Building A Vacuum Tube Amplifier, Part 3

[randyc]

Here's some more photos of the 40 watt amplifier, these taken during testing. (Chronologically, these belong between Parts 1 and 2, the previous photos that I posted. I am permitted only 4 photos at a time and I just noticed that these got omitted during the original posting.)

Pulse testing a bread-board of the output stage allowed me to determine whether my choice of 6JN6 tubes was a good one for the power amplifier. Pulsing allowed me to measure the maximum output power in a manner that did not stress the tubes, the output transformer or my home-built high voltage power supplies. (I did have to build this extra, small chassis but the time spent was well worth it in terms of the confidence engendered !)




Confirming preamp performance and bias points. The (very expensive) instrument at the upper right of the photo is a spectrum analyzer and was later used to measure amplifier distortion under full drive and noise. Just behind the amplifier chassis on the work bench, a one quart paint can is visible. It is my design of a high-power load. Various commercial versions are offered to allow the amplifier to be driven at full power but permit the volume to be adjusted to a friendlier level.

My version consists of two 50 watt resistors, simulating speakers, sealed in a bath of cooking oil. There's a tiny thermometer (a meat cooking thermometer, actually) installed in the assembly. It's there so that I could confirm that the internal temperature didn't exceed my expectations. (Under 50 watt RMS drive conditions, the temperature of the oil bath exceeds 220 degrees F.) Total cost was about $6.00 U.S. - I think that the commercial versions are pretty expensive !





The first real test drive. The speaker (which is located too far to the right to appear in the photograph) is a JBL D-130 in a Fender Showman cabinet. It sounded really nice. The final configuration is two tens in an ash cabinet to match the head cabinet. The rear panel of the amplifier has an impedance switch to allow either 4 ohm or 8 ohm configurations of speakers. The switch can be seen in Part 2 photos.




Hope that you enjoy the photos !

cheers

[604bourne123]

the guitar looks inviting

[Jazzarian]

Hope that you enjoy the photos !

cheers

Looks like my ES175. Is it an L4? The top looks more like spruce than maple, that's why I ask.

Why the cooking oil? Wouldn't a simple 8 ohm load resistor work fine, free standing?

What did the spectrum analyzer reveal?

[randyc]

Looks like my ES175. Is it an L4? The top looks more like spruce than maple, that's why I ask.

Why the cooking oil? Wouldn't a simple 8 ohm load resistor work fine, free standing?

What did the spectrum analyzer reveal?

You're correct, the guitar is an L-4CES, you're also correct about the top. The rest of the guitar is tiger-flame maple. The back is really beautiful. It's a far prettier guitar than my L-5 CES (but the L-5 is the one with the unbelievable tone).

I first tried the resistive loads in air and they almost glowed when I cranked up the power to 50 watts. I didn't trust the resistance values to be stable and correct at those elevated temperatures, which would have introduced error in my power measurements. I used a small fan to cool them for a while but the noise bothered me.

I was thinking about water-cooling the load but was worried about the water boiling and producing internal pressure in the paint can. My wife suggested cooking oil. It's cheap, compounded for high temperatures and a fairly good thermal conductor with reasonable specific heat (although water is better).

No doubt that I could have purchased high power loads that would have been suitable in free air. But this was an inexpensive alternative and I had a lot of fun with the small project. (Real nerdy fun, like performing finite element analysis on the load to estimate the internal temperatures.)

Using the spectrum analyzer, I measured the harmonic power levels up to the twentieth harmonic and then integrated all of the levels to give overall distortion. That's how I came up with 40 watts at 2.5% distortion and 45 watts @ 5% disxtortion.

(Incidentally, I measured the Fender Bassman at the same time using the spectrum analyzer and the same load. It produced, as advertised, 50 watts at 5% distortion.)

Lacking a real distortion analyzer, the spectrum analyzer technique works fine, although time consuming. During test and alignment, I used a simpler distortion measuring method: A - B on the dual trace oscilloscope, comparing input signal to output signal. If the phase is properly aligned (signals exactly 180 degrees out of phase, input to output) and amplitude scaling adjusted properly, this gives good, real-time measurements.

Probably would have worked just fine for final distortion measurements as well. My background is RF/microwave so I instinctively turn to the spectrum analyzer for frequency, power, noise, spurious and harmonic measurements ...

cheers

[Jazzarian]

You're correct, the guitar is an L-4CES, you're also correct about the top. The rest of the guitar is tiger-flame maple. The back is really beautiful. It's a far prettier guitar than my L-5 CES (but the L-5 is the one with the unbelievable tone).

I first tried the resistive loads in air and they almost glowed when I cranked up the power to 50 watts. I didn't trust the resistance values to be stable and correct at those elevated temperatures, which would have introduced error in my power measurements. I used a small fan to cool them for a while but the noise bothered me.

I was thinking about water-cooling the load but was worried about the water boiling and producing internal pressure in the paint can. My wife suggested cooking oil. It's cheap, compounded for high temperatures and a fairly good thermal conductor with reasonable specific heat (although water is better).

No doubt that I could have purchased high power loads that would have been suitable in free air. But this was an inexpensive alternative and I had a lot of fun with the small project. (Real nerdy fun, like performing finite element analysis on the load to estimate the internal temperatures.)

Using the spectrum analyzer, I measured the harmonic power levels up to the twentieth harmonic and then integrated all of the levels to give overall distortion. That's how I came up with 40 watts at 2.5% distortion and 45 watts @ 5% disxtortion.

(Incidentally, I measured the Fender Bassman at the same time using the spectrum analyzer and the same load. It produced, as advertised, 50 watts at 5% distortion.)

Lacking a real distortion analyzer, the spectrum analyzer technique works fine, although time consuming. During test and alignment, I used a simpler distortion measuring method: A - B on the dual trace oscilloscope, comparing input signal to output signal. If the phase is properly aligned (signals exactly 180 degrees out of phase, input to output) and amplitude scaling adjusted properly, this gives good, real-time measurements.

Probably would have worked just fine for final distortion measurements as well. My background is RF/microwave so I instinctively turn to the spectrum analyzer for frequency, power, noise, spurious and harmonic measurements ...

cheers

My ES175 has flame maple front back and sides. However the back of my Super V CES is mindblowing in comparison. Gibson spared no expense on that guitar (built by Triggs & Hutchins).

The wife had a great idea. Sure beats PCB laden oil! I'm sure the General Electric company would agree, having spent so much to clean up the Hudson River. Does the wife know about dielectric properties?


Great idea using the Spectrum Analyzer like that. Out to the 20th harmonic? Gotta love it. Tubes great on the even harmonic distortion. Hell of a gadget to have around the house. My local ACE hardware doesn't have any to rent for the day.

[randyc]

Great idea using the Spectrum Analyzer like that. Out to the 20th harmonic? Gotta love it. Tubes great on the even harmonic distortion. Hell of a gadget to have around the house. My local ACE hardware doesn't have any to rent for the day.

No, my wife doesn't know anything about dielectrics except in a medical context. She was Director of Nursing at a local hospital until cancer struck. (She got sick three years before I did.)

I watched EBay for a long time, finally got the analyzer for 10% new cost. It was damaged during shipment but fortunately, old H-P equipment is designed well, comes apart easily and is easy to repair. Got it working in a few hours.

A better buy came around last year on EBay, an HP-8640B ultra-low-noise generator, 500 kHz to 1100 MHz, AM, FM, ext mod, built-in frequency counter, the works. I got it for $300, 3% of new cost ! There were three others there the same day, all went for around $1000 ... go figure !

The deals on test equipment and machine tools are out there, just like guitars. I constantly look for deals on guitars (I hope that my wife never reads this). It may take several years to find one and the deal may not be on the exact model I wanted but the price is usually the one that I had in mind.

Unfortunately, my children have little to no interest in any of this stuff. Eventually my wife may give it away, LOL !

cheers

[Jazzarian]

Unfortunately, my children have little to no interest in any of this stuff. Eventually my wife may give it away, LOL !

cheers

That's sad. Engineering, like jazz guitar is hard, few people will put forth the effort. Few people appreciate the effort it takes.

When I lived in NJ there was a Sam Ash Music nearby, one of George Benson's sons worked there. One day I asked him if he or any of his brothers played guitar. Not one did. One of the brothers dabbled in drums, but that's it.


Hey, best of luck on the health issues!

[lpdeluxe]

I constantly look for deals on guitars (I hope that my wife never reads this). It may take several years to find one and the deal may not be on the exact model I wanted but the price is usually the one that I had in mind.

Unfortunately, my children have little to no interest in any of this stuff. Eventually my wife may give it away, LOL !

cheers

Someone said, "I hope my wife doesn't sell my guitars for what I told her I paid for them!"

My will specifies that my kids will have my stuff appraised before they start carting "all that old stuff" off to the Goodwill store.

[randyc]

Someone said, "I hope my wife doesn't sell my guitars for what I told her I paid for them!"

My will specifies that my kids will have my stuff appraised before they start carting "all that old stuff" off to the Goodwill store.

vis-a-vis your next sentence, that's a great idea !

P.S. and OT: didn't you say you played pedal steel? I have a ca 1928 Weissenborn Style 1 (apparently David Linley loves them) in immaculate condition with original canvas case. What would it be worth, any idea ?

[lpdeluxe]

P.S. and OT: didn't you say you played pedal steel? I have a ca 1928 Weissenborn Style 1 (apparently David Linley loves them) in immaculate condition with original canvas case. What would it be worth, any idea ?

Man, I don't know, but I can ask the dealer here in town who sold me my '30s Regal Model 19 -- he deals in acoustic instruments (and in fact, told me last month that a local luthier is building Weissenborn clones). In a couple of days I'll get to talk to him. I've known him for years, and last fall his wife (a fiddler) and I played in the orchestra pit for a local U production of Big River (me on harp).

I USED to play pedal steel, before foot and knee surgery (relics of my motorcycle racing days) put an end to that. Still, it hurt when I sold the MSA single-neck I had -- I love the sound of pedal steel!

I don't even want to guess, but my FIRST impulse was to offer to trade my Samick L5CES copy...I know, you've got that lovely L4....

[randyc]

Man, I don't know, but I can ask the dealer here in town who sold me my '30s Regal Model 19 -- he deals in acoustic instruments (and in fact, told me last month that a local luthier is building Weissenborn clones). In a couple of days I'll get to talk to him. I've known him for years, and last fall his wife (a fiddler) and I played in the orchestra pit for a local U production of Big River (me on harp).

I USED to play pedal steel, before foot and knee surgery (relics of my motorcycle racing days) put an end to that. Still, it hurt when I sold the MSA single-neck I had -- I love the sound of pedal steel!

I don't even want to guess, but my FIRST impulse was to offer to trade my Samick L5CES copy...I know, you've got that lovely L4....

Many thanks, LPD, it would be nice to have an approximate value to add to my homeowner's insurance list.

I've already used up my allotment of 4 photos for this post, I'll add a new post with the photo and the little bit about the guitar that I know. Perhaps you could also chip in since you know a lot more about it ..

cheers

[Jazzarian]

Many thanks, LPD, it would be nice to have an approximate value to add to my homeowner's insurance list.

I've already used up my allotment of 4 photos for this post, I'll add a new post with the photo and the little bit about the guitar that I know. Perhaps you could also chip in since you know a lot more about it ..

cheers

BTw, do you own a "tube tester", or have some way of "matching" tubes?

[randyc]

BTw, do you own a "tube tester", or have some way of "matching" tubes?

No, I don't - it's not clear to me that I need one, since the plate and screen voltages in the old, commercial tube testers are unlikely to suit modern applications. My experience with musical instrument amplifiers suggests that it's common practice for the makers to operate plate and screen at much higher levels than tube manufacturer recommendations

I built several high-voltage supplies, a single-channel, current-limited +400 volt version for plates and a dual channel, current limited +300 and -120 volts for screen and grid bias application. Using these and a normal 6.3 VAC filament supply, I just plug the tubes into sockets on a convenient breadboard chassis and characterize them.

The small breadboard chassis in the first photo - the one that I built for pulse testing - works great for general purpose tube characterization. Everything about tubes is relatively simple, compared to their solid-state counterparts.

The adjustable high voltage supplies are the key ! I looked all over the internet, trying to find lab supplies capable of 400 volts and 200 mA, they're just not being made any more. I finally found a couple in a surplus warehouse somewhere in the mid-west. The guy wanted LOTS of money and when I suggested that we negotiate the price, he suggested that I go away

So I designed and built my own high voltage lab supplies and learned a lot in the process. I can provide schematics for the bold-hearted but this would NOT be a project for people who are not firmly based in solid-state design, especially MOSFET design.

If anyone is interested, I can probably photograph the pages of my lab book and get a legible schematic ... note that I assume no responsibility or liabiliity for the design and function of the power supplies. These are tricky and potentially lethal devices, not designed like low voltage lab supplies at all !

[Jazzarian]

No, I don't - it's not clear to me that I need one, since the plate and screen voltages in the old, commercial tube testers are unlikely to suit modern applications. My experience with musical instrument amplifiers suggests that it's common practice for the makers to operate plate and screen at much higher levels than tube manufacturer recommendations

I built several high-voltage supplies, a single-channel, current-limited +400 volt version for plates and a dual channel, current limited +300 and -120 volts for screen and grid bias application. Using these and a normal 6.3 VAC filament supply, I just plug the tubes into sockets on a convenient breadboard chassis and characterize them.

The small breadboard chassis in the first photo - the one that I built for pulse testing - works great for general purpose tube characterization. Everything about tubes is relatively simple, compared to their solid-state counterparts.

The adjustable high voltage supplies are the key ! I looked all over the internet, trying to find lab supplies capable of 400 volts and 200 mA, they're just not being made any more. I finally found a couple in a surplus warehouse somewhere in the mid-west. The guy wanted LOTS of money and when I suggested that we negotiate the price, he suggested that I go away

So I designed and built my own high voltage lab supplies and learned a lot in the process. I can provide schematics for the bold-hearted but this would NOT be a project for people who are not firmly based in solid-state design, especially MOSFET design.

If anyone is interested, I can probably photograph the pages of my lab book and get a legible schematic ... note that I assume no responsibility or liabiliity for the design and function of the power supplies. These are tricky and potentially lethal devices, not designed like low voltage lab supplies at all !

Geez, when you think about it, there isn't much need of such supplies for consumer electronics. CRTs now replaced with low voltage LCD screens, no need for "triplers" and high voltage transfomers. Most everything today runs on relatively low voltages.

PC CPUs core voltages are like 1.2 volts these days. Utterly amazing.

Great job making your own supplies. Any big "Darlington" blocks on heatsinks?

BTw, long ago I used to work on the "VSCF" for the F/A-18. "Variable Speed Constant Frequency" solid state generators. Very cool stuff. You'd have loved it. First time I got to see switching power supplies.

[randyc]

Geez, when you think about it, there isn't much need of such supplies for consumer electronics. CRTs now replaced with low voltage LCD screens, no need for "triplers" and high voltage transfomers. Most everything today runs on relatively low voltages.

PC CPUs core voltages are like 1.2 volts these days. Utterly amazing.

Great job making your own supplies. Any big "Darlington" blocks on heatsinks?

BTw, long ago I used to work on the "VSCF" for the F/A-18. "Variable Speed Constant Frequency" solid state generators. Very cool stuff. You'd have loved it. First time I got to see switching power supplies.

You're right, there's little need for these supplies, the market is tiny, so the folks that have the old lab supplies want big $$$ for them. I still fume at the attitude of the guy that I talked to about his used hi-voltage supplies !

No Darlington pairs in the supplies, the main design issues were thermal (as one would expect) and sensing/control - sensing small voltage variations in high voltages and level shifting them down to low voltage control circuits (consequent loss in loop gain). I fried more than a few 1000 volt MOSFETs, creeping up on a final design. Ended up with forced air cooling ... a painful learning experience since I no longer have colleagues to consult with regarding these things.

Got out my thermodynamic texts and did a lot of number crunching after blowing the fifth MOSFET. Didn't believe the results so I downloaded an inexpensive FEA program from the internet. The free program didn't have enough nodes to do a complete analysis but the simple analysis was enough to suggest that my original thermal calculations were not conservative enough and the newer estimates were more accurate.

(The downloaded FEA program was sort of impressive, though. A year or so later, I bought an upgraded version for about $300 from the same people. The programmers are Chinese, so the program is not real user friendly for English speakers but I've learned my way around ...)

Yeah, I have some hardware on the F/A-18 and several other airframes, interesting problems involved. My career was military/aerospace oriented. (Even after I retired, I did consulting work in this area, mostly writing proposals for new military RFP's.)

cheers

[Jazzarian]

You're right, there's little need for these supplies, the market is tiny, so the folks that have the old lab supplies want big $$$ for them. I still fume at the attitude of the guy that I talked to about his used hi-voltage supplies !

No Darlington pairs in the supplies, the main design issues were thermal (as one would expect) and sensing/control - sensing small voltage variations in high voltages and level shifting them down to low voltage control circuits (consequent loss in loop gain). I fried more than a few 1000 volt MOSFETs, creeping up on a final design. Ended up with forced air cooling ... a painful learning experience since I no longer have colleagues to consult with regarding these things.

Got out my thermodynamic texts and did a lot of number crunching after blowing the fifth MOSFET. Didn't believe the results so I downloaded an inexpensive FEA program from the internet. The free program didn't have enough nodes to do a complete analysis but the simple analysis was enough to suggest that my original thermal calculations were not conservative enough and the newer estimates were more accurate.

(The downloaded FEA program was sort of impressive, though. A year or so later, I bought an upgraded version for about $300 from the same people. The programmers are Chinese, so the program is not real user friendly for English speakers but I've learned my way around ...)

Yeah, I have some hardware on the F/A-18 and several other airframes, interesting problems involved. My career was military/aerospace oriented. (Even after I retired, I did consulting work in this area, mostly writing proposals for new military RFP's.)

cheers

Ever think about making "audiophile" stereo amps? You're 90% there already. MOSFET or tube. You can pretty much name your own price because only rich people buy such, or people like me long before I had a mortgage :}

Ditto with high end guitar amps. Let's say you had a high end "Fender" clone, and Dumble/Boogie clone and a MOSFET jazz amp.

I say all this because DaddyLoveHandles has an archtop luthier down the street. $10,000 archtop variety, apparently.

[randyc]

I have seen some of the silly-expensive vacuum tube amplifiers that are being offered. I don't think that I would have the stones to sell low-fidelity copies of RCA circuits from the 1940's to people who don't know anything about the subject, it would seem unkind or something.

But that's just MY opinion, if others see this as their rice bowl and if people are willing to pay, well then ...

I sort of have the same opinion about hand-made guitars. The standard, to me at least, is the L-5CES/Super 400 guitars, I just can't see what improvements could be made to the sound of these instruments that would justify paying double/triple/quadruple the (already expensive) price.

Because of the workmanship ? The time spent making the guitar ? To a typical consumer, that probably wouldn't justify cost, our economic system operates on competition, where mass-producers are favored.

But I might justify paying these prices as ART objects

And, if one is a working musician and deriving ones income solely from music, might as well get the best equipment that money can buy because it's tax deductible (depreciated over five years, IIRC). I've frequently marveled at folks who drive $40,000 autos and try to save money by playing $300 guitars. If one is serious about music, then one should be prepared to invest in one's future, right ?

High end stereo amplifiers might be a different story. I've looked at them on the internet and some have very impressive specifications (unlike musical instrument amplifiers). I think that they may be priced closer to real value.

It also seems to me that the consumers of such equipment are much better informed about what they are buying than the average musician, who might rely more on word of mouth and buzz-word advertising than high-income stereo gear consumers (who tend to also have higher levels of education, higher incomes and a higher level of skepticism about their purchases).

None of the above is meant to be pejorative, I'm not assigning "values" to any group or individual, hey, it's all good, right ? I get worse about yacking as time goes by .. any excuse to project opinions that only I really ascribe to, LOL !

[Jazzarian]

I have seen some of the silly-expensive vacuum tube amplifiers that are being offered. I don't think that I would have the stones to sell low-fidelity copies of RCA circuits from the 1940's to people who don't know anything about the subject, it would seem unkind or something.

But that's just MY opinion, if others see this as their rice bowl and if people are willing to pay, well then ...

I sort of have the same opinion about hand-made guitars. The standard, to me at least, is the L-5CES/Super 400 guitars, I just can't see what improvements could be made to the sound of these instruments that would justify paying double/triple/quadruple the (already expensive) price.

Because of the workmanship ? The time spent making the guitar ? To a typical consumer, that probably wouldn't justify cost, our economic system operates on competition, where mass-producers are favored.

But I might justify paying these prices as ART objects

And, if one is a working musician and deriving ones income solely from music, might as well get the best equipment that money can buy because it's tax deductible (depreciated over five years, IIRC). I've frequently marveled at folks who drive $40,000 autos and try to save money by playing $300 guitars. If one is serious about music, then one should be prepared to invest in one's future, right ?

High end stereo amplifiers might be a different story. I've looked at them on the internet and some have very impressive specifications (unlike musical instrument amplifiers). I think that they may be priced closer to real value.

It also seems to me that the consumers of such equipment are much better informed about what they are buying than the average musician, who might rely more on word of mouth and buzz-word advertising than high-income stereo gear consumers (who tend to also have higher levels of education, higher incomes and a higher level of skepticism about their purchases).

None of the above is meant to be pejorative, I'm not assigning "values" to any group or individual, hey, it's all good, right ? I get worse about yacking as time goes by .. any excuse to project opinions that only I really ascribe to, LOL !

I like high end audio. For the types of speakers I tend to like (ribbons, electro-static), tube amplifiers aren't a good way to go. I've got a pair of Apogee Stages, a full range ribbon speaker.

I had a friend of mine who's expensive tube amp fried one of his Theil speakers. Something shorted, and the speaker didn't like being subjected to a DC voltage. Since the Apogee company doesn't exist anymore, I'd better steer clear of tubes!

I suspect many "audiophile" tube amps are really fancy incarnations of circuits from the old RCA and Westinghouse cookbooks. Someone like yourself with metal bending and cabinet making abilities would be set to make some dandy amps, maybe with fancy chrome plated transformers.

I've got one of the ultimate "L5s" on the planet, a Super V CES (L5 with Super400 neck). It would be very hard to top its sound, but other archtops with floating pickups have great sounds too. I'd love a Campellone in the 6 grand range with floating pickup.

[randyc]

I've got one of the ultimate "L5s" on the planet, a Super V CES (L5 with Super400 neck). It would be very hard to top its sound, but other archtops with floating pickups have great sounds too. I'd love a Campellone in the 6 grand range with floating pickup.

Oh yes that's a real dandy of a guitar and, maybe one day, one of those will follow me home, too !

And I do love looking at hand-crafted masterpieces (like a Campellone) but I don't see one in my future. If I plug my single-pickup L-5 Wes into any amplifier, tube or transistor ... I'm there. It just doesn't get any better for me. Except maybe for a vintage Johnny Smith or your Super V ... or another nice Guild.

I guess I like all of them and I'd probably buy ANYTHING nice as long as the price was right, boutique models included

[Jazzarian]

Oh yes that's a real dandy of a guitar and, maybe one day, one of those will follow me home, too !

And I do love looking at hand-crafted masterpieces (like a Campellone) but I don't see one in my future. If I plug my single-pickup L-5 Wes into any amplifier, tube or transistor ... I'm there. It just doesn't get any better for me. Except maybe for a vintage Johnny Smith or your Super V ... or another nice Guild.

I guess I like all of them and I'd probably buy ANYTHING nice as long as the price was right, boutique models included

I didn't realize you had a Wes model.

You know why Wes liked a single pickup? No pickup switch to get in the way of your thumb! I kid thee not, I bang the switch whenever I attempt octaves with a thumb. Kind of gets in the way of playing in an exuberant fashion when concentrating on not banging the pickup switch.

Wes would always wear out the top of his L5, until Gibson placed the plastic heart where he kept is other 4 fingers. Oh how I loved his tone so. There's some fine Wes DVDs out there now, great to have.

I too love Gibson Johnny Smiths, LeGrandes, Super400s.
They'll always be a place in my heart for them.

[randyc]

You know, I ALWAYS wondered about that little heart, I thought that maybe it had some association to him. The simple answer never occurred to me.

cheers