Dear all,
for my job I have to experiment with a particular tube power amplifier, it will be used for a frequency range of 20Hz 300kHz and it should be rated 600W.
The particular thing is that the output is at a constant voltage of 240V (sine of course).
Now, I'm thinking of using KT150 in class B since distortion is not a key factor.
600W @ 240VRMS means a load of 96ohm and a current of 2.5ARMS.
considering a transformer efficiency of 80%, the power in the primary should be Pp=1.2*600W = 720W.
considering the class B efficiency around 70% I can extimate the power dissipated by the tubes as 216W so using four tubes they should be able to dissipate 54W each.
Now, I'd like to use 700Vdc, in this way the primary of the transformer will see 1400Vpk swing, the RMS primary voltage is about 992V, the formula for the primary power can be written as Ppri=(Vpri)^2/Zpri and so Zpri=1370ohm. Stupid mistake! the transformer will se 496VRMS and Zpri = 342ohm.
My doubt is that in class B operation the secondary load will be reflected on one tube per cycle so each tube will see Zaa/4, should I use a Zaa=4*Zpri?
And also, when one tube is switched off and the other is conducting, on the switched off coil will appear an additive potential due to the 1:1 transformer made by the two primary windings? Does my tubes should be rated for max anode voltage of 1400V?
thank you very much
best regards
luix
for my job I have to experiment with a particular tube power amplifier, it will be used for a frequency range of 20Hz 300kHz and it should be rated 600W.
The particular thing is that the output is at a constant voltage of 240V (sine of course).
Now, I'm thinking of using KT150 in class B since distortion is not a key factor.
600W @ 240VRMS means a load of 96ohm and a current of 2.5ARMS.
considering a transformer efficiency of 80%, the power in the primary should be Pp=1.2*600W = 720W.
considering the class B efficiency around 70% I can extimate the power dissipated by the tubes as 216W so using four tubes they should be able to dissipate 54W each.
Now, I'd like to use 700Vdc, in this way the primary of the transformer will see 1400Vpk swing, the RMS primary voltage is about 992V, the formula for the primary power can be written as Ppri=(Vpri)^2/Zpri and so Zpri=1370ohm. Stupid mistake! the transformer will se 496VRMS and Zpri = 342ohm.
My doubt is that in class B operation the secondary load will be reflected on one tube per cycle so each tube will see Zaa/4, should I use a Zaa=4*Zpri?
And also, when one tube is switched off and the other is conducting, on the switched off coil will appear an additive potential due to the 1:1 transformer made by the two primary windings? Does my tubes should be rated for max anode voltage of 1400V?
thank you very much
best regards
luix
Last edited:
With 4 tubes perhaps 300W.For more, higher voltage or more tubes.
Forget 70% with tubes, more like 50%.With triode mode even less.
Yes, the tubes see allmost double the supply voltage.
You could consider six EL509 (or PL509) in a serial PP without output transfo with a power supply of +500 and -500volts (rated 1A).No transformer loss and no peak voltage.
Mona
How is this one grasp you? Transformer Zpri/Zsec is about 6 / 100 ohms
Attachments
Last edited:
i hope PRR chimes in and talk sense with you.......
this is a 300 watt monoblock done by Patrick Turner, i suggest that you study it
carefully....300W-Monobloc
IMHO, paralleling a lot of tubes to get to 300 watts is not a job to be taken on by a newbie...
u would look for tubes that can do that in pairs....4-250 comes to mind....
this is a 300 watt monoblock done by Patrick Turner, i suggest that you study it
carefully....300W-Monobloc
IMHO, paralleling a lot of tubes to get to 300 watts is not a job to be taken on by a newbie...
u would look for tubes that can do that in pairs....4-250 comes to mind....
Attachments
You can use 8 pairs of KT150's in class AB without stressing the tubes. This is what the Audio Research Reference 750 SE power amp uses and can provide 750W with low distortion and up to 850W at onset of clipping.
Tubes can swing above B+ supply. Zpri is Zaa. No need to use tubes specified for 1400V, the supply is at the center tap.
I think your biggest problem is how to get 600W from 20Hz to 300 KHz. Never seen a transformer like that.
Tubes can swing above B+ supply. Zpri is Zaa. No need to use tubes specified for 1400V, the supply is at the center tap.
I think your biggest problem is how to get 600W from 20Hz to 300 KHz. Never seen a transformer like that.
Last edited:
20 hz to 300 khz ... this is asking for trouble.
A transformer able to deliver 20hz at 3000 watt will certainly be unable to push more than 10khz at the same power.
A 300khz transformer able to deliver 3000 watts will certainly be unable to push 20hz at the same power.
Doubling , // amps and speakers, 600 watt requirement easy become a 300 watt, then a 150 watts then a 75 watts and will sound better.
A transformer able to deliver 20hz at 3000 watt will certainly be unable to push more than 10khz at the same power.
A 300khz transformer able to deliver 3000 watts will certainly be unable to push 20hz at the same power.
Doubling , // amps and speakers, 600 watt requirement easy become a 300 watt, then a 150 watts then a 75 watts and will sound better.
> 20Hz 300kHz
That is hardly likely. Talk to a very experienced winder.
Oh, wait: 300 Hz ?? A shaker-table? That would be do-able. Transformer loss at real 20Hz would be large, but 80% gives a fair allowance. More tubes will cover transformer shortcomings (to a point).
Or an "AC wall outlet" to test all cases, except 16Hz was in use far into the 20th century, and aircraft did (may still do) sport 400Hz AC.
> extimate the power dissipated by the tubes as 216W
This work has been DONE. The highest number on 6550/KT88 factory datasheets is 100 Watts (few %THD). 720W means seven pairs. Even if you assume a KT150 is 1.5 times a 6550/KT88 (not proven), that is five pair. Not two pair ("four tubes").
Moving from "class AB" to "class B" reduces the idle dissipation but does not increase the maximum output.
Relaxing the "distortion", but still expecting "sine-like" output, may get another 20% or a bit more. I would just leave that in reserve.
As the KT150 is very expensive, not second-sourced, and uncertain long-term availability, and tube costs will NOT be a deciding factor in this amplifier, and coming up short may be very disappointing, I would pencil-in eight pair of 6550/KT88 and then address the other large problems.
I used to tour with ten pair 8417 (hi-gain 6550) as five 150W channels. So the same as your rig, but as three milk-crates, and 50Hz-10KHz power bandwidth seemed not just acceptable but exemplary. (Response extended past 20KHz but not at full 150W/quad).
This is actually an "easy" (well, much easier) thing in transistors. 720W of tranny-amp is a pawnshop item. That much power at 100V is available in the Installed Sound (ball-parks) catalogs. 1KVA 100V:240V transformer is not exotic. Likely OK 50Hz-5KHz. But past that is really another world. Iron gets to be a drag. If you truly want 300KHz, air-core works.
You could also rig a full-bridge off a 340V DC supply and pull 240V RMS across the two outputs. This could even be done without any transformers (but talk of transformerless power supplies is banned here, with reason). 300KHz used to be out-of-sight for high-power linear transistors. In MOSFETs, not so much, but it needs a well-designed driver. "zero" to 300Hz is no-sweat.
> when one tube is switched off and
This design/build really should be left to someone with experience. Someone who knows we "always" run power tube B+ right near rating.
Yes, for that assumed 700V supply, the plates will swing from 700 to 100 to 1300 and back. The "Plate Voltage rating" *knows* you will be swinging this way. While there may be more damage at the 1300V peak, there is less damage at the 100V dip. It "averages out". If your plate supply voltage is the rated spec, it will be fine. 6550 was rated to stand 600V B+ in demanding broadcast (video tape motor servo) work. KT88 may have a higher number. Today I am sure both are the same guts in different bottles and using the highest-Volt plate-stuff available.
(For symmetric signals. TV sweep tubes ran highly asymmetric swings, 250V supply, 90V dips, and peaks to 2000V. Even at that, "audio power" tubes worked OK, except the octal base is not good for that much. 6BG6 _is_ a 6L6GC with plate moved to top-cap. (Actually the 6BG6 is a re-think 6L6 with cap, and the 6L6GC is a 6BG6 without top cap.))
That is hardly likely. Talk to a very experienced winder.
Oh, wait: 300 Hz ?? A shaker-table? That would be do-able. Transformer loss at real 20Hz would be large, but 80% gives a fair allowance. More tubes will cover transformer shortcomings (to a point).
Or an "AC wall outlet" to test all cases, except 16Hz was in use far into the 20th century, and aircraft did (may still do) sport 400Hz AC.
> extimate the power dissipated by the tubes as 216W
This work has been DONE. The highest number on 6550/KT88 factory datasheets is 100 Watts (few %THD). 720W means seven pairs. Even if you assume a KT150 is 1.5 times a 6550/KT88 (not proven), that is five pair. Not two pair ("four tubes").
Moving from "class AB" to "class B" reduces the idle dissipation but does not increase the maximum output.
Relaxing the "distortion", but still expecting "sine-like" output, may get another 20% or a bit more. I would just leave that in reserve.
As the KT150 is very expensive, not second-sourced, and uncertain long-term availability, and tube costs will NOT be a deciding factor in this amplifier, and coming up short may be very disappointing, I would pencil-in eight pair of 6550/KT88 and then address the other large problems.
I used to tour with ten pair 8417 (hi-gain 6550) as five 150W channels. So the same as your rig, but as three milk-crates, and 50Hz-10KHz power bandwidth seemed not just acceptable but exemplary. (Response extended past 20KHz but not at full 150W/quad).
This is actually an "easy" (well, much easier) thing in transistors. 720W of tranny-amp is a pawnshop item. That much power at 100V is available in the Installed Sound (ball-parks) catalogs. 1KVA 100V:240V transformer is not exotic. Likely OK 50Hz-5KHz. But past that is really another world. Iron gets to be a drag. If you truly want 300KHz, air-core works.
You could also rig a full-bridge off a 340V DC supply and pull 240V RMS across the two outputs. This could even be done without any transformers (but talk of transformerless power supplies is banned here, with reason). 300KHz used to be out-of-sight for high-power linear transistors. In MOSFETs, not so much, but it needs a well-designed driver. "zero" to 300Hz is no-sweat.
> when one tube is switched off and
This design/build really should be left to someone with experience. Someone who knows we "always" run power tube B+ right near rating.
Yes, for that assumed 700V supply, the plates will swing from 700 to 100 to 1300 and back. The "Plate Voltage rating" *knows* you will be swinging this way. While there may be more damage at the 1300V peak, there is less damage at the 100V dip. It "averages out". If your plate supply voltage is the rated spec, it will be fine. 6550 was rated to stand 600V B+ in demanding broadcast (video tape motor servo) work. KT88 may have a higher number. Today I am sure both are the same guts in different bottles and using the highest-Volt plate-stuff available.
(For symmetric signals. TV sweep tubes ran highly asymmetric swings, 250V supply, 90V dips, and peaks to 2000V. Even at that, "audio power" tubes worked OK, except the octal base is not good for that much. 6BG6 _is_ a 6L6GC with plate moved to top-cap. (Actually the 6BG6 is a re-think 6L6 with cap, and the 6L6GC is a 6BG6 without top cap.))
Last edited:
Thank you all for your reply, the amp I'm going to design (if it is a feasible idea) have not an exact purpose, we want to see if tubes have advantages in robustness against BJT and MOSFET.
Transistors usually blow up in a microsecond under peak stress conditions instead I believe that tubes have a high thermal inertia so it is much slower to blow up and it will be easy to protect the output stage from damange.
What are you saying to me is that in any case transformer cannot have a so wide frequency response, have I understandt well?
It would be better to use an OTL topology?
The problem using small tubes like 6550 is that I need a lot of matched pieces so the price for building a prototype will grow up too much, I prefer to use bigger tubes but in small quantities.
What about of using RF power tetrodes like 4CX250?
PS I'm not new in tube based amplifier, I've designed and build amps from 10W single ended to 100W PP 4x6550, this is the first time I'm going in so high power direction.
Transistors usually blow up in a microsecond under peak stress conditions instead I believe that tubes have a high thermal inertia so it is much slower to blow up and it will be easy to protect the output stage from damange.
What are you saying to me is that in any case transformer cannot have a so wide frequency response, have I understandt well?
It would be better to use an OTL topology?
The problem using small tubes like 6550 is that I need a lot of matched pieces so the price for building a prototype will grow up too much, I prefer to use bigger tubes but in small quantities.
What about of using RF power tetrodes like 4CX250?
PS I'm not new in tube based amplifier, I've designed and build amps from 10W single ended to 100W PP 4x6550, this is the first time I'm going in so high power direction.
- Status
- Not open for further replies.