Enjoy the benefits of a premium membership!
Upgrade Account
-
TalkBass has been independent since 1998. Add your voice.
Create a free account to reply to discussions, view embedded media, and browse with fewer display ads.
Join free Log in
Want zero display ads or expanded classifieds tools? Compare plans.
SVT-CL: Mixing Power Tubes
- Thread starter willeb
- Start date
okcrum
in your chest
If you're looking at the tubes in amp they're offset a bit so you can see how the trios are arranged like bowling pins.
Invalid Link Removed
No, you're confusing the total bias on each side of the teeter-totter with each pair of tubes that does the push-pull, or up-and-down, for each half of the wave form. You still must match each pair, so that the wave form is matched. That is the weight or gain of each pair of power tubes. Then once you have filled up each side with matched pairs, then you adjust the bias to get the optimum performance out of the group of tubes on each side of the push-pull.
If I have six tubes, and three have relatively higher gain characteristics in the tolerance, and three have relatively lesser gain, if I put the three high gain tubes on one side, in the "1" sockets as described, and the three lesser gain tubes in the "2" side, at the output I will have a distorted wave form, and no biasing can bring that completely back. If I raise the bias on the lesser gain side to compensate, all I will do is pour more current through the tubes and lessen their life as they try to do as much work as the three higher gain tubes on the other side to balance the wave form.
On the other hand, if I rearrange the tubes so I have the same number of high and low gain tubes on each side, matching each pair, say a high, not-so-high, and low in the three #1 sockets, and a corresponding set of high, not-so-high, and low in the three #2 sockets, then each side will work together, and process each half of the signal, at roughly the same overall rate of gain, or amplification. The output wave form will be symmetrical, and then bias can be used properly to accomplish the fine adjustment and balance the overall gain characteristics of the amp, and that's where all this discussion of "three - not two" comes in to play.
So for each #1 tube of a certain gain characteristic, there needs to be a "matching" #2 tube of a similar gain characteristic, so the bias can do its job. That's what I mean for working in pairs. Again, read page 29 of the RCA tube book to see the basic push-pull circuit. All an SVT does is to put three of these circuits together, in pentode mode instead of triode mode as set forth in the book. Here is the basic pentode circuit, with the extra grids:
http://www.enjoythemusic.com/magazine/sound_practices/2/triode_pentode.htm
It says EL34, but all pentodes, including 6550s, have the same basic pin-out and wiring; the difference is in the plate voltage supplied by the power transformer and the bias to the grids to match the gain characteristics of the particular tube.
Now, compare that push-pull circuit to the SVT schematic I posted earlier. It is the same thing, multiplied by three on each side. So there are three sets of two tubes, and all three on one side are tied together to a common bias, and the three on the other side are tied together to another common bias, just like the teeter-totter picture I posted. They work in pairs, one of each pair on each side of the push-pull circuit, with multiple pairs for more power, and bias in groups of all the tubes on one side or the other to do the fine balance on the waveform.
Don't confuse what takes to bias an amplifier with the basic theory on how it works: push-pull with each tube of each pair of tubes working on one half of the wave form, and the number of pairs multiplied, in this case by three, to get more power out of the amp. If this is difficult to understand, then we need to discuss an even more fundamental review of how a waveform, with its positive and negative halves, works to make a complete waveform our ears perceive as sound, and why it is more efficient, and why you can get more power, to have a separate power tube take care of each half of the wave form, rather than Class A operation where one tube takes care of both halves of the waveform.
Here is the fundamental problem in this discussion: some time ago, I had a similar private discussion with another forumite, who has an electrical engineering degree. He confessed that tube technology is not taught in engineering schools anymore, so the basic concepts of how a push-pull tube circuit works are not taught. This leads to confusion on why the tubes must be matched in pairs, one on each side of the circuit, and then why you can tie the common bias for the tubes on each side of the circuit together and only have two bias controls.
Again, read the book.
Last edited:
Each "pair" in a six tube A/B, A/B1 can not be separated from each other unless they are gain stages feeding the next section (you wind up with a single pair at the load). If their product is additive not progressive the "three" on the "push" or "pull" is a product of the three, not individual as you are suggesting. They are six tubes, no longer in pairs.
The biggest trouble I see with your position is you are looking at a circuit with a pair of tubes and forgetting that it is not a circuit without including the output transformer. To triple that circuit requires tripling the output transforms or primary winds in a single transformer.
The biggest trouble I see with your position is you are looking at a circuit with a pair of tubes and forgetting that it is not a circuit without including the output transformer. To triple that circuit requires tripling the output transforms or primary winds in a single transformer.
Last edited:
No, it doesn't. I am not "forgetting" anything. The power transformer is made sturdy enough to feed all six tubes, three tied together on one side of the push-pull, and three on the other side. You still match the six tubes a pair at a time to make sure you don't overload one side of the amp or the other by putting all the highest gain tubes on one side or the other, but match them across each side of the amp, in pairs.
http://www.duncanamps.com/technical/valvematch.html
http://www.premierguitar.com/articles/tube-matching-ndash-needed-or-not-1
Invalid Link Removed
So you're suggesting that balancing them in pairs ensures the triplets on each side are balanced. And that is pretty much exactly what everyone else is saying. The net on each side needs to be balanced. You are simply suggesting a means to achieve that.
Here are how the two trios are grouped in a CL or VR. Each group of three are tied together essentially in parallel.
Last edited:
Doesn't matter "how sturdy" the OT is for this discussion. You need six matching tubes, not three different matched pairs. You can compensate for one triplet against the other but not within the "push" or "pull" sides when using a single different tube. That would be like trying to balance your past wages for your current expenditures.
I think you understand what I am saying. Yes, the net on each side needs to be balanced.
The reason for the discussion about pairs started with the OP saying he blew one tube. Since the tubes had age on them, you replace the pair, not just one tube. Then the discussion happened from there.
When to replace a single tube, and when to replace the pair:
http://www.audiocircle.com/index.php?topic=89231.0
You have it backwards. If you try to compensate three on one side against three on another, you have to over-bias one side and you kill tubes. Read the threads in the links and read the book. The only reason I mentioned temporarily installing an odd tube for the one that blew was to finish a gig with a lesser chance of damage to the amp than leaving the blown tube in. Then when the gig is done, you pull the temporary and its mate from the other side and replace both with a matched pair. Read the book and read the linked threads.
If you purchase a matched pair, then, to use your wage analogy, you are rebalancing income with expenditures with each pair of tubes, the + and - equal out. There may be slightly more or less overall income and expense, but the ledger balances, and that is the desired goal. If you don't purchase matched pairs, and have unbalanced tubes, then you either have too much income or too much expense from one side of the push-pull to the other, and the ledger is out of balance.
More links:
Invalid Link Removed
Why sequential matched pairs are better than a matched quad or sextet:
http://www.ax84.com/bbs/dm.php?thread=342990 and the quote out of the thread:
"Theoretically, 2 matched pairs should be as good as a matched quad. Even better - because one might match a quad so that only each "PP side" is matched to the other and the tubes are not interchangeable. So we could view 2 matched pairs as a "interchangeably matched quad.""
This particular thread then goes off on a tangent of using two dissimilar tubes on each side of the push-pull, but the quote is valid: when purchasing tubes, purchase sequential matched pairs so if one goes, you only replace the pair, not the entire six, which the remaining four are still good.
Obviously, if you want the biggest, baddest, loudest that the amp can deliver, purchase a matched sextet every time. But most people don't have that kind of cash. I didn't when I was younger, and that's why I sold the SVT head I used to own.
To get the best average, as economical as possible performance out of the amp, purchase the tubes in matched pairs, and when a tube on one side blows, replace it and its mate on the other side of the push-pull with a matched pair, tweak the bias if necessary, keeping the other four good tubes running, and rock on.
Last edited:
I was the OP.
This appears most people here are in violent agreement with you. The disconnect seems to be semantic.
And for what it's worth, I replaced one on each side and pulled the original one to us in a pinch.
Sorry I just skimmed through, but didn't see anything directly pertaining output tube circuits employing more than a single pair (directly).
Yes. Great picture. 123 is one side of the push pull, and 456 is the other side. So, if #1 blew, you would pull it and #4, replace those two tubes with a matched pair, tweak the bias if necessary, and rock on, since the remaining four are still good and matched across to each other. And so forth with #2 and #5 as a pair; #3 and #6 as a pair.
Last edited:
Keep reading, especially the later threads I linked that do have more to do with matched quads and why. The book and the first links do focus on the pairs, the fundamental theory of a push-pull circuit, and why it is necessary to balance each pair. Again, an SVT power section is nothing more than three pairs with the "push" side of three all tied together and the "pull" side of three tied together to a single bias control and output transformer lead. To keep each side balanced, match each pair ongoing: 1 & 4; 2 & 5; 3 & 6.
Last edited:
I may read to see if anything sways my views from the Doctorates that taught me when this was still current technology.
Sorry did find anything other than occasional bad information (may be to "simplify") and personal conjecture. I'm starting to think this is more an OCD issue than a splitting hairs issue? The "explanation" of bias as it applies to crossover (notch) distortion was especially grossly inadequate.
Well, I am not perfect, but I know you need to get the pairs together.
http://www.talkbass.com/threads/is-a-matched-sextet-for-svt-necessary.531377/
Bassman Paul sums it up: "Matching in pairs, as long as one tube from each of the pairs is on opposite sides of the OPT balances the current on each side of the OPT, BUT one pair will do more work that the next." This is exactly the point I am making: matching pairs.
1) This whole thread is what to do when one tube blows to get by until new tubes can be purchased.
2) Push-pull amplifiers work with pairs of tubes tied together, one pushes, the other pulls. To get more power, add pairs like is done on an SVT.
3) Balancing the current draw is what we are trying to achieve
4) Yes, replacing all tubes at once is optimal, but not usually economically possible, so a pair at a time replaced, knowing that the pairs may load slightly differently.
Now, I'm done. I've given all the resources I can find about push-pull amplifiers working in pairs, one push, one pull. All the push tubes are tied together to a common bias and all the pull tubes are tied together for a common bias.
Everyone on this forum needs to have a fundamental course on how push-pull amplifiers work, if they are going to continue to discuss it. I did all those years ago, so if I forget a detail here and there, so be it. The fundamentals are as they are.
When your SVT tubes blow up because you overloaded one side with the higher gain tubes, and did not balance them as three pairs, remember I warned you.
http://www.talkbass.com/threads/is-a-matched-sextet-for-svt-necessary.531377/
Bassman Paul sums it up: "Matching in pairs, as long as one tube from each of the pairs is on opposite sides of the OPT balances the current on each side of the OPT, BUT one pair will do more work that the next." This is exactly the point I am making: matching pairs.
1) This whole thread is what to do when one tube blows to get by until new tubes can be purchased.
2) Push-pull amplifiers work with pairs of tubes tied together, one pushes, the other pulls. To get more power, add pairs like is done on an SVT.
3) Balancing the current draw is what we are trying to achieve
4) Yes, replacing all tubes at once is optimal, but not usually economically possible, so a pair at a time replaced, knowing that the pairs may load slightly differently.
Now, I'm done. I've given all the resources I can find about push-pull amplifiers working in pairs, one push, one pull. All the push tubes are tied together to a common bias and all the pull tubes are tied together for a common bias.
Everyone on this forum needs to have a fundamental course on how push-pull amplifiers work, if they are going to continue to discuss it. I did all those years ago, so if I forget a detail here and there, so be it. The fundamentals are as they are.
When your SVT tubes blow up because you overloaded one side with the higher gain tubes, and did not balance them as three pairs, remember I warned you.
This thread seems to be getting out of hand. For the record in any of the tube SVTs there is no such thing as a pair. bias is adjusted, in most, by adjusting one side and then the other. If a single tube goes down you need to match a tube to the others on that side and adjust the bias appropriately. If you cannot match then you NEED to replace the triple on that side and redo the bias to suit.
Also for the record I went to engineer school when tubes were current technology and transistors were a science that would never last!!
Also for the record I went to engineer school when tubes were current technology and transistors were a science that would never last!!
