I find the question relevant, as the availability of capacitors are in the 400VDC range, not 450 or 500 VDC where the values are much less per volume size (sic!). I work a lot with maximizing output of valve amps, and used a lot of HIP5600 in the past to control B+ etc. It is always about space, heat (transfer) and optimizing El34s, as it brings the supressor grid on a separate pin instead of connecting to the cathode.
Hi HowardJohnstone,
I don't know what part of the world you are in, but 450 VDC is a very common rating in axial and radial caps. Some 500 VDC types and you can get 600V capacitors. CE Manufacturing produces several in the 475, 500 and 525 V ratings. Multisection caps that are very space efficient on a chassis.
For a tube, most run at 400VDC + at higher power levels. An EL34 likes higher voltage and lower current, I much prefer the 6CA7 which is about equivalent (you can plug them right in). Because a 6CA7 is a beam power tube, it is more efficient and has lower plate resistance. All pluses. The screen dissipates less power because it is in a "shadow" from the control grid.
I service and design tube power amps as well as solid state. All I'm going to say is that tubes are fun to play with, a different flavour.
-Chris
I don't know what part of the world you are in, but 450 VDC is a very common rating in axial and radial caps. Some 500 VDC types and you can get 600V capacitors. CE Manufacturing produces several in the 475, 500 and 525 V ratings. Multisection caps that are very space efficient on a chassis.
For a tube, most run at 400VDC + at higher power levels. An EL34 likes higher voltage and lower current, I much prefer the 6CA7 which is about equivalent (you can plug them right in). Because a 6CA7 is a beam power tube, it is more efficient and has lower plate resistance. All pluses. The screen dissipates less power because it is in a "shadow" from the control grid.
I service and design tube power amps as well as solid state. All I'm going to say is that tubes are fun to play with, a different flavour.
-Chris
Would you say individuals working with the big DHTs like GM70, 211 etc are more on the rare side?
Well - yeah!
Honestly, working with early generation tubes in single ended mode doesn't make any sense at all. Yes, it has a "sound", that's distortion and you could get the same juice with P-P 6BQ5 and a distortion pedal for "that sound". These tubes are horrible expensive, low efficiency and low performance. But I guess some folks out there like steam engines too.
In the 1920's, they figured out that P-P had lower distortion and higher power using the same tubes. After that the only place you found single ended was in cheap radios and phonographs (I had a couple as a kid). Directly heated filaments have so many disadvantages it isn't funny. It was the only way they knew how to do it when those tubes were designed. For the later RF versions, the requirements were different and they were not designed for audio at all. Indirectly heated cathode surfaces were a massive step forward.
Anyway, when people get excited about DHT single ended stuff, all you can do is shake your head. But it's fun. If you aren't serious (read realistic), go for it! If you want the best sound ... try class "D" and an effects pedal. Much cheaper and probably lower distortion (and I don't like class "D").
Just my opinion, but I apprenticed on this stuff and serviced enough of it through the years.
Honestly, working with early generation tubes in single ended mode doesn't make any sense at all. Yes, it has a "sound", that's distortion and you could get the same juice with P-P 6BQ5 and a distortion pedal for "that sound". These tubes are horrible expensive, low efficiency and low performance. But I guess some folks out there like steam engines too.
In the 1920's, they figured out that P-P had lower distortion and higher power using the same tubes. After that the only place you found single ended was in cheap radios and phonographs (I had a couple as a kid). Directly heated filaments have so many disadvantages it isn't funny. It was the only way they knew how to do it when those tubes were designed. For the later RF versions, the requirements were different and they were not designed for audio at all. Indirectly heated cathode surfaces were a massive step forward.
Anyway, when people get excited about DHT single ended stuff, all you can do is shake your head. But it's fun. If you aren't serious (read realistic), go for it! If you want the best sound ... try class "D" and an effects pedal. Much cheaper and probably lower distortion (and I don't like class "D").
Just my opinion, but I apprenticed on this stuff and serviced enough of it through the years.
i hear ya. i do like hearing the old guard speak of their experiences; keeps me from going down unneeded rabbit holes
Well, if you're doing and exploring for the fun of it - great! Experience is a great thing. But people sell this stuff for many thousands on the premise of "purity" and the way the music should be heard. That's not okay.
Common sense is your best friend. In this case, look at historical equipment production from the early days on up to the late 1970's. Through that time period, more expensive and better performing equipment was designed a certain way, cheaper stuff another (think tube table radio). In the 1980's advertising started to lie more and more blatantly and it was mostly built on audio fashion. Think of more expensive, heavy Marantz, Sansui and McIntosh along with other great brands. Then in the 80's, everyone lost their way. Today some have it together. Other stuff is contract manufactured garbage and a story, or low volume production not designed well - but it has a story!
Common sense is your best friend. In this case, look at historical equipment production from the early days on up to the late 1970's. Through that time period, more expensive and better performing equipment was designed a certain way, cheaper stuff another (think tube table radio). In the 1980's advertising started to lie more and more blatantly and it was mostly built on audio fashion. Think of more expensive, heavy Marantz, Sansui and McIntosh along with other great brands. Then in the 80's, everyone lost their way. Today some have it together. Other stuff is contract manufactured garbage and a story, or low volume production not designed well - but it has a story!
The 400 and 450V caps are the “regular” ones, without any special treatments or manufacturing technique. They are high volume because of their utility in the primary sides of SMPS’s therefore the performance per dollar ratio tends to be high. They simply have to be competitive. That’s perhaps the best reason to use 450V caps if you can. Sure, you can get good 500-600V caps but you will pay a premium for them, because people using them are playing with tubes. Expensive tubes.
The main reason for me doing anything with tubes is the fun. My best, and at this point biggest, tube amplifier doesn’t “sound” like a tube amplifier at all. Not that I was disappointed, but I was shocked more than anything. I figured I’d build the biggest practical tube amp with common parts, even if they are on the expensive side. I designed it using the same basic principles I would for building one of my rave PA amplifiers, and I GOT THE SAME RESULT. I was expecting a “bigger“ version of this little 6BQ5 push pull using small open frame transformers I had once, or maybe as good as that old Fisher 400, and going to use it for bass guitar. Ended up being too good for that so I revamped it into a 19“ rack form factor and built another.
B+ caps were 4x470 uF/450V Nichicon snap ins, in series-parallel. And I bought them from Tanner’s before he retired and closed shop - probably sitting on a shelf for a decade or more before the amps were built. Same ones were in the low pass section of the crossover for the speakers used with these beasts - a dual 15” 3-way built many years ago, gigged with, and later decommissioned for a 2x12/2” pair. They were good caps.
The main reason for me doing anything with tubes is the fun. My best, and at this point biggest, tube amplifier doesn’t “sound” like a tube amplifier at all. Not that I was disappointed, but I was shocked more than anything. I figured I’d build the biggest practical tube amp with common parts, even if they are on the expensive side. I designed it using the same basic principles I would for building one of my rave PA amplifiers, and I GOT THE SAME RESULT. I was expecting a “bigger“ version of this little 6BQ5 push pull using small open frame transformers I had once, or maybe as good as that old Fisher 400, and going to use it for bass guitar. Ended up being too good for that so I revamped it into a 19“ rack form factor and built another.
B+ caps were 4x470 uF/450V Nichicon snap ins, in series-parallel. And I bought them from Tanner’s before he retired and closed shop - probably sitting on a shelf for a decade or more before the amps were built. Same ones were in the low pass section of the crossover for the speakers used with these beasts - a dual 15” 3-way built many years ago, gigged with, and later decommissioned for a 2x12/2” pair. They were good caps.
Hi wg-ski,
For sure. Better transformers, better circuits and better tubes = more accurate sound. Nothing wrong with E-I frame transformers, they just don't have end bells. However the better transformers are made with thinner laminations of much better steel. The layering (interleave) is also controlled better (and designed better!). This makes a massive difference, good output transformers are well worth the extra cost. You can also use more feedback with those.
Many of those 450 V caps were carryovers from the earlier days before SMPS supplies. We already had them being made for industrial supplies and the designs from the "old tube days". The technology was repurposed and updated.
Yep, those 600 V caps used to be axial, and they were expensive for sure. That's why you see series connected capacitors. An old high power amp trick, like Hiwatt amps using KT88 banks, and almost anything else to save money. Very practical.
For sure. Better transformers, better circuits and better tubes = more accurate sound. Nothing wrong with E-I frame transformers, they just don't have end bells. However the better transformers are made with thinner laminations of much better steel. The layering (interleave) is also controlled better (and designed better!). This makes a massive difference, good output transformers are well worth the extra cost. You can also use more feedback with those.
Many of those 450 V caps were carryovers from the earlier days before SMPS supplies. We already had them being made for industrial supplies and the designs from the "old tube days". The technology was repurposed and updated.
Yep, those 600 V caps used to be axial, and they were expensive for sure. That's why you see series connected capacitors. An old high power amp trick, like Hiwatt amps using KT88 banks, and almost anything else to save money. Very practical.
I designed and built a number of GM70 SE amplifiers operating at plate voltages of 1 - 1.2kV and achieved as much as 30W with very benign distortion characteristics without global feedback. (2% thd at 25W) I retired my pair and sold them to a friend because of the problem with heat, and power consumption, but they served well for 13 years. The amps that replaced them amortized based on power savings in about one year.
I run a combination of class D for subs and woofers and class A SE dht (EML 300B) at 450V for a measly 6.5W per channel x4 to run my horns.
The GM70 and 845 are extremely linear as are the 572 and its derivatives, but low mu power triodes are hard to drive.
I have worked on amps ranging from 200V B+ to +2700V, most commonly below 500V if I didn't design it. 🙂
I run a combination of class D for subs and woofers and class A SE dht (EML 300B) at 450V for a measly 6.5W per channel x4 to run my horns.
The GM70 and 845 are extremely linear as are the 572 and its derivatives, but low mu power triodes are hard to drive.
I have worked on amps ranging from 200V B+ to +2700V, most commonly below 500V if I didn't design it. 🙂
lol!
Just think Kevin. 30 watts with maybe 430 VDC with lower distortion and a much lower retube bill. You can do that with 7591A (pushing them), 6L6GC easy or 6CA7. Those would all last ages. Power consumption? Much more reasonable. They just don't look nearly as cool. Some manufacturers claimed 30 watts with 7189 outputs. Hmmm. Doubt it.
At 1.2 KV, electricity is actively searching for a path to ground. Components are a lot more expensive too. Costs mount rapidly as the voltage increases, then when things go wrong .... they really go wrong! Spectacularly sometimes. Energy goes up as the square of voltage doesn't it? I forget the exact multiplier.
By the way, I hated working on early CRT TVs. I refused. Anode of 25 KV to 40 KV. You don't have to touch anything, it comes to get you. B&W was maybe 12 KV. Once I watched an arc jump out to nail my arm, inches away. It was supposed to have been discharged by the tech.
Just think Kevin. 30 watts with maybe 430 VDC with lower distortion and a much lower retube bill. You can do that with 7591A (pushing them), 6L6GC easy or 6CA7. Those would all last ages. Power consumption? Much more reasonable. They just don't look nearly as cool. Some manufacturers claimed 30 watts with 7189 outputs. Hmmm. Doubt it.
At 1.2 KV, electricity is actively searching for a path to ground. Components are a lot more expensive too. Costs mount rapidly as the voltage increases, then when things go wrong .... they really go wrong! Spectacularly sometimes. Energy goes up as the square of voltage doesn't it? I forget the exact multiplier.
By the way, I hated working on early CRT TVs. I refused. Anode of 25 KV to 40 KV. You don't have to touch anything, it comes to get you. B&W was maybe 12 KV. Once I watched an arc jump out to nail my arm, inches away. It was supposed to have been discharged by the tech.
I wasn't going to use tubes either until a 6922 absolutely destroyed an opa2134 for sound quality driving a mosfet power amp.
12ax7 and 6922 work very well on 100 to 150v B+ supplies. Easy to get from 50w ss amp secondary with a cap voltage tripler.
12ax7 and 6922 work very well on 100 to 150v B+ supplies. Easy to get from 50w ss amp secondary with a cap voltage tripler.
The most ive played with was 200 volts and that was on some 6C33Cs and 6AS7G i got a crazy deal on some GM70s and have a box of 16. Been pondering if i want to make something w them but just havent had to time. i do like the idea of playing around in 200 - 500 volt range and make some OTL but have always been curious about the bigger DHT. I may experiment with them but not sure if i want to go down that rabbit hole. I live in Texas so not worried about heat haha
Hi stocktrader200,
If you mean you preferred the 6922 - okay. Any good op amp is far better for sound quality.
I was Canadian Factory warranty for Counterpoint. Their preamps had extremely low distortion (SA-x000 series over the 1000) and low noise. Other brands did as well, more complicated circuits. I have an SA-3000 and some classic tube preamps (Fishers, Eicos etc ...), I'll take a solid state or op amp based unit for my better systems thank you. I do like a nicely done tube pre, but they are not higher in sound quality.
If you mean you preferred the 6922 - okay. Any good op amp is far better for sound quality.
I was Canadian Factory warranty for Counterpoint. Their preamps had extremely low distortion (SA-x000 series over the 1000) and low noise. Other brands did as well, more complicated circuits. I have an SA-3000 and some classic tube preamps (Fishers, Eicos etc ...), I'll take a solid state or op amp based unit for my better systems thank you. I do like a nicely done tube pre, but they are not higher in sound quality.
OTL is a massively bad plan. Talk about the wrong technology for the job.
Do me a favor. Look up the peak current for a speaker with some phase angle (not resistive). Now look at the spec sheet for the tube. Right. How many in parallel do you want to use? That is why the output transformer was invented. Those old engineers were pretty darned smart! Now think about what happens when one of those paralleled tubes says goodbye to this reality.
Do me a favor. Look up the peak current for a speaker with some phase angle (not resistive). Now look at the spec sheet for the tube. Right. How many in parallel do you want to use? That is why the output transformer was invented. Those old engineers were pretty darned smart! Now think about what happens when one of those paralleled tubes says goodbye to this reality.
oh trust me, i know the drawbacks of an OTL. I do love the sound though!!! Pros and cons w all of these and that includes SS too. i'm still in the journey of seeing what i want to just focus on
Last edited:
A lot of industrial equipment used rectified 240 or 277. That takes a…. 450V cap. They used to use soup can CG types back then. Still have some of them. 1970’s date codes and I bet they’ll still work.
Actually, the maximum peak current for the loudspeaker IS drawn at zero (or damn close to it) phase angle. For a speaker driver with a high ratio of Qms to Qes, the maximum phase deviation CAN be quite high, and the impedance still below nominal. Just not at the absolute minimum. It will cause the current to go locally high when vg1<<0, so the tube can’t deliver its maximum current. Then there is distortion. In a solid state amp it will try to deliver maximum current while Vce is still high, and if your transistors aren’t up to the task…. Kabang.
Hi Blk Dynamite,
Had to service some. No, they really don't do well with low impedance. Mostly because a linear tube is a high impedance device. Excessive current peaks do nasty things to tube cathodes. That's before you see red plates because the average power may still be low.
You probably like the sound because they use higher feedback ratios. Transformers limit how much feedback you can utilize due to phase margin. So basically an OTL may have low distortion before it hits current peaks.
Solid state has far greater forgiveness for current spikes, and any decent SS amp uses current limiting to prevent damage. Then of course there is the even lower possible distortion at way higher power levels. Naw, I can't say they suffer from similar issues at all. Guess why OTL manufacturers aren't successful long term? It's all about performance to price ratio. Own one and eventually you'll cry "UNCLE!". Any high power tube amp suffers the same high maintenance cost issues, but a modest power rating on OTL is a high power rating in a tube amp using an output transformer using the same tubes. In other words, an output allows the same power output with far fewer tubes.
A CJ Premier One used 6 x KT88 per channel for 250 watts each (600V B+). Okay, I'll take a Marantz 500, Yamaha PC2002 or Bryston 4B cubed any day, all day long over that. Now imagine that in OTL.
Hi wg_ski,
Resistive the current is in phase with voltage, so peak current would be at Vpk. With high reactance you get substantial currents at zero, but the current peak can occur anywhere in relation to the voltage peak. With a pure reactance, your current peak would be at the zero crossing.
When designing solid state, BJT especially, that is what an SOA curve is all about. When C-E voltages are high, many high power, TO-204 transistors may only manage 1 ampere. I haven't looked recently because I'm used to old parts. Maybe my stuff is extremely robust now! lol! Anyway, many people are shocked when they actually look at the curves, a couple hundred volts and you may be in the vicinity of 100 mA or lower. As a result, I went to mosfets for HV regulators, those might survive a short.
Had to service some. No, they really don't do well with low impedance. Mostly because a linear tube is a high impedance device. Excessive current peaks do nasty things to tube cathodes. That's before you see red plates because the average power may still be low.
You probably like the sound because they use higher feedback ratios. Transformers limit how much feedback you can utilize due to phase margin. So basically an OTL may have low distortion before it hits current peaks.
Solid state has far greater forgiveness for current spikes, and any decent SS amp uses current limiting to prevent damage. Then of course there is the even lower possible distortion at way higher power levels. Naw, I can't say they suffer from similar issues at all. Guess why OTL manufacturers aren't successful long term? It's all about performance to price ratio. Own one and eventually you'll cry "UNCLE!". Any high power tube amp suffers the same high maintenance cost issues, but a modest power rating on OTL is a high power rating in a tube amp using an output transformer using the same tubes. In other words, an output allows the same power output with far fewer tubes.
A CJ Premier One used 6 x KT88 per channel for 250 watts each (600V B+). Okay, I'll take a Marantz 500, Yamaha PC2002 or Bryston 4B cubed any day, all day long over that. Now imagine that in OTL.
Hi wg_ski,
Resistive the current is in phase with voltage, so peak current would be at Vpk. With high reactance you get substantial currents at zero, but the current peak can occur anywhere in relation to the voltage peak. With a pure reactance, your current peak would be at the zero crossing.
When designing solid state, BJT especially, that is what an SOA curve is all about. When C-E voltages are high, many high power, TO-204 transistors may only manage 1 ampere. I haven't looked recently because I'm used to old parts. Maybe my stuff is extremely robust now! lol! Anyway, many people are shocked when they actually look at the curves, a couple hundred volts and you may be in the vicinity of 100 mA or lower. As a result, I went to mosfets for HV regulators, those might survive a short.
What I meant is that the minimum impedance occurs at zero phase. The maximum phase deviation always occurs somewhere other than minimum impedance. They physical current peak can be anywhere, even the zero crossing (or damn close to it). If the driver or speaker system has a minimum of 2.8 ohms, the phase angle will be zero there. When you hit the max phase deviation of say 60 degrees, the impedance will be above 2.8 ohms. But maybe still below 4, especially if Qms is high (high mechanical energy storage). The speaker will draw more current at zero phase (for a given voltage) than it does working into 60 degrees. The two points will occur at different frequencies. Especially in the solid state forum there is a common misconception that reactive loads can cause more current draw than Vpeak divided by Zmin, but that is NOT physically possible. The amplifier may indeed have difficulty delivering the current demanded by that reactive load, but it’s not because transient impedance somehow falls at or below Zmin. It’s just too low for the amplifier at that phase angle, whatever it is.