Is it worth it to build discrete if I want say just 60W RMS into 4 ohms?
If anyone has done some comparisons I'd love to hear some opinions.
I have built a 60W chip amp in the past and compared it to a similar 60W Tube power amp. The tubes won in this particular experience, but I'm curious about a chip amp vs a well designed discrete solid state power section.
If anyone has done some comparisons I'd love to hear some opinions.
I have built a 60W chip amp in the past and compared it to a similar 60W Tube power amp. The tubes won in this particular experience, but I'm curious about a chip amp vs a well designed discrete solid state power section.
Last edited:
Discrete amps
I build a few LM3875 gainclones(Audiosector) in the past using black gate caps, TDK potentiometers, the best I could afford. the sound was very good and to be honest, I like it very much. By the way I used 20000uf per rail in the power supply. I was very happy. then one day here in the forum a member said "Wait until you build a well designed discrete amplifier" ...then I got my feet wet and went to the discrete way,I build many and most of them (90%) sounded better than the LM3875 amp kit. Now I can say, I am not going back to build amps using IC's like any STK's, LM's etc. First. I find the discrete amplifiers are more reliable, easy to work with and if you get the right design...you will end up with a killer amp. I can say easily that discrete amplifiers have more AUTHORITY than those chipamps.
Would I go back and build a "chipamp" for my personal use? the answer is NO WAY!!!. Look at it this way, a well designed chipamp will sound better against a bad designed discrete amp
. But if you compare a Well designed chipamp against a Well designed discrete amp...the discrete will be the winner hands down.
Jump in to the water and get your feet wet...build a discrete amp! You will be glad you did
I build a few LM3875 gainclones(Audiosector) in the past using black gate caps, TDK potentiometers, the best I could afford. the sound was very good and to be honest, I like it very much. By the way I used 20000uf per rail in the power supply. I was very happy. then one day here in the forum a member said "Wait until you build a well designed discrete amplifier" ...then I got my feet wet and went to the discrete way,I build many and most of them (90%) sounded better than the LM3875 amp kit. Now I can say, I am not going back to build amps using IC's like any STK's, LM's etc. First. I find the discrete amplifiers are more reliable, easy to work with and if you get the right design...you will end up with a killer amp. I can say easily that discrete amplifiers have more AUTHORITY than those chipamps.
Would I go back and build a "chipamp" for my personal use? the answer is NO WAY!!!. Look at it this way, a well designed chipamp will sound better against a bad designed discrete amp
. But if you compare a Well designed chipamp against a Well designed discrete amp...the discrete will be the winner hands down.Jump in to the water and get your feet wet...build a discrete amp! You will be glad you did
The My-Ref design of a chip amp is a different animal.
From what I understand it is a hard to beat design when it comes to chipamps.
It is what is called a composite opamp design and its specs show it to be better than just your common chipamp build, as it typically has a much lower THD.
I have read some very good comments about using a composite input design and a paralleled power output stage.
I haven't heard one yet, But I will be sometime soon.
FWIW
jer
From what I understand it is a hard to beat design when it comes to chipamps.
It is what is called a composite opamp design and its specs show it to be better than just your common chipamp build, as it typically has a much lower THD.
I have read some very good comments about using a composite input design and a paralleled power output stage.
I haven't heard one yet, But I will be sometime soon.
FWIW
jer
The reason therefore could be the lower thermal power compression in each integrated output power buffer devices of the TPA6120.
For Jeff Rowland is the use of chip amps the fact, that there are no fakes on the marked in opposite to discrete power devices, particularly power BjT's.
Maybe he's right - go to
http://www.diyaudio.com/forums/solid-state/115281-counterfeit-transistors.html
Counterfeit Transistors
and post #4 about
http://www.diyaudio.com/forums/chip...ons-tda7293-tda7294-lm3886-etc-available.html
Last edited:
Hell to the yes!
Chip amps are a quick 'n' dirty solution to problems such as the audio deck of a COMM receiver. Fidelity not required, so it makes no difference.
Chip amps always represent some one else's idea of how it should sound. A discreet implementation is your idea as to how it should sound. I prototyped a design that sounds pretty good, much better than many a Big Box offering, using my own ideas.
^^
Definitely give it a go. Just make sure that, once it's completed, connect a dummy load. That way, you can make the adjustments (if any) to prevent speaker busting DC offsets. Also, bring up the power with a Variac, and o'scope while keeping an eye on the current draw of the finals to watch out for RF oscillations. With hollow state, the worst that can happen is finals going red plate, so you power off and fix it. A solid state amp can go Chernobyl on you before you can power off. That can set off a chain reaction that poofs lots of transistors, take out voltage regulators, and generally cause a whole bunch of grief.
It's always a good idea to test, test, and test some more as each stage is completed as opposed to slapping everything together and hitting the "On" switch only to discover something's not right, then having to figure out what's not working right.
Definitely give it a go. Just make sure that, once it's completed, connect a dummy load. That way, you can make the adjustments (if any) to prevent speaker busting DC offsets. Also, bring up the power with a Variac, and o'scope while keeping an eye on the current draw of the finals to watch out for RF oscillations. With hollow state, the worst that can happen is finals going red plate, so you power off and fix it. A solid state amp can go Chernobyl on you before you can power off. That can set off a chain reaction that poofs lots of transistors, take out voltage regulators, and generally cause a whole bunch of grief.
It's always a good idea to test, test, and test some more as each stage is completed as opposed to slapping everything together and hitting the "On" switch only to discover something's not right, then having to figure out what's not working right.
60W into 4ohms resistor is not the same as 60W into 4ohms speaker !
National tell us that the peak output current is guaranteed to be >7Apk when the 3886 chipamp is kept cold (Tc=25degC and low duty signal).
60W into 4r0 requires an output of 21.9Vpk and 5.48Apk. This is easily achieved by a cold 3886.
But a reactive speaker fed with fast starting or fast stopping Audio Signals can demand currents that exceed three times that drawn by a 4r0 resistor. i.e. a 4ohms speaker can demand >16Apk.
This cannot be guaranteed by a cold chipamp. When the chipamp is warm this guaranteed 7Apk drops to a lower level and when hot even lower still.
A hot 3886 cannot properly drive a reactive 4ohms speaker load due to excessive current clipping into typical low impedance and low sensitivity speakers.
I never recommend chipamps for 4ohms duty.
I do recommend chipamps for 8ohms and greater, if they do not demand excessive current.
National tell us that the peak output current is guaranteed to be >7Apk when the 3886 chipamp is kept cold (Tc=25degC and low duty signal).
60W into 4r0 requires an output of 21.9Vpk and 5.48Apk. This is easily achieved by a cold 3886.
But a reactive speaker fed with fast starting or fast stopping Audio Signals can demand currents that exceed three times that drawn by a 4r0 resistor. i.e. a 4ohms speaker can demand >16Apk.
This cannot be guaranteed by a cold chipamp. When the chipamp is warm this guaranteed 7Apk drops to a lower level and when hot even lower still.
A hot 3886 cannot properly drive a reactive 4ohms speaker load due to excessive current clipping into typical low impedance and low sensitivity speakers.
I never recommend chipamps for 4ohms duty.
I do recommend chipamps for 8ohms and greater, if they do not demand excessive current.
Last edited:
Hat tip to AndrewT for addressing the real issue for any amp . . . application, application, application. As with any amp you must match power to sensitivity (and the desired acoustic output), and consider carefully the effect of reactive loads.
The 3886, because of it's excellent (thermal) bias tracking, has negligible crossover distortion . . . in that regard it sounds just like Class A or Class D. It has at worst inaudible distortion of any sort over its "normal" operating range . . . just don't ask too much power from it. It is not the "amp of choice" for a 4 ohm subwoofer, or a complex crossover that drops to 2 ohm at any frequency (that only an idiot would design such a crossover doesn't stop it from happening). OTOH it is as good as it gets for driving most tweeters, and most midrange drivers as well. It is essentially flawless at up to its rated power into a single voice coil.
The only reason to build a discrete AB amp is if you are faced with both a power requirement and a reactive load that can't be met by either a chip amp or a Class D amp (and in that case you'd be better off re-designing the speaker/crossover instead).
Bottom line is that any amp that "sounds different" from a 3886 at 5-10 Watts is itself flawed (and if you "like" the way it sounds it's because you "like" the distortion it introduces). When you need "big power" use Class D . . . there are several clean and simple Class D designs with similarly low harmonic and no crossover distortion that can put hundreds of Watts into almost anything.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.