Power Tube Specs, performance and longevity

[new2Krell]

Hi all, I would love to hear your comments in connection with power tubes, particularly regarding specs / matching, real world measurements, and tube life.

Over the years, it has become apparent to me that not all power tubes perform as expected in my amplifiers. Worth mentioning perhaps, I have no means of testing the power tube specs other than measuring current draw in my amps (trim pots set to minimum plate current). For the record, my power amps use EL34’s and I believe the terminology for the amps bias system to be, “fixed bias – adjustable”.

First point: Matched sets sometimes perform very well (similar), sometimes not. Using the above method on one occasion three tubes of a matched set (tested on an AVO) pulled around 25/26ma, but the fourth 42ma (unusable – amp recommendation 26 – 29ma, I also bias at the lower end as that produces the best sound, to my ears) My reserve NOS Svetlana EL34’s have also revealed similar anomalies. One set with test readings: PC 67.0 / TC 8.5, PC 67.0 / TC 8.5, PC 71.6 / TC 9.0, PC 71.6 / TC 9.0 give readings in my amps of 18.6, 20.8, 21.0 and 23.1 ma. Not as dramatic as the previous example and all usable tubes, but not exactly as closely matched as the test results would suggest? Another set I have: 88 / 9.4, 86 / 9.5, 88 / 9.5 and 86 / 9.5 give readings of 24.4, 23.1, 28.7 and 29.8 in the amplifiers. Two are fine for use, two are simply, “over the top” as far as I am concerned. Note, I have no idea what equipment these last two sets were tested on.

Why is there such a variation with “matched” tubes when installed in amplifiers – surely matched tubes should have a similar current draw?

For me, the ideal would be tubes that draw around 25/26 ma at idle in my amplifiers (min bias setting). There is no real issue with tubes that read below that figure - I simply adjust the bias to suit, but......

This leads to my second point / question: should / will all tubes (same manufacturer / type) perform the same when biased to the same current draw, or will a tube that naturally has a 26ma current draw, be any different to a tube that has a natural idle current draw of say 16ma, and needs to be biased up?

Third, and final question: if a tube starts out with a lower “natural” current draw and has to be forced up to obtain the required current draw / tube output, does this have any detrimental effect on the longevity of the tube (it is, after all, working harder - in my mind at least)?

I have tried searching for threads with information along these lines, but have not managed to find anything. It would be great to hear comments from knowledgeable members. Many thanks in advance

 

[AMR / iFi audio]

Well, I'll try to address some of the questions:
1. Matched tubes should perform the same, meaning they should have similar bias points. If they don't, then either the matching was not conducted properly, or one of them got damaged in transport.
2. They should have the same bias current at the same bias points. Up to 5-10%. If the difference is bigger, then you can make it up by regulating bias, but that shouldn't be a difference of 26mA to 16mA.
3. Not likely in most cases. Excessive heat is what makes tube's life shorter. If you need to crack the bias voltage to get more current, then it is okay. It's the bias current and plate voltage that makes up tube's power dissipation. So no, it won't have a negative effect.

 

[new2Krell]

Thanks for your comments AMR. You have largely confirmed what I was thinking about the test results and "my" results.
Good to hear there should be no detrimental effects to longevity for tubes that need to be "wound up" to the required level - that clearly wasn't my line of thought (logic), so really great news.

Clearly some of the tubes, in those "sets" mentioned, fall welll outside 5 - 10%

The only option then is to use tubes that give similar results in my amps, which is something I have done in the past. Some tubes have had a colour code rather than quoted figures, and I have always assumed it fine to mix and match these. When tubes are considerably different to the norm (ie. others with the same colour code) I presume, just because the current draw may be the same or similar to tube with another colour, these tubes wouldn't really be matched? Or does having the ability to adjust the bias largely negate the necessity to use pre-matched tubes anyway - as long as I use tubes that fall within 5-10% at the same bias point in my amps?

I don't know much about tube testing, other than there are various testers available. A quick search suggests testing can be done at various voltages, even with the same device.
One thing I didn't, and possibly should have mentioned earlier - my amps run in excess of 500volts (plate). Is it possible when tubes are tested at lower voltages differences appear or are exaggerated at higher voltages? Or will tubes tested at lower voltages remain broadly inline as the voltage rises? Just wondering if this could potentially be another reason for discrepancies.

Much appreciated!

 

[AMR / iFi audio]

Tubes should be fairly similar in the beginning, but with time, their parameters will start to drift. If one of the tubes is an outlier in the beginning it can run out of adjustment range before the others do. Also their bias is a factor of their electrical parameters and you want them close together so your amplifier performs as good as it can.

 

[retired farmer]

When one goes should I change the whole set?

 

[retired farmer]

part of the op was asking about life span. I also would be intrigued by what people think about life span. The canary ca 160s i have are an el 34 based tube monoblock. I bought it and the store handed me a set of electro harmonix tube set. Not sure but three of them.went in a fairly short amount of time. Thinking the tubes were likely used. For better or worse I just changed the three bad ones and re baised them all. I have a new set on order. I tried gold lion kt77s in those amps I didn't care for that. I actually really like the sound of the electro harmonix tube set. I ordered a matched quad of kt88 from sohpia they are advertised as long life what ever that translate s to. As that amp was used I want a set as backup. If I like the sohpia ones the stock tubes will go in the drawer as backup. Put a set of tad 6l6 tubes in another set of monoblocks I don't like them that well so waiting for the new set to be delivered and the tags will go into the drawer as backups. Any ideas of life span I would be interested in. I read some places up to 4000 hours and another place I was reading 10000 hours on the sohpia tubes.

 

[adrianywu]

Matching plate current at a fixed bias point is pretty pointless, especially if your amp have adjustable bias to match the current of the different output tubes. And the measurements done by your supplier are not necessarily at the operating point of your amp. If you want to match tubes, you need a curve tracer, such as this one: https://www.essues.com/etracer/
The longevity of the tubes depend on how they are run. Does your amp pre-heat the cathodes before applying B+ voltage ? Failure to do so will result in cathode stripping and shorter life. Cathode bias is also gentler on the tubes than fixed bias, since there is cathode degeneration to avoid very high plate current. With the E-tracer, you can also measure leakage current and dispose of tubes that are leaky and prone to failure before it occurs. Few suppliers will do that for you, and it is often necessary to buy more tubes than you need and choose the best ones if you want optimal working condition.

 

[new2Krell]

I would say it depends. I have recently had a failure, though I am not sure whether it was the tube at fault or a resistor (I am actually leaning toward a resistor) - the tube red plated. The tubes had been in there a while, I couldn't say the number of hours, so on this occasion I played safe and changed them all, after replacing the resistor. If I had 3 tubes fail out of 4, I would simply replace them all and keep the remaining one as a spare.
Re longevity, I have owned my amps (EL34 monoblocks) for over 35 years and during this time have only experienced 2 or maybe 3 tube issues. I consider tubes in my amps to have a pretty good longevity, but they are biased quite cool. The previous tube set were biased at 27.5mA. So far, the best perfomance with the new set comes at a slightly lower 26.7 mA, either way the figures only equate to around 53-55%. In the early years I tried a variety of tubes and settled on the Winged C maybe around 20 years ago. This is the first issue I have had involving one of this type and, as mentioned, I am not even sure it was really the tube at fault.

Matching plate current at a fixed bias point is pretty pointless, especially if your amp have adjustable bias to match the current of the different output tubes. And the measurements done by your supplier are not necessarily at the operating point of your amp. If you want to match tubes, you need a curve tracer, such as this one: https://www.essues.com/etracer/
The longevity of the tubes depend on how they are run. Does your amp pre-heat the cathodes before applying B+ voltage ? Failure to do so will result in cathode stripping and shorter life. Cathode bias is also gentler on the tubes than fixed bias, since there is cathode degeneration to avoid very high plate current. With the E-tracer, you can also measure leakage current and dispose of tubes that are leaky and prone to failure before it occurs. Few suppliers will do that for you, and it is often necessary to buy more tubes than you need and choose the best ones if you want optimal working condition.

Interesting Adrian! I have often noticed "matched" tubes do not always seem to behave (draw similar plate current) when installed in my amps. Not a great problem as I have adjustable bias, but your comment may go some way to suggesting why this is.
An issue I had replacing this last set was, the initial replacements just didn't sound as they should. I had a pair that drew around 22mA and a pair around 25mA. I put the 22s together in one amp and the 25s in the other (biased them to the same level), but even after several hours (days even) the sound just wasn't right. I located another pair of 22mA tubes and replaced the 25's. No matter what I did, trying all levels of bias, and switching tubes with each other, I just couldn't get them to perform as they should. I found yet another pair of 22's and, swapping them one by one, I found two culprits. I have no way of testing them so no idea what the issue is, but something is certainly odd about these two tubes (I have never had this issue before). They have been put on one side, together with the old tube that red plated, to be tested by someone at some point. It sounds like a curve tracer test you mentioned may be more revealing.

 

[adrianywu]

I would say it depends. I have recently had a failure, though I am not sure whether it was the tube at fault or a resistor (I am actually leaning toward a resistor) - the tube red plated. The tubes had been in there a while, I couldn't say the number of hours, so on this occasion I played safe and changed them all, after replacing the resistor. If I had 3 tubes fail out of 4, I would simply replace them all and keep the remaining one as a spare.
Re longevity, I have owned my amps (EL34 monoblocks) for over 35 years and during this time have only experienced 2 or maybe 3 tube issues. I consider tubes in my amps to have a pretty good longevity, but they are biased quite cool. The previous tube set were biased at 27.5mA. So far, the best perfomance with the new set comes at a slightly lower 26.7 mA, either way the figures only equate to around 53-55%. In the early years I tried a variety of tubes and settled on the Winged C maybe around 20 years ago. This is the first issue I have had involving one of this type and, as mentioned, I am not even sure it was really the tube at fault.



Interesting Adrian! I have often noticed "matched" tubes do not always seem to behave (draw similar plate current) when installed in my amps. Not a great problem as I have adjustable bias, but your comment may go some way to suggesting why this is.
An issue I had replacing this last set was, the initial replacements just didn't sound as they should. I had a pair that drew around 22mA and a pair around 25mA. I put the 22s together in one amp and the 25s in the other (biased them to the same level), but even after several hours (days even) the sound just wasn't right. I located another pair of 22mA tubes and replaced the 25's. No matter what I did, trying all levels of bias, and switching tubes with each other, I just couldn't get them to perform as they should. I found yet another pair of 22's and, swapping them one by one, I found two culprits. I have no way of testing them so no idea what the issue is, but something is certainly odd about these two tubes (I have never had this issue before). They have been put on one side, together with the old tube that red plated, to be tested by someone at some point. It sounds like a curve tracer test you mentioned may be more revealing.

If you match tubes based on the plate current at one particular DC operating condition, two tubes might have the same current at that point but the curves might be completely different. You can test them at a different bias voltage and see if the current still matches. At least that way, you will have two points on the curve that matches. As you know, if you just have one point, it tells you nothing about the slope of the curve. If you have two tubes with very different curves at each phase of a push-pull amp, the output distortion will increase. Matching plate current is really just to ensure the output transformer does not saturate due to a net DC current. It does not do anything to the distortion.

 

[retired farmer]

So what am I paying for when I buy matched pairs,matched quads, and matched octets,? Or am I just giving the seller extra money

 

[retired farmer]

Also I am wondering if the higher voltage is part of my tube life problems? I am thinking that is a likely possibility. Regards

 

[retired farmer]

The voltage measures 123 at my house

 

[adrianywu]

Also I am wondering if the higher voltage is part of my tube life problems? I am thinking that is a likely possibility. Regards

You would have to ask the seller what he means by matched pair.

 

[new2Krell]

If you match tubes based on the plate current at one particular DC operating condition, two tubes might have the same current at that point but the curves might be completely different. You can test them at a different bias voltage and see if the current still matches. At least that way, you will have two points on the curve that matches. As you know, if you just have one point, it tells you nothing about the slope of the curve. If you have two tubes with very different curves at each phase of a push-pull amp, the output distortion will increase. Matching plate current is really just to ensure the output transformer does not saturate due to a net DC current. It does not do anything to the distortion.

Understood! Like RF, I have always bought matched tubes and until now I have not had any issues like this. I guess there is matching and there is matching.

The voltage measures 123 at my house

Over here the traditional mains voltage was 240, within Europe 220. I moved to Europe in the late 80's taking my amps with me. The power amps were fine but the pre just wouldn't work on the lower voltage. Fortunately, the transformer in the amp was pre-wired to work off 220v - it just needed re-wiring. Having returned to the UK I had the pre re-wired again for 240v and all was fine, until a smoothing cap failed in one of the power amps. The caps were rated at 500v, but the amps were running at over 500v. On start up I have seen 563v, hence I replaced the caps with 600v versions. The voltage here tends to fluctuate between 240 and 250. I have seen it as high as 252v on occasion, more often than not it hovers around the 245 mark +/- 2v. 245v is the reference point I use for setting the bias and the amps generally vary from 510 - 515v.
Yes, voltage can be an issue. The winged C's are good for 800v plate, but the cathode voltage is getting right on the limit. Having said that, as mentioned previously, possibly only 1 tube failure in over 20 years and even that may not have been the tube itself. I suspect it is not likely to be a voltage issue. I would check the specs of your tubes and compare with actual amp voltages. BE CAREFUL if you are messing around inside the case trying to take measurements - IT COULD BE LETHAL! If you are not sure, better to get a knowledgeable person to check for you. You may need to source different tubes, but feel it unlikely. No harm in trying other tubes though (same type), a bit of rolling!

If the tubes are biased hot, that will reduce lifespan.

 

[new2Krell]

A quick correction to an error in my last post, it's the screen grid voltage that is cloes to tube spec limits, not cathode voltage. Apologies for the brain fade. (it was getting late).

 

[WLGMuzza]

Very interesting discussion. I have been in dispute with the local agent for a very well known manufacturer of high end tube amplifiers. The estimated average tube life for my power amplifier is 3,000 hours on KT88 tubes due to some fancy power on sequencing and autobias.

I've had 3x sets of tubes, two supplied by the ampliifer manufacturer, and one set supplied and tested by a reputable tube dealer.

None have gotten substantively close to the advertised, and statement in the manual, 3,000 hours, with the first set that came with the amplifier only making it to 1900 hours before total replacement.

The reason given by the manufacturer is 'tube variability despite selling me 2x sets of factory tested and burnt-in items. It was also suggested to me that the latest tubes (those manufacturered in the past 2-3 years) are less reliable than those produced in the prior 5-10 years.

Let the buyer beware......

YMMV

Cheers

 

[AMR / iFi audio]

Very interesting discussion. I have been in dispute with the local agent for a very well known manufacturer of high end tube amplifiers. The estimated average tube life for my power amplifier is 3,000 hours on KT88 tubes due to some fancy power on sequencing and autobias.

I've had 3x sets of tubes, two supplied by the ampliifer manufacturer, and one set supplied and tested by a reputable tube dealer.

None have gotten substantively close to the advertised, and statement in the manual, 3,000 hours, with the first set that came with the amplifier only making it to 1900 hours before total replacement.

The reason given by the manufacturer is 'tube variability despite selling me 2x sets of factory tested and burnt-in items. It was also suggested to me that the latest tubes (those manufacturered in the past 2-3 years) are less reliable than those produced in the prior 5-10 years.

Let the buyer beware......

YMMV

Cheers

Well, what are the estimated tube life periods stated by the tube manufacturers? OF the set that you have ordered from a other vendor.

 

[WLGMuzza]

Well, what are the estimated tube life periods stated by the tube manufacturers? OF the set that you have ordered from a other vendor.

I guess that's the rub, I don't have those details, instead relying on the tubes tested and certified by the manufacturer to reach their claimed average life.

I paid over double the price for those tubes compared to a set from an independent retailer. The average life is less than 1900 hours, not close to 3,000 hours, compared to the 'non-authorised' version at 1400 hours.

Do EH publish tube life figures? I suspect not given the variable operating parameters across different applications.

 

[retired farmer]

@ WLGMuzza which tu e numbers were you having the sort life with? EL34's? I would find that interesting as the electro harmonix EL34's I they didn't seem to last to long that being said the dealer put a used set into a new amp for me. But very short life. I actually like the way they sound.

Regards Tom

 

[WLGMuzza]

@ WLGMuzza which tu e numbers were you having the sort life with? EL34's? I would find that interesting as the electro harmonix EL34's I they didn't seem to last to long that being said the dealer put a used set into a new amp for me. But very short life. I actually like the way they sound.

Regards Tom

A set of EH KT88's, a set of EH 6550's, and a set of Preferred Series KT88's.