For about 100-500wat output, maybe with different frequency responses (I'm designing a BIG 2-4k tri-amped system) perferbably with 0.1% distortion (you can always try lol). I was looking in to ones used for transmitting but most of the one's I've seen have about 5% distortion, Svetlana have some explicitly described as possible class A useage.
I was thinking get some ones that would be overpowered and reduce the operation conditions to reduce distortion? Cheers for any help on this I proper want to do it!
I was thinking get some ones that would be overpowered and reduce the operation conditions to reduce distortion? Cheers for any help on this I proper want to do it!
You want ridiculous power, you need a big transmitting tube. My favorite, the 833A. You will need a big power supply, a good OPT and some serious cooling since the filament alone eats 100 watts. I could get an easy 200 watts in A2, 50 watts in A1 came at 0.9% distortion. To get to 0.1% you will need NFB and a really really good OPT.
http://www.tubelab.com/833SE.htm
http://www.tubelab.com/833SE.htm
tubelab.com said:
Yeman I've already found your page, it's well informative! The lnik's not working right now though
And Bandersnatch the GM100 looks the most plausible right now since it's got the highest output but there might be more. I'm probably going to ring all the valve manufacturers in the UK (because that's where I'm from and it'll be easier to co-ordinate supplies) and ask their opinions.
I was also going to make it OTL for full range response (Atmosphere so they can get 2Hz out of their OTL stuff)
If you can't find triodes big enough, Fake it!
Though for reasons of grid leak at very low plate voltage,
I had to abuse the triode's plate more like a screen in UL...
Whatever, its OTL except for the cap... I'm working on it.
This is largely based upon Nelson Pass' Zen + Aleph.
Bent around a Triode's 4th circuit. Plate in > Cathode out...
35W ain't too bad for one half of a 12AU7 in SE.
-----------------------------
Look to Tubelab's Super 45 for further inspiration:
http://www.tubelab.com/SuperTubeSE.htm
Though for reasons of grid leak at very low plate voltage,
I had to abuse the triode's plate more like a screen in UL...
Whatever, its OTL except for the cap... I'm working on it.
This is largely based upon Nelson Pass' Zen + Aleph.
Bent around a Triode's 4th circuit. Plate in > Cathode out...
35W ain't too bad for one half of a 12AU7 in SE.
-----------------------------
Look to Tubelab's Super 45 for further inspiration:
http://www.tubelab.com/SuperTubeSE.htm
Attachments
That changes everything. If you are using conventional speakers with the typical 8 ohm impedance you need to source (and sink) a lot of current. Several AMPS are required for the 100 watt range. This isn't going to happen with any reasonable transmitting tube, and is going to require several low impedance tubes in parallel, or some solid state assistance.
Yes Atmasphere gets a low cutoff frequency without output transformers but it takes a big bunch of 6AS7's in parallel to do it. The 6AS7 is the lowest cost tube capable of handling some current. The next step up is the 6336A or the 6528, these get expensive. Beyond that there is the 7241 or 7242, big bucks though.
If you are using electrostatic speakers it is possible to drive them directly with sweep tubes, but then you don't need a 2Hz corner frequency unless you have really big speakers.
Just a few thoughts - it seems that ur rather completely unclear as to what it is that you actually want to do.
All of these big high power transmitting tubes require very very high voltage and almost all of them also have a comensurately high plate impredance that means a difficult to build output transformer.
Then too these tubes are not designed as Class A tubes, so you have to run them at a substantially lowered plate voltage and adjust the bias - while that may get you into the range of class A It surely will not give you anything near 500watts of class A, even in PP. You actually have to derate the tube substantially compared to its Class B spec.
Then you also have to drive them with something substantial in most cases... depends a bit on the tube's gain, but it's not going to be a 12AX7 usually.
Suggest that you think through your project more fully and narrow down the real world specs that you need to acheive before deciding that you want or need?
All of these big high power transmitting tubes require very very high voltage and almost all of them also have a comensurately high plate impredance that means a difficult to build output transformer.
Then too these tubes are not designed as Class A tubes, so you have to run them at a substantially lowered plate voltage and adjust the bias - while that may get you into the range of class A It surely will not give you anything near 500watts of class A, even in PP. You actually have to derate the tube substantially compared to its Class B spec.
Then you also have to drive them with something substantial in most cases... depends a bit on the tube's gain, but it's not going to be a 12AX7 usually.
Suggest that you think through your project more fully and narrow down the real world specs that you need to acheive before deciding that you want or need?
Tubelab inspiration: You test your wares behind a perspex screen ? Good thought.
I always thought I was ridiculously dangerous without one and often just use ear protectors and safety glasses typ protection of old fashioned CRT implosion. Not only is this tube stuff dangerous, worst are those massive block plastic power mosfets and IGBts which I use for traction SMPS work. YOu need a helmet for those.
Stay alive.
richy
I always thought I was ridiculously dangerous without one and often just use ear protectors and safety glasses typ protection of old fashioned CRT implosion. Not only is this tube stuff dangerous, worst are those massive block plastic power mosfets and IGBts which I use for traction SMPS work. YOu need a helmet for those.
Stay alive.
richy
Electrolytics generate lot of nasty smoke and yellow sparks when plugged into an outlet. Some 30 years ago my roommate once put one under the bed when we selebrated a New Year with our girlfriends. Switching on a "firework" he ignited a bed dress, so we had nothing but Champagne to extinguish the fire...
It IS Lexan. The shield in that particular test was meant to seperate ME from the highly unsafe 1500 volts running through Radio Shack clip leads.
I use a smaller piece of Lexan whenever I am playing with ANYTHING that is plugged into the wall socket. Mosfets go bang in a most spectacular way. I have even placed an inverted coffee mug over them to contain the spray. The Lexan shield will be needed with my new power supply since it is hungry and likes to eat things. It blew the guts out of an electrolytic cap in a most spectacular fashion late last year. 650 volts at 1.7 amps will do that. A fire extinguisher is always at hand too, although I have never had to use it. I have stunk up the house real good a few times in the last 30 years.
Yep, done that too. 2N3773 VS wall outlet, wall outlet wins.
Woah cheers some very informative responses! To reiterate the general spec I'm aiming for:
Class A (linear as linear can be)
Full frequency response (possibly output transformerless)
(Several) hundred watts of power
Low distortion (<0.1%, because dreams can come true)
I'll get it done eventually, because it is theoretically possible...
This site's pretty useful http://www.vt52.com/diy/myprojects/amps/gm100/gm100.htm
arranging them in to a bank and co-ordinating cooling would be a nightmare though!
Class A (linear as linear can be)
Full frequency response (possibly output transformerless)
(Several) hundred watts of power
Low distortion (<0.1%, because dreams can come true)
I'll get it done eventually, because it is theoretically possible...
This site's pretty useful http://www.vt52.com/diy/myprojects/amps/gm100/gm100.htm
arranging them in to a bank and co-ordinating cooling would be a nightmare though!
Regarding the Russian 6C33; I've found these triodes notoriously variable and difficult to match. Anything else is now getting exotic, price wise not viable.
Kenpeter is right. Bodge it with multiple parallel pairs or triplets.That would greatly ease the o/p tranny and power supply design.
richy
Kenpeter is right. Bodge it with multiple parallel pairs or triplets.That would greatly ease the o/p tranny and power supply design.
richy
Test Electrix said:
May I ask, What specific application is this Huge and very power-hungry amp intended to do??
Class 'A' is the mot power wasteful 'format' thats why when higher powers than a few tens of watts are needed, class AB etc are used....
...........and co-ordinating cooling would be a nightmare though!...........
Well, There you have it!--All Class A will generate vast amounts of heat in comparison to the output power--for small amps of less than say, 20W O/P this aint too much of a problem--But Hundreds of watts--Worth keeping in mind, a Class A dissipates its Maximum heat when there is No signal input/O/P..?!!
Perhaps I should ask, Why Class A and why such a high power output TOGETHER??
Yes, Absolutely Anything is possible, But Impractical or just stupidly expensive, when there are other more viable options....
When looking at OTL scenarios, worth remembering that high output powers are Very hard on the valve --They are having to pass currents well in excess of the maker ratings, and while they will do this happily for a while, using an OTL for extended periods at its maximum O/P will render the valves in the power-stages useless in a matter of several hours..
I have an OTL, a Circlotron running a pair of 6C33C per channel. This will produce 56W into 8 ohm loads.--..Sounds just great--even at high O/P--but,-- But I wouldnt expect it to last for long at full power....
IF you do end up using 6C33 worth burning them in well (24 hours just heater, then 400mA at 60-80V for 3-4 hours) before even considering 'Matching', Dont bother static matching, better to match Gm, and trim each valve for current independently in the amp--always worked for me....
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.