Mark Levinson No23 repair help

OPS stands for OutPut Stage, or in this very case especially the large TO3 transistors on your heatsink.
Now take the example in the peak of the LS output, almost 70Volt and 35Amp.
The power supply om the OPS is 86Volt, so when supplying the 35A, there is 86-70=16Volt on the TO3’s.
16V x 35A = 560 Watt.
That’s what you see on the red line at that peak moment.
Do the same calculation for an output voltage of 40Volt and 20Amp.
(86-40)x20= 920 Watt.
So altough the LS voltage is lower, the OPS has to consume more power.

Hans
 
Now I know how the OPS and the resistive load are working together. It is rather simple with resistive load as the phase stays the same but becomes difficult when you add the inductance (real loudspeaker) like in the previous plots. It changes the phase of the current compared to the voltage.
 
Hello Hans,

I made the multimeter measurement. I added 800Hz sine signal to the input of the amp and measured with multimeter with 10A range at the speaker terminals. When there was 20mV at the input and 2.5A at the multimeter I stopped because I think the transformer started to hum and I did not want to continue. Am I right that when multimeter is measuring 2.5A, my amp is consuming 2,5A X 86V = 215W? Is it normal that in such high power levels the transformer starts to hum?
 
The situation got much better when I shorted the xlr input pins 1 and 3. Then I could go higher levels without huge hum. Attached is my measurement. First by accident I used square wave signal(blue curve). I could go up to 5.3A when the heat became too high and the amp turned itself off. After cooling I used sine wave(red curve) and I could go up to 4.7A when the protection circuit came into play and reduced the current to around 3.6A. 4.7A is slightly lower than your 6A simulation.

Br,
Hannu
 

Attachments

  • ml23_output_current.png
    ml23_output_current.png
    25.8 KB · Views: 152
Hello Hans,

I made the multimeter measurement. I added 800Hz sine signal to the input of the amp and measured with multimeter with 10A range at the speaker terminals. When there was 20mV at the input and 2.5A at the multimeter I stopped because I think the transformer started to hum and I did not want to continue. Am I right that when multimeter is measuring 2.5A, my amp is consuming 2,5A X 86V = 215W? Is it normal that in such high power levels the transformer starts to hum?

Hi Hannu,
Yes, 2.5A on the meter is 215Watt under the condition that your meter measures RMS and that your meter is still accurate at 800Hz.
And no, it is absolutely not normal that your transformer starts humming at that power level, so I'm a bit suspicious to what was really happening.
Without a scope it will be hard to know what is going on.

Hans
 
The situation got much better when I shorted the xlr input pins 1 and 3. Then I could go higher levels without huge hum. Attached is my measurement. First by accident I used square wave signal(blue curve). I could go up to 5.3A when the heat became too high and the amp turned itself off. After cooling I used sine wave(red curve) and I could go up to 4.7A when the protection circuit came into play and reduced the current to around 3.6A. 4.7A is slightly lower than your 6A simulation.

Br,
Hannu
Thx for the measurement.
I don't understand why shortening pin 1 to 3 improves things.
That points in the direction of some oscillation taking place with pin 3 open, which would explain the humming of your transformer.
So what you measured was obviously just a fraction of what really happened.
But nevertheless, there is nothing wrong with shorting pin 3 to gnd.

Looking at your graph, the first time you tried, you could not only achieve a higher current as the second attempt, but also under a different slope and that's very weird.
Slope should be exactly the same in both cases.
That brings me back to the question: how accurate is your multimeter at 800Hz and what is measured, RMS or average value.
The other question is the value of the internal resistance.
The slope of your orange line suggests 50mV*Gain/4.8Amp = 0,05*21.3/4.8 =0,22R, so lets say 0.2R and not 100mV/10A=0,01R that was expected.
That corresponds quite well with your measurement, assuming your meter measures rms, because 4.8A rms corresponds to 4.8*sqrt(2)=6.8A peak.

Hans
 
Hannu,

I have repeated your measurements in my sim model assuming a 0.2R multi-meters resistance.
Simulated are 7 amplitude steps at 800Hz from 10mVrms to 70mVrms input voltage.
The left part of the image below showing the output current for these resp. 7 steps.
As can be seen, from 60mVrms and higher the current limiter comes into action.

In the right part of the image, the rms current is shown as a function of the rms input voltage.
This is what you have supposedly measured and have shown in your graph.

Hans
 

Attachments

  • 800Hz.jpg
    800Hz.jpg
    308.8 KB · Views: 130
Nice to see that the simulations and measurements are matching almost. That current waveform really shows how fast the protection circuit comes to play. I am wondering what happens in higher frequencies? Can the protection circuit drop the current as fast as here if the frequency is like 10kHz?
How about in real life if listening music, which frequencies will hit to the protection circuit first? I have understood that low frequencies require more power in delivering music. I know it is not very obvious when listening to music because these levels are so high but just curious...
 
How about in real life if listening music, which frequencies will hit to the protection circuit first? I have understood that low frequencies require more power in delivering music. I know it is not very obvious when listening to music because these levels are so high but just curious...
Down to 2R load, protection will never come into action at any audio frequency while playing music, which is a big improvement over the original version.

Hans
 
Yes it does behave the same, see the short spike at the left.
Current goes up, gets interrupted by the protection and goes down.
It goes up again to a stable level when input voltage rises to 3.1V.
The difference with the 800Hz image is that the used input voltage went up to only 70mV.

Hans
 
That is true. I forgot that you have 3.1V in the input.

I noticed that for some reason in the right amp the output transistors at negative side are not well balanced in idle anymore. One of the voltage over 0.22ohm resistors is 20mV as the others have 9-10mV. Do I kill this transistor if I make the same multimeter test? How is it possible that I have now so different results? I don't think I have made errors in measurements.
 
Hello Hans,

I need again help to repair my right amp.
The problem is that there is a short somewhere in the output stage.
I have connected the lamps to the VCC unregs. the pot R124 is completely open(max ohms). The voltage over the lamps are at the positive side -76V and at the negative side 73V. At R45 I have 1.3V/0.4V and at R50 I have 0.5/-0.4V.

I did also a check for the WH.1 and WH.5 without VCC unreg(exept the points between R144/R39 and R14/R126 which was connected to VCC unreg).
Your post#11 I measured:
1. 0V/-1.86V
2. 1.89V/0V
So I assume no problems there.

I cannot find any shorts from the TO-3 output transistors or transistors just before them.

I am out of ideas but that R45 negative side is my suspect that something is not right there.

PS. I have also ordered some parts for the PCB upgrade but I would like to repair this first.
 
Hello Hans,

I need again help to repair my right amp.
The problem is that there is a short somewhere in the output stage.
I have connected the lamps to the VCC unregs. the pot R124 is completely open(max ohms). The voltage over the lamps are at the positive side -76V and at the negative side 73V. At R45 I have 1.3V/0.4V and at R50 I have 0.5/-0.4V.

I did also a check for the WH.1 and WH.5 without VCC unreg(exept the points between R144/R39 and R14/R126 which was connected to VCC unreg).
Your post#11 I measured:
1. 0V/-1.86V
2. 1.89V/0V
So I assume no problems there.

I cannot find any shorts from the TO-3 output transistors or transistors just before them.

I am out of ideas but that R45 negative side is my suspect that something is not right there.

PS. I have also ordered some parts for the PCB upgrade but I would like to repair this first.
Hello Handel,
Weird that with R124 fully open, that there is still current flowing in the output transistors, that´s definitely wrong.
Measured Voltage on R45 should be +/- 1.2Volt and on R50 should be +/-0.6Volt in normal operation, but with the pot fully open quite a bit lower.
Checking WH1-WH5 without Vcc unreg is not leading to useful measurements
Start by shorting WH1 to WH5 and check that the LS output is ca 0 Volt and that no voltage is measured over the lamps.
When however current flows, tell me the voltages on R45, R50 and LS.

Hans

.
 
  • Like
Reactions: 1 user