I'm looking for some existing designs for a 25 WPC stereo power amp.
This is kind of interesting in that it can't be MORE than 25 WPC, since the speakers to be used are known to blow if you exceed that power level.
The amp is to be used as the power amplifier for a Behringer digital modelling guitar amp, which only has line level outputs.
It will be driving a pair of 25 watt rated Celestion greenback speakers.
I've been looking around and it seems that my options for buying such an amp, the way I want it, are limited.
My requirements are 25 WPC, stereo, push-pull, and it has to fit into a 1 RU height rack enclosure.
Additionally, since I have the components on hand and want to put them to good use, I would prefer to use the following devices:
Drivers: MJE 15034 and MJE 15035
Outputs: MJW 21195 and MJW 21196.
Class of amplification: I don't think that in this case it's really all that important, as this amplifier will be amplifying a guitar amp signal that will typically be heavily distorted.
And, I want it to be a simple design, easy to build.
Making up a PC board is easy enough once I have the design. I have access
to board makers and I have my own small machine shop, so all I need is
a good circuit to start from.
I freely admit, while I'm a decent technican, I'm no designer. That's why I'm posting here. And, someone here probably already has a ready-made design that fits the bill nicely or can be adapted to it.
CJ
This is kind of interesting in that it can't be MORE than 25 WPC, since the speakers to be used are known to blow if you exceed that power level.
The amp is to be used as the power amplifier for a Behringer digital modelling guitar amp, which only has line level outputs.
It will be driving a pair of 25 watt rated Celestion greenback speakers.
I've been looking around and it seems that my options for buying such an amp, the way I want it, are limited.
My requirements are 25 WPC, stereo, push-pull, and it has to fit into a 1 RU height rack enclosure.
Additionally, since I have the components on hand and want to put them to good use, I would prefer to use the following devices:
Drivers: MJE 15034 and MJE 15035
Outputs: MJW 21195 and MJW 21196.
Class of amplification: I don't think that in this case it's really all that important, as this amplifier will be amplifying a guitar amp signal that will typically be heavily distorted.
And, I want it to be a simple design, easy to build.
Making up a PC board is easy enough once I have the design. I have access
to board makers and I have my own small machine shop, so all I need is
a good circuit to start from.
I freely admit, while I'm a decent technican, I'm no designer. That's why I'm posting here. And, someone here probably already has a ready-made design that fits the bill nicely or can be adapted to it.
CJ
You should look for an amplifier with +/-25 Volt DC supply rails
See:
http://sound.westhost.com/project3b.htm
See:
http://sound.westhost.com/project3b.htm
Last edited:
Distortion does create more energy vs. RMS sine waves, that's true.
This should be looked at from a voltage limiting perspective.
It is also expected that the 25 watt rated speakers will handle the output of a 25 watt (undistorted) amplifier and also handle its output with any amount of hard clipping.
It's not at all unusual for a set of four of these speakers to last a lifetime when connected to a 100 watt Marshall all-tube amp cranked all the way, with extreme levels of distortion. So if four of these can handle 100 watts plus distortion, then one can handle 25 plus distortion.
However, the classic Marshall amp that they'd normally be used with has a tube output stage, transformer coupled, and runs a quartet of EL34s, 6550s, or 5881s.
It's true that the operator is generally responsible for blowing up equipment, but
in this particular application, it's considered normal to be able to run the amp at full
output in heavy distortion and nothing should blow up.
Now, I'm not going to be driving the solid state power amp to clipping. It needs to have
enough headroom to avoid that. But it does have to have enough headroom to handle
the full output level of the preamp and also be able to drive the speakers to their design
limit of 25 watts RMS.
CJ
This should be looked at from a voltage limiting perspective.
It is also expected that the 25 watt rated speakers will handle the output of a 25 watt (undistorted) amplifier and also handle its output with any amount of hard clipping.
It's not at all unusual for a set of four of these speakers to last a lifetime when connected to a 100 watt Marshall all-tube amp cranked all the way, with extreme levels of distortion. So if four of these can handle 100 watts plus distortion, then one can handle 25 plus distortion.
However, the classic Marshall amp that they'd normally be used with has a tube output stage, transformer coupled, and runs a quartet of EL34s, 6550s, or 5881s.
It's true that the operator is generally responsible for blowing up equipment, but
in this particular application, it's considered normal to be able to run the amp at full
output in heavy distortion and nothing should blow up.
Now, I'm not going to be driving the solid state power amp to clipping. It needs to have
enough headroom to avoid that. But it does have to have enough headroom to handle
the full output level of the preamp and also be able to drive the speakers to their design
limit of 25 watts RMS.
CJ
I agree. As a general rule, I like to run speakers at about 60% rated RMS power as the maximum. IOW, ~90Wrms into a 150Wrms speaker. This prevents the destruction and quick wear of the reflex and spider. Besides, once you reach this level of diaphram motion, it becomes non-linear and creates its own distortion regardless of the signal it is receiving.
Last edited:
This is a different application and requires a hard limit on the power amplifier. 25 watts RMS is that limit. I'm only looking for a 25 WPC stereo amp design, and more power is NOT wanted.
I have plenty of larger power amps on hand already. The point is to NOT be able to destroy the speakers with too much power.
CJ
I have plenty of larger power amps on hand already. The point is to NOT be able to destroy the speakers with too much power.
CJ
Member
Joined 2009
Paid Member
I was aiming for this power level with my TGM amplifiers, I used +/- 26V rails.
http://www.diyaudio.com/forums/solid-state/140461-tgm-amplifier.html
http://www.diyaudio.com/forums/solid-state/145652-tgm2-amplifier.html
http://www.diyaudio.com/forums/solid-state/167369-designing-tgm3-output-triples.html
for casual listening I prefer the first one. For other applications I now prefer the last one.
http://www.diyaudio.com/forums/solid-state/140461-tgm-amplifier.html
http://www.diyaudio.com/forums/solid-state/145652-tgm2-amplifier.html
http://www.diyaudio.com/forums/solid-state/167369-designing-tgm3-output-triples.html
for casual listening I prefer the first one. For other applications I now prefer the last one.
driving greenbacks with a chimp would be like running a Ferrari on diesel fuel (if it were possible). These old 'speakers need a real valve amp to push them, with 6L6GC output. Ideal with Gibson axes.
The reason most 'speakers fail is due to the glue holding the voice coil on its former melts, this is a heating problem. So an amplifier pushing out 25 watts (rms) can put out more power if it is run with a waveform which produces more heating due to limiting (bigger area under the waveform).
For this reason, a smaller output amplifier could be more damaging than a bigger one (which won't be limiting). So 25 watts is a best guess, its not gospel truth, and if you put in 26 watts, it won't throw out fire and brimstone!!
The reason most 'speakers fail is due to the glue holding the voice coil on its former melts, this is a heating problem. So an amplifier pushing out 25 watts (rms) can put out more power if it is run with a waveform which produces more heating due to limiting (bigger area under the waveform).
For this reason, a smaller output amplifier could be more damaging than a bigger one (which won't be limiting). So 25 watts is a best guess, its not gospel truth, and if you put in 26 watts, it won't throw out fire and brimstone!!
Hi CJ,
At the power levels you're playing with, the MJE 15034 and MJE 15035 will serve nicely for output transistors. It would be a complete waste to use MJW 21195 and MJW 21196 transistors, they are not going to buy you anything in this application at all. At 4 amperes, you should be fine on your fixed (and known load of 8 ohms). I'm sure you must also have something like an MJE15032 and MJE15033 on hand as well. They would suit your application even better.
Now the big question. How clean do you want the amp to be? Personally, I do not care for chip type amps, but that's just me. If you want a design that is really clean, simply build the Symasym (known quantity) and adjust component values for the lower supply voltage. There is no shortage of small amplifiers with "relaxed specifications" you can use. Any of the small practice amps from Roland or Traynor would serve well there. Again, remain with the TO-220 packages for your output.
Why do you have those 350 VDC drivers on hand and not the matching MJE15032 and MJE15033 parts? The ones you have are a full 350 VDC at lower current. From tech to tech, you'll want the higher current rating for your drivers.
-Chris
At the power levels you're playing with, the MJE 15034 and MJE 15035 will serve nicely for output transistors. It would be a complete waste to use MJW 21195 and MJW 21196 transistors, they are not going to buy you anything in this application at all. At 4 amperes, you should be fine on your fixed (and known load of 8 ohms). I'm sure you must also have something like an MJE15032 and MJE15033 on hand as well. They would suit your application even better.
Now the big question. How clean do you want the amp to be? Personally, I do not care for chip type amps, but that's just me. If you want a design that is really clean, simply build the Symasym (known quantity) and adjust component values for the lower supply voltage. There is no shortage of small amplifiers with "relaxed specifications" you can use. Any of the small practice amps from Roland or Traynor would serve well there. Again, remain with the TO-220 packages for your output.
Why do you have those 350 VDC drivers on hand and not the matching MJE15032 and MJE15033 parts? The ones you have are a full 350 VDC at lower current. From tech to tech, you'll want the higher current rating for your drivers.
-Chris
I have those drivers and output chips on hand because I got them as spares for repair of a Krell power amplifier which, as luck turned out, didn't need any output devices, just a huge pile of new caps.
I also got those devices for free from the OnSemi free samples program. (Up to five free devices per type, per week, free, just pay shipping.) You should join this program!
I was (and still am) considering building a high power homebrew amp using those devices, following a fully balanced Class AB topology, but figured that since they're way over spec'ed for a little 25 WPC amp, well, why not go ahead and use them? I'm of the opinion that an amp can't be overly robust in its design and construction, so
if you consider that I really like the Krell school of amp over-design, you'll see why
I'd want to use these robust 350 volt, 200 watt rated output devices and let them run
almost at idle while meeting my power output requirements.
I've elected to take an easier path and just bought a Topping TP-21 amplifier for 60 bucks. It's 14 watts per channel into 8 ohms, probably plenty for my needs, and I can
retrofit it into a 1 RU enclosure easily enough. It was cheap enough that I can afford to do some experiments.
I can use the devices I have on hand for a larger amp project in the future. Something
more suitable for their design application, very Krell-like. Several pairs of complementary devices, low current per device, high total wattage.
CJ
I also got those devices for free from the OnSemi free samples program. (Up to five free devices per type, per week, free, just pay shipping.) You should join this program!
I was (and still am) considering building a high power homebrew amp using those devices, following a fully balanced Class AB topology, but figured that since they're way over spec'ed for a little 25 WPC amp, well, why not go ahead and use them? I'm of the opinion that an amp can't be overly robust in its design and construction, so
if you consider that I really like the Krell school of amp over-design, you'll see why
I'd want to use these robust 350 volt, 200 watt rated output devices and let them run
almost at idle while meeting my power output requirements.
I've elected to take an easier path and just bought a Topping TP-21 amplifier for 60 bucks. It's 14 watts per channel into 8 ohms, probably plenty for my needs, and I can
retrofit it into a 1 RU enclosure easily enough. It was cheap enough that I can afford to do some experiments.
I can use the devices I have on hand for a larger amp project in the future. Something
more suitable for their design application, very Krell-like. Several pairs of complementary devices, low current per device, high total wattage.
CJ
Cj
Tnx for info I will check what I have here.......and post it tomorrow.
For info....those guitar speakers are capable to handle much more power. If they would be intended for ,,normal,, use would be rated some 100w. Handling 25w of guitar signal is something else because signals are heavily distorted, clipped, compressed,.....so that's why are rated ,,only,, 25W. Power amp for such purposes must handle such signals too. It must be very robust well cooled.....and big output transistors are wise and obvious choice. Such amps have in most cases quite high supply voltage but small power transformer to avoid early clip and get more headroom.....A simple diode or led ,,soft limiter ,, is quite usual on input. But I must agree with Cats Squirrel that tube amp is the right thing to drive greenbacks. I will look for something anyway....
Best regards
Taj
Tnx for info I will check what I have here.......and post it tomorrow.
For info....those guitar speakers are capable to handle much more power. If they would be intended for ,,normal,, use would be rated some 100w. Handling 25w of guitar signal is something else because signals are heavily distorted, clipped, compressed,.....so that's why are rated ,,only,, 25W. Power amp for such purposes must handle such signals too. It must be very robust well cooled.....and big output transistors are wise and obvious choice. Such amps have in most cases quite high supply voltage but small power transformer to avoid early clip and get more headroom.....A simple diode or led ,,soft limiter ,, is quite usual on input. But I must agree with Cats Squirrel that tube amp is the right thing to drive greenbacks. I will look for something anyway....
Best regards
Taj
Hi CJ,
I know about those programs, and I'm very careful not to abuse them. I sample to design with parts I don't have, or no one has. Never for anything else. If the program is abused (and they have tightened all of them up within the last 5 years especially), they will remove it. Please don't abuse these programs and buy your parts when you need more than a couple. Otherwise, you'll ruin it for the people who really do need the program, and the program was designed for those people. As you are in service, you are not allowed to use these parts for repairs that you charge for either. The temptation would be strong to do this. It's only your moral character that can keep you from cheating (stealing really). After all, these parts are sent to you with the condition that they are not resold.
I guess I'm not into electronic waste. I've been in service for over 35 years now, and seen a lot of everything. Wasting parts really doesn't work for me, nor do I cut things to the bone either. There is a happy medium you can exist in. As for Krell, I know better rather than to chase excess. In the earlier days, they could melt coins, but their gear sounded awful - very grainy. It's your cuppa, and that's fine. You should examine more designs.
Now, the drivers you do have are good for 350 VDC, the outputs for 250 VDC. Does this make sense to you? Drivers rated at the same 250 VDC, but at twice the current (8A as opposed to 4 A) would be far more useful. It doesn't take an engineer to figure that one out. So, if you use them, just make sure the parts will not be destroyed in your application. It's good to remember that output transistors suffer beta droop as the current rises (they all do - BJTs). That means the current demands from your driver stage increases at a rate that isn't linear, but higher by some factor.
-Chris
I know about those programs, and I'm very careful not to abuse them. I sample to design with parts I don't have, or no one has. Never for anything else. If the program is abused (and they have tightened all of them up within the last 5 years especially), they will remove it. Please don't abuse these programs and buy your parts when you need more than a couple. Otherwise, you'll ruin it for the people who really do need the program, and the program was designed for those people. As you are in service, you are not allowed to use these parts for repairs that you charge for either. The temptation would be strong to do this. It's only your moral character that can keep you from cheating (stealing really). After all, these parts are sent to you with the condition that they are not resold.
I guess I'm not into electronic waste. I've been in service for over 35 years now, and seen a lot of everything. Wasting parts really doesn't work for me, nor do I cut things to the bone either. There is a happy medium you can exist in. As for Krell, I know better rather than to chase excess. In the earlier days, they could melt coins, but their gear sounded awful - very grainy. It's your cuppa, and that's fine. You should examine more designs.
Now, the drivers you do have are good for 350 VDC, the outputs for 250 VDC. Does this make sense to you? Drivers rated at the same 250 VDC, but at twice the current (8A as opposed to 4 A) would be far more useful. It doesn't take an engineer to figure that one out. So, if you use them, just make sure the parts will not be destroyed in your application. It's good to remember that output transistors suffer beta droop as the current rises (they all do - BJTs). That means the current demands from your driver stage increases at a rate that isn't linear, but higher by some factor.
-Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.