Class A sounds better, but...
For a drum machine?
I am assuming that you're talking about for stage use. Go class B. Drum machines are producing an entirely synthetic sound, meaning that you're not trying to reproduce anything in high fidelity. The heat and inefficiency of class A amps aren't worth it for live concerts. The sound quality (especially through 'pro' speakers) would be completely lost. Use class A or AB at home and B (or D) on the road.
Grey
For a drum machine?
I am assuming that you're talking about for stage use. Go class B. Drum machines are producing an entirely synthetic sound, meaning that you're not trying to reproduce anything in high fidelity. The heat and inefficiency of class A amps aren't worth it for live concerts. The sound quality (especially through 'pro' speakers) would be completely lost. Use class A or AB at home and B (or D) on the road.
Grey
2nd that...
Nice to see you back around grey. I agree with what you said... it's not worth the efficiency penalty for sound reinforcement use (and I won't even say anything about drum machines, even though I play beautiful acoustic drums that are full of great real wood tone
)
Does anyone know of any class D designs that are around the 200-400 Watt range? Something like that would be excellent for this application- efficient, little heat, light weight.
Travis,
If you're interested in hearing music that sounds as real and natural as possible, you probably won't want anything other than a class A. It would be a total waste for a drum machine. It would be very useful for a monitor amp in a studio, or something like that where you want to hear every last little detail.
Steve
Nice to see you back around grey. I agree with what you said... it's not worth the efficiency penalty for sound reinforcement use (and I won't even say anything about drum machines, even though I play beautiful acoustic drums that are full of great real wood tone
)Does anyone know of any class D designs that are around the 200-400 Watt range? Something like that would be excellent for this application- efficient, little heat, light weight.
Travis,
If you're interested in hearing music that sounds as real and natural as possible, you probably won't want anything other than a class A. It would be a total waste for a drum machine. It would be very useful for a monitor amp in a studio, or something like that where you want to hear every last little detail.
Steve
I would have to say that the difference in complexity comes from the fact that in A/B and B you have to use a push pull output, and usually feedback (where you feed a little of the amps output back to the input for reduction of distortion) to get rid of crossover distortion (the imperfect matching of each half of the wave when one set of output transistors switches off and the other set switch on). You can get into stability problems when using feedback.
Class A is the "american muscle car" of amps: brute force, big displacement. The other designs are the "honda" approach: technology to make more out of less. Or, at least, this is my opinion of it.
Each has it's place
Class A is the "american muscle car" of amps: brute force, big displacement. The other designs are the "honda" approach: technology to make more out of less. Or, at least, this is my opinion of it.
Each has it's place
http://users.ece.gatech.edu/~mleach/lowtim/
http://users.ece.gatech.edu/~mleach/superamp/
both are nice projects.
http://users.ece.gatech.edu/~mleach/superamp/
both are nice projects.
AudioFreak said:
I highly approve of the recommendation of the Leach Amp. It is a well documented project the many people have built on this forum. The pcbs are cheap also, $25 for 2 from prof. leach.
I have built 1 two channel leach amp, 1 three channel leach amp, and 1 one channel leach superamp. They are fairly cheap, and easy to make.
--
Brian
gte619j@prism.gatech.edu
Leach cooling problem
I have also made quite many Leach amps. Four one-chan units and now almost complete one 2-chan unit. Problem is... I have bought two 1000mm pieces of heatsink equal to Fischer SK70 (page A78 in Fischer catalog) and have used it this way:
* Two one-chan units, both have single 6" piece and a Papst fan with thermostat. Both have standard 2+2 power transistor configuration.
* One one-chan unit, 10" piece and two cheaper and noisier fans, same fine thermostat with three variables (idle speed of the fan, turning point for acceleration as temp. rises, steepness of the acceleration). This unit has 3+3 transistors in the power stage.
* One one-chan unit with two 6" pieces. This unit has no fan, and has 4+4 transistors in the power stage.
* The two-chan unit I will complete soon will have two 10" heatsinks, and 3+3 power transistors in both channels.
The SK70 is specified as 0.67K/W at 6" and my guesstimate is 0.4K/W at 10" length. I have now one channel working in the 2-chan unit, idling at 200mA, that ought to be 22W of heat. Sadly, in 25 deg. C ambient, the heatsink is about 45 deg. C which means the heatsink performs more than twice as bad as it should.
I also know why it's so. The heatsink is about 60*120*250mm and having it stand on the 60*120mm end would allow air to flow freely. My housings for these amplifiers are 120mm high (this newest 140mm high) and so I have the heatsinks resting on their long (yet narrower) edges in all my amplifiers.
I feel a bit bad, but cant't fix it anymore. Yet even the first fan-less unit (one chan, 8 power transistors, two 6" heatsinks) has survived quite a techno torture to a 2 Ohm load (two Adire Tempests, all coils in parallel), but still losing half of the cooling capacity bothers me. So please, everyone trying a custom housing for any design, keep this in mind.
Also remember that even the most silent of all fans makes some noise. See Papst catalog for yourselves. I used 4412FGL and it has a rubber plate between it and the case's bottom, yet the screws that attach it to the heatsink carry the noise. The heatsinks are bolted directly to the case's bottom in all three units with fan(s). There's a 1" space between the sink and the fan to prevent too much turbulence and associated noise. I didn't want to afford an expensive aluminium housing, so all these are made of 0.75mm galvanized steel. Yes, it bends a little.
-Kimmo Sundqvist
I have also made quite many Leach amps. Four one-chan units and now almost complete one 2-chan unit. Problem is... I have bought two 1000mm pieces of heatsink equal to Fischer SK70 (page A78 in Fischer catalog) and have used it this way:
* Two one-chan units, both have single 6" piece and a Papst fan with thermostat. Both have standard 2+2 power transistor configuration.
* One one-chan unit, 10" piece and two cheaper and noisier fans, same fine thermostat with three variables (idle speed of the fan, turning point for acceleration as temp. rises, steepness of the acceleration). This unit has 3+3 transistors in the power stage.
* One one-chan unit with two 6" pieces. This unit has no fan, and has 4+4 transistors in the power stage.
* The two-chan unit I will complete soon will have two 10" heatsinks, and 3+3 power transistors in both channels.
The SK70 is specified as 0.67K/W at 6" and my guesstimate is 0.4K/W at 10" length. I have now one channel working in the 2-chan unit, idling at 200mA, that ought to be 22W of heat. Sadly, in 25 deg. C ambient, the heatsink is about 45 deg. C which means the heatsink performs more than twice as bad as it should.
I also know why it's so. The heatsink is about 60*120*250mm and having it stand on the 60*120mm end would allow air to flow freely. My housings for these amplifiers are 120mm high (this newest 140mm high) and so I have the heatsinks resting on their long (yet narrower) edges in all my amplifiers.
I feel a bit bad, but cant't fix it anymore. Yet even the first fan-less unit (one chan, 8 power transistors, two 6" heatsinks) has survived quite a techno torture to a 2 Ohm load (two Adire Tempests, all coils in parallel), but still losing half of the cooling capacity bothers me. So please, everyone trying a custom housing for any design, keep this in mind.
Also remember that even the most silent of all fans makes some noise. See Papst catalog for yourselves. I used 4412FGL and it has a rubber plate between it and the case's bottom, yet the screws that attach it to the heatsink carry the noise. The heatsinks are bolted directly to the case's bottom in all three units with fan(s). There's a 1" space between the sink and the fan to prevent too much turbulence and associated noise. I didn't want to afford an expensive aluminium housing, so all these are made of 0.75mm galvanized steel. Yes, it bends a little.
-Kimmo Sundqvist
My leach amp barely evens warm on the heatsinks.
Here are some pictures:
Leach Amp pics
--
Brian
gte619j@prism.gatech.edu
Here are some pictures:
Leach Amp pics
--
Brian
gte619j@prism.gatech.edu
Now let's be ashamed. Here are some pictures of my amplifiers. I don't own a digital camera, otherwise there'd be much more.
http://www.cs.helsinki.fi/u/kmtsundq/yleiskuva.html
The text is in Finnish, and translates this way...
General view at the nonfinished
This has been modified in following ways: 1. The heatsink on the relay card has been swapped for a bigger one, 2. added a 230V factory-made (and awfully dim) green light (a "power on" signal) and 3. a red led connected in parallel with a 1N4007 and then in series with 150nF cap to signal the operation of Rod Elliot's soft-start circuit (residing on the "relay card" a bit behind the capacitors in the picture). 4. The thing has feet, discs cut out of birch (about as thick as my wrist) at the summer cabin, 2cm high, 5. the case has a top, held in place by 7 screws, having about 1000 holes in an 20*20cm area (all 4mm in diam.) for cooling (all rounded with no burrs left), 6. a three-pole "sugarcube" connector both inside and outside on the rear panel for attaching a power meter (I own four Velleman K4306, and the three poles are earth, +15V and speaker out).
Additionally, there's a recipe for catastrophe inside attached (glued) to the front panel... a small board with a relay (and a delay for it) that switches mains for the 545VA transformer.
When I turn on any of my Leach amplifiers...
1. Switch on the front panel powers a small 63VA 18+18V transformer.
2. The blue relay closes immediately, so 545VA transformer is also powered, but in series with ceramic resistors that are 90 Ohm and 34W combined.
3. White/transparent relay, driven by Rod Elliot's soft-start delay circuit, bypasses the 34W (actually two 17W 180 Ohm) resistors.
When switching off...
4. After turning off the front panel switch, white/transparent relay disconnects in less than 0.5 seconds. Same applies during a brownout/blackout or other mains failure, so when power comes back, the soft-start delay starts over again.
5. After about 4 seconds, the blue relay also disconnects.
This way, I can use a 2.5A slow-blow fuse (could maybe use even a fast one) for the big transformer rated 230V 545VA (which comes to 2.37A). Before adding the second relay, that was having the switch power both transformers, I lost a couple of fuses when switching off the amp.
http://www.cs.helsinki.fi/u/kmtsundq/releko.html is a close-up of the relay card. And I'm not going to translate that.
This one above was one of the two first one-chan units with Papst fans. Below are pictures of the unfinished (now almost finished) fourth one-chan fan-less Amp.
http://www.cs.helsinki.fi/u/kmtsundq/l4kuvat/rivat.jpg
http://www.cs.helsinki.fi/u/kmtsundq/l4kuvat/l4apuko.jpg
http://www.cs.helsinki.fi/u/kmtsundq/l4kuvat/mainkortti.jpg
(There you can see my modifications for 4+4 power transistors)
http://www.cs.helsinki.fi/u/kmtsundq/l4kuvat/leiaut.jpg
(The picture's name is a joke on how 'layout' is pronounced)
Also, if you plan to use led lights on 110..240V AC, I recommend using a diode bridge and a big enough resistor. With a capacitor instead of a resistor you will just eventually destroy the led in one single charge peak.
Those eight 1N4007's there in the l4apuko.jpg are diode bridges for the front panel's leds.
-Kimmo Sundqvist, ashamed enough but not too much because of the build quality or maybe lack of it
http://www.cs.helsinki.fi/u/kmtsundq/yleiskuva.html
The text is in Finnish, and translates this way...
General view at the nonfinished
This has been modified in following ways: 1. The heatsink on the relay card has been swapped for a bigger one, 2. added a 230V factory-made (and awfully dim) green light (a "power on" signal) and 3. a red led connected in parallel with a 1N4007 and then in series with 150nF cap to signal the operation of Rod Elliot's soft-start circuit (residing on the "relay card" a bit behind the capacitors in the picture). 4. The thing has feet, discs cut out of birch (about as thick as my wrist) at the summer cabin, 2cm high, 5. the case has a top, held in place by 7 screws, having about 1000 holes in an 20*20cm area (all 4mm in diam.) for cooling (all rounded with no burrs left), 6. a three-pole "sugarcube" connector both inside and outside on the rear panel for attaching a power meter (I own four Velleman K4306, and the three poles are earth, +15V and speaker out).
Additionally, there's a recipe for catastrophe inside attached (glued) to the front panel... a small board with a relay (and a delay for it) that switches mains for the 545VA transformer.
When I turn on any of my Leach amplifiers...
1. Switch on the front panel powers a small 63VA 18+18V transformer.
2. The blue relay closes immediately, so 545VA transformer is also powered, but in series with ceramic resistors that are 90 Ohm and 34W combined.
3. White/transparent relay, driven by Rod Elliot's soft-start delay circuit, bypasses the 34W (actually two 17W 180 Ohm) resistors.
When switching off...
4. After turning off the front panel switch, white/transparent relay disconnects in less than 0.5 seconds. Same applies during a brownout/blackout or other mains failure, so when power comes back, the soft-start delay starts over again.
5. After about 4 seconds, the blue relay also disconnects.
This way, I can use a 2.5A slow-blow fuse (could maybe use even a fast one) for the big transformer rated 230V 545VA (which comes to 2.37A). Before adding the second relay, that was having the switch power both transformers, I lost a couple of fuses when switching off the amp.
http://www.cs.helsinki.fi/u/kmtsundq/releko.html is a close-up of the relay card. And I'm not going to translate that.
This one above was one of the two first one-chan units with Papst fans. Below are pictures of the unfinished (now almost finished) fourth one-chan fan-less Amp.
http://www.cs.helsinki.fi/u/kmtsundq/l4kuvat/rivat.jpg
http://www.cs.helsinki.fi/u/kmtsundq/l4kuvat/l4apuko.jpg
http://www.cs.helsinki.fi/u/kmtsundq/l4kuvat/mainkortti.jpg
(There you can see my modifications for 4+4 power transistors)
http://www.cs.helsinki.fi/u/kmtsundq/l4kuvat/leiaut.jpg
(The picture's name is a joke on how 'layout' is pronounced)
Also, if you plan to use led lights on 110..240V AC, I recommend using a diode bridge and a big enough resistor. With a capacitor instead of a resistor you will just eventually destroy the led in one single charge peak.
Those eight 1N4007's there in the l4apuko.jpg are diode bridges for the front panel's leds.
-Kimmo Sundqvist, ashamed enough but not too much because of the build quality or maybe lack of it
Why is class A so popular ?
Because it alows low feedback, low openloop gain, simple designs and a nice impression to your friends "wow.. a Class A amplifier.. must be pretty good then huh.."
I measured my output power yesterday.. didn't need more than 3 Watt's at any time...
think about it..
Because it alows low feedback, low openloop gain, simple designs and a nice impression to your friends "wow.. a Class A amplifier.. must be pretty good then huh.."
I measured my output power yesterday.. didn't need more than 3 Watt's at any time...
think about it..
In a band situation.. yes you will !
.. things rises up pretty fast ...
I play the guitar myself and nothing beats the acoustical-feedback-sustain of a 50Watt marshall tube powerstage clipping constanly to 90Watt rms through a 100dB efficient 4x12"stack .. yep.. people go deaf that way! What I mean is that in a pratice room or stage situation things are quite different and you need lots of power.
.. things rises up pretty fast ...
I play the guitar myself and nothing beats the acoustical-feedback-sustain of a 50Watt marshall tube powerstage clipping constanly to 90Watt rms through a 100dB efficient 4x12"stack .. yep.. people go deaf that way! What I mean is that in a pratice room or stage situation things are quite different and you need lots of power.
Why Class AB in non-power situations?
Hi,
I thought this might be an appropriate place to post this question.
I have any number of tape recorders, both cassette and reel to reel.
I noticed that the final stage of the line amps -- so the ones leading to "LINE OUT" RCA jacks -- all contain complementary PNP / NPN transistors in what looks like push-pull configurations. That is both so in older machines with discrete transistors as well as newer ones with ICs, like this one used in a TASCAM:
http://www.njr.com/semicon/PDF/NJM4558_NJM4559_E.pdf
Why would one do so? Do not push-pulls introduce crossover distortion where the transistors turn off and on and require a lot of feedback to correct it? I thought push-pull configurations were used in power stages because it is far more efficient to do so than using a Class A design. But in a low power "Line Out" stage is this not unnecessary?
Push-pull probably also helps with a lower output impedance perhaps?
In sum, I am confused why this seems to be done universally in commercial products.
I will appreciate any thoughts.
Hi,
I thought this might be an appropriate place to post this question.
I have any number of tape recorders, both cassette and reel to reel.
I noticed that the final stage of the line amps -- so the ones leading to "LINE OUT" RCA jacks -- all contain complementary PNP / NPN transistors in what looks like push-pull configurations. That is both so in older machines with discrete transistors as well as newer ones with ICs, like this one used in a TASCAM:
http://www.njr.com/semicon/PDF/NJM4558_NJM4559_E.pdf
Why would one do so? Do not push-pulls introduce crossover distortion where the transistors turn off and on and require a lot of feedback to correct it? I thought push-pull configurations were used in power stages because it is far more efficient to do so than using a Class A design. But in a low power "Line Out" stage is this not unnecessary?
Push-pull probably also helps with a lower output impedance perhaps?
In sum, I am confused why this seems to be done universally in commercial products.
I will appreciate any thoughts.
I think there are many discrete designs with single-ended class-A line outputs, the QUAD-33 pre-amplifier for example. Push-pull class-AB may be useful in professional equipment that has to be able to drive 600 ohm loads, or for line outputs that are also used as headphone outputs.
Op-amps almost invariably have class-(A)B outputs because they are intended to be more or less universal, and some users want a low supply current and others a large output driving capability.
Op-amps almost invariably have class-(A)B outputs because they are intended to be more or less universal, and some users want a low supply current and others a large output driving capability.
What you're showing is an IC op-amp, and they're almost always class AB (that's "almost" Class B, but with the output transistors biased so that they're both on during some part of the waveform - yes, this takes more power than "true" Class B but it reduces crossover distortion greatly (even before applying feedback), and still takes a lot less power than Class A.
An equivalent (as far as low impedance output and output signal voltage swing) Class A circuit would take substantially more power and get warm, even thought it's only a "line level" output.
Yes, Class AB circuits have crossover distortion, but in a well-designed circuit (including many or most IC op-amps) that distortion is pretty low, and in higher-end op-amps even insignificant and hard to measure.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.