Hi everyone. Bit of a funny question. I want to make a Class A amp from a retired amp chassis and transformer and would like to know which of the classic circuits would work. The amp had a reasonably hefty transformer for a Class AB. According to the back of the amp it is rated to 230W (240V). I'm assuming then that it is roughly 230VA. I measured the secondary AC output voltage as 32V.
I have limited knowledge here but all of the Class A designs that I have looked at - Hiraga le class A, Hiraga "le monstre", Death of Zen, JLH, various Pass Labs - they all seem to use DC voltage of anywhere from 12V to 27V. What is the issue with using higher voltage? After rectification I might be looking at abount 45V DC (32.5V × 1.4)? Why cant these amps just be set to lower amps and higher voltage? Do they rely on high cirrent to be stable?
I don't need high output - 8 or 10 watts would be fine. The amp has a reasonable heat sink and good ventilation i would be augmenting this with 1-2 fans.
Using what i have will give me a head start when it comes to quality passive components if I can make it work.
Thanks.
I have limited knowledge here but all of the Class A designs that I have looked at - Hiraga le class A, Hiraga "le monstre", Death of Zen, JLH, various Pass Labs - they all seem to use DC voltage of anywhere from 12V to 27V. What is the issue with using higher voltage? After rectification I might be looking at abount 45V DC (32.5V × 1.4)? Why cant these amps just be set to lower amps and higher voltage? Do they rely on high cirrent to be stable?
I don't need high output - 8 or 10 watts would be fine. The amp has a reasonable heat sink and good ventilation i would be augmenting this with 1-2 fans.
Using what i have will give me a head start when it comes to quality passive components if I can make it work.
Thanks.
Hi.
32 volts for hiraga is high, you could consider to go regulated that would lower volts to match hiraga and also clean up the dc, hiraga needs clean dc otherwise there will be audible hum.
A ripple eater like jlh would also lower voltage a few volts.
Don't forget heavy heatsinking.
The 32 volts mentioned are probably unloaded, loaded with 2-3 amps it should be a tad lower
32 volts for hiraga is high, you could consider to go regulated that would lower volts to match hiraga and also clean up the dc, hiraga needs clean dc otherwise there will be audible hum.
A ripple eater like jlh would also lower voltage a few volts.
Don't forget heavy heatsinking.
The 32 volts mentioned are probably unloaded, loaded with 2-3 amps it should be a tad lower
You could set a classA amp to run at higher voltage and lower current but it'll only stay in classA with a higher impedance load than a normal 8ohm speaker. If you have an output transformer you can achieve this but otherwise you're pretty much stuck with the current and volts which correspond to a standard 8ohms load. Given you're fine with ~10watts then for a single supply, you'll need about 27V with an 8ohm load.
15-20W Class AB Audio Amplifier
A design with class-A performance but reduced thermal dissipation
by J. L. Linsley Hood
(Wireless World, July 1970)
The Class-A Amplifier Site - JLH Class-AB Amplifier
A design with class-A performance but reduced thermal dissipation
by J. L. Linsley Hood
(Wireless World, July 1970)
The Class-A Amplifier Site - JLH Class-AB Amplifier
What do you have in mind? Class A - Single Rail voltage or +- GND?
you can achieve higher output with higher voltage, what of course will pull higher HEAT DISSIPATION with it. and we are talking lots of heat. Consider the double Amount of HEAT as you have Output power. And I believe this is also true for Current Source A. As the Transistor is always switched on. So big Heatsinks are a MUST and it sure will heat up your Listening room.. but Sound quality is above all other..
If you like to get an decent Class A which is acceptable, then go for one of Nelson Pass F Series,, the latest I was listening to was F7 and it impresses because of using only few Elements and the Sound is also Class A.. I think you may even be able to order the DIY Kit.. using per side 4 transistors 2 Jfet and 2 Mosfets, 7 Resistors per side only. Then also PSU of that F7 is made to cope with the Amp that it can deliver it's Power without Failure..
I think this is your Best Choice, designing your own is also an option but I can assure you it will use up time and money where otherwise all is set already for you..BTW the F7 does not heat up that much, you will always be able to touch the heat sinks..
Hope this helps.
Regards Chris
you can achieve higher output with higher voltage, what of course will pull higher HEAT DISSIPATION with it. and we are talking lots of heat. Consider the double Amount of HEAT as you have Output power. And I believe this is also true for Current Source A. As the Transistor is always switched on. So big Heatsinks are a MUST and it sure will heat up your Listening room.. but Sound quality is above all other..
If you like to get an decent Class A which is acceptable, then go for one of Nelson Pass F Series,, the latest I was listening to was F7 and it impresses because of using only few Elements and the Sound is also Class A.. I think you may even be able to order the DIY Kit.. using per side 4 transistors 2 Jfet and 2 Mosfets, 7 Resistors per side only. Then also PSU of that F7 is made to cope with the Amp that it can deliver it's Power without Failure..
I think this is your Best Choice, designing your own is also an option but I can assure you it will use up time and money where otherwise all is set already for you..BTW the F7 does not heat up that much, you will always be able to touch the heat sinks..
Hope this helps.
Regards Chris
Many AB class amplifiers can be converted to class A. Massive heatsink (or active cooling) are needed. Large capacitors, or a stabilized power supply, or SMPS.
There is an exact relationship between power, load resistance, supply voltage and current consumption for class A. Also, thermal power.
Le Ultra-low Distortion Class-A Amplifier de L. Nelson-Jones : Wireless World (03/1970)
nj10w.pdf
nj10w.pdf - Google Drive
There is an exact relationship between power, load resistance, supply voltage and current consumption for class A. Also, thermal power.
Le Ultra-low Distortion Class-A Amplifier de L. Nelson-Jones : Wireless World (03/1970)
nj10w.pdf
nj10w.pdf - Google Drive
I think I have a very basic understanding. The current has to be set high enough for the transistors to be "switched on" all of the time? At a set current, higher voltage means more power and more heat lost. Does that essentially mean that my transformer with limited VA rating and relatively high voltage secondary mean that it isn't very suited to a Class A amp, even of low power output?
So would i be right in saying that i either need to reduce the voltage via regulated supply (giving off heat) or run the transistors at higher voltage (giving off heat). Either way starting with a higher voltage will reduce efficiency?
My plan is literally to use this amp during our winter to warm the living room! If I am paying for heating then I may as well get some sweet class A sound out of it. We do have young kids though and i don't want it to be a burns risk if they manage to touch it.
My plan is literally to use this amp during our winter to warm the living room! If I am paying for heating then I may as well get some sweet class A sound out of it. We do have young kids though and i don't want it to be a burns risk if they manage to touch it.
That indeed qualifies for (enough) basic understanding..
What you mentioned about the max. VA and de Vac I'd suspect the transformer a bit under dimension-ed for class A operation even at 8 Ohm load.
But you can always decide to have the first watts in class A and the rest in AB, fully using the capabilities of the transformer. In which case you can set the current between 500mA and 1A.. You won't have to sink too much heat too.🙂
Still not clear if the transformer is 32-0-32 Vac or 32Vac with center tap, that would be 16-0-16 Vac.
Suppose 32-0-32 Vac, to supply two channels (left and right).
The rectified Vac will not become the optimistic ( * 1.4), usually it's 1.2 (1.3 at best) in class AB amp's, but due to iron and copper losses count 1.0 to 1.1 as factor, yielding in +/- 37Vdc rails. That sums up to 74Vdc total. Maximum drain from the transformer is 230VA / 2*32Vac = 3.5A (includes some losses), but rectified will stay at 3.5A as no more current can flow from the transformer or the Vac dips considerable.
Power from the supply available: 74 * 3.5 = 260W (dc).
Two channels, 130W per channel.
Reserve 10W for the front stages, leaves 120W for the class A end stage.
At +/- 37Vdc, that's 1.6A bias.
Power dissipation without program ('audiable delight'): 120W
Check the thermal specs of the heatsinks before continuing. Forced air flow can help, but tornado force is not much better then a stiff breeze. Both are noisy and use 'transformerpower' too.
Theoretical maximum efficiency of a class A amp is 25%, one might say 30W can be reached.
However...
Bias is 1.6A. That's also the peak output current (to remain in class A, no shut off).
On a 8Ω load this gives (P = Iexp2 * R) 20W^, 12.8V^.
Listening to sinewaves (a bit boring after some time), the effective values become ( both * 0.7) 9V~/1.1A~ which is 10W at 8Ω.
One cannot drive lower impedances then 8Ω, because there is no current available anymore from the transformer. If you have to drive say 4Ω loads, you have to start with 5W/8Ω up to 10W/4Ω.
So, 10W class A is a goal to achieve, but given the numbers another transformer with less voltage and more current would be recommended for such a project.
Impedences of loudspeakers are not 'fixed', that is to say that the dynamic respons on a almost perfect voltage source (what an amplifier should be) is more demanding then the static measurements predicts. With static impedances going up and down, dynamic imedances requires a serious marging from this not so perfect as expected voltage source, the involved amplifier. And funny enough, the lowest impedance is often near the most energetic part of the program (the said delight), ranging from some 100 to 300 Hz.
Recalculate everything if the assumption of 32-0-32Vac differs from the actual output from the transformer taps.
Suppose 32-0-32 Vac, to supply two channels (left and right).
The rectified Vac will not become the optimistic ( * 1.4), usually it's 1.2 (1.3 at best) in class AB amp's, but due to iron and copper losses count 1.0 to 1.1 as factor, yielding in +/- 37Vdc rails. That sums up to 74Vdc total. Maximum drain from the transformer is 230VA / 2*32Vac = 3.5A (includes some losses), but rectified will stay at 3.5A as no more current can flow from the transformer or the Vac dips considerable.
Power from the supply available: 74 * 3.5 = 260W (dc).
Two channels, 130W per channel.
Reserve 10W for the front stages, leaves 120W for the class A end stage.
At +/- 37Vdc, that's 1.6A bias.
Power dissipation without program ('audiable delight'): 120W
Check the thermal specs of the heatsinks before continuing. Forced air flow can help, but tornado force is not much better then a stiff breeze. Both are noisy and use 'transformerpower' too.
Theoretical maximum efficiency of a class A amp is 25%, one might say 30W can be reached.
However...
Bias is 1.6A. That's also the peak output current (to remain in class A, no shut off).
On a 8Ω load this gives (P = Iexp2 * R) 20W^, 12.8V^.
Listening to sinewaves (a bit boring after some time), the effective values become ( both * 0.7) 9V~/1.1A~ which is 10W at 8Ω.
One cannot drive lower impedances then 8Ω, because there is no current available anymore from the transformer. If you have to drive say 4Ω loads, you have to start with 5W/8Ω up to 10W/4Ω.
So, 10W class A is a goal to achieve, but given the numbers another transformer with less voltage and more current would be recommended for such a project.
Impedences of loudspeakers are not 'fixed', that is to say that the dynamic respons on a almost perfect voltage source (what an amplifier should be) is more demanding then the static measurements predicts. With static impedances going up and down, dynamic imedances requires a serious marging from this not so perfect as expected voltage source, the involved amplifier. And funny enough, the lowest impedance is often near the most energetic part of the program (the said delight), ranging from some 100 to 300 Hz.
Recalculate everything if the assumption of 32-0-32Vac differs from the actual output from the transformer taps.
Last edited:
From: Wireless World, March 1970.
Efficiencies above 25% are class A/AB rated (max up to some 80%), not class A.
Just (try to) pull more current from the transformer then specified.
Your retired amp, I'm curious to what brand/model it is.. pics are welcome too, especially nudes.
Output voltage swing (pk-pk) = Vcc - { Vce.sat3 + Vbe4 + Vce.sat6 + (I + Î) * R11}
Also, power output (sinewave) = (output voltage swing)
2 pk-pk / (8 * Rload)
Since Vout (r.m.s.) = Vp-p / 2
-
Vout (pk-pk) =
!"#$%&')(
%*+),+-),+
Vcc = ./0!123 456
17
4)89:<;!=>?
@ A593 + Vbe4 + Vce.sat6 + (I + Î) * R11, minimum.
The standing current must exceed Vpk-pk / (4 * Rload) in order to achieve the required voltage swing,
and for its satisfactory safety margin it should exceed Vcc / (4 * Rload).
Taking typical values for the circuit given using an 8 ohm load, and 10 W output level, we get:
Vcc = BCDEGFHE)I JKHFMLI EHFHE
I KHFGNE)I OEHFHE
I POQ)RE
I KCSTJU!V
Imin = 28 / (4 * 8) = 875 mA (a value of 900 mA being actually used.)
For a 3 ohm load and 10 W output we get figures of 19.5 V for Vcc, and 1.63 A for Imin. (Total power
31.8 W, 31.5% efficient).
For a 15 ohm load and 10 W output we get figures of 36V for Vcc, and 0.6 A for Imin. (Total power
21.5 W, 46.4% efficient).
From these figures it is apparent that the rise in Vce.sat and Vbe figures with the current used in a 3 ohm
amplifier seriously reduces the overall efficiency. In the case of the 15 ohm load on the other hand, the
efficiency is not far short of the theoretically possible figure of 50% for a class A stage. The efficiency of
the 8 ohm stage is 39.8%.
Details of value changes for 3 ohm, and 15 ohm circuits are given with the constructional details below.
Better to watch the script (some distortion). Calculation for 10 watt 3: 8: 15 ohms)))
Also, power output (sinewave) = (output voltage swing)
2 pk-pk / (8 * Rload)
Since Vout (r.m.s.) = Vp-p / 2
-
Vout (pk-pk) =
!"#$%&')(
%*+),+-),+
Vcc = ./0!123 456
17
4)89:<;!=>?
@ A593 + Vbe4 + Vce.sat6 + (I + Î) * R11, minimum.
The standing current must exceed Vpk-pk / (4 * Rload) in order to achieve the required voltage swing,
and for its satisfactory safety margin it should exceed Vcc / (4 * Rload).
Taking typical values for the circuit given using an 8 ohm load, and 10 W output level, we get:
Vcc = BCDEGFHE)I JKHFMLI EHFHE
I KHFGNE)I OEHFHE
I POQ)RE
I KCSTJU!V
Imin = 28 / (4 * 8) = 875 mA (a value of 900 mA being actually used.)
For a 3 ohm load and 10 W output we get figures of 19.5 V for Vcc, and 1.63 A for Imin. (Total power
31.8 W, 31.5% efficient).
For a 15 ohm load and 10 W output we get figures of 36V for Vcc, and 0.6 A for Imin. (Total power
21.5 W, 46.4% efficient).
From these figures it is apparent that the rise in Vce.sat and Vbe figures with the current used in a 3 ohm
amplifier seriously reduces the overall efficiency. In the case of the 15 ohm load on the other hand, the
efficiency is not far short of the theoretically possible figure of 50% for a class A stage. The efficiency of
the 8 ohm stage is 39.8%.
Details of value changes for 3 ohm, and 15 ohm circuits are given with the constructional details below.
Better to watch the script (some distortion). Calculation for 10 watt 3: 8: 15 ohms)))
L.Nelson-Jones: At the same time, as JLH1969 in the same magazine WW 1970.
There are also suggestions for design.
There are also suggestions for design.