Generally speaking, no. Grossly mismatched tubes can raise the distortion in a push pull amp like the SPP regardless of how they are biassed. Also grossly mismatched tubes can cause volume imbalance between the two channels in a SE amp.
Truly matched tubes however rarely exist. Often many matched pairs or matched quads sold may be "matched" only because the readings on a tube tester were close to each other. Perfect matching would require a curve tracer, and that would only offer a good match at that time. There is no guarantee that the tubes would stay matched for long since they often wear at different rates.
Someone who is trying to do tube matching right will run every tube in a test or burn in rack for a few days before matching, them reject the outliers before selling them. There are some vendors who do this.
I bought a bunch of "gently used" "previously auditioned" tubes about a dozen years ago. Each had a sticker on them with the matching number written on it from a popular Tube Store. They were all good tubes and many are still kicking today, but when I decided to match up 2 pair recently for a big push pull amp, the numbers didn't mean much.
I am of the belief that good matching is needed only when tubes are connected directly in parallel, or if they are being run close to their spec limits in a push pull amp. There are of course plenty of people who would disagree with me though.
Truly matched tubes however rarely exist. Often many matched pairs or matched quads sold may be "matched" only because the readings on a tube tester were close to each other. Perfect matching would require a curve tracer, and that would only offer a good match at that time. There is no guarantee that the tubes would stay matched for long since they often wear at different rates.
Someone who is trying to do tube matching right will run every tube in a test or burn in rack for a few days before matching, them reject the outliers before selling them. There are some vendors who do this.
I bought a bunch of "gently used" "previously auditioned" tubes about a dozen years ago. Each had a sticker on them with the matching number written on it from a popular Tube Store. They were all good tubes and many are still kicking today, but when I decided to match up 2 pair recently for a big push pull amp, the numbers didn't mean much.
I am of the belief that good matching is needed only when tubes are connected directly in parallel, or if they are being run close to their spec limits in a push pull amp. There are of course plenty of people who would disagree with me though.
Thank you, Tubelab! This also answers my question regarding volume imbalance I experienced with two 6L6GC tubes I was running in the SSE I own. They were close in gm measurement on my Hickok tube tester, but I don't have a setup for the curve tracer.
I am now motivated to setup a curve tracer just because I've got a lot of tubes. Not as many as you.
Thanks again!
I am now motivated to setup a curve tracer just because I've got a lot of tubes. Not as many as you.
Thanks again!
Hi George - out of interest, what do you use for testing tubes?
I have this notion that I could build a generic power supply, and then build some form of tube tester that would make use of the variable supplies for heater, grid, screen and plate, and hence would just be used to match bases to test conditions, and execute tests.
Or, perhaps a home brew tube tester that also can be used as a generic power supply, if that is a more logical approach.
I have this notion that I could build a generic power supply, and then build some form of tube tester that would make use of the variable supplies for heater, grid, screen and plate, and hence would just be used to match bases to test conditions, and execute tests.
Or, perhaps a home brew tube tester that also can be used as a generic power supply, if that is a more logical approach.
Years ago I lost my faith in any type of popular tube tester. Even a good Sencore Gm tube tester failed to find tubes with high distortion unstable bias or several other defects that would render them useless in the circuit. So I sold all my tube testers.
If I have a bunch of tubes to test, or a specific need, I will rig up something that will show me what I need to know as quickly as possible. Examples:
I had a bunch of used 5842's and WE417's and a tube dealer that wanted to purchase them for good money....if they were known good. In this case I took a working TSE amplifier board wired the inputs in parallel, removed the output tubes, and connected it up to a bench power supply. I popped in two tubes to be tested, and measured the THD at the output tube grid pin and the plate voltage. I measured one tube while the other was warming up. A scope trace showed the output overlaid on the input to display the gain. Tubes with high THD, low gain, out of line plate voltage, or were slow to warm up were rejected. The few with the lowest THD were kept, the rest sold.
12Asomething7's get tested in a SSE board so I can observe the match between the sections.
Output tubes get rigged up into a suitable test amp and tested at full rated dissipation and beyond. Careful observation of the plate in a dark room when pushed to the edge of red plate will find tubes with poor alignment (uneven plate heating). These will fail prematurely in a high dissipation application.
There are other tests like the pencil whack for microphonics, which I use as needed.
I have a curve tracer, but it is somewhat limited in application especially for oddball connections like screen and dual drive, or power levels beyond 30 watts. In these cases I just use power supplies and meters......
Note dumm blond arithmetic errors can make tubes explode....that pesky decimal point thing. Feeding a poor little tube 400 watts when 40 was already beyond its 24 watt rating makes a mess. It's been 6 months or so and I'm still finding shards of glass in the carpet with my feet.
If I have a bunch of tubes to test, or a specific need, I will rig up something that will show me what I need to know as quickly as possible. Examples:
I had a bunch of used 5842's and WE417's and a tube dealer that wanted to purchase them for good money....if they were known good. In this case I took a working TSE amplifier board wired the inputs in parallel, removed the output tubes, and connected it up to a bench power supply. I popped in two tubes to be tested, and measured the THD at the output tube grid pin and the plate voltage. I measured one tube while the other was warming up. A scope trace showed the output overlaid on the input to display the gain. Tubes with high THD, low gain, out of line plate voltage, or were slow to warm up were rejected. The few with the lowest THD were kept, the rest sold.
12Asomething7's get tested in a SSE board so I can observe the match between the sections.
Output tubes get rigged up into a suitable test amp and tested at full rated dissipation and beyond. Careful observation of the plate in a dark room when pushed to the edge of red plate will find tubes with poor alignment (uneven plate heating). These will fail prematurely in a high dissipation application.
There are other tests like the pencil whack for microphonics, which I use as needed.
I have a curve tracer, but it is somewhat limited in application especially for oddball connections like screen and dual drive, or power levels beyond 30 watts. In these cases I just use power supplies and meters......
Note dumm blond arithmetic errors can make tubes explode....that pesky decimal point thing. Feeding a poor little tube 400 watts when 40 was already beyond its 24 watt rating makes a mess. It's been 6 months or so and I'm still finding shards of glass in the carpet with my feet.
No, but when I lived in Florida I did all my testing in white room, but there were no black curtains, and the (radio) station was torn down when the land became too valuable for such activities.
Testing tubes for glow in an amp is done with no input at all for a SE, or any class A amp (like the SSE) since that is the worse case for tube dissipation.
A class AB amp needs to be cranked to the point of maximum dissipation with a sine wave. That is usually somewhere between half and 3/4 of full power output.
It's weird, I plugged the two 6L6GC tubes that caused the volume imbalance in the Tubelab SSE amp into another homemade SET amp I have- and there is no volume imbalance.
This leads me to think that it must be something in the Tubelab SSE amp?? I bought the amp used and tested the tubes on my Hickok tube tester and B&K 707 tester. They tested good on the B&K, about 87%, and Gm measurements were close on the Hickok, 3400 and 3600.
I'm going to replace the cathode resistors in the SSE amp. Does that sound like a good place to start? The solder joints look a little discolored and I don't know how hard this amp was used. I have replaced the caps on the board. Otherwise, I haven't a clue as to what to look for.
This leads me to think that it must be something in the Tubelab SSE amp?? I bought the amp used and tested the tubes on my Hickok tube tester and B&K 707 tester. They tested good on the B&K, about 87%, and Gm measurements were close on the Hickok, 3400 and 3600.
I'm going to replace the cathode resistors in the SSE amp. Does that sound like a good place to start? The solder joints look a little discolored and I don't know how hard this amp was used. I have replaced the caps on the board. Otherwise, I haven't a clue as to what to look for.
Attachments
The cathode bypass capacitor will affect the stage gain - a bad one will lower it; I'm not sure that the cathode resistor will affect stage gain.
I haven't read the whole thread, but are you assuming the gain imbalance is in the power stage, or have you verified that?
Do you have an oscilloscope or a sensitive ac voltmeter so you can follow ac signal through the amp and see what is happening with it?
edit: for some reason, I see only ONE cathode resistor on your amp. Where is the other one? On the other side of the board? Why? 680 ohms is probably too high for 6L6GC - I would think you want 560 - 510, maybe even 470 ohms ...
edit2: I would also take a close look at the three terminal connector on the side with no visible cathode resistor - it's probably OK, but looking at the picture, it could also be that both ends of the jumper wire are in the same terminal which is not OK; it would leave the screen floating and effect (affect?) the stage gain.
I haven't read the whole thread, but are you assuming the gain imbalance is in the power stage, or have you verified that?
Do you have an oscilloscope or a sensitive ac voltmeter so you can follow ac signal through the amp and see what is happening with it?
edit: for some reason, I see only ONE cathode resistor on your amp. Where is the other one? On the other side of the board? Why? 680 ohms is probably too high for 6L6GC - I would think you want 560 - 510, maybe even 470 ohms ...
edit2: I would also take a close look at the three terminal connector on the side with no visible cathode resistor - it's probably OK, but looking at the picture, it could also be that both ends of the jumper wire are in the same terminal which is not OK; it would leave the screen floating and effect (affect?) the stage gain.
Last edited:
IIRC you stated in a different thread that the difference in gain followed the output tubes when they were switched. If that is the case changing every part in the amp will have little effect.
The SSE is a zero feedback design. This means that the gain is largely affected by the characteristics of the tubes themselves. It the amp used a lot of global feedback the gain of the amp is determined mostly by the feedback components, and not the tubes. If your other amp uses feedback, that would explain this behavior.
The SSE is a zero feedback design. This means that the gain is largely affected by the characteristics of the tubes themselves. It the amp used a lot of global feedback the gain of the amp is determined mostly by the feedback components, and not the tubes. If your other amp uses feedback, that would explain this behavior.
Determined that the B+ voltage is 473.4 volts. R17 =38.7V and R27=38.1V. That calculates to 56.3 mA/output tube (GE 6L6GC tubes). This gives me 24.49W per tube.
Next, I checked the 12AT7 tube. Guess what? One side of the 12AT7 tube is low! Popped in a NOS RCA 6201 tube with black plates.
Both channels now play warm, lush, smooth, silky music. Bingo, bango, duck soup. My family thinks I'm nuts, even my musician son. But, I can only play the stereo.
Also, I connected the output transformers UL connections and I am satisfied with the amp!!! No background noise at all. Staging and seperation is perfect. Dynamics are great. The sound is projected into the listening room. Volume is satisfyingly loud and I plan to add switches for UL and CFB.
This weekend, there will be a magic show where I will build the TSE-II board I bought from a friend who ordered two SPP boards and this TSE-II board from Tubelab. I bought some Electra Print OPTs and have some 45 tubes for it. I'm trying to build my retirement amps. I also discovered that the predecessor forum for the TSE-II contains all the information needed. Maybe I'll write a posting titled Tubelab for Dummies?
All the parts are on a spreadsheet someone put together. The OPTs suggested are right there. The only thing I have to do is create a CAD model of the chassis and make them on a CNC milling machine.
I'm not an EE (but know how to spell it), so Tubelab made this easy. Just gotta spend the time to read through his website and the DIY Audio posts.
I'm so glad that Tubelab extended his 5 year mission and has gone boldly where no tube has gone before!!
Happy Thanksgiving!
Next, I checked the 12AT7 tube. Guess what? One side of the 12AT7 tube is low! Popped in a NOS RCA 6201 tube with black plates.
Both channels now play warm, lush, smooth, silky music. Bingo, bango, duck soup. My family thinks I'm nuts, even my musician son. But, I can only play the stereo.
Also, I connected the output transformers UL connections and I am satisfied with the amp!!! No background noise at all. Staging and seperation is perfect. Dynamics are great. The sound is projected into the listening room. Volume is satisfyingly loud and I plan to add switches for UL and CFB.
This weekend, there will be a magic show where I will build the TSE-II board I bought from a friend who ordered two SPP boards and this TSE-II board from Tubelab. I bought some Electra Print OPTs and have some 45 tubes for it. I'm trying to build my retirement amps. I also discovered that the predecessor forum for the TSE-II contains all the information needed. Maybe I'll write a posting titled Tubelab for Dummies?
All the parts are on a spreadsheet someone put together. The OPTs suggested are right there. The only thing I have to do is create a CAD model of the chassis and make them on a CNC milling machine.
I'm not an EE (but know how to spell it), so Tubelab made this easy. Just gotta spend the time to read through his website and the DIY Audio posts.
I'm so glad that Tubelab extended his 5 year mission and has gone boldly where no tube has gone before!!
Happy Thanksgiving!
Last edited:
Yes sir, those are Electa Print OPTs and a Hammond PT. Jack sent the OPTs in less than a week.
Yeah, I need another amp, too. Just ask my former accountant wife.
I'm going to use a pair of CX-345 tubes, see if there's anything to the 45 tube fuss. Also building a AN 300B kit. I'm trying to build my retirement amps, planning to sell off the gear I don't use anymore.
Yeah, I need another amp, too. Just ask my former accountant wife.
I'm going to use a pair of CX-345 tubes, see if there's anything to the 45 tube fuss. Also building a AN 300B kit. I'm trying to build my retirement amps, planning to sell off the gear I don't use anymore.
Hah - I thought so; can't see much in that dim box, but the build quality looks pretty good from what I can see. At all events I've heard nothing but good things about them, and can't wait to hear your impressions.
The AN kit sounds intriguing too. The only tube amp kit I've built to date is an ST-70 clone (the Tubes4HiFi version). The remainder have all been scratch builds because - frankly - I love the challenge. I'll add it helps to have a full woodworking shop and access to metal lathes and milling machines!
And yes, I think there is something to the 45 fuss. I'll admit to harboring some pretty severe doubts for over 20 years before finally succumbing to the urge about five months ago and chunking a pair in my breadboarded "Borg" TSE-II (seen below with "2A300Bs"). Heck, I had it set up for 2.5 V filaments anyway, so all I had to do was crank the bias down to get ~30 mA plate current at ~9.5 W plate dissipation. It sounded incredible.
The AN kit sounds intriguing too. The only tube amp kit I've built to date is an ST-70 clone (the Tubes4HiFi version). The remainder have all been scratch builds because - frankly - I love the challenge. I'll add it helps to have a full woodworking shop and access to metal lathes and milling machines!
And yes, I think there is something to the 45 fuss. I'll admit to harboring some pretty severe doubts for over 20 years before finally succumbing to the urge about five months ago and chunking a pair in my breadboarded "Borg" TSE-II (seen below with "2A300Bs"). Heck, I had it set up for 2.5 V filaments anyway, so all I had to do was crank the bias down to get ~30 mA plate current at ~9.5 W plate dissipation. It sounded incredible.
Attachments
Mr+Zenith--- That sure is pretty! There are a lot of people who don't appreciate the craftsmanship that goes into your amp. People here do...
Jack at Electra Print said he likes point to point amps. I haven't done this long enough to have a real preference. I do appreciate Tubelab and the PCBs/information and sharing his experimentally derived data about tubes.
Brings up another question- I changed the 12AT7 yesterday and moved the SSE to another location in my listening room today. No sound from the left channel. I then changed the amp back to where it was originally, and noticed one side of the 12AT7 tube was glowing brighter on one side. Turned the amp off and plugged in another 12AT7 tube and everything is good again.
The tubes I have are all old, NOS. They were given to me by a neighbor who has since passed away. My tester is at a friends house and he's out of town until late Sunday. I intend to test both halves of the two 12AT7 tubes I've changed out on this amp.
It doesn't make sense that the heater voltage would suddenly spike on just the 12AT7 tube, nor that it would only affect one side. I'll also check the 6 volts at the socket.
Do these type driver tubes normally fail like this? Again, these are old NOS tubes. I'm over 60 now, and I've also got a lot of wear and tear.
Jack at Electra Print said he likes point to point amps. I haven't done this long enough to have a real preference. I do appreciate Tubelab and the PCBs/information and sharing his experimentally derived data about tubes.
Brings up another question- I changed the 12AT7 yesterday and moved the SSE to another location in my listening room today. No sound from the left channel. I then changed the amp back to where it was originally, and noticed one side of the 12AT7 tube was glowing brighter on one side. Turned the amp off and plugged in another 12AT7 tube and everything is good again.
The tubes I have are all old, NOS. They were given to me by a neighbor who has since passed away. My tester is at a friends house and he's out of town until late Sunday. I intend to test both halves of the two 12AT7 tubes I've changed out on this amp.
It doesn't make sense that the heater voltage would suddenly spike on just the 12AT7 tube, nor that it would only affect one side. I'll also check the 6 volts at the socket.
Do these type driver tubes normally fail like this? Again, these are old NOS tubes. I'm over 60 now, and I've also got a lot of wear and tear.
Thanks, Duke - that's a massive compliment considering that was my breadboarded version! The final build can be seen here (post 483 if your browser fails to take you to the correct one). It wound up being my exhibit at Burning Amp this year, and was very well received.
Regarding PCB vs. point-to-point, I've done both. A well-designed PC board like George's can be a wonderful thing, offering easy layout and a quick build. However, you're usually bound to conforming the final chassis to the board's layout - which isn't hard to do, but can be tedious. That said, point-to-point offers complete freedom in your layout options, and even a certain artistry can be achieved if done well (for instance, I like building my ground buses "1920's style" using no. 12 copper wire). It does take time though. Sonically, I can't discern any difference between the two.
Sorry to hear of the half-demise of your 12AT7. This isn't unheard of in the 12xx7-types when wired for 6 V operation. Like many members' tube caches, mine is primarily used tubes and there's no telling what sort of abuse they endured before I acquired them. And I highly suspect that some of the tubes I purchased as NOS "ain't" NOS. You pays your money, and you takes your chances. Glad you had another on hand though.
The final build can be seen here (post 483 if your browser fails to take you to the correct one). It wound up being my exhibit at Burning Amp this year, and was very well received. Regarding PCB vs. point-to-point, I've done both. A well-designed PC board like George's can be a wonderful thing, offering easy layout and a quick build. However, you're usually bound to conforming the final chassis to the board's layout - which isn't hard to do, but can be tedious. That said, point-to-point offers complete freedom in your layout options, and even a certain artistry can be achieved if done well (for instance, I like building my ground buses "1920's style" using no. 12 copper wire). It does take time though. Sonically, I can't discern any difference between the two.
Sorry to hear of the half-demise of your 12AT7. This isn't unheard of in the 12xx7-types when wired for 6 V operation. Like many members' tube caches, mine is primarily used tubes and there's no telling what sort of abuse they endured before I acquired them. And I highly suspect that some of the tubes I purchased as NOS "ain't" NOS. You pays your money, and you takes your chances. Glad you had another on hand though.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.