I just scored an extremely well regulated 250W power supply that will yield 28.43VDC unloaded and is rated up to 10 Amps (with forced cooling). I also have loads of NPN power transistors. I am thinking a marriage of the two is a great idea. I like the idea of quasi-comp. output stages, and classA is not my favorite.
So: I am hoping to build a quasi-complementary output, 2 or 2.1 channels class B or AB amplifier. With that sort of current available, what sort of power should I shoot for, per channel?
But more importantly, what would YOU do?
So: I am hoping to build a quasi-complementary output, 2 or 2.1 channels class B or AB amplifier. With that sort of current available, what sort of power should I shoot for, per channel?
But more importantly, what would YOU do?
linear or switcher ?
if it's a linear supply and "extremely well regulated" you may want to get rid of the regulation -- there is a lot of overhead in a commercially designed supply which can be dispensed with -- as an example,the HP 6129C power supply, designed to produce 50V @5amps has secondary rails of +/- 75V. It really is an amplifier in disguise.
disconnect the rectifiers and examine the transformer secondary voltages -- then report back.
switchers are different -- but the transformer can often times be unsoldered and rewound for the secondary voltage you need -- of course you will have to modify the control loop, etc.
if it's a linear supply and "extremely well regulated" you may want to get rid of the regulation -- there is a lot of overhead in a commercially designed supply which can be dispensed with -- as an example,the HP 6129C power supply, designed to produce 50V @5amps has secondary rails of +/- 75V. It really is an amplifier in disguise.
disconnect the rectifiers and examine the transformer secondary voltages -- then report back.
switchers are different -- but the transformer can often times be unsoldered and rewound for the secondary voltage you need -- of course you will have to modify the control loop, etc.
Jackinnj:
Linear regulation, I think. I haven't reverse engineered it yet and the model number escapes me at the moment. At a glance, it's an opamp driving a TIP31 to a VRef, driving a 2n3055, driving six! 2n3773s. TO-3 Rectifiers apparently. The main filter caps (3x13000uF ! ) are rated for 50V, I haven't measured the secondaries directly yet. It's rated to be 60% efficient... and 0.03% regulation. The noise unloaded was 0.04mVAC. That level of regulation really makes me happy... it's even got a smashingly good spec. with a sudden 50% load increase. But the efficiency... oh boy. I'll have to think about that. The transformer is rated for about 600+VA input, so maybe 500VA out wouldn't be out of order... well now there's another bug in my ear...
If I took out the regulation, there's a half-dozen 150W output devices right there. Oh boy, more NPNs
Blmn:
I was thinking of "or less" myself!... So let's get the calculator... 28V, rails, bridged into 2 ohms... would be... 
Then a couple of low-power channels for stereo.
PaulB:
Chances of getting another of these for free range from slim to none. It's a $180 supply new AFAIK, and this one was apparently dropped and retired due to a dented chassis. ClassA is all good except for the Texas summers. I don't need a 500W heater on my bench when it's 110*F outside!
Linear regulation, I think. I haven't reverse engineered it yet and the model number escapes me at the moment. At a glance, it's an opamp driving a TIP31 to a VRef, driving a 2n3055, driving six! 2n3773s. TO-3 Rectifiers apparently. The main filter caps (3x13000uF ! ) are rated for 50V, I haven't measured the secondaries directly yet. It's rated to be 60% efficient... and 0.03% regulation. The noise unloaded was 0.04mVAC. That level of regulation really makes me happy... it's even got a smashingly good spec. with a sudden 50% load increase. But the efficiency... oh boy. I'll have to think about that. The transformer is rated for about 600+VA input, so maybe 500VA out wouldn't be out of order... well now there's another bug in my ear...
If I took out the regulation, there's a half-dozen 150W output devices right there. Oh boy, more NPNs
Blmn:
I was thinking of "or less" myself!
... So let's get the calculator... 28V, rails, bridged into 2 ohms... would be... Then a couple of low-power channels for stereo.
PaulB:
Chances of getting another of these for free range from slim to none. It's a $180 supply new AFAIK, and this one was apparently dropped and retired due to a dented chassis. ClassA is all good except for the Texas summers. I don't need a 500W heater on my bench when it's 110*F outside!
The regulation is so good because it's based on an LM723CN regulator. That regulator is useable for positive or negative voltages and inspires me to look for a higher-voltage type of similarly-great regulator for the *big* transformer I have for a dual supply for The Lab. Anyway, the LM723 has a max input voltage of 40VDC so there is potentially up to that much but likely less than 35-38V.
for a dual supply for The Lab. Anyway, the LM723 has a max input voltage of 40VDC so there is potentially up to that much but likely less than 35-38V.
I know at this stage of design it's all numbers grabbed from the air... but where did you come up with 50Vpp? I think you are maybe thinking of bridged "normal" amps with 28V rails but please confirm this for me.
And, how do 50V and 10A combine to make 160W? (really, no kidding what do you mean by this? )
And, how do 50V and 10A combine to make 160W?
(really, no kidding what do you mean by this? )
Yeap, I´m thinking about bridging the amp, so if you consider Vcc equals 28.5 you can consider (2 x 28.5v - some losses) over the load (peak to peak). So, just do the math and you will find near 17.7 v RMS over the load.
If you consider near 55% the efficiency of the amp, over 28.5vcc you will find near 10A of current.
regards,
If you consider near 55% the efficiency of the amp, over 28.5vcc you will find near 10A of current.
regards,
OK, this is good stuff...
50Vpp =approx. 35VRMS... a 4 ohm load is 2 to each side...
But then I am nearly out of current! an ampere or two only left! (VERY round numbers here! )
I guess that's still, say, uh, about 10-25W left over for the other two channels, each and a little headroom.
25x2 + 160x1... not shabby. I will revive this thread next week after I measure the transformer's secondaries...the power supply is in my garage and the "safety" setup for testing is here at work. wasting that 40% of the transformer's output is starting to bother me more and more, especially putting some possible ballpark numbers on power out of the amplifier
Blmn, thanks alot for your input. I would appreciate it if you will stick this out with me.
50Vpp =approx. 35VRMS... a 4 ohm load is 2 to each side...
But then I am nearly out of current!
an ampere or two only left! (VERY round numbers here! )I guess that's still, say, uh, about 10-25W left over for the other two channels, each and a little headroom.
25x2 + 160x1... not shabby. I will revive this thread next week after I measure the transformer's secondaries...the power supply is in my garage and the "safety" setup for testing is here at work. wasting that 40% of the transformer's output is starting to bother me more and more, especially putting some possible ballpark numbers on power out of the amplifier
Blmn, thanks alot for your input. I would appreciate it if you will stick this out with me.
No,
We are talking about peak to peak, not peak values.
If you have available 28.5 volts, in a bridge you will put over the load more or less 25 volts per side of the bridge (I´m being a little bit conservative here).
So, the amp will deliver near 50 volts peak to peak to the load or 50/(2*sqrt(2))=17.7 volts RMS.
and 17.7^2/2= 157W continuous
or near 80W at 4 ohms etc...
Regards
We are talking about peak to peak, not peak values.
If you have available 28.5 volts, in a bridge you will put over the load more or less 25 volts per side of the bridge (I´m being a little bit conservative here).
So, the amp will deliver near 50 volts peak to peak to the load or 50/(2*sqrt(2))=17.7 volts RMS.
and 17.7^2/2= 157W continuous
or near 80W at 4 ohms etc...
Regards
ohhhhh. nevermind me, this will be my first scratch-built amplifier design, and I've never had to do all these calculations for myself.
But don't worry, this is not a total hand-holding expedition for you! I am also very eager to learn to do the next one by myself (or with just a few hints ) so I do some research when it becomes clear (to me, at long last) that I don't know what I am talking about...
I found this right off, and of course you are right.
So there is also potential for 8.8VRMS out of a non-bridge channel yes? =maybe 8, 16 or 30-odd (approx) watts, depending on load... which is not too shabby for just the stereo part.
But don't worry, this is not a total hand-holding expedition for you! I am also very eager to learn to do the next one by myself (or with just a few hints
) so I do some research when it becomes clear (to me, at long last) that I don't know what I am talking about...I found this right off, and of course you are right.
So there is also potential for 8.8VRMS out of a non-bridge channel yes? =maybe 8, 16 or 30-odd (approx) watts, depending on load... which is not too shabby for just the stereo part.
Yes,
There is, but, in an non bridged configuration you have to deal with low impedances to achieve more power.
Anyway, it´s possible and I think the best way for you in this case is working on chip amps. Easy to build and in the this range of power and voltages. A little research in the forum will give you lots of information to start ok.
The Rodd Eliot´s site would be fine for the basic too (a google search will find it)
Best regards,
There is, but, in an non bridged configuration you have to deal with low impedances to achieve more power.
Anyway, it´s possible and I think the best way for you in this case is working on chip amps. Easy to build and in the this range of power and voltages. A little research in the forum will give you lots of information to start ok.
The Rodd Eliot´s site would be fine for the basic too (a google search will find it)
Best regards,
Excellent advice...
But using either a chipamp or a pre-made design would be rather not the point...
I have this transformer begging to be used, and a ton of components to use with it. I want to make *my own* design...not make a direct copy from someone else. Anyhow, the biggest hurdle for me seems to have been overcome, in selecting the ouput topology. I am thinking I will go one of two ways with the front end... either D. Self's "blameless", suited to my available power, or else something closer to "el cheapo" of the first variety. Then again, I may just use an opamp and avoid the question altogether!
I put up those last numbers assuming about 2, 4 or 8 ohms load.
As for being entirely pigheaded about not using someone else's design; if I ever get around to it, the "big" transformer will go into a P68 high-power version, suited to the just-over 70VDC I would get out of it... maybe two, since it's all of a 1500VA transformer! If not, it will be a Very High Power, variable voltage, variable current, with short circuit protection and fold-back current limiting, bench power supply. Can you say 15V 90A? (etc. )
/edit: oh yeah! It looks like the secondries are 34.6 and 37.3VAC, hopefully with one being an extension of the other (longer wire, another terminal) or else I will have to start thinking about *really* high power... anyhow, It seems with this set of voltages I can go anywhere from about 8W to around the transformer's rated 660!
The regulator section is looking destined for the scrap bin.
Thanks!
But using either a chipamp or a pre-made design would be rather not the point...
I have this transformer begging to be used, and a ton of components to use with it. I want to make *my own* design...not make a direct copy from someone else. Anyhow, the biggest hurdle for me seems to have been overcome, in selecting the ouput topology. I am thinking I will go one of two ways with the front end... either D. Self's "blameless", suited to my available power, or else something closer to "el cheapo" of the first variety. Then again, I may just use an opamp and avoid the question altogether!
I put up those last numbers assuming about 2, 4 or 8 ohms load.
As for being entirely pigheaded about not using someone else's design; if I ever get around to it, the "big" transformer will go into a P68 high-power version, suited to the just-over 70VDC I would get out of it... maybe two, since it's all of a 1500VA transformer! If not, it will be a Very High Power, variable voltage, variable current, with short circuit protection and fold-back current limiting, bench power supply. Can you say 15V 90A? (etc. )
/edit: oh yeah! It looks like the secondries are 34.6 and 37.3VAC, hopefully with one being an extension of the other (longer wire, another terminal) or else I will have to start thinking about *really* high power... anyhow, It seems with this set of voltages I can go anywhere from about 8W to around the transformer's rated 660!
The regulator section is looking destined for the scrap bin.
Thanks!
Ok,
Usually the lower voltage terminal is the high current output, so, lets consider just it.
If this transformer has two secondaries (34.6v + 34.6v), you have good luck. It means more or less 50v/5A symmetrical (+50v and -50v), and it fits ok for Self´s designs.
If not, or if you have just one secondary you have to think about a bridge configuration or a design using non symmetrical power supply usually with a big capacitor at the output.
You can do many things with the additional 37v output. Some mosfets designs could use it. See at http://www.diyaudio.com/forums/showthread.php?threadid=36378 for an example.
Regards,
Usually the lower voltage terminal is the high current output, so, lets consider just it.
If this transformer has two secondaries (34.6v + 34.6v), you have good luck. It means more or less 50v/5A symmetrical (+50v and -50v), and it fits ok for Self´s designs.
If not, or if you have just one secondary you have to think about a bridge configuration or a design using non symmetrical power supply usually with a big capacitor at the output.
You can do many things with the additional 37v output. Some mosfets designs could use it. See at http://www.diyaudio.com/forums/showthread.php?threadid=36378 for an example.
Regards,
I think I need to clarify this a bit...!
There appears to be a pair of separate windings, but I need to break out the ohmmeter to verify this. On one side of the secondary side of the transformer is a skinny (18AWG? ) set of wires that are around 10VAC, but they are not what interest me right now. The other side of the secondary side has three big FAT terminals with two about 12AWG or 14AWG wires on each, that go to the same place on the PCB, so I think it's just for ease of soldering...easier 2x12 than 1x10 for soldering to a spade terminal! anyhow, the fat outputs are laid out so:
1Black 2white 3black
and I still haven't done all the reverse-engineering but it looks like they are the main difference between the outputs of the power supply, which is designed to put out 24 or 28 volts, depending on where a nearby (FAT! ) jumper block is placed. the readings go
1-2 ~3V
2-3 ~34V
1-3 ~37V
so I think this may be one winding with two taps for the different voltages. If I want more than double-digits of Watts out of these amplifiers it will be only by bridging if this is the case, which is not really a problem for me. I am already thinking of the big-capacitor route because of the single-supply I think I have.
Note: most of my MOSFETs are for CRT flyback switches, not audio.
There appears to be a pair of separate windings, but I need to break out the ohmmeter to verify this. On one side of the secondary side of the transformer is a skinny (18AWG? ) set of wires that are around 10VAC, but they are not what interest me right now. The other side of the secondary side has three big FAT terminals with two about 12AWG or 14AWG wires on each, that go to the same place on the PCB, so I think it's just for ease of soldering...easier 2x12 than 1x10 for soldering to a spade terminal! anyhow, the fat outputs are laid out so:
1Black 2white 3black
and I still haven't done all the reverse-engineering but it looks like they are the main difference between the outputs of the power supply, which is designed to put out 24 or 28 volts, depending on where a nearby (FAT! ) jumper block is placed. the readings go
1-2 ~3V
2-3 ~34V
1-3 ~37V
so I think this may be one winding with two taps for the different voltages. If I want more than double-digits of Watts out of these amplifiers it will be only by bridging if this is the case, which is not really a problem for me. I am already thinking of the big-capacitor route because of the single-supply I think I have.
Note: most of my MOSFETs are for CRT flyback switches, not audio.
but want to confirm.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.