I'm in the process of building my second amplifier and trying to clear up some things that I didn't quite figure out the first time around. During my first project, I sized my transformers by copying from similar schematics. This time I'd like to go about this a little better informed.
Can anyone tell me general rules-of-thumb for sizing output transformers? Is there a calculation I can do? The rating on a Hammond 1645 is 30W, but I see it in TUT3 being used for a 100W Marshall clone. What gives?
Can anyone tell me general rules-of-thumb for sizing output transformers? Is there a calculation I can do? The rating on a Hammond 1645 is 30W, but I see it in TUT3 being used for a 100W Marshall clone. What gives?
Power limit depends on resistance and, for low frequencies, saturation. You can get 30W at 20Hz or 120W at 40Hz from the same transformer, as far as saturation limits go.
Resistance means it'll heat up more if you run too much power through it for long periods, probably not an issue with a hi-fi listener.
If in doubt, find one rated for what your tubes will put out...
Tim
Resistance means it'll heat up more if you run too much power through it for long periods, probably not an issue with a hi-fi listener.
If in doubt, find one rated for what your tubes will put out...
Tim
Are you referring to the load resistance of the tubes, resistance of the primary windings, or neither?
I read one article that suggested matching the primary impedence with the impedence of the tube but (excusing my lack of knowlege) when I look at the tube data sheets I don't see impedence listed. Am I barking up the wrong tree?
Also, are there any resources you would recomend for getting a clearer understanding of the relationship between tubes and transformers?
Thanks,
Drew
I read one article that suggested matching the primary impedence with the impedence of the tube but (excusing my lack of knowlege) when I look at the tube data sheets I don't see impedence listed. Am I barking up the wrong tree?
Also, are there any resources you would recomend for getting a clearer understanding of the relationship between tubes and transformers?
Thanks,
Drew
oversimplification but...
I have been told to take your plate resistance of the output tube, say a triode with plate resistance of 1000 ohms and multiply that by a factor of between 2 to 4 times. So,
1000 x 2 = 2000 primary impedance of transformer (SE)
1000 x 4= 4000 " " "
the larger the factor, the lower the distortion or said another way, you trade power for distortion and vice versa. You get more power using the 2000 impedance but higher distortion.
Play with some loadlines at varying impedances for any given tube. Of course, these examples are for single ended operation. For push pull, well that's another story....
Hope this helps,
Rick
I have been told to take your plate resistance of the output tube, say a triode with plate resistance of 1000 ohms and multiply that by a factor of between 2 to 4 times. So,
1000 x 2 = 2000 primary impedance of transformer (SE)
1000 x 4= 4000 " " "
the larger the factor, the lower the distortion or said another way, you trade power for distortion and vice versa. You get more power using the 2000 impedance but higher distortion.
Play with some loadlines at varying impedances for any given tube. Of course, these examples are for single ended operation. For push pull, well that's another story....
Hope this helps,
Rick
For triodes, best power comes at 3 * Rp.
For pentodes, it's something like (Vb - Vsat) / (Isat / 2), with Ib = Isat / 2 (Vb = B+ supply; Vsat = 20 to 300V or more, depends on type; Isat = saturation current (current level on Vg = 0V plate curve at Vsat) and Ib = bias current). 'Least that's how I calculate it.
Specifics: resistance and impedance are different things. Resistance of a transformer typically means the DCR of the winding, which might be on the order of 200 ohms for the primary (obviously it depends on the transformer). This factors in to efficiency, but has nothing to do with impedance matching to the plate circuit. The *impedance* is what the tube is interested in, and is what the transformer is designed for. Which is also only a guideline: due to the way it works, the actual figure depends on the load connected. For instance, you might have a 5k:8 ohm transformer. With a 12 ohm load connected, the primary looks like 7.5k. Primary and secondary are thus connected by the impedance ratio. This is also why you have impedance taps for different speakers.
Tim
For pentodes, it's something like (Vb - Vsat) / (Isat / 2), with Ib = Isat / 2 (Vb = B+ supply; Vsat = 20 to 300V or more, depends on type; Isat = saturation current (current level on Vg = 0V plate curve at Vsat) and Ib = bias current). 'Least that's how I calculate it.
Specifics: resistance and impedance are different things. Resistance of a transformer typically means the DCR of the winding, which might be on the order of 200 ohms for the primary (obviously it depends on the transformer). This factors in to efficiency, but has nothing to do with impedance matching to the plate circuit. The *impedance* is what the tube is interested in, and is what the transformer is designed for. Which is also only a guideline: due to the way it works, the actual figure depends on the load connected. For instance, you might have a 5k:8 ohm transformer. With a 12 ohm load connected, the primary looks like 7.5k. Primary and secondary are thus connected by the impedance ratio. This is also why you have impedance taps for different speakers.
Tim
Some manufacturers don't reccomend different secondary taps.You have the best results only when they're parallel!
I order my own with only one tap at 6 ohm.
Notice that most Hi-Fi speakers rated at 8ohm are normally btw 5 to 7 ohm.The same happens with speakers rated at 4 ohm.
(At the first purpose is normally one 8R woofer and one 8R tweeter and at the second purpose two parallel 8R woofers with one 8R tweeter).
Menno van der Veen (Plitron)has done an experiment with different speakers and he builds his OPTs at 5ohm.
GKlab,where I bought my OPTs, recommends at 6ohm.
And,you can connect at this tap,speakers that are rated 4R or 8R!
Without any major differences.
The core of an OPT doesn't change a lot if it's 20W P-P or 40W P-P.It's the turns ratio that changes and the diameter of the wires used.
Some manufacturers have some articles on they're sites about this topic.I think Lundahll but I don't remember,well.
I order my own with only one tap at 6 ohm.
Notice that most Hi-Fi speakers rated at 8ohm are normally btw 5 to 7 ohm.The same happens with speakers rated at 4 ohm.
(At the first purpose is normally one 8R woofer and one 8R tweeter and at the second purpose two parallel 8R woofers with one 8R tweeter).
Menno van der Veen (Plitron)has done an experiment with different speakers and he builds his OPTs at 5ohm.
GKlab,where I bought my OPTs, recommends at 6ohm.
And,you can connect at this tap,speakers that are rated 4R or 8R!
Without any major differences.
The core of an OPT doesn't change a lot if it's 20W P-P or 40W P-P.It's the turns ratio that changes and the diameter of the wires used.
This is a very complex problem 'cause the relation btw tube or tubes and OPT depends from speakers,too.
Some manufacturers have some articles on they're sites about this topic.I think Lundahll but I don't remember,well.
> Can anyone tell me general rules-of-thumb for sizing output transformers?
For hi-fi: output iron bigger than power iron.
Guitar amps often use smaller transformers than hi-fi. Several reasons. For one, guitarists don't look at distortion numbers. Also, they don't play the "bigger=better" game that hi-fi geeks do. Also remember that guitar only goes to 80Hz, so any 20Hz spec is pointless. (Bass goes to 42Hz, but Fender-bass speakers usually don't.) For another, guitarists even like an amp that changes sound at high outputs. And most practically: guitar amps have to be lugged from gig to gig, so they want to be lighter than hi-fi amps!
> The rating on a Hammond 1645 is 30W, but I see it in TUT3 being used for a 100W Marshall clone. What gives?
The 30W rating is fairly hi-fi. Feed it 100W at 30Hz, it will be pretty bent. 100W at 80Hz will have just a slight edge. 100W at 42Hz won't come out "perfect", but that is good because it throws harmonics up above 50-60Hz, the cut-off of a bass speaker with high efficiency in a cabinet one person can handle.
For instrument amps, the best way to size the tranny is to imitate an existing design you like.
Load impedance:
For an ideal triode, ignoring ratings, maximum power is 2*Rp. But at higher impedance, distortion falls faster than power, so 2X to 5X Rp is common. And power falls so slowly in this range that it isn't real important whether you think an "8Ω" speaker is 5 or 6 or 10 ohms. (It will be around 50 ohms in the bottom of its range: speaker "matching" is all approximate.)
In a real triode, you often run out of plate dissipation before you hit rated plate voltage, so loads much higher than 2*Rp give maximum power.
For pentodes: Sch3mat1c gave the theory for a single pentode. Similar considerations solve the push-pull pentode, except Ib can be less than half of Isat (Class AB). But Isat depends on Vg2, which may be adjusted in design, trading higher Isat against lower grid drive voltage.
In general, you should start with the tube data-sheet suggestions. These show the tube in a good light, without exceeding ratings, a good starting place. You will discover that guitar-amps often push tubes far beyond the recommended conditions: the book values assume long life in fixed installations, guitar-amps work only a few hours a day and hopefully for-pay, so are willing to trade tube-life against more power. So don't be afraid to go somewhat beyond the book-rating, but don't go as far as Fender did in the 1960s when they could get special-spec versions of tubes. Some of the things Fender did to the poor 6V6, no Chinese 6V6 will tolerate for long.
> when I look at the tube data sheets I don't see impedence listed.
hmmmm... no in-line images here? marked-up 6L6 data sheet showing load impedance
For hi-fi: output iron bigger than power iron.
Guitar amps often use smaller transformers than hi-fi. Several reasons. For one, guitarists don't look at distortion numbers. Also, they don't play the "bigger=better" game that hi-fi geeks do. Also remember that guitar only goes to 80Hz, so any 20Hz spec is pointless. (Bass goes to 42Hz, but Fender-bass speakers usually don't.) For another, guitarists even like an amp that changes sound at high outputs. And most practically: guitar amps have to be lugged from gig to gig, so they want to be lighter than hi-fi amps!
> The rating on a Hammond 1645 is 30W, but I see it in TUT3 being used for a 100W Marshall clone. What gives?
The 30W rating is fairly hi-fi. Feed it 100W at 30Hz, it will be pretty bent. 100W at 80Hz will have just a slight edge. 100W at 42Hz won't come out "perfect", but that is good because it throws harmonics up above 50-60Hz, the cut-off of a bass speaker with high efficiency in a cabinet one person can handle.
For instrument amps, the best way to size the tranny is to imitate an existing design you like.
Load impedance:
For an ideal triode, ignoring ratings, maximum power is 2*Rp. But at higher impedance, distortion falls faster than power, so 2X to 5X Rp is common. And power falls so slowly in this range that it isn't real important whether you think an "8Ω" speaker is 5 or 6 or 10 ohms. (It will be around 50 ohms in the bottom of its range: speaker "matching" is all approximate.)
In a real triode, you often run out of plate dissipation before you hit rated plate voltage, so loads much higher than 2*Rp give maximum power.
For pentodes: Sch3mat1c gave the theory for a single pentode. Similar considerations solve the push-pull pentode, except Ib can be less than half of Isat (Class AB). But Isat depends on Vg2, which may be adjusted in design, trading higher Isat against lower grid drive voltage.
In general, you should start with the tube data-sheet suggestions. These show the tube in a good light, without exceeding ratings, a good starting place. You will discover that guitar-amps often push tubes far beyond the recommended conditions: the book values assume long life in fixed installations, guitar-amps work only a few hours a day and hopefully for-pay, so are willing to trade tube-life against more power. So don't be afraid to go somewhat beyond the book-rating, but don't go as far as Fender did in the 1960s when they could get special-spec versions of tubes. Some of the things Fender did to the poor 6V6, no Chinese 6V6 will tolerate for long.
> when I look at the tube data sheets I don't see impedence listed.
hmmmm... no in-line images here? marked-up 6L6 data sheet showing load impedance
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.