Nelson Pass said:we are very
unlikely to be requiring 1000 watts at 20,000 Hz.
Intersonics, the company that brought Servo drive subwoofers to the world. builds HF devices used to simulate zero gravity. They use 15 and 20K signals to float small objects like raisins etc for anti gravity experiments with things like metallic glass etc that cant be duplicated in a gravity environment......I wonder how much power it takes to float a small object with a 20K signal????
Zc -
Thinking again....
Lemme see, somewhere back there BarmyCanine...er, SmarmyDog asked a few questions. I didn't have any notes with me at the time, but I managed to locate part of the stuff for this project and I will herewith attempt a reply:
--Current requirement for the filament supply. First off, the 6SN7 claims to want .3A. There are two 6SN7s, so that's .6A. That's not too bad. The 6550 is a little more hungry. According to the sheets, it gobbles 1.6A. Each. Now, this is where you need to pay attention...we're not going to worry about the heater/cathode potential for the cascode tubes, seeing as how they're only 50V off of ground and that's well within the spec for the 6SN7. V1 and V9 (the current source 6550 tubes; the ones on the bottom) are also right at ground. Go ahead and tap them into the same filament supply. So now we're up to .3+.3+1.6+1.6= 3.8A total. Naturally, you're going to want to have some elbow room, so get something capable of at least 7 or 8 amps...say 50VA as an absolute minimum. Bear in mind that filaments are like class A amps--they draw that much current all the time, and I do mean all the time. Better to have too much VA than not enough. Filament transformers tend to run warmish, and for good reason.
But, Grey, what about V2 and V8? Glad you asked. Those will need their own filament supply. You're looking at 3.2A draw there, but you'll need to tag it to the midpoint between the rail and ground. A simple resistive divider will do. Let's say two 100k 1W resistors. Join either the positive or negative to the midpoint of the divider--doesn't matter which.
--350V current: Trivial. You're looking at less than 5mA. Use a small isolation (1:1) power transformer with a voltage doubler rectifier and you're good to go.
--375V current: This is where I bent my original plan a bit. 375V is going to be a little tough to get with a voltage doubler. You'll probably need to get a separate transformer for this one. The current draw per follower is about 60mA, so the two outputs together will draw about 120mA. A 100 to 120VA transformer will be ideal here. If you want to use this transformer to power the differential, it's fine by me. The extra current draw will be negligible and you can knock the voltage down with regulators or resistors.
I don't believe I showed the two output stage rails tied together on the schematic. Do so. It will make the power supply's job much easier. The two sides' current draw will sum together as DC. Trust me, we like DC current draw...a lot. Power supplies can deliver DC comparatively easily.
--Regulators or CLC? Either one. Granted, I can be a bit fanatical about power supplies, but that's just me. You'll be fine with a normal power supply.
--General notes. 6550s are single pentodes (or tetrodes, depending on who you ask...just to complicate things, we're using them as triodes). Each is an independent unit. You're well within spec on all parameters, so if you want to goose the output current, feel free to do so. Bear in mind that the 6550 plate is only rated for 30-odd watts of heat. You can push this limit a little harder than you can the average power MOSFET, but don't overdo it. Let's take 25W as a maximum. The plates are rated for 600V or so, so the voltage isn't going to be a problem. I don't rememer what the maximum current is right off the top of my head--I'll have to go look that one up. But there's leeway.
The 6SN7s, however, are twin triodes. Think about the 2SK389. It's a twin JFET, right? Wrap it in a glass envelope and call it a tube. I remember the first time I started thinking seriously about solid state. I looked around for the twin devices. I kept looking...and looking...and looking... Whaddaya mean there ain't no twin devices? They ought to be common as grass! Well, they aren't. They're scarce, either as bipolar or FET. It's a flippin' crime! Tubes? Oh, yeah! Now, we're talkin'. All the common audio small signal tubes are twins, and they're tightly matched. Almost impossibly so by solid state standards. So don't worry about having to buy a million and match them. Every 6SN7 comes pre-matched. Cool, eh?
Okay, so what does this mean in terms of this circuit?
I'd recommend making one tube V3 & V7 and the other V4 & V6. You can use them "vertically" if you want (V3 & V4 as one tube, and V6 & V7 as the other) and it will work just fine, but this will tilt the match in your favor.
If there's anything I missed, holler. I'll probably reply by Thanksgiving.
Hopefully.
Grey
--Current requirement for the filament supply. First off, the 6SN7 claims to want .3A. There are two 6SN7s, so that's .6A. That's not too bad. The 6550 is a little more hungry. According to the sheets, it gobbles 1.6A. Each. Now, this is where you need to pay attention...we're not going to worry about the heater/cathode potential for the cascode tubes, seeing as how they're only 50V off of ground and that's well within the spec for the 6SN7. V1 and V9 (the current source 6550 tubes; the ones on the bottom) are also right at ground. Go ahead and tap them into the same filament supply. So now we're up to .3+.3+1.6+1.6= 3.8A total. Naturally, you're going to want to have some elbow room, so get something capable of at least 7 or 8 amps...say 50VA as an absolute minimum. Bear in mind that filaments are like class A amps--they draw that much current all the time, and I do mean all the time. Better to have too much VA than not enough. Filament transformers tend to run warmish, and for good reason.
But, Grey, what about V2 and V8? Glad you asked. Those will need their own filament supply. You're looking at 3.2A draw there, but you'll need to tag it to the midpoint between the rail and ground. A simple resistive divider will do. Let's say two 100k 1W resistors. Join either the positive or negative to the midpoint of the divider--doesn't matter which.
--350V current: Trivial. You're looking at less than 5mA. Use a small isolation (1:1) power transformer with a voltage doubler rectifier and you're good to go.
--375V current: This is where I bent my original plan a bit. 375V is going to be a little tough to get with a voltage doubler. You'll probably need to get a separate transformer for this one. The current draw per follower is about 60mA, so the two outputs together will draw about 120mA. A 100 to 120VA transformer will be ideal here. If you want to use this transformer to power the differential, it's fine by me. The extra current draw will be negligible and you can knock the voltage down with regulators or resistors.
I don't believe I showed the two output stage rails tied together on the schematic. Do so. It will make the power supply's job much easier. The two sides' current draw will sum together as DC. Trust me, we like DC current draw...a lot. Power supplies can deliver DC comparatively easily.
--Regulators or CLC? Either one. Granted, I can be a bit fanatical about power supplies, but that's just me. You'll be fine with a normal power supply.
--General notes. 6550s are single pentodes (or tetrodes, depending on who you ask...just to complicate things, we're using them as triodes). Each is an independent unit. You're well within spec on all parameters, so if you want to goose the output current, feel free to do so. Bear in mind that the 6550 plate is only rated for 30-odd watts of heat. You can push this limit a little harder than you can the average power MOSFET, but don't overdo it. Let's take 25W as a maximum. The plates are rated for 600V or so, so the voltage isn't going to be a problem. I don't rememer what the maximum current is right off the top of my head--I'll have to go look that one up. But there's leeway.
The 6SN7s, however, are twin triodes. Think about the 2SK389. It's a twin JFET, right? Wrap it in a glass envelope and call it a tube. I remember the first time I started thinking seriously about solid state. I looked around for the twin devices. I kept looking...and looking...and looking... Whaddaya mean there ain't no twin devices? They ought to be common as grass! Well, they aren't. They're scarce, either as bipolar or FET. It's a flippin' crime! Tubes? Oh, yeah! Now, we're talkin'. All the common audio small signal tubes are twins, and they're tightly matched. Almost impossibly so by solid state standards. So don't worry about having to buy a million and match them. Every 6SN7 comes pre-matched. Cool, eh?
Okay, so what does this mean in terms of this circuit?
I'd recommend making one tube V3 & V7 and the other V4 & V6. You can use them "vertically" if you want (V3 & V4 as one tube, and V6 & V7 as the other) and it will work just fine, but this will tilt the match in your favor.
If there's anything I missed, holler. I'll probably reply by Thanksgiving.
Hopefully.
Grey
monster amp
Nelson, thanks for the reply regarding the current. I just found my hi-voltage insulated gloves, several transformers out of a WW 2, P-51 and my rubber mats, and my leftover SK output devices from the Aleph projects.
I just need snow to motivate me to start heating my basement with tubes. dave
Nelson, thanks for the reply regarding the current. I just found my hi-voltage insulated gloves, several transformers out of a WW 2, P-51 and my rubber mats, and my leftover SK output devices from the Aleph projects.
I just need snow to motivate me to start heating my basement with tubes. dave
monster amp
Mr. Grollins--I never thought I would build an amp that is bigger than my 2 Newfoundlands (150lbs each) but your monster amp even if I build it to a smaller scale will dwarf my Aleph 1.2's.
I will need my newfies hooked up to a harness with big wheels on the amp to help me move it around. In all seriousness, I saw your old schematic for a 10watt xa amp based on a NP design.
I am building some line arrays (I have a thread in loudspeakers) which will have a 96db efficiency due to the 24 drivers per side and I am wondering if you everbuilt the xa and if it would work on my LA speaker. This speaker project has opened up the possibility to using low watts and tubes. Typically I listen to music between 75 to 85 db and occasionally up to 95db (not for long)I am also interested in building the j-fet alpeh. My home built Aleph 1.2 monsters can then be retired to rock and roll. This will probably help save the worlds oil and coal reserves due to the Aleph's penchant for a big electricity appetite.
I have had a problem in the past with a snowbound Newfie wanting to lay next to or on a nice hot Aleph--if I put the amp in the speaker encloseure than that won't happen! My listening room is in a semi-fixed up 30x60x9 basement. dave (daly2k)
Mr. Grollins--I never thought I would build an amp that is bigger than my 2 Newfoundlands (150lbs each) but your monster amp even if I build it to a smaller scale will dwarf my Aleph 1.2's.I will need my newfies hooked up to a harness with big wheels on the amp to help me move it around. In all seriousness, I saw your old schematic for a 10watt xa amp based on a NP design.
I am building some line arrays (I have a thread in loudspeakers) which will have a 96db efficiency due to the 24 drivers per side and I am wondering if you everbuilt the xa and if it would work on my LA speaker. This speaker project has opened up the possibility to using low watts and tubes. Typically I listen to music between 75 to 85 db and occasionally up to 95db (not for long)I am also interested in building the j-fet alpeh. My home built Aleph 1.2 monsters can then be retired to rock and roll. This will probably help save the worlds oil and coal reserves due to the Aleph's penchant for a big electricity appetite.
I have had a problem in the past with a snowbound Newfie wanting to lay next to or on a nice hot Aleph--if I put the amp in the speaker encloseure than that won't happen! My listening room is in a semi-fixed up 30x60x9 basement. dave (daly2k)
My monster amp? Nah, I just stuck some parts together that might serve as a front end. The thing is modular enough that I (or someone else) may take a whack at another front front end later. Maybe solid state...maybe another tube one. You never know.
The 10W amp was the Mini-A, the 'xa' was the Aleph-X, which was more like 40W/ch. As to suitability of either amp for a given speaker or listnening environment, that's always open to question. My feeling is to have enough power to handle peaks without clipping. I use a 10:1 ratio as my rule of thumb, meaning that if my average listening level is to be 1W, then I want 10W--minimum--in the amp. Several other factors weigh in, as well. One is the impedance of the load. The Mini-A, as published, didn't have enough bias current to handle, say, a 4 Ohm load. The Aleph-X did. Now, that's not to say that either circuit couldn't be modified to produce an arbitrary amount of current, and I've played such games myself.
If you run into impedance problems, you could always do something along the lines of two Mini-As per array, each handling half the drivers. The amps are small, cheap to build, and easy, since you don't have to match the output devices. (Matching the inputs is still a good idea.)
I'm not sure that placing an amplifier--especially a class A circuit--inside a speaker is a good idea. It will seriously interfere with heat dissipation. Consider placing the heatsink(s) on the back where they can breathe freely.
Grey
EDIT: If I ever built an xa? Hmmm. Methinks you've never read the Aleph-X thread.
The 10W amp was the Mini-A, the 'xa' was the Aleph-X, which was more like 40W/ch. As to suitability of either amp for a given speaker or listnening environment, that's always open to question. My feeling is to have enough power to handle peaks without clipping. I use a 10:1 ratio as my rule of thumb, meaning that if my average listening level is to be 1W, then I want 10W--minimum--in the amp. Several other factors weigh in, as well. One is the impedance of the load. The Mini-A, as published, didn't have enough bias current to handle, say, a 4 Ohm load. The Aleph-X did. Now, that's not to say that either circuit couldn't be modified to produce an arbitrary amount of current, and I've played such games myself.
If you run into impedance problems, you could always do something along the lines of two Mini-As per array, each handling half the drivers. The amps are small, cheap to build, and easy, since you don't have to match the output devices. (Matching the inputs is still a good idea.)
I'm not sure that placing an amplifier--especially a class A circuit--inside a speaker is a good idea. It will seriously interfere with heat dissipation. Consider placing the heatsink(s) on the back where they can breathe freely.
Grey
EDIT: If I ever built an xa? Hmmm. Methinks you've never read the Aleph-X thread.
monster amp
jacco, my speakers are still in construction, I can post a picture when they are finished--probably a while yet because of the wood work that I need to do.
GRollins, you are right I have never "studied" the alephX thread but I will now to get some ideas. thanks for the feedback. dave
jacco, my speakers are still in construction, I can post a picture when they are finished--probably a while yet because of the wood work that I need to do.
GRollins, you are right I have never "studied" the alephX thread but I will now to get some ideas. thanks for the feedback. dave
I realized that I have not thanked MikeW, whose kind & timely donation of a healthy-sized filament transformer saved the day. My usual filament power supply went down at a most inopportune moment, and within a day or two this big honkin' transformer shows up on my doorstep. Cool, eh?
Now, the really amazing thing is that Mike didn't know that my filament supply had croaked--he just sent the thing on his own initiative. Must have been a disturbance in the Force or something.
At any rate, everybody needs to doff their hats whilst facing in the Boston direction and say,"Thank you, Mike."
Thanks, Mike. Couldn't have done it without you, man.
Grey
P.S.: For those who might be interested in SF, Mop-Up is in the Jan/Feb issue of Analog.
Now, the really amazing thing is that Mike didn't know that my filament supply had croaked--he just sent the thing on his own initiative. Must have been a disturbance in the Force or something.
At any rate, everybody needs to doff their hats whilst facing in the Boston direction and say,"Thank you, Mike."
Thanks, Mike. Couldn't have done it without you, man.
Grey
P.S.: For those who might be interested in SF, Mop-Up is in the Jan/Feb issue of Analog.
I have no idea what kind of currents we are going to face in the monster amp. Since I’m not involved in the development but would like to contribute in some way, I might have an interesting item to add for the final layout.
Here’s a massive ampere meter. As you see the range is 800A and the diameter is 18cm (+/-7.1inch).
It weighs 2.3Kg and if the shipping costs are reasonable, I’m prepared to donate and ship the thing to the ‘final assembly plant’.
/Hugo
Here’s a massive ampere meter. As you see the range is 800A and the diameter is 18cm (+/-7.1inch).
It weighs 2.3Kg and if the shipping costs are reasonable, I’m prepared to donate and ship the thing to the ‘final assembly plant’.
/Hugo
Attachments
OMG! Do you have six of these (or know where I can get ‘em)? I suspect the Environmental Protection Agency would pay us a visit if we biased them to 400A, but a snazzy new brushed aluminum faceplate with a PASS-ified scale would be oh-so-cool…
PS: Nice CD too! I have her on 5.1 SACD. A definite favorite.
PPS: I’ll try and post a construction update some time this coming week.
-Casey Walsh
PS: Nice CD too! I have her on 5.1 SACD. A definite favorite.
PPS: I’ll try and post a construction update some time this coming week.
-Casey Walsh
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.