Rerurn to Romy the Cat's Site

Audio For Dummies ™
Topic: Uncomfortably comfortable tube testers

Page 1 of 3 (45 items) 1 2 3 »


Posted by Romy the Cat on 01-07-2007

Paul, another my spin a little off the subject but since it has to do with your current dealing with ML2 I would express it here.

Since you use some tube gear then why you agree to be a slave of random circumstances? Tubes have problems – they go out off their good operation gradually…  and it is the bitch.  When they do when the filaments are still glowing and the tube could care the current (many variations are possible, deepens of the tube) but the tube sounds “like shit”. It actually good when the tube sounds like shit – juts replace it and forget it, however the problem is the most of the tubes do not sound like shit but rather have a fraction of their sound “like shit”. It is very hard to detect it subjectively that the tube is going and you need a tube tester to do it.

It looks like your are deep into the audio it is does not look that tomorrow you will switch to stamps collecting or to chess playing. If so, then get yourself a tube tester or any other means to test tubes objectively and reliably. The alternative might be terrible. Some people have their equipment operated for year with worn tubes, the tubes that might have some tonal and “energetic” deficiency but the people juts do not know bout it. We demand objective cardiogram in order to detect a health condition of heart. Tubes are the hart of tube amps and why do not treat them with the same demands?

Rgs,
Romy the Cat

PS: this thread is split from: Initial thoughts about new/old Lamm ML2s

Posted by Paul S on 01-07-2007
Wise words that mirror my own renewed concerns, Romy.

Any recommendations for a particular tester, anyone?  The last time I put my hands on a tester to use it was in a drugstore, one of the many places you could find them back in the day when radios and TVs were almost all tubed.  Oh, and it was as big as a two-drawer filing cabinet on stilts...

As much as I do NOT want to OD on DIY, I have long wanted to be able to test my tubes for myself, both to extend my purchasing options and to add that capability for process-of-elimination trouble shooting, just like having a good DMM or dB meter, etc.

Although I am in no way "committed to tubes", I suppose I have to face the fact that my aural tastes have long tended that way.  In any case, I don't see the ML2s getting superceded any time soon...

Best regards,
Paul S

Posted by Romy the Cat on 01-07-2007

I do not know, you have to decide to yourself what kind tubes and the most important HOW you would like to test them. For different tubes, differently used the different things are important.  I had a few testers and all of them were OK but all of them had some “twists” that made me to wish something else and did not allow me to do what I wanted.

Hickoks are nice, particularly the 580A (allow to measure a small currents) there are many others. But most of the testers kind of too generic and too … uncomfortably comfortable. I need to test a specific tube, twice a year at the very specific mode, specific operation conditions and get the very specific answers. I do not need to have those stupid English Dialers and user-friendly operations. I’m not a pharmacy in 1968 after all.

Generally the best bet (after you nailed down what you want to do with your specific tubes) is to build a very simple own tube tester where the necessary for you tube will be sitting in the very defined operation mode, identical to the way in which it used in your equipment. (BTW, do you know that all of those tube testers drive plates with AC?).

I personally resolved the problems for myself by going for a half-ass solution. I got Hickok TV-2 and slightly modified it. My Hickok TV-2 can push quite a lot of current now and it measures the current in real time. Still, it does not allow me to measure the tube’ gain but I have other ways to do it. Some people feel that TV-2 it too complex and too slow in operation. I disagree. It is not complex it does exactly what it meant to do. Yes it is not as “plug and play” as some other testers but I use my tester a few times a year, so I’m not in hurry… Also, I use tubes only in my phone and my power amps. With phono is a different story but in my Super Melquiades I have made essentially a tube testing capacity right into the amp. Knowing bias for a new output tube at a giver current and knowing where it go with time it is possible semi-say what is going on with the tube. I way “semi” because the 6C33C is a freaky tube and has it’s own idiosyncrasies…. My biggest problem to resolve those idiosyncrasies is… the price of the 6C33C. Since I buy them for $5.5 I can afford to be very liberal with them…

BTW, if you need to teat juts small tubes then you might consider the contemporary digital testers. I have seen some very cool testers but I never own one….

The Cat

Posted by Paul S on 01-07-2007
Deja vu:  I seem to remember getting about this far on the subject any number of times.  Not only do I have fairly limited applications, but I also noticed that many of the testers do not even allow the kind of testing I would want to do, or each is limited to one or two of the several functions I need.

I remember some of your photos showing your DMM hooked into a test(er) loop.

And, no, I never realized the damn things used AC!  How weird is that?!

But few of those portable testers test at actual working voltages, etc.

Have you seen this one?: www.amplitrex.com/at1000.html

As close as I've seen so far to covering what I want, but a $200 adaptor is required for the 6C33C, which pushes it to $2,600, right out of the chute.

I could just go ahead and pay Lamm's price for lots of critical ML2 tubes and still have enough left over to make a nice (tax deductable) contribution to the GOP 2008 war chest, maybe get myself a dinner invitation...

Best regards,
Paul S

Posted by Romy the Cat on 01-07-2007

Perhaps you are right. Buying out tubes from manufacture is a secure way to go. Still, I abandoned this practice in 2002 when Lamm sold me a pair of 12AX3 for his L2. When I got them one of the tubes was with paces of metal riddling inside. I figure out that it was probably damaged in shipping, although I very much doubt as was quite well packed. I send the tube back and Lamm sent me a replacement. The replacement tube did not go off. I send the tube back again. Lamm sent me another replacement but this time the new tube was heavily USED with the walls quite burned out. (On this type of tubes it is a very good indictor that the tube should be trashed – the 12AX3 in L2 has life-span less then 6 month). I disregarded the fact the I paid near $40 for the pair of tubes that somewhere else were $3.5 and decided juts discount the entire story, presuming that it would be too pathetic on my side sending the damn tubes back again. I bought the tubes elsewhere and forgot about it. Then, in a few days when I received from Lamm’s wife an email asking me to pay for the replacement shipment (!!!!!) I said to myself that it was it and it was the very last time when consider buying anything form Lamms.

I moral of the story for me was – whatever they (manufactures) do and whatever bogus assurances they provide – I do not care about them. It is my responsibility to keep my preamp operational. After all, their units serve them only once – when we pay for them. Since then, the units serves us and I need my tubes were driven by voltages instead of egos and by currents instead of “somebody’s reputation”.  BTW, the friend of your Dennis Boylev when we were exchanging insulting emails with him unambiguously implied that the only people who “reach the level of audio knowledge” are those who “listen to what I have to say”. It would not be too difficult to imagine the direction what I suggested him to go. Do you think it is only him?

In the end: if you but some very expensive and very rare tubes who specialize in digging out some rarities and see then at premium pries then you have a remote assurance that the tube would be what was promised (at least mutual conductance and cathode emission at cold, not pre-burned tubes). However, when you deal with “everyday tubes” no one will care to verify their usefulness. Perhaps I am wrong with my attitude but sine you got a good tube tester and very basic knowledge how to test tubes then whatever attitude you have because confidence deriving form your own practice instead of somebody’s else egomania. What? The $300 extra dollars for the tube tester? Another couple years of GOP’s “management” of this country and $300 will the price we will be paying for a slice of pizza… So presume that you exchange the tube tester today, the one that will be serving you all your life, to the slice of pizza in 2008….

The caT

Posted by jessie.dazzle on 01-07-2007

If you don't already know of it, here is a link to interesting German site (in English as well) with loads of info on Tube testers :

http://www.jacmusic.com/Tube-testers/index.html

If you scroll down to this phrase :
"SHORT OVERVIEW OF TUBE TESTER FEATURES IN EXEL FORMAT", there is a link to an Excel chart listing features of several testers.

I will try linking it via html :
http://209.85.135.104/search?q=cache:-wZJjydu-NkJ:www.jacmusic.com/Tube-testers/overview.xls

... If I read correctly, a few of the testers appear to use DC plate current (?).

jd*


Posted by hagtech on 01-07-2007
As usual, they leave out the VacuTrace, by far the best modern day tube tester.  Same with the Alan Douglas book (who claims to be an expert in the field).  They both mention the Sofia, however - a machine plagued with problems - as the epitome of testing.  Yet they only sold about 4 of them and have been out of production for over ten years.

Other testers you can buy new are available from transcendent sound, vacuum tube valley, maxi-matcher, and (at least a few years ago) new sensor.

jh

Posted by Romy the Cat on 01-07-2007

 hagtech wrote:
As usual, they leave out the VacuTrace, by far the best modern day tube tester. 
Interesting, Jim.

I was not in the market for tube tester but I looking at your tracer I fugue out that I might sell a few my cheaper testers (that I need to anyway somewhere) and get yours.  I did not deal with tracers and I hardly know if the trace would give me necessary information. I do not need to chaise operations point – it is have discovered and fixed. As I understand that with a tube getting worn it pick some distortions and alinearity and the trace might show it up.

Still, I wonder if the tracer like your would be able to do what I need. I care less about trunsconductance but I need to assess the level of cathode poisoning and the actual gain of the tubes. Will your tracer be able to pump 250-300mA at 230V? Will it be able to supple up to 100-110V of bias? Also, if it possible to have your tracer not with 2 receptacles but with 4 or 5 had-wired receptacles (transmitting tube, dual triode, single triode (1), single triode (2), penthode). I would have no problem to do it myself theoretically…

Another point. I deal with Russian 6C33C that have absolutely random quality. The new tubes could have completely different operation and I more concern about the progress of each given tubes with time then compliance to a selected trace pattern. So, dos it men that I need to use one of those commentary scopes with memory and compare the traces externally?

Other things. Is it possible to intermit microphoneiks based upon the trace when I knock the tube?

Now the crazy things. I know you professionals would feel that I am an idiot but I did not did not fine any more reliable way to say anything objectively about the tube gain. Having six 6C33C running parallel in 6 channels you might understand that it is critical for me.  Seeing the 6C33C that with identical mutual conductance, identical current and identical cathode emission produce +/_2dB I was wondering how do not convert my amplifier into equalizer… So, the only way how I do it is to set the identical current for all tubes (disregarding the amount if bias as the relation between current and bias in 6C33C is very unpredictable), load the tubes against the normal operational load and then driving reference current thought the grid, measuring the actual gain. I am actual was thinking to add this capacity into my Hickok TV-2 but I do not mind to pay money to somebody who have already took care bout it. So, will the trace or any other operation mode on your tracer be able to objectively tell about the tubes gains against a given load? If not, then would it be possible to add this functionally for extra charge?

Rgs,
Romy the Cat

Posted by hagtech on 01-08-2007

 cat wrote:
Will your tracer be able to pump 250-300mA at 230V? Will it be able to supple up to 100-110V of bias?


Not quite.  I designed it for preamp tubes.  But then I realized I could push it to cover a lot of the more popular output tubes such as EL34, 6550, and even 300B.  The VacuTrace is limited in how hard it can drive a power tube.  It'll deliver 200mA to plate at up to 400V  (really about 380V under load), and only -70V of bias.  This does NOT cover the full range of the bigger tubes.  It will, however, give you a pretty good idea of the condition of the tube and often covers the operating point.  I show this example of a 300B. 



The red overlay is actual photo of VacuTrace output.  It is indeed limited, but does cover a big swath of operating region.  As far as hard-wiring adapter cards, that is pretty straightforward as I sell "blank" ones. 

Russian 6C33C that have absolutely random quality


I have not tried this tube yet.  It might push the limits of the tester.  The 6.3V heater current supply is nominally rated at 5 amps.  It might be ok to push it to 6.6 amps.  I have to think about this, make sure there is some headroom in the transformer core.  Easiest thing would be to fire up a tube and see what happens.  Send me one and I can trace it for you (of course I will return it). 

I need to use one of those commentary scopes with memory


You definately need at least a cheap analog scope, with XY capability.  As for measuring tubes over time, I hadn't thought of that.  Do you think the curves are shifting?  Sure, a tube warms up.  To measure degradation over some time, I think the easy way is to just take two photos.  Maybe you know somebody with a camera?  Ar ar.  One before and one after.  Or you can take spot measurements of current, transconductance, and plate resistance and some specified operating point.  The VacuTrace is limited to withing its range.  If you try to drive it too hard the internal amplifiers protect themselves and go into a "chirping" mode. 

It does not work to test microphonics on this machine.  Nothing shows up on the scope. 

reference current thought the grid, measuring the actual gain


The VacuTrace does something the others don't.  It measures output conductance.  Or one over plate resistance.  This value will give you a comparison on loading.  A tube with higher plate resistance will obviously have lower output gain into a load.  So you are correct, there is more involved than just transconductance.

jh

Posted by Romy the Cat on 01-08-2007

 hagtech wrote:
I have not tried this tube yet.  It might push the limits of the tester.  The 6.3V heater current supply is nominally rated at 5 amps.  It might be ok to push it to 6.6 amps.  I have to think about this, make sure there is some headroom in the transformer core.  Easiest thing would be to fire up a tube and see what happens.  Send me one and I can trace it for you (of course I will return it).

I do not think that heater supply is a problem. The 6C33C has 6.5A at 6.3V

http://www.goodsoundclub.com/PDF/6C33C.pdf

but I usually very liberal with abusing tube testers. My Hickok TV2 is rated at 6.3V at 3A if I’m not mistaken but after a few out of hearing a half of 6C33C it has absolutely cool transformer. In fact when I was making the Hickok be able to handle the 6C33C I had power resistors smoke (I replaced all of them with more powerful) but I never had problem with filaments. In the worst case it always possible to feed the tube heaters form externals transformer… My biggest question is: will a trace be able to tell me the “state of the mind” of a given tube.  If you would to burn you time like I might send you two 6C33C – a new one and the burned out (you would not need to return it) and it would be interesting it the traces would be interpretable. Unquestionably to have a trace for the very given tube is “nice to have” but does it have any practical meaning for the semi-armature, purely “usability” tasks that I need to resolve with this tube?

 hagtech wrote:
The VacuTrace does something the others don't.  It measures output conductance.  Or one over plate resistance.  This value will give you a comparison on loading.  A tube with higher plate resistance will obviously have lower output gain into a load.  So you are correct, there is more involved than just transconductance.

Actually this is very cool feature for my low frequency channel in Super Melquiades. A change current in plate changes impedance of the plate and with consequentially affect the loading. In the past it was kind of balance between how hard you can drive the tube with currant and where the transformer core stop to handle it. With Super Milq LF transformer, gaped at 450mA I can go as high as I wish but staring at some level the sound stop to be changing. I wonder it that “level” would be different for different tubes and will another 6C33C’s plate have different impedance at given current, I am sure it will. Well, it is possible to circulate by “other ways” theoretically….

 hagtech wrote:
So you are correct, there is more involved than just transconductance.

BTW, I was in past contemplating to add to my tube test the “real gain” mode. I was thinking to rectify and filter plate current and feeding the greed with 1000Hz for a simple generator to measure AC on the loaded tube’s plate… It world give me the “no BS gain”… Frankly speaking it is the only one of the two real world parameters that I care (cathode emission, relive to what it was at the very same tube when it was new is the second one)

Rgs,
Romy

Posted by hagtech on 01-08-2007
 romy wrote:
send you two 6C33C – a new one and the burned out


Please do.  I have interest in the result myself and don't mind spending the time.  I can post photos for each tube.

a trace for the very given tube is “nice to have” but does it have any practical meaning?


That can only be answered by the end user.  Suprisingly, I actually sell more units to afficianados than dealers or manufacturers.  You have to be a serious tube guy to even consider such a machine.  Many are satisfied with a go/no-go type measurement.  However, there are many of us who live in the analog world and cannot divide a grey scale.

feeding the greed with 1000Hz for a simple generator to measure AC on the loaded tube’s plate


Yes, that's exactly how I measure the transconductance.  Except I use a 500Hz square wave.  Much easier to sample and hold the peak level of a square wave than a sinewave.  Sometimes the voltages can get pretty small, and doing analog division can be noisy.  Oh wait, it is different than what you suggest, as I load the plate with (close to) zero ohms.

jh

Posted by Romy the Cat on 01-08-2007

 hagtech wrote:
Please do.  I have interest in the result myself and don't mind spending the time.  I can post photos for each tube.

Sure, let do it. Email me your address and I will send them to you. (Do you have sockets?)

 hagtech wrote:
That can only be answered by the end user.  Suprisingly, I actually sell more units to afficianados than dealers or manufacturers.  You have to be a serious tube guy to even consider such a machine.  Many are satisfied with a go/no-go type measurement.  However, there are many of us who live in the analog world and cannot divide a grey scale.

It is not about the “go/no-go” but the recognition of what a given person would like to get out measurement. Jim as I understand, and correct me if I’m wrong, the trace is a characteristic if the tube design, dimensions, sixes, allocations of the electrons clouds, materials and zillion other things. If a person has two dozens of 12Ax7 or a dozen of different 2A3 tubes, made in different years, by different manufacturers and by different contraction techniques and materials then trace is very nice objective way to differentiate the tube performance. However, in my case, what I use the very same tube in the very fixed operation mode the trice would not be indicative, would be? Well, it possible that a new burned tube will have trace moved but would it say anything about the tube state? I really do not know at this point  as I never traced them.
 hagtech wrote:
Yes, that's exactly how I measure the transconductance.  Except I use a 500Hz square wave.  Much easier to sample and hold the peak level of a square wave than a sinewave.  Sometimes the voltages can get pretty small, and doing analog division can be noisy.  Oh wait, it is different than what you suggest, as I load the plate with (close to) zero ohms.

Hm… actually I was thinking to load the tube with the actual load  (with choke), the one that it will see in amps… Well, theoretically if your tracer gives exact number of transconductance and the exact number of plate impedance then it is possible easily to circulate gain. Still, and it is what I afraid – did you come across a situation when the estimated gain is not the same what your dB meter shows after you power amp’s secondary?

Rgs,
Romy the Cat PS: BTW, I had more on it at: http://www.goodsoundclub.com/TreeItem.aspx?postID=2277

Posted by hagtech on 01-08-2007
Sadly I have never spent the time to try and correlate the sonics of a tube with the trace characteristics.  Ok, sometimes it is obvious.  A flat line is a dead tube.  Normally I measure and compare both sections of dual triodes.  I need them matched in my circuits.  Very well matched for balanced operation, reasonably matched for channels (for single-ended circuits I split the halves between channels (another long subject)).  I grade each tube.  Not so good tubes hit the unused bin (I throw away quite a few).

Not well matched tubes often show the characteristic of one side having a funny shape.  It's like the curves are more linear, start stronger near zero bias, but then stay flat and weaken at higher currents.  At least compared to the good tube.  I assume the funny shape would correlate to bad sound.  But like I said, I never spent the time to listen.  I just chuck 'em. 

Another test is to watch the tube as it warms up.  You don't want to push this test too hard as it can lead to cathode stripping.  As the heater warms up, I apply plate voltage.  The curves come up but max out at the available electron limit - the curves have an "S" shape.  The tops go flat.  This flat continues to rise as the cathode reaches operating levels.  Eventually I cannot see it anymore.  It all happens in ten seconds.  Now, I have noticed that one old and very used tube I had would not be strong and held onto the "S" shape.  A real good indication there was very little emission left.

Having done this for many years, I would guess that you can pick out a bad one, but not pick out a good one.  Does that make sense?  A bad tube will be obvious.  But two good tubes will measure fine and operate fine, but not guarantee sonics.  One may sound way better than the other.  In circuit they will have the same electrical performance, same risetime, same distortion, same gain.  But to the ear they can be world's apart.

estimated gain is not the same


Well, I've only done two amplifier designs.  One PPT, the other SET.  I ran the math based on average curves, not specific tubes.  So I don't think I can really answer the question.  Not only that, I think I made a mistake on the PP calculations.  The gains came out right, but not the output impedance.  The first OPTs I used had the specified turns ratio.  But the transient performance sucked.  I plopped in a very similar tranny (or so I thought) with a lower turns ratio and now everything came together.  Huge difference.  But the output impedance was as low as I had calculated for the other tranny.  So I dunno. 

jh

Posted by Romy the Cat on 01-09-2007

Jim, actually I see your PO Box at your site, is it the place where I need to send the tubes?

I undusted what you are saying. It is easy to pick the obviously bad tubes but it is much more complicated to have a tube that is “a half way there” and be able via testing it to assess its sonic merit. The tubes degrade gracefully and in some instances completely unpredictably. I have seen perfectly measured (at least tranceconductance and gain) power tube that sounded extremely poor and at the same time, among the small signal tubes, I have seen burned-to-death tubes that sounded better then anything new. Go figure…

I think each tube brand/type, like humans, has own way to go out. Sure, there are some common denominators and tendencies and testers or perhaps the tracers might portray those tendencies. Still, I feel the most valuable in tube testing is not the testing procedure or testing equipment but an ability of off an individual to conduct necessary test and to intrepid the result in context of his/her auditable practice. I have very limited grip on this interpretation, probably if I use more expensive tubes then I would treat them with more respect. Still I am sure that “testing understanding” are obtainable skills and I do not mind to develop them…

The caT

Posted by Romy the Cat on 01-10-2007

Jim, I have sent you the tubes – it might be an interesting journey.

I send you:

1) 6C33C that was working with both cathodes approximately a year. It is not dead year but it is deep into retirement age. Run it at 220V and 200-250mA full-tube, or around 135-150mA if you use a single cathode. It might need bais from minus 65V to minus 110V on grid for the full curent.

2) 6C33C new. It is brand new tube, never works. Market with word “new”. Be advised that since it is “random” 6C33C it might not necessary push more current at the given bias then the use tube. Still, most likely it will need to close the tube more to give the necessary current. I did not measure it.

3) 6E5P, new. This is the Milq’s driver. Take a look what it is. Use it at 180V-200V and ~15-20mA. It will be around -3.4V in grid

4) 6E5P, heavily used, it could be easily spotted as its walls looks like after Hiroshima.

5) A kinky 6E6P-DR (it is shorter and has a gold string inside). It is an equivalent of 6E5P, the same constriction, only with “western” pins layout. This tube was designed to meet military specification and made with expensive materials and to very high QA specs.  I use it in the Super Milq HF channel. The tube that you got is very fresh and might be conceded as “new”. It has “brownish” wall but it is not the use of the tube in the field. All “DR” tubes were pre-burned for 200 hours in the plant and then were retested before they ever left the manufacturing facility. Use the 5E6P-DR in the same way as the 6E5P.

Everything it very straightforward with those tube tubes. If you never dealt with 6C33C then be advised that there is not such a thing as too mach to heat the tube. If you look on-line or talk with people then you might see that the Morons complain that 6C33C has tendency to overheat anode with fixed bias. It is incorrect. The tube in order to care power should be really hot, deep hot. Do not give to her more then 30W before first 15 minutes, more then 45W before 30 minutes. After one hour of heating you might dissipate on it 55-60W and the tube will do perfectly fine. I made Milq to have 2 min delay before I do B+ do not stress the amps with high signal during the first half hour-45 min. Well, the amp sound not good in beginning anyhow…

Rgs,
Romy

Posted by hagtech on 01-10-2007
around 135-150mA if you use a single cathode


Hey, I never thought of this.  You can run 1/2 the tube at a time!  This actually puts it within the limits of the VacuTrace.  I can test each side separately.

jh

Posted by Romy the Cat on 01-11-2007

 hagtech wrote:
Hey, I never thought of this.  You can run 1/2 the tube at a time!  This actually puts it within the limits of the VacuTrace.  I can test each side separately.

..and while you do so you might discover one of the mysteries of the 6C33C – the tube that is running with both cathodes has each cathode loosing it’s emission with different tempo and each half is loosing it’s transconductance with different pace. It does not necessary has any influence to sound but it says prefabs about the bad quality production of Russian tubes – when each cathode has a different amount of the cockroaches legs on it. BTW, the 6C33C with one cathode it is how it being used in the Super Melquiades HF and MF channels.

The Cat

Posted by Paul S on 01-11-2007
the less you know about...

So, how do we go about screening tubes for audio, particularly audio signal tubes?  The only person I can think of who claims to have found a specific testing paramerer for this is Roger Modjeski, of Music Reference and RAM (tube) fame, who says his tests for the "RAM Factor" weed out the bad/yield the good in terms of sonics.  I have used lots of his tubes, and they have always been good-sounding versions of whatever (generally-inexpensive) brand they started life as, but I cannot say that RAM tubes I tried sounded better than or even as good as various (expensive) NOS tubes I compared them to, in the same application.  Although I ass-u-me the NOS tubes in question were "tested" and "matched" by the seller, none claimed Ram Factor-like screening, which means that in my experience the particular brand and vintage of a tube may trump RAM screening of "lesser" tubes; ie, good bloodlines will out.

But this still does not solve the problem of what we should test for and how do we do it for a given tube.  Since this thread started I have been looking at testers, trying to land on one that appears to be worth the trouble, for my applications, which are strictly limited to audio.  What a maze!  In addition to the link I posted back up the thread I ran across this one:  www.tubesontheweb.com/matic.htm   Still, so far, all of them look like variants of the same testers that most of the decent vendors are already using.

With respect to the "decent vendors", I have for many years now averaged about 25% bad tubes from all sources, and I only buy from sources who at least claim they test and match, etc.  Does this mean that somehow the testers are dropping the ball on 25% of the tubes they test (if you can accept that they really do test them)?  

Germane to this thread,  how is any testing I do going to improve on this situation?

Is there any agreement on what is relevant?  Does anyone really know?

How many test and which tests does it take to adequately screen one's hi-fi tubes, and how many testers are required to run these tests?

I am more confused about this now than was when I started reading more about it.

Best regards,
Paul S

Posted by hagtech on 01-12-2007
Tubes came today. 

solve the problem of what we should test for


I don't think you can get one tester to do everything.  Not only does one need the basic parameters such as transconductance, but you have shorts, gas, leakage, microphonics, secondary emission, etc.  All kinds of relevant stuff to check.  The question is what to check for.  I think the way to answer this is with experience.  And RAM is way ahead of most people in this regard.  Has he explained what the 'factor' consists of?

With my limited experience, I would suggest looking at grid current.  Measure it versus bias at different operating points.  What you will find interesting is that there is a point where it passes through zero.  That's usually where I set my operating point (it's more of a locus of points).  I modified a VacuTrace adapter card and added a 10k resistor on the grid, then used a DVM to measure voltage across it. 

For me, I use the tester to weed out bad tubes, then match and grade the good ones.  Those with funny shaped curves get lost.  My amplifiers ship with these modern production tubes that I screen.  Can't remember the last time I had a DOA.  In fact, maybe never.  But hey, I'm not really a big player like Lamm.  It's ok for me to ship with Sovtek or similar new stock tubes because customers almost always roll them.  No reason for me to put money into something they replace.  I guess the point I want to make is that tested tubes should be 99% acceptable.  So what I buy from the factory gives me about a 30% reject rate.  Same as you.  So it's about 70% ok coming in, 99% going out.

jh

Posted by hagtech on 01-12-2007

The sockets for 6c33 should be here in a few days.  Meanwhile, I fired up the 6ep5s.  I labeled them ABC.  The B sample ('83) with gold wire has a short or something.  The other two look fine A ('63) has a little more output than C ('62) for the given bias.  These are just random shots.  Settings were 1V/step on the grid and a G2 voltage of 187.  Anything in particular you would like measured?  Transconductance at some bias?  I need to work out a way to get display lines on the scope to show up, then we'll have calibrated mA/div info.

"A" triode mode:



"A" tetrode:



"C" triode:



"C" tetrode:



jh


Page 1 of 3 (45 items) 1 2 3 »