While studying the subject of SE amps, i see that most DHT amps use an interstage transfo instead of a coupling cap, and this is never the case with a se amp using indirect heated tubes. Is there a technical reason for that that escapes me, or is it just because of the higher costs of that transfo? DHT tubes are mostly more expensive than others, so that may be the case...
Probably partly nostalgia, but also DHT usually have quite low mu so need the extra boost from an interstage transformer. As soon as valves started having higher gain the transformer could be dropped. A coupling cap is cheaper, smaller, wider bandwidth, less distorting, less prone to hum pickup etc.
Well, i don't agree, i heared recently on a demonstration a custom DHT amp with a 26 trough a (Tango) interstage transfo to a 45 with again a Tango output transfo, and that beated all amps i heared before on sound (and i heared a lot). Sadly nobody makes those tubes anymore and they are rare and expensive. That guy also had a double 6N1P driver/EL84 SE also with tango transfo at the end and that, altough a wonderfull amp on it's own, did not came near in my opinion. Both were listened on the same speakers and in the same room with the same source.
But each to his own off course, this is a subjective opinion.
Last edited:
There is no technical reason. Amps using indirectly heated outputs generally tend to be more budget oriented, so this may be one reason.
Well, i don't agree, i heared recently on a demonstration a custom DHT amp with a 26 trough a (Tango) interstage transfo to a 45 with again a Tango output transfo, and that beated all amps i heared before on sound (and i heared a lot). Sadly nobody makes those tubes anymore and they are rare and expensive. That guy also had a double 6N1P driver/EL84 SE also with tango transfo at the end and that, altough a wonderfull amp on it's own, did not came near in my opinion. Both were listened on the same speakers and in the same room with the same source.
But each to his own off course, this is a subjective opinion.
Your disagreeing with facts. Just because you like the transformer sound dosnt mean it's more acurrate.
Ok, so a 6922-EL34 SE aiming for 5w with a good interstage transfo may work well. I want to design and build one like that when i'm sure enough of my knowlodge of tube amps (not yet) so that may be a route to go...
Who's saying i'm aiming for accuracy. I'm aiming for a sound i love.
And i believe that solid state amps in class A (like the First Watt designs) are way more accurate than any tube amp with transfo's that color, so ...
Before this gets into a religious war, consider…
From the dawn of the vacuum valve era until about the late 1930s, the only 'foils' a manufacturer had at her disposal to create rolled-tube capacitors was either copper or tin. There really was no cheap aluminum foils, especially ultra-thin types. As a consequence, given also there were absolutely no plastics (for ultra-thin dielectrics), creating a capacitor with even single-digits worth of microfarads was an expensive proposition.
That's "first".
Second, as noted above, the mu of valves wasn't very high in the "big bottle" era. They performed quite nicely all in all, but high gain wasn't expected of them. The tolerances weren't present in the manufacturing lines to get millimeter-or-under cathode-to-grid–1 spacing. However, building a bigger bottle was easy, since for most intents and purposes the envelope was created from repurposed vacuum standard incandescent light bulb glass-blowing equipment. Bigger was easy. And it turned out that there were thousands of cottage industries eager to make whatever transformer(s) hit the fancy of the valve amplifier designing bunch.
Moreover, you could get at least a factor of 2 gain with an interstage. Even factors of 3, 4 or 5 weren't unheard of. The speakers of the day were so pathetic, frequency-response wise (and the available source-material also so pathologically non-high-fidelity), that if there were interwinding capacitance issues that'd limit amplifier bandwidth, well … it largely didn't matter. Bigger bottles + interstage step-up transformers got the job done. Also, without very much complexity, you also "got for free" the phase-inverter function with an interstage transformer, for increasing output power via push-pull designs.
No losers there! Win, win, win.
But also as noted above, with the "taming of the bottle", the relatively rapid development of specific-purpose tube-making equipment, came smaller more durable, less finicky "bottles". Came indirect heated cathodes to totally tame A/C hum (and extend life of tubes.) Came aluminum foil wound capacitors (even before electrolytics); then wet-type electrolytic dielectric aluminum foil capacitors were discovered, and rapidly transformed (couldn't pass up the pun) a revolution in power-supply mass-and-ripple reduction without needing fork-lifts to carry around electronic equipment.
The designers, subsequent to the expensive-capacitor and lots-of-interstage-transformer era, embraced the much lighter weight, smaller, higher gain tubes, lower power requirements and especially the magnetic-flux balancing push-pull idea for output power. Because the interstage transformer so handily could deliver, if wound thoughtfully, nearly full-audio-spectrum bandwidth and a bit of voltage gain and for-free complimentary signal inversion, of all transformers they remained designed-in for a long time. Well into the 1950s.
But by the late 1950s and especially early 1960s, “bottles were cheap” and transformers had become expensive, relatively. Capacitors were terribly inexpensive all in all. And resistors were “nearly free”. Hence designs edged toward minimizing transformers and special bias power supplies, and simultaneously toward “cathode bias + electrolytic” automagic methods that'd serve the consumer well, even tho' the plethora of bottle makers were notably loathe to actually make same-numbered bottles “to actual spec”.
Self-adjusting “good enough” circuits rose to rule the scene.
Because 99% of the consumers couldn't identify a bias screw from a toggle switch.
Nor could they be counted on not to screw “it” up.
With today's electron-beam evaporated aluminum-on-ultra-thin-film capacitors, we have “arguably” near-perfect capacitor solutions, for prices that are — in the historical context — breathtakingly low. And while us old fuddy-duddies like our valves, we also have a whole constellation of semiconductor devices which can markedly assist the non-valve side of nominally tube-amplifier design. We're basking in saunas steamed with bourbon.
Just saying.
GoatGuy
From the dawn of the vacuum valve era until about the late 1930s, the only 'foils' a manufacturer had at her disposal to create rolled-tube capacitors was either copper or tin. There really was no cheap aluminum foils, especially ultra-thin types. As a consequence, given also there were absolutely no plastics (for ultra-thin dielectrics), creating a capacitor with even single-digits worth of microfarads was an expensive proposition.
That's "first".
Second, as noted above, the mu of valves wasn't very high in the "big bottle" era. They performed quite nicely all in all, but high gain wasn't expected of them. The tolerances weren't present in the manufacturing lines to get millimeter-or-under cathode-to-grid–1 spacing. However, building a bigger bottle was easy, since for most intents and purposes the envelope was created from repurposed vacuum standard incandescent light bulb glass-blowing equipment. Bigger was easy. And it turned out that there were thousands of cottage industries eager to make whatever transformer(s) hit the fancy of the valve amplifier designing bunch.
Moreover, you could get at least a factor of 2 gain with an interstage. Even factors of 3, 4 or 5 weren't unheard of. The speakers of the day were so pathetic, frequency-response wise (and the available source-material also so pathologically non-high-fidelity), that if there were interwinding capacitance issues that'd limit amplifier bandwidth, well … it largely didn't matter. Bigger bottles + interstage step-up transformers got the job done. Also, without very much complexity, you also "got for free" the phase-inverter function with an interstage transformer, for increasing output power via push-pull designs.
No losers there! Win, win, win.
But also as noted above, with the "taming of the bottle", the relatively rapid development of specific-purpose tube-making equipment, came smaller more durable, less finicky "bottles". Came indirect heated cathodes to totally tame A/C hum (and extend life of tubes.) Came aluminum foil wound capacitors (even before electrolytics); then wet-type electrolytic dielectric aluminum foil capacitors were discovered, and rapidly transformed (couldn't pass up the pun) a revolution in power-supply mass-and-ripple reduction without needing fork-lifts to carry around electronic equipment.
The designers, subsequent to the expensive-capacitor and lots-of-interstage-transformer era, embraced the much lighter weight, smaller, higher gain tubes, lower power requirements and especially the magnetic-flux balancing push-pull idea for output power. Because the interstage transformer so handily could deliver, if wound thoughtfully, nearly full-audio-spectrum bandwidth and a bit of voltage gain and for-free complimentary signal inversion, of all transformers they remained designed-in for a long time. Well into the 1950s.
But by the late 1950s and especially early 1960s, “bottles were cheap” and transformers had become expensive, relatively. Capacitors were terribly inexpensive all in all. And resistors were “nearly free”. Hence designs edged toward minimizing transformers and special bias power supplies, and simultaneously toward “cathode bias + electrolytic” automagic methods that'd serve the consumer well, even tho' the plethora of bottle makers were notably loathe to actually make same-numbered bottles “to actual spec”.
Self-adjusting “good enough” circuits rose to rule the scene.
Because 99% of the consumers couldn't identify a bias screw from a toggle switch.
Nor could they be counted on not to screw “it” up.
With today's electron-beam evaporated aluminum-on-ultra-thin-film capacitors, we have “arguably” near-perfect capacitor solutions, for prices that are — in the historical context — breathtakingly low. And while us old fuddy-duddies like our valves, we also have a whole constellation of semiconductor devices which can markedly assist the non-valve side of nominally tube-amplifier design. We're basking in saunas steamed with bourbon.
Just saying.
GoatGuy
I used to use Hammond 124B for driver/phase splitter and while I liked the sound of it, the scope revealed the HF transient response was terrible. 10kHz square looked like shark fins! I switched to cathodyne and the results are orders of magnitude better, but since music doesn't contain square wave (well, not often anyway) it's generally good enough. It's just all of the points that DF96 made and the added cost for a quality part that really drive a nail in the coffin for me. Still, the idea of making an amplifier that only used coils and tubes is pretty cool IMHO. Only for the "wow" factor though.
Always love reading those posts, GoatGuy. Very informative as usual.
Always love reading those posts, GoatGuy. Very informative as usual.
Interstage transformers have lots of advantages when driving output tubes. It has been discussed in numerous threads. As for phase-splitters...i have failed to find an equally good sounding solution using active splitters.
As for voltage gain - i think it is a generally bad idea to use an interstage for this.
As for voltage gain - i think it is a generally bad idea to use an interstage for this.
One technical reason why IT's are mostly found with DHT's is that these are generally the most linear devices and work fine without feedback. Applying global feedback with two transformers in the loop can be quite challenging and sometimes impossible. Of course less linear devices could be used with local feedback and it always seems to be a waste for most people....not for me. I am just building such an amp!
There are indirect types (both triodes and triode-strapped pentodes/tetrodes) that show very good linearity but are only found in the low power range. The best low power types are very good drivers for sure (several are Russian types and are also cheap!) and they are often found in amplifiers. I am not aware of indirect type power triodes (more than 10-12 W plate dissipation) that are very linear except for the 6HS5 and its equivalents but this needs to work at least around 900-1000V anode voltage to get some useful power and the output transformer for it is a challenge. There might have been some medium-high power ones in the past but to me they are extinct. Maybe there are some made for non audio applications but don't seem to be common...
P.S.
I do not recommend interstage transformers with step-up, except for few cases of splitters like 1:1+1 designed for very specific applications. They usually bring more troubles than advantages. For me (moderate) step-up transformers have only place at the input of the amplifier (and driven by a suitable pre-amp or source).
There are indirect types (both triodes and triode-strapped pentodes/tetrodes) that show very good linearity but are only found in the low power range. The best low power types are very good drivers for sure (several are Russian types and are also cheap!) and they are often found in amplifiers. I am not aware of indirect type power triodes (more than 10-12 W plate dissipation) that are very linear except for the 6HS5 and its equivalents but this needs to work at least around 900-1000V anode voltage to get some useful power and the output transformer for it is a challenge. There might have been some medium-high power ones in the past but to me they are extinct. Maybe there are some made for non audio applications but don't seem to be common...
P.S.
I do not recommend interstage transformers with step-up, except for few cases of splitters like 1:1+1 designed for very specific applications. They usually bring more troubles than advantages. For me (moderate) step-up transformers have only place at the input of the amplifier (and driven by a suitable pre-amp or source).
Last edited:
Do you really think that the Hammond is a reference to draw general conclusions? They are honest transformers, some type are also quite good, but not really a reference. Cost is subjective.
I haven't tried that. You might need some compensation. I am using cathode feedback only with a special winding in the transformer.
Another solution for you might be: for the power stage you could just use the Schade fb between plate and grid of the power tube exactly as was presented in the original work. At the end of that article there is the 6L6 power stage driven by a SE driver + splitting transformer and could use the plate-to-grid fb also at the driver. The input stage might not be the usual one with tiny tubes and will have to work a bit harder but don't think is a problem if it has to swing moderate signal.
P.S.
Send me a pm with your specific application. We might see if you could use your IT in a more "creative" way.
Another solution for you might be: for the power stage you could just use the Schade fb between plate and grid of the power tube exactly as was presented in the original work. At the end of that article there is the 6L6 power stage driven by a SE driver + splitting transformer and could use the plate-to-grid fb also at the driver. The input stage might not be the usual one with tiny tubes and will have to work a bit harder but don't think is a problem if it has to swing moderate signal.
P.S.
Send me a pm with your specific application. We might see if you could use your IT in a more "creative" way.
Last edited:
Try to understand two types of interstage. one with magnetization current and other without. (Small amount of current with leakage curent of the C and grid curents of the tube are exist but it is very low...) Interstage transformer is not a driver, the part transforming source from prim to sec with square of transfer ratio, or square of ratio of turns (without losses...) In that transformation not the voltage or impedance going into transformation but capacitance too...
So with step up of say 1:4, the capacitance transforms from primaru 14 times. And that is causing potential damage at high end of BW. For good response at low end the inductance of primary should be high or matched with driver source. the internal resistance of tube. As more inductance we need (for the higher Ri tube driving TX, as much number of turns we have resulting in increase of capacitance at the primary and squared ratio multiplying to the secondary.
When we have magnetization currents trough the primary (without cancelation from the secondary or other extra layer for instance), the things are getting more complicated. So when You use interstage transformer, try to use 1:1 or slight step down, higher L(Hy) of primary, lower capacitance, lower leakage inductance, potent drive tube with lower Ri, and with as low of reactive load on secondary... So when ordering or considering interstsge transformers please note that is from vital importance to have more datas, like Lp(Hy), Ls, Rdc-primary, Rdc-secondary, Cp, Cs, Coupling factor, magnetic gap value and so. To determine the best driver tube for interstage. Unfortunately most of manufacturers are not giving that datas. Only transfer ratio is not enough. and My opp is that with proper interstage sound is much better And for Push pull spliting phase, interstage almost is a must, and good and proper interstage will perform better against any phase splitter.
So with step up of say 1:4, the capacitance transforms from primaru 14 times. And that is causing potential damage at high end of BW. For good response at low end the inductance of primary should be high or matched with driver source. the internal resistance of tube. As more inductance we need (for the higher Ri tube driving TX, as much number of turns we have resulting in increase of capacitance at the primary and squared ratio multiplying to the secondary.
When we have magnetization currents trough the primary (without cancelation from the secondary or other extra layer for instance), the things are getting more complicated. So when You use interstage transformer, try to use 1:1 or slight step down, higher L(Hy) of primary, lower capacitance, lower leakage inductance, potent drive tube with lower Ri, and with as low of reactive load on secondary... So when ordering or considering interstsge transformers please note that is from vital importance to have more datas, like Lp(Hy), Ls, Rdc-primary, Rdc-secondary, Cp, Cs, Coupling factor, magnetic gap value and so. To determine the best driver tube for interstage. Unfortunately most of manufacturers are not giving that datas. Only transfer ratio is not enough. and My opp is that with proper interstage sound is much better
And for Push pull spliting phase, interstage almost is a must, and good and proper interstage will perform better against any phase splitter.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.