Ian:
I found that the Zener on the problematic board was not acting like a Zener. It takes about 22V VCC for the Zener to reach 5.1V, but after that the Zener voltage on the problematic board continued increasing as VCC increased.
Since the Zener on the good board was running about 5.6V, I picked a matching Zener from my stock and installed it on the problematic board. (I also note that online photos of Version 3 of the L12 board look like the Zener has been increased to 5V6). With the replacement there has been a significant improvement in the noise, with only a little hum remaining which I attribute to the old original main electrolytics in the Sony chassis.
Thanks again for your help.
I found that the Zener on the problematic board was not acting like a Zener. It takes about 22V VCC for the Zener to reach 5.1V, but after that the Zener voltage on the problematic board continued increasing as VCC increased.
Since the Zener on the good board was running about 5.6V, I picked a matching Zener from my stock and installed it on the problematic board. (I also note that online photos of Version 3 of the L12 board look like the Zener has been increased to 5V6). With the replacement there has been a significant improvement in the noise, with only a little hum remaining which I attribute to the old original main electrolytics in the Sony chassis.
Thanks again for your help.
Aha! So it was the zener. That's an unusual fault as we don't find many power amplifier designs that use them for regulation. It's great that you checked it out properly and described the problem in detail. Better still, it seems that all is working now without much cost apart from your time and a bit of frustration. Nice sleuthing work too - enjoy 
Colleagues, let's make a list of improvements for this set. So: I have a capacitor in the feedback 1000mkf ( written on Board revision 4) . Change to 220mkf Elna silmik. The input solid-state capacitor is changed to a jumper and put the polypropylene at 10mkf at the input terminal (I think more than enough capacity). Anything else?
Last edited:
I received a pair of these today. Bought the complete modules with heat sinks. Soldering seems fine. Heat sinks seem a bit small, but will see how hot they get. However the surface on the heat sinks was neither flat or smooth, so I ended up resurfacing them. Also put 'clamps' on the transistors instead of using the normal screws, because one of the holes was not deep enough, so one of the screws snapped when I tried to tighten it.
Only mod so far is 'HBR:s'. Drilled out the trough-plating from ground plane to signal ground and put a 10ohm resistor between them on the solder side on the pcb.
Will add bias pot:s before i connect them too.
-Has anybody done the mod to disconnect the RC circuit on the input diff stage (I think?) as LJM recommended? https://www.diyaudio.com/forums/sol...-cfp-output-amp-120w-2-8r-10.html#post3841662
'I have a suggestion. Regarding L12-2.
With scissors, cut off the capacitance. Its model for 101.
L12-2 will behave better.'
-Has anybody tried Calvin's tweak on the compensation?
Can't wait to connect these and listen to them! I like the MX50SE (especially in the lows/mids), but wanted to try this since many like it, and it should be better at driving low impedance loads.
Only mod so far is 'HBR:s'. Drilled out the trough-plating from ground plane to signal ground and put a 10ohm resistor between them on the solder side on the pcb.
Will add bias pot:s before i connect them too.
-Has anybody done the mod to disconnect the RC circuit on the input diff stage (I think?) as LJM recommended? https://www.diyaudio.com/forums/sol...-cfp-output-amp-120w-2-8r-10.html#post3841662
'I have a suggestion. Regarding L12-2.
With scissors, cut off the capacitance. Its model for 101.
L12-2 will behave better.'
-Has anybody tried Calvin's tweak on the compensation?
Can't wait to connect these and listen to them! I like the MX50SE (especially in the lows/mids), but wanted to try this since many like it, and it should be better at driving low impedance loads.
Attachments
Last edited:
Ok, first impression is good, but not what I expected. I like the mids and highs, but I miss some 'control and dynamics' in the bass.
The problem is that I was expecting a MX50SE on steroids (I like the sound of those, especially bass/mid with easy loads), but the sound of the 12-2 is not similar. The MX50 has a gut punch feeling to the sound, and the 12-2 is sweet-talk in comparison
It would still be nice if someone could share experiences on the mods i mentioned above.
The problem is that I was expecting a MX50SE on steroids (I like the sound of those, especially bass/mid with easy loads), but the sound of the 12-2 is not similar. The MX50 has a gut punch feeling to the sound, and the 12-2 is sweet-talk in comparison
It would still be nice if someone could share experiences on the mods i mentioned above.
Last edited:
I have adjusted bias according to Calvin's pdf, that says 35-40mA/7,7-8,8mV on the common resistors (total current). I was running +-37VDC with 'oversized' caps.
Also tried bipolar electrolytics on the DC cap for feedback, and film cap on the input dc blocking. I will probably do some more with these later, but for now they are 'on the shelf'.
Testing MX50X2 at the moment. Sounds good and has good control of the bass, but MX50SE with the Sanken output transistors sounds sweet to my ears from bass to treble. A low budget allrounder.
Also tried bipolar electrolytics on the DC cap for feedback, and film cap on the input dc blocking. I will probably do some more with these later, but for now they are 'on the shelf'.
Testing MX50X2 at the moment. Sounds good and has good control of the bass, but MX50SE with the Sanken output transistors sounds sweet to my ears from bass to treble. A low budget allrounder.
Last edited:
I saw these L12-2 boards being sold as 30W class a amps..
30W 8Ω Class A Amplifier Board 2-Ch Power Amplifier Class A Amp Board A30-L12 | eBay
Gave me an idea to try +-18VDC supply voltage and bias them high(I have a pair of switched supplies and nice heat sinks). But the question is how well will they work with such a low supply voltage? Will I lose more than I gain? Maybe it has been simulated?
What do you think?
30W 8Ω Class A Amplifier Board 2-Ch Power Amplifier Class A Amp Board A30-L12 | eBay
Gave me an idea to try +-18VDC supply voltage and bias them high(I have a pair of switched supplies and nice heat sinks). But the question is how well will they work with such a low supply voltage? Will I lose more than I gain? Maybe it has been simulated?
What do you think?
Last edited:
You need a different type of bias controller to run amplifiers in class A reliably. My suggestion is don't try it with this compact board that mounts virtually flat on the heatsink, which will cause some heat problems. Start with a proven class A design if that's what you want when complete, since the really good class AB sound character of this design will not translate to class A operation anyway .
Big transformer and big capacitance! I would say minimum 40mF for two channels.
I have been testing them with +-37VDC with two beer can size capacitors (I think over 1F each). The transformer is from a scrapped sony surround receiver, so I'm not sure about the rating. I would say 'medium size'
Anyway, I have not been testing them to clipping, only moderate levels, and then voltage drop on the rails was around 1V.
I would say minimum 40mF for two channels.I have been testing them with +-37VDC with two beer can size capacitors (I think over 1F each). The transformer is from a scrapped sony surround receiver, so I'm not sure about the rating. I would say 'medium size'
Anyway, I have not been testing them to clipping, only moderate levels, and then voltage drop on the rails was around 1V.
I was just reading about class A in Self's book, and as I understand it, he got really nice results biasing a CFP class B into class A. He made some sophisticated bias circuit, but kept the standard circuit as a backup too.. Makes me think the class B bias setup should be ok. Bias will probably not be ultra stable, but it seems many of the 'proper' class A designs bias vary quite a bit with temperature too.
Maybe a temperature switch or similar on the heat sink to cut power would be good just in case it it decides to get out of control temperature wise..
I think the PCB can be 'lifted' from the heat sink a bit bending the legs of the output transistors, and if they are mounted vertically, there will be airflow around the PCB too.
I feel more motivated to try again, after reading about it.. (even if I don't understand everything Self is writing)
There are a couple obvious mistakes in the schematic.
I don't see any output protection so be sure you never short it.
Just to add some info to the thread. I had idea that lowering resistance on emitter resistors would help bass 'control' as I'm not so happy with the bass from these, but did not notice any diffrence with 0,08ohm, so I went back to 0,15ohm that was supplied with the boards.
The following idea was to increase the capacitors in the RC supply to the low current stages. 200uF (diameter restriction on the board) good quality caps was tried, but I can't say I noticed any difference.
Now I think heavy bias with lower supply voltage might be the next experiment, maybe to run them as tweeter amps in active setup.
The following idea was to increase the capacitors in the RC supply to the low current stages. 200uF (diameter restriction on the board) good quality caps was tried, but I can't say I noticed any difference.
Now I think heavy bias with lower supply voltage might be the next experiment, maybe to run them as tweeter amps in active setup.
A small capacitor (~1nF) between the bases of the power transistors can help cross-over at the expense of power supply rejection. But Darlingtons do not provide access to the base of the last transistors so this may not help. Darlingtons are a compromise in linear applications. I have simulated lots of Miller cancelation schemes but CFP is never as fast and clean as cross-coupled EFs. New fast transistors could probably make it "good enough".
In fact, it is a very mature circuit. From L12-2
He works in class AB mode.
There are many people. Ask me repeatedly what type A recommendation I have.
I have repeatedly stressed that virtually all analog amplifiers can work in class A.
The premise is that you need to lower your voltage and increase the static current.
Add enough radiators.
To meet the needs of those who like class A amplifiers.
I simply changed the static current of L12-2. It's just a change of resistance.
It works very well. But I have to explain. If it works in class AB. For example, L 12-2
The effect will still be good. There won't even be a significant difference.
It's very good. Because of its circuit. The distortion is very small. Whether working in AB. or A.：p
Thank you for the reply, and nice to see you are back here on the forum!
Are you working on new audio circuit designs too?
So this is verified to be stable long term, and working well in class a?
I definitely have to try this
Another question to LJM:
Do you have some theory why I find the low frequency 'control/dynamics' better with your MX50SE than with the L12-2?
This is with 8ohm speakers, and normal listening level, maybe 10W max.
Are you working on new audio circuit designs too?
So this is verified to be stable long term, and working well in class a?
I definitely have to try this
Another question to LJM:
Do you have some theory why I find the low frequency 'control/dynamics' better with your MX50SE than with the L12-2?
This is with 8ohm speakers, and normal listening level, maybe 10W max.
Last edited: