Studio Reference Amplifier

Thanks Miles. I found an another mistake....Here is corrected schematic
 

Attachments

  • sch.JPG
    sch.JPG
    501.7 KB · Views: 1,280
Hi, first test with SR60:

I have only made a few very short tests. Operating voltage is 50V, the current is 40mA, offset is 200mV, although I have matched the input transistors 😕
The poti does not change anything.
The sine at 1 kHz looks good. Unfortunately, I have no more time this weekend. I will continue next week.

It is exciting again with every new amplifier 🙂

regards Olaf
 

Attachments

  • SR60_Test_01.jpg
    SR60_Test_01.jpg
    418.2 KB · Views: 1,002
  • SR60_Test_02.jpg
    SR60_Test_02.jpg
    282.8 KB · Views: 426
  • Sinus_1kHz.jpg
    Sinus_1kHz.jpg
    241.6 KB · Views: 334
Hi, first test with SR60:

I have only made a few very short tests. Operating voltage is 50V, the current is 40mA, offset is 200mV, although I have matched the input transistors 😕
The poti does not change anything.
The sine at 1 kHz looks good. Unfortunately, I have no more time this weekend. I will continue next week.

It is exciting again with every new amplifier 🙂

regards Olaf

Nice work, 100R pot is for BIAS you can add 220R pot to set DC offset.
Regards
 

Attachments

  • APEX SR60.JPG
    APEX SR60.JPG
    332.5 KB · Views: 691
Hi

I was interested in this design and I input into LTSpice to see how well the SR60 does before I consider building one. I am getting oscillation depends on how the input driving condition. Attached is the simulation file, I use the transistor model file by Bob Cordell. I change a little bit by adding two extra output pairs. Also, I use BC547 instead of MPSA42 as I don't have the model.

I also run the THD test comparing with a typical Blameless amp, SR60 is about 15dB higher in 2nd and 3rd harmonics when I ran 1K, 5K, 10K and 20KHz at 1.4V input. I use 4ohm load, that's the reason I use 3 pairs of output transistors to avoid beta droop and lower crossover distortion. I run the transistor at 100mA per pair and use .22ohm emitter resistor.

Please tell me what I did wrong.

Thanks
 

Attachments

Go back to single output pair for the sim. Get that working first before you complicate it.

100mA through 0r22 Re gives 22mVre. This should be close to optimal ClassAB biasing.
What current does the sim say is flowing though the two VAS/TIS transistors?

No base stoppers on the output pair.
No Output Zobel.
What capacitance on the output caused the oscillation?
 
Go back to single output pair for the sim. Get that working first before you complicate it.

100mA through 0r22 Re gives 22mVre. This should be close to optimal ClassAB biasing.
What current does the sim say is flowing though the two VAS/TIS transistors?

No base stoppers on the output pair.
No Output Zobel.
What capacitance on the output caused the oscillation?

Same thing.

I attached exactly what you asked. I attached also the .asc for gain phase plot and the graph of the loop gain plot. Phase margin at loop gain=0dB is 0. At about -30dB( which is close to open loop gain=0dB), phase is -140deg. That's not stable.

There is no capacitance at the output. Base stop can only help particular for multiple pairs.

It make very little difference between 3 pairs of output transistors or one pair.
 

Attachments

note that the gain is less steep below 3MHz.
Then begins to steepen and has become very steep as it crosses 0dB @ ~4.4MHz.
This virtually guarantees oscillation.
You need to move the knee up in frequency to beyond 4.4MHz
That can be by using a faster device that is currently introducing just a bit too much phase delay, or by introducing a zero to flatten off the roll off in the 3 to 5MHz range.
The same device could be speeded up by changing it's Ic. fT is very dependant on Ic.
 
note that the gain is less steep below 3MHz.
Then begins to steepen and has become very steep as it crosses 0dB @ ~4.4MHz.
This virtually guarantees oscillation.
You need to move the knee up in frequency to beyond 4.4MHz
That can be by using a faster device that is currently introducing just a bit too much phase delay, or by introducing a zero to flatten off the roll off in the 3 to 5MHz range.
The same device could be speeded up by changing it's Ic. fT is very dependant on Ic.

It is obvious that in real life, this amp is stable.😕 This is a popular DIY amp. I tend to trust the real life result than the simulation.
 
OLAF

Did you make scope measurement in the amplifier?
Could you please share it with us?

I also got oscillation in simulation and I believe 33nF + 22R is a big RC to put between amplifier inputs (personal opinion).

I remake the compensation arrangement and I add a predriver pairs in output stage. With these mods I got it stable in simulator and THD of 0.000267% with 55V rails.

sds/ronaldo
 
Hi, I have only measured the sine between input and output and that is clean (see picture). I do not think that the amplifier oscillates. Power consumption of the circuit and temperature of the transistors is as expected. Only the offset is a problem. With 470R instead of 180R the offset drops from -200mV to -120mV. At 1k I still have -100mV. There is something wrong. Can be, I still have a layout error or a transistor is not ok.
Otherwise no progress and tonight nothing more. 🙂

regards Olaf
 
Olaf
Thank you about your answer. I could not see the picture.

I suppose there is big differences in your input devices. Can you match pairs?
I suppose it will resolve your problem.

I attached B285 file. It is a commercial product from 80's with this topology. Good sounding in my personal opinion. Also attached my version of SR60

SDS/Ronaldo
 

Attachments

Last edited: