Anybody have any good ideas for building a good 30-50 watt tube integrated amp. Something with fast transparent midrange, great soundstage and fairly tight bass. Simple design too.
Hoping to find something which will sound as good as the best current commercial offerings?
After doing a lot of reading I don't feel confident that there is.
Some of the projects may be neat to build, But do we end up with a great amp when we are finished or just a fair sounding amp?
Hoping to find something which will sound as good as the best current commercial offerings?
After doing a lot of reading I don't feel confident that there is.
Some of the projects may be neat to build, But do we end up with a great amp when we are finished or just a fair sounding amp?
I can't recommend any particular design from my own experience. I'm messing about with EL34 triode-connected PP, which gives only 15 watts or so.
However, I assure you that a home-built amp can be every bit as good as any commercial amp and probably better. Commercial amps are usually either compromised by cost considerations or are horrendously expensive (sometimes both!). Don't get taken in by the commercial hype.
However, I assure you that a home-built amp can be every bit as good as any commercial amp and probably better. Commercial amps are usually either compromised by cost considerations or are horrendously expensive (sometimes both!). Don't get taken in by the commercial hype.
If you want a genuine 30W - 50W amplifier of superlative quality, it's not going to be all that simple and it will certainly be big and heavy. The quality will be largely determined by the output transformers (LF distortion due to core saturation and HF ringing due to leakage inductance and stray capacitance). HF distortion will be down to the driver. A simple driver will be more likely to cause slewing distortion because of inadequate current to charge the output stage's input capacitance. Too many coupling capacitors will cause blocking and prolong distortion when (not, if) the amplifier is overloaded. Reducing the number of coupling capacitors implies DC coupling and almost always requires extra power supplies.
The difference with the commercial designs is that the cost of development is amortized over hundreds of production amplifiers, so they can justify spending the time to fettle the amplifier and screw the very best performance out of the cheapest parts. Or, at least, that should be the case.
The difference with the commercial designs is that the cost of development is amortized over hundreds of production amplifiers, so they can justify spending the time to fettle the amplifier and screw the very best performance out of the cheapest parts. Or, at least, that should be the case.
PP 7591s in pentode mode will yield the low end of your power range. 7591s are easy to drive, which makes a 2 stage amp feasible. The volume control is right at the I/P.
GOOD "iron" is a must, as was previously pointed out. Weight will be a factor, as choke I/P filtration should be considered for the B+ PSU. 7591 idle current will be low (Class "AB" operation). For max. linearity, pentode mode "finals" require regulated screen grid B+.
A rough "sketch". 12AT7 differential splitter/driver drives the "finals". Loop NFB is applied to non-inverting I/P of splitter driver. Inside the NFB loop, the only caps. are those that couple the splitter to the finals. CCS in the splitter's tail optimizes performance. Overall PSU requirements are 3 rails: B+, B-, and C-.
Jim McShane points out that the combo of a 12AT7 and PP finals yields a GOOD sounding "waterfall" distortion spectrum.
GOOD "iron" is a must, as was previously pointed out. Weight will be a factor, as choke I/P filtration should be considered for the B+ PSU. 7591 idle current will be low (Class "AB" operation). For max. linearity, pentode mode "finals" require regulated screen grid B+.
A rough "sketch". 12AT7 differential splitter/driver drives the "finals". Loop NFB is applied to non-inverting I/P of splitter driver. Inside the NFB loop, the only caps. are those that couple the splitter to the finals. CCS in the splitter's tail optimizes performance. Overall PSU requirements are 3 rails: B+, B-, and C-.
Jim McShane points out that the combo of a 12AT7 and PP finals yields a GOOD sounding "waterfall" distortion spectrum.
labjr said:
If one of your concerns is building an amp that is a "known quantity" in comparison to commercial amps, a good suggestion would be to build a replica Dynaco ST-70. It delivers 30-35 watts/channel.
The ST-70 was sold as a commercial amp until the early 90s. There's even a Stereophile review of it, including measurements, on their site. It's still a very popular amp today because it sounds so good. There are at least six companies (Curcio, DiyTube, Welborne, Triode Electronics, VTV, Van Alstine) who sell PCBs with upgraded driver circuits, so you have a lot of choice, and so many people have built these that the support is there. You can also buy PCBs for the stock circuit, but the tube it uses is getting rare and most of the upgraded designs have better performance.
If you can find an old ST-70 locally to reuse the transformers, which are still comparable to the best commercial iron, this is also one of the more inexpensive amp projects you can build. If you can't, Ned at Triode Electronics sells replica ST-70 transformers for a reasonable price.
I checked out the Stereophile review of the ST-70 II . Not the greatest review.
Most of the kits seem like they are just alright sounding.
I wonder about the new designs coming out of china. Like the new Prima Luna EL34 integrated. That is getting some great reviews and the circuit seems pretty simple. Or the Cayin stuff.
Or does anyone have a circuit for any of the Quicksilver amps.
I want to build something I know will be every bit as good as the best commercially available amps.
Most of the kits seem like they are just alright sounding.
I wonder about the new designs coming out of china. Like the new Prima Luna EL34 integrated. That is getting some great reviews and the circuit seems pretty simple. Or the Cayin stuff.
Or does anyone have a circuit for any of the Quicksilver amps.
I want to build something I know will be every bit as good as the best commercially available amps.
what about something like this...
http://www.everestaudio.com/price-pages/sophia-kt88-amplifier.htm
mated to a preamp like this or even a simpler preamp
http://www.everestaudio.com/price-pages/harmonix-reimyo-cat-777-vacuum-tube-control-amplifier.htm
http://www.everestaudio.com/price-pages/sophia-kt88-amplifier.htm
mated to a preamp like this or even a simpler preamp
http://www.everestaudio.com/price-pages/harmonix-reimyo-cat-777-vacuum-tube-control-amplifier.htm
John,
I'm point man on a LOW budget project. Another of this site's moderators, Planet10, is involved.
El Cheapo Project
My concept for the 7591 integrated is to take "El Cheapo" and feed it a TON of steroids.
Since we are talking about adjustable fixed bias for the "finals", the 7591 grid leak resistors need to be on the small side, say 47 KOhms. Maintaining the gain of the splitter/driver leads to a large load resistor, say 470 KOhms. That, in turn, leads to a high B+ rail voltage for the 'T7s. Pete Millett's "trick" of getting 2 B+ rails out of a single CT winding looks good. Take a Hammond 715 power trafo or something similar. Bridge rectify the entire winding with series connected pairs of UF4007s forming the ground connection and 6AX4 damper diodes forming the B+ connection. Filter as you see fit. This rail is high voltage/low current (12 mA. draw) that overcomes the drop in the large 'T7 load resistors. The CT of the rectifier winding gets connected to both plates of a 5AR4. The cathode of the 5AR4 gets connected to a LC section and bleeder resistor. Follow with paired LC sections (1/channel). This yields somewhat in excess of 400 V. for driving the 7591s. Stacked VR150 and VR105 gas diodes provide the tightly controlled screen grid B+. Use a separate regulator stack in each channel.
This is not a bargain basement project. Use 100 KOhm DACT stepped attenuators at the I/P. The high pass cap. should be larger than that in El Cheapo; a turnover freq. of about 17 Hz. seems right. Better "iron" will not saturate as fast as the T/E stuff in El Cheapo, but some protection is still indicated.
On the O/P "iron" front, custom stuff with 10% cathode feedback tertiary windings would be nice. Not having to rely totally on loop NFB to obtain an adequate damping factor can't hurt.
"Regular" inductive wirewound resistors will be used as 'T7 loads. Given the significant number of turns, enough HF peaking might be present to eliminate the need for RF chokes to complete phase compensation requirements.
Use TO220 case FETs in the CCSes. This project is not about pinching pennies. Maximizing sonic potential is the goal here.
Things are dicey with my previous thinking. That HIGH voltage/low current rail is going to be up near 1.5 KV. DANGER!! 
Maybe there's a better way. The object is to keep the gain of the splitter/driver up. Something SY stated in a concurrent PP EL34 thread rings a bell. Cathode follower drive of the "finals" would allow "normal" values for the 'T7 load resistors.
PP EL34
Heck, I'm a pragmatist. If 1 KVdss FETs don't "fry", they too could be used as voltage followers driving the 7591s. This idea is lifted from svokke's mike preamp thread.
svokke's mike preamp
The Hammond power trafo previously mentioned is fine. Use it to feed a GZ37 and choke I/P filter. Every last mA. the trafo is capable of can be used.
Maybe there's a better way. The object is to keep the gain of the splitter/driver up. Something SY stated in a concurrent PP EL34 thread rings a bell. Cathode follower drive of the "finals" would allow "normal" values for the 'T7 load resistors.
PP EL34
Heck, I'm a pragmatist. If 1 KVdss FETs don't "fry", they too could be used as voltage followers driving the 7591s. This idea is lifted from svokke's mike preamp thread.
svokke's mike preamp
The Hammond power trafo previously mentioned is fine. Use it to feed a GZ37 and choke I/P filter. Every last mA. the trafo is capable of can be used.
Svokke provided a link in his mike preamp thread to a page of MOSFET uses. I'm replicating that link here.
MOSFET "Follies"
IMO, there's considerable potential in FET voltage followers. This is a theme I've "harped" on before. In carefully selected situations, a combination of tubes and FETs will do a better job than either device type alone. "Keep your eyes on the prize." In this case, the prize is GOOD sound.
MOSFET "Follies"
IMO, there's considerable potential in FET voltage followers. This is a theme I've "harped" on before. In carefully selected situations, a combination of tubes and FETs will do a better job than either device type alone. "Keep your eyes on the prize." In this case, the prize is GOOD sound.
labjr said:
It is also a different amplifier than the original ST70 which is what I believe Wodgy was referring to by the list of mods he noted. The original is not without flaw but it one you will find many like. It is the amp I started with and I'd be happy with one as my sole choice. It is kind of like that first true love.
I would also concur that basing an integrated off an original ST70 would not be a bad way to go. The output transformers used on the ST70 were very good and improved front-end mods abound. There certainly isn't a "best" integrated amp choice under any circumstance though.
I have noticed that there is a ton of ST-70 mods out there. I guess because there was a glut of them made.
However, if one were starting from scratch would this be the circuit to build, or are there better alternatives?
Are the original ST-70 outputs as good as some of the best available today?
Would one take the ST-70 circuit, buy some Hammond outputs and build it? Not that Hammond makes the best transformers.
I just wonder if some are biased because of the low cost of the ST-70 project. Should I really expect to hear great sound from Triodes $59 transformers?
I think that saying it's personal preference would not be correct here. I had an old Fisher 400 receiver that I replaced with some brand-name 300B SET amps and the difference like hearing a recording vs hearing the live event.
However, if one were starting from scratch would this be the circuit to build, or are there better alternatives?
Are the original ST-70 outputs as good as some of the best available today?
Would one take the ST-70 circuit, buy some Hammond outputs and build it? Not that Hammond makes the best transformers.
I just wonder if some are biased because of the low cost of the ST-70 project. Should I really expect to hear great sound from Triodes $59 transformers?
I think that saying it's personal preference would not be correct here. I had an old Fisher 400 receiver that I replaced with some brand-name 300B SET amps and the difference like hearing a recording vs hearing the live event.
Was the Fisher reconditioned? If not, the comparison is COMPLETELY unfair. Of course, 300B stuff can be VERY nice. However, Fisher receivers in proper repair are nothing to sneeze at.
Even with an updated driver circuit, the ST70 does not have enough gain to stand by itself as an "integrated". That's why I suggested the 7591 previously. The 7591 was designed to compete with the EL34. Fisher and Scott were highly competent. Both of them used the 7591/7868/6GM5 (Octal/Novar/Noval) to considerable advantage. In a carefully worked out circuit, both the EL34 and 7591 work WELL.
There are at least 2 sources for "exact" replicas of the ST70 O/P trafos: Triode Electronics and MagneQuest. Both adhere to the original design, but use some "Teflon" insulated wire that rates to be a slight improvement over the original. I don't think Hammond stuff is as good.
FWIW, my biases run towards keeping NFB loops short and as cap. free as can be. The chances of success are better, IMO. When multiple caps. are in the loop and/or the loop encompasses several stages, all sorts of things, like phase shift oscillation, can go wrong. I have no pretensions of being able to lift the late, great, Stew Hegeman's slide rule.
A favorite EL34 PP recommendation for refitting an ST70 or... not really that much different from Eli's El Cheapo.... i'd consider the power supply a starting point.
from http://www.vacuumstate.com/schematics.htm
dave
from http://www.vacuumstate.com/schematics.htm
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
dave
The cascoded 6922 differential phase splitter circuit Dave (Planet10) provided is nice. However, with the EL34s triode wired, 15 W. is about the limit. Also, I think that circuit runs Class "A"; so, O/P power will be even lower.
The cascode splitter circuit could be altered to use ultralinear topology and adjustable fixed bias. Then, a 30 W. O/P with the EL34s operating in Class "AB" is a reasonable expectation. As is done in "El Cheapo" a small amount of loop NFB needs to be applied to the non-inverting I/P of the splitter. Use the HF peaking trick by substituting inductive Dale wirewound resistors for the Mills non-inductive parts.
Unfortunately, gain will still be inadequate for use without a line level preamp that raises the signal by a few dB.
The cascode splitter circuit could be altered to use ultralinear topology and adjustable fixed bias. Then, a 30 W. O/P with the EL34s operating in Class "AB" is a reasonable expectation. As is done in "El Cheapo" a small amount of loop NFB needs to be applied to the non-inverting I/P of the splitter. Use the HF peaking trick by substituting inductive Dale wirewound resistors for the Mills non-inductive parts.
Unfortunately, gain will still be inadequate for use without a line level preamp that raises the signal by a few dB.
I wonder if someone makes a driver board with extra gain to make it usable as an integrated amp.
I've done quite a bit of reading lately about all the driver boards available. Everyone thinks theirs is the best one. I wonder which one is really best.
I think all of them maintain the original Ultralinear operation. But isn't triode operation better sounding?
I think this guy is using a passive attenuator for his ST-70
http://community.webshots.com/photo/354063825/363206616lQTYyt
I've done quite a bit of reading lately about all the driver boards available. Everyone thinks theirs is the best one. I wonder which one is really best.
I think all of them maintain the original Ultralinear operation. But isn't triode operation better sounding?
I think this guy is using a passive attenuator for his ST-70
http://community.webshots.com/photo/354063825/363206616lQTYyt
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.