacuum Tube Valley makes a tester that tests similar parameters. I once built a mock-up of a typical Fender preamp stage that could select between the two triodes so a studio could test and evaluate their 12AX7/12AT7 family tubes. They actual ran a CD player through it and sent the signal to a monitor amplifier to actual listen to the differences between different tubes. They could even hear the difference in the sound of different brands of 12AX7 tubes. This would be pretty much the same as when you put the tubes in the mixer to check them. When I repair vintage amplifiers, I will try several different types of 12AX7 tubes for the best sound (some amplifiers like certain 12AX7's over others).
 
 The other tester we use is a New London Instruments Model 901A Transconductance Analyzer. This tester is set up by using a tube manual for particular tube being tested. All parameters (plate voltage, grid bias, etc.) can be adjusted and monitored. You can actually plot tube curves on this device.
 
 We also have a Military TV-3B/U and some TV-7D/U's that we use for quick go/no-go tests. These are basically military versions of Hickok mutual conductance testers. (I repaired and calibrated many of these when I was in the Navy). If the tube tests defective on these tester, they are further evaluated on the George Kaye and New London testers
 
 Another interesting fact regarding small signal tubes (especially 12AX7 types) is that the tubes have been modified slightly to satsify the needs of the musical instrument amplifier and high end audio industries (the biggest users). Some of these changes are to reduce noise and microphonics in high gain amplifiers or improve linearity of the tube. Some of these changes have shiffted the tube curves slightly or altered the geometry of the tube, causing them to test weak or defective on simple emission testers. I would say that placing the tubes in the mixer was the most thorough and valid test, because you could actually hear what was going on (an emission tester will not detect noise or a roll-off of high or low frequencies). As stated in the RCA Receiving Tube Manual (RC-30) in the section on tube testing, the best test of a tube is to place in the actual circuit in which it will be used.

[FYL]

While I fully agree on amateurs "testing" tubes using crappy uncalibrated gear, Alan Otto makes too broad generalizations. For instance :

Emission testers such as [...] B&K Dynajet types

B&K produced emission as well as transconductance testers in their Dynajet line. I routinely use a 650 which happens to be one of the finest and fastest transconductance testers ever manufactured.

[EL34]

You would have to take that up with him.

This only has to do with me accepting tube returns.
I don't accept any home tube tester results as a valid reason to returns tubes

[PRR]

The issue is "Emission testers". These were NEVER good tests. Especially for small-signal tubes like 12AX7, and the BIG signal tubes like EL34 in guitar-amp duty.

The ONLY excuse for Emission Test is that it is cheap and it will quickly find high-high-hour tubes which are actually worn. When you pull 23 tubes out of a TV set that has run 9 hours a night 7 nights a week, there is a good chance of finding a "dead" or "weak" tube, which -might- actually be a problem.

Hmmmmm..... ya know, those cheap testers are -biased- to read "poor". The main use was to test used tubes and SELL new tubes. Whether an actual radio-repair shop (who does not mind replacing tubes) or the Drugstore Tube Tester with a rack of new FOR SALE tubes underneath.

I have an emission tester. Might be a Mighty Mite.  It found a hot-short in an old tube, saved an amp's butt. I tap the emission test button and see if the meter moves -anywhere- near the "OK" zone; complete lack of reading means I mis-set the switches, the tube-type, or the tube is totally dead. But I don't give a hoot if the reading is only halfway up to "OK"; I know that most audio systems will never require "full emission".

And prolonged "emission testing" WILL damage the tube.

You could plausably decline to accept returns which have been through an Emission Tester. The tube may have been fine before they "tested" it for an hour.

It can be hard to know what kind of tester the customer used. You might simply refuse returns based -only- on a tube-tester reading.

I still say: breadboard a Champ so you can probe cathode voltages. If most 12AX7 show 1.5V-1.7V, and you got one which reads 0.9V, it is dubious. AND if it reads 1.6V but hisses like a snake-worship church, sumthin's wrong. You can put 12AT7 and 12AU7 in the same socket; the "normal" voltage will be different and gain will be down, but way-out tubes will be obvious when compared with others. All the Octal Audio Power tubes will "work" in a Champ's 6V6 socket if you wire it right. True, a 6550 may run fine at happy-6V6 11W and turn sick at 41W, but that's rare.

OTOH; New Sensor "should" understand that most of the market is ignorant idiots. They "should" gimmick the tubes so that they read "OK" on the $19 Sencore Cheapo tester.

BTW, IIRC: the pair of Sovtek 6550s I got from you had handwritten numbers which appeared to be current and transconductance. That told me that someone had lit them up with specific G1 G2 P voltages, read current, and wiggled to read transconductance. To me, this is first-order PROOF the tube was "good" when tested. I don't know what the "right" numbers should be (I don't know what the the test conditions were), but I am sure that if they tested a run and got 61 63 58 13 60 210 66... that the "13" and "210" tubes would be culled-out.

[EL34]

I have only had two guys want to return tubes this year
It hardly ever happens.

Isn't it strange that both guys had listened to the tubes with a tube tester instead of with their ears?

I am sure that their ears would have found something wrong also after pulling the bad tube from the tube tester and listening for the defect.

[supro66]

I have always got very good tubes from you


I test them on my newest state of the art Tube tester




KISS

Keep It Simple Stupid

http://www.tone-lizard.com/Tube_Testers.html

[FYL]

This only has to do with me accepting tube returns.
I don't accept any home tube tester results as a valid reason to returns tubes

And you're 100% right.

[EL34]

This is how I test tubes  ;D

[FYL]

Others use different testers:



From Black Swamp Audio

 8)

[RicharD]

I once matched up a small bucket of output tubes on my Hickok.  When I put them in a real amp running at a much higher B+, they didn't match anymore.  Time spent=lesson learned.  I almost never use my Hickok anymore.  I only use it to check my Doug tubes.   :P

I haven't bought a whole lot form Doug, but I've never gotten a defective part from him either.

[mackie2]

TUBE TESTERS ALL:

I BUILT THE MOTHER OF ALL TESTERS (BRUTUS I)  FOR POWER  TUBES FOR AN AMP MFG.  IT WAS ACTUALLY USED TO AGE THE TUBES, OCTAL AND 9 PIN, FOR WHAT EVER BURN IN PERIOD SELECTED, IE: 36HR.  ALL 12 TUBES ARE ON THE SAME O/T AT OPERATING B+, IE:400V.  PRECISION HIGH QUALITY PARTS USED THROUGHOUT.  VAR BIAS, FIL, B+, TUBE CURRENT LOGGED AND MONITORED THROUGH THE BURN PERIOD.  YOU WOULD BE SURPRISED HOW MUCH THE CURRENT MFG TUBE VARY AFTER BURN.
THIS TYPE OF BURN IN USED TO BE USED FOR TEST EQUIPMENT TUBES BACK IN THE  50'S.  HOWEVER, THE  SOUND OF THE TUBES
MFG TODAY STILL HAVE TO PLEASE THE EAR IN AN AMP FOR THE FINAL TEST

MACKIE2
 

[bnwitt]

I once matched up a small bucket of output tubes on my Hickok.  When I put them in a real amp running at a much higher B+, they didn't match anymore.  Time spent=lesson learned.  I almost never use my Hickok anymore. 

Even matching tubes on my Maxi matcher which has 400 volts plate voltage is no guarantee they'll match when put in an amp that has 465V or 500V.  In the amp is the best way to check for matching.

[Greasehorse]

My KD-1575 was gone through by Kennedy within a couple of years. I have tested maybe 50 tubes on it since. But, so far, if it says it's bad it is bad...

[craneclimber]

This is a little off topic.......just dusted off my tube testers (2) after reading thru this forum, to confirm what I remember - that they do test for mutual conductance. They've been sitting for a while and I'm sure I'll get around to cleaning them up soon.
Anyone familiar with Seco models 107B and 107C?
Just looking for opinions as to their reputation for accuracy.
Thanks

[FYL]

Anyone familiar with Seco models 107B and 107C?

No direct experience. They seem to be nice B&K knock-offs working on very low voltages.

http://www.tone-lizard.com/Seco.html

[craneclimber]

FYL - Thanks for the link, I've been to the tone-lizard site before but had not thought of checking it for tube testers. Good reading!

[samato]

I'm just getting started with tube amp building/repair and I don't know much about testing tubes or any other parts for that matter.

Is there a simple way, without any special tools other than maybe a DMM, to make sure tubes are safe to put into an amp for further testing? 

It makes sense to me that the best way to test would be with the tubes in the circuit they will live in where you can use your ears and a DMM to check real numbers but I just want to make sure I don't damage things by putting in a bad tube.

[simonallaway]

Is there a simple way, without any special tools other than maybe a DMM, to make sure tubes are safe to put into an amp for further testing? 

It makes sense to me that the best way to test would be with the tubes in the circuit they will live in where you can use your ears and a DMM to check real numbers but I just want to make sure I don't damage things by putting in a bad tube.

Exactly. The method that has been proposed many times here, and seems to make much more sense than a testing device is to place them in circuit such as a champ. Then it's as 'real' as it gets.

[samato]

Okay, let me re-phrase my questions.

Is it possible for a tube to be bad in such a way that running it in a working amplifier will cause damage to the amp? 

If so, is there a way to test for this prior to putting the tube in an amp without special tools? 

If special tools are required, what are they and how does one use them?

[FYL]

Is there a simple way, without any special tools other than maybe a DMM, to make sure tubes are safe to put into an amp for further testing? 

You may check for cold shorts between elements using a basic DMM. Build a small rig fitted with sockets and accessible testing points - directly using probes on subminis and novals is pretty silly. You should get at least 2M for small signal toobs and >500K for large bottles.

[samato]

Is there a simple way, without any special tools other than maybe a DMM, to make sure tubes are safe to put into an amp for further testing? 

You may check for cold shorts between elements using a basic DMM. Build a small rig fitted with sockets and accessible testing points - directly using probes on subminis and novals is pretty silly. You should get at least 2M for small signal toobs and >500K for large bottles.

Thanks. 

Are shorts between tube elements the only things with a tube that are likely to damage an amp?  Also, I guess the next step beyond what you described above would be to build something that would allow me to test for shorts while the filament is warm - or it that overkill? 

[PRR]

Preamp tubes can't blow-up an amp, only fry their plate resistor.

Power tubes can do damage. And a hot-short is possible: that's how my VT-40 smoked. And that's maybe the ONE thing that $20 testers like Eico 667 ARE good for: heat the cathode and test for shorts. This test is quite safe (for tubes rated >100V, which is mostly what we use). And the first time the $20 tester avoids a $100 PT smoke-out it is worth it (even if you pay $60 collectors-price for that cheap 667).

[FYL]

Are shorts between tube elements the only things with a tube that are likely to damage an amp? 

Direct damage, yes, at least for power tubes. Other defects can have an indirect effect.

Also, I guess the next step beyond what you described above would be to build something that would allow me to test for shorts while the filament is warm - or it that overkill? 

Get a small tube tester, as PRR suggested. But don't expect to pay only $60 for a working 667...

[samato]

I've read about using a car battery charger to make a tube tester but I'm having a hard time finding that info again.  Can anyone point me in the right direction or explain it to me?

What I think I'd like to do is use a battery charger connected to a pair of octal tube sockets on some sort of chassis.  One socket would provide 6.3V to the filament pins for testing power tubes and the other socket would send 5V to the filament pins for testing rectifier tubes - or is this not necessary for rectifier tubes?  I would just use this to test for hot shorts.  Further testing and testing of 9 pin tubes would be done in the amp.

If this seems logical/doable, can anyone help me figure out how to make this?  I don't know exactly what type of charger to use or how to get it from 12V to 6.3V/5V.  I read you tap it before the rectifier but I wouldn't know how to do this.  I'm also wondering if a cheap float charger like this one would work:

http://www.harborfreight.com/automatic-battery-float-charger-42292.html

I have one and it works well for it's intended purpose.  I also have one like this, though not from Harbor Frieght, but I wouldn't really want to take it apart so it can't be used as a battery charger anymore:

http://www.harborfreight.com/10-2-55-amp-6-12-volt-battery-charger-engine-starter-66783.html?utm_term=66783&utm_medium=cse&utm_source=googlebase

[FYL]

I've read about using a car battery charger to make a tube tester but I'm having a hard time finding that info again.  Can anyone point me in the right direction or explain it to me?

AFAICT some vintage filament testers could also be used as battery chargers, but not the other way around.

[samato]

I've read about using a car battery charger to make a tube tester but I'm having a hard time finding that info again.  Can anyone point me in the right direction or explain it to me?

AFAICT some vintage filament testers could also be used as battery chargers, but not the other way around.

Why, too much current for the tube from a battery charger?

[stratele52]

Battery charger with a very small load like a tube can send too much voltage on fillement and burn them.

 You can use some " wall" transformer to do that

[HotBluePlates]

One note which goes back to the original point of this sticky:

If you listen to the guy who wrote the book on tube testers (Alan Douglas, Tube Testers and Classic Electronic Test Gear, Aug 1, 2000), he'll tell you that even some of the best commercial tube testers will only provide an indication of a tube's health.

Further, it is difficult to calibrate them, such that several samples of the same model of tester will read the same tube differently. That means the specific number assigned to the tube's Gm by one tester will likely not match a second tester, even if both are properly working and "calibrated".

So, outside of characterizing a tube using your own power supplies and test signals, you will likely never have absolute true numbers given by a tester. If needed, we can discuss some simple technical reasons why this is true (at least on the Hickok models, with which I'm familiar).

That's one of several reasons why the real test of a tube is in the circuit in which it will be operated.

[stratele52]

I agree with you HotBluePlate,

tubes tester don't work at high voltage and same spec as an Guitar amp; not usefull for guitar tubes amp.

[HotBluePlates]

The plate voltage is not the significant issue.

The first main issue is that 5 tube testers of the same model by the same manufacturer could give 5 different readings for the same tube, and yet all be working correctly. That is the reason many with deep knowledge of tube testers will say they only provide an indication of a tube's health. You still have to know how a good tube will read on your tester, regardless of what the roll chart and manual might tell you.

The second big issue is the nature of the signal applied to the grid to measure transconductance (or "dynamic mutual conductance"). The common Hickok circuit applies an a.c. signal which does not change when the tube is changed. So an output tube might have a 5vac signal applied to the grid as a test signal, but so does a 12AX7; 5vac is way too big an input signal for meaningful measurement for that tube. Note the exact applied voltage differs for different models.

The third issue is the one you raise, which is operating point/plate voltage. Tubes must have tolerance for variation of their characteristics, just to enable practical manufacturing and an affordable product. We don't know that this tolerance enables a meaningful comparison of performance of two different tubes at one operating condition, on the basis of a measurement at a very different operating condition.

It is possible to conduct precise and accurate measurement of a vacuum tube's characteristics; it is very difficult to perform that measurement on every tube type out there with the minimum setup and speed-of-use that most tube testers offer. Accuracy/applicability is traded away for speed in most tube testers.

[John]

I have been browsing the tube testers on ebay to get an idea of price. Besides the Eico 667, any other cheap, simple, easy-to-use testers?
 All I'm interested in is something that checks for shorts as PRR mentioned. I'd hate to pop a tube in, start powering up and blow an old PT/OT that up to then was good.

Even at 60/70 bucks for the 667, if it saves me one piece of iron it'd be worth it.

Thanks!

[FYL]

The 667 is excellent when it comes to measuring leaks and shorts. If you can find one for $60 or so, get it.

[John]

There's a big difference in price between the 667 and 666. Tone Lizard has them both listed as being usable for guitar tubes. Thoughts anyone?

Remember, I'm not looking to make Doug's life miserable trying to return tubes  I just don't want a shorted (old) tube frying something else.

[Willabe]

John, I'd stay away from the _ 666 _         


                  Brad     

[John]

Yeah, I saw it called the satan tester. Maybe that's the price difference?

[Willabe]

Yeah, I saw it called the satan tester. Maybe that's the price difference?

               


            Brad   

[FYL]

There's a big difference in price between the 667 and 666. Tone Lizard has them both listed as being usable for guitar tubes. Thoughts anyone?

Both are OK when it comes to test leakage and shorts.

[John]

FYL, thanks for your replies, I appreciate it!

I'm also pondering building a little board that will just heat the cathodes - should be able to just run those off a battery? -, allowing me to check for shorts with my DMM... PRR's post gave me that idea.

[Stefan_E]

Hi from Sweden, first post btw :)

I built the RAT tube tester: riodeel.com/tester.htm, works pretty good. Max 300 Vp/Vs though. If i had known better when i built it i would have modded to get higher voltages, easy fix but mind that it may need other transistors to accomodate for those voltages. Anyway, a good tube tester and fun build. Learned a lot about tubes in the making of it. One good thing is that you hook up your multimeter for readings, and i trust my Fluke's over analog built in meters.

It tests gm and mu under pretty "real" conditions and even comes with a 1kHz tone generator, could always hook up your own tone gen if you need more detailed testing. Make sure to cool the voltage regulators good (the IRF820's or similar) since they drift with temperature.

When it comes to matching power tubes there are a lot of different opinions on how it should be made. Dynamic, i.e working with a tone on the grid, or static, i.e with grid fixed. One is good for providing equal curves when working, the other good for low hum/equal idle currents. Which do you want? :) I would think the "static" testing is best for guitar tube amps. After all, some of these even are deliberately off regarding positive/negative output from the PI to provide more "3d" sound (Dumble anyone?).

Anyway, i agree - go with your ears and use the tester for ball park and basic tests, can't go wrong with that

[HotBluePlates]

When it comes to matching power tubes there are a lot of different opinions on how it should be made. Dynamic, i.e working with a tone on the grid, or static, i.e with grid fixed. One is good for providing equal curves when working, the other good for low hum/equal idle currents. Which do you want?

I would think some kind of dynamic test would be more informative, especially at various output levels. See this article, starting at paragraph 4.

[cesout]

I picked up a Tektronix 576 some years ago and recently built a test fixture for it to test tubes. It has two octal sockets (A/B) and a single 9pin socket for dual triodes. The fixture has connections for heater and screen supplies. The screen supply is needed when running the power tubes.

It uses the 576's Left/Right switch to put each device under test. Using the curve tracer, I can get tc curves like in the data sheets. It is useful for tube matching and general tube testing. Since it will operate the tube through it's entire operating range, far more rigorous than any basic tube tester. It will put up to 1.5KV on the plate so running tubes at rated voltage is not an issue.

With pre tubes, you would still need to test for microphonics of course. But that is simple - plug it in to the amp.

Update:

Here are some pics of it in action. These triodes are as close as you can find them, virtually identical. I've purchased 'matched' triodes that are off by .5 ma - boo hiss. Working on PS for screen supply to test power tubes.

[Merlin]

Another mini tube tester project that might interest some people:
http://valvewizard.co.uk/valvetester.html

[km6xz]

The Tek 576 is somewhat of a white-elephant on the used test gear market you can find one in the US for a very reasonable price. I've had the 575 before, and still have the old classic 570 that is going for an arm and leg....but is only good for tube curves. I use a Fairchild 6200B on my workbench for solid-state and tube curves. It has a anode sweep to 1,000 volts and steps current or voltage for the control element. I built up a small screen supply and heater in a self contained socket fixture that could be used within 24 inches of the curve tracer. That was the plan anyway, there was too much cable capacitance so the traces became loops. So now, I just use the external screen and heater supply, and use a tiny plug-in socket adapter plugged in to the front panel terminals of the tracer. That got rid of the looping.

I find that few of the modern tubes test to the specs of the old RCA tube manual, lots of 6L6 style tubes labeled as such but not very close to the definition of a 6L6. Matching is a joke except at some static low level fixed anode voltage. When building a new project I start with the tube curves as measured and forget the tube manual. Tubes sold as the same designation are all over the map and should not be called the same. Whatever one is, is OK, as long as we know what the specs are so we can build around the specs. For repair shops or manufacturers who are going through thousands of tubes are OK if they stick with style but it would be trouble if there were multiple sources of the same type tube. When I had a large 15 tech shop for pro audio throughout the 1990s until 2004 when I moved from the country, we used only one "brand" and only a few styles within the single brand just so we could predict the results of a repair and guarantee the operation of a repaired unit. Some brands were just plan poor in consistency and we could not warranty the work is a customer brought in their own tubes which were often expensive and poor.  Groove Tube were the worst for being inconsistent and they were more than other brands which got the whole stock from the exact same plant in Russia or China.   
I live across town from the real Svetlana and visit the plant sometimes. They have really cut back on their low power tube production and raised prices so I can't depend on them for the power tubes I liked of theirs:6550's, EL34 and 6L6xx.
Yes, a curve tracer is sure handy now that tubes are so variable.

[twbranch]

In my experience the best and only real way to test tubes is in the circuit it is being used in. Tube testers are 70 year old technology in my opinion. With that said, I would only buy the George Kaye model (if I could find one). It is built around an actual circuit used in amps (at least I am told, I have never seen one). My home system started out as a Fender bassman clone amp that has grown to include other amps with it. It is set up on a book case with switches so that the general circuitry can be used by all testing amps. It is a mess and takes up space but it started out as small project and as I gained more knowledge of tubes it grew with my knowledge. BTW it has been growing for over 15 years now. If only I could make it smaller.

[LHPcope]

For those of you that may be interested I am attaching the URL for the full calibration procedure for the Hickok 539C.  Part 15 gives the rundown for checking the tester's gm.  This procedure should be adaptable to other transconductance type testers.

http://www.desmith.net/NMdS/Electronics/Hickok/Hickok%20539C%20Calibration%20V4.1.pdf