Need advise on this amp...

[Jan Dupont]

Jojo
I only see one 35 Volt supply on the schematic you posted......

I do not have a schematic of amp I'm must happy with at the moment, because it is still "a projeck", but it is very similar to the one that you are interested in.. ..
However I'm using whats called 3-deep-darlington in the output
(a predriver transistor, a driver transistor and 4 output transistors per channel)

MJ15003/MJ15004 are still available in Scandinavia without problems, and will be for the next year or more....
I've have not heard about that they will be obsolete???
But you can use MJ15022/MJ15023 instead.

There are two ways to meassure the output:
The wrong way: using only a 8/4/2 Ohm resistor (as in the specs you see from most manufactures).
For making a correct measurement of the power output you need a 8/4/2 Ohm resistor in series with a coil (app. 3uH) and a capacitor (app. 10uF) in parallel and then meassure the output with an oscilloscope. The true output is the one you can meassure just before the amp clipps

[janneman]

Jan,

I saw your advise on the DC offset control. But please note that the DC loop gain is about 3 in the amp, so the offset will be limited to less than the intrinsic LM301 offset, which probably is below 10mV anyway. Surely no need to add offset control (which will drift anyway).

On another note, do you suggest to parallell the 10uF to the load resistor for power measurements?? This spells disaster is you measure longer then a few seconds at full power or at elevated frequencies, IMO.

Other Jan (Didden)

[Jan Dupont]

Janneman Thanks for joining us.

About DC-offset: I have very good experience with this, and in my amp the DC-offset is below 1 mV on the output!!!
10 mV on the opamp x Voltage gain of the rest of the amp gives a lot of DC on the output......

Regarding the meassurement of output power I can see that it could be missunderstod... Thanks for pointing this out!!

The coil and the capacitor should be connected in parallel and THEN in series with the resistor... NEVER EVER CONNECT A CAPACITOR DIRECTLY OVER THE OUTPUT TERMINALS!!!!!!!


The above mentioned method is very similar to a loudspeaker.

[janneman]

Jan,

Yes, now I understand the 10uF.

But I differ on the offset. The LM301A (most loose spec'd version) has a typical 2mV, max 7.5mV offset. And the DC loop gain in this amp is 1x (not 3x as I said before, there is 100% DC feedback to inv input, sorry about that goof). So my 10mV is very, very worst case. I say, putting in the offset control is unneccessary and can make matters worse by drift, potmeter failure etc.

I am sure you are aware that the LM301A is a very tired opamp. Not wanting to go to a faster one, because that can upset the rest of the amp, something like a TL084 or an LF411 is perfectly feasible (you probably can't get the 301 anymore); the latter will bring the uncompensated offset close to 1mV anyway.

Jan Didden

[JojoD818]

ACD,

The supply voltage is single +50v only not 35volts. You see, that is why I want to learn how to measure an amp's output power. I just can't believe that amp can do 250W on an 8 ohm load. However, I don't have any access to an oscilloscope, is there any other way to do this?

Janneman,

Thank you for joining us. The TLO84 is available here in my country but will it just be as easy as a swap or must I change anything from the circuit.

Thanks
Jojo

[Jan Dupont]

Janneman,

Sorry regarding the DC-offset. You are right!! I missed the C2 capacitor….
In this case it’s a matter of believes. I like the possibility to adjust the DC-offset, however it is not necessary in this circuit, as you correctly have pointed out.
I think this mistake comes from that I only like to work with DC-coupled amps.

Regarding the choice of opamp there are many types on the marked. If you were building a “common amp” with standard components I would recommend the LF356 (I know that this is an old type, but it has proven to be very rough and it have a fair “sound”). The LF356 is also very cheap. Later you can always change to a more expensive (and maybe better ???) one. But there are no logic in using a high cost opamp and then use standard components in the rest of the circuit. The sound is no better than all the components together. I prefer to use well-known components upper frequency limit of no more than 4 MHz, as this can save you a lot problems with self-oscillation (which f.ex. can be caused by the PCB layout and wiring)

Jojo,
I have looked closer to this 250W circuit and it’s what called a Class H circuit (or bridged H-amp).
This is a technique developed to run an amp from a single supply. If the supply is +50V, each side of the output is designed to “rest” in the middle at+25V (with reference to ground) and with 0V input. Along with the input signal, one amp goes positive on the output and the other negative and then creating a maximum of +48V on one side of the speaker and 2V on the other side (= total of 46V). The rest is just math: 46V : 8 Ohm speaker = 5.75A. 46V x 5.75A = 264W. This is only true if the power supply can deliver the current!!! You can also use these this example to calculate the output power of the amp you wanted to build. However this is the “Asian” (no offends) way to rate the output power…. The correct way is to measure it in real!!!

[JojoD818]

ACD,

I agree with you on the issue of "Asian" power ratings. That is why I don't just trust any amp I find in my country. Shops here "tend" to overrate a power amps output power.

As I've said before, I have built a couple of those 250w bridged amp and they are really loud!

Now, my next project will be the first schematic (150W) that I 've attached. I just hope to find the original parts first. If you do have other suggestions such as possible substitutes to the parts please let me know.

Jojo

[janneman]

Originally posted by ACD
[snip]Regarding the choice of opamp there are many types on the marked. If you were building a “common amp” with standard components I would recommend the LF356 (I know that this is an old type, but it has proven to be very rough and it have a fair “sound”). The LF356 is also very cheap. Later you can always change to a more expensive (and maybe better ???) one. But there are no logic in using a high cost opamp and then use standard components in the rest of the circuit. The sound is no better than all the components together. I prefer to use well-known components upper frequency limit of no more than 4 MHz, as this can save you a lot problems with self-oscillation (which f.ex. can be caused by the PCB layout and wiring)[snip]

Agreed. Indiscriminately upgrading to the latest, fastest opamp brings a high risk of stability problems as you say. Yes, the 356 would also be OK here I guess, have no experience with it myself.


Jan Didden

[Jan Dupont]

Janneman

You should try the LF356.
I use app. 30 a year on experiments, and have never burned one off (ok maybe one or two )

[JojoD818]

My amp is up and running! Well, I just used a single pair of Mj15024/MJ15025 for the output for testing and it worked ok. Though I had to change a resistor value in the biasing circuit due to a lower voltage rail that I used.

Next step would be to add more output devices or maybe trying out other transistors such as the 2SC2922.

To those who replied to this post, many thanks to you!