L20.5 AP SYS Test data My design

I designed this amplifier. The original page is lost.
There are a lot of people asking for the parameters of my design. I re - update the web page.
To facilitate their acquisition


  • IMG_20180425_135259.jpg
    399.5 KB · Views: 3,234
  • 205-1.jpg
    590.4 KB · Views: 3,289
  • 205-5.jpg
    629.2 KB · Views: 3,233
I designed this amplifier. The original page is lost.
There are a lot of people asking for the parameters of my design. I re - update the web page.
To facilitate their acquisition


  • L20-5频率响应.jpg
    95.7 KB · Views: 4,166
  • L205-2017.jpg
    182.5 KB · Views: 4,005
  • IMD-.jpg
    112.2 KB · Views: 2,157
  • thd+d-freq.jpg
    113 KB · Views: 2,086
It's very low distortion. The main reason is because it uses a CFP output structure.
CFP can refer to DOUGLAS SELF.
Can be applied to other general circuit structure. But note that CFP is easy to self oscillation.
To ensure the stability of the instrument for measuring strictly.
Because CFP almost no crossover distortion, so L20.5 very low quiescent current. No heat.
But it's smaller than CLASS A amplifier distortion.
The distortion of 50 watts is about 0.001-0.002%.
It was more than I made LME49810 amplifier. Of course also includes many I have been making the amplifier.


  • 无标题.jpg
    70.6 KB · Views: 2,996
Hi, LJM! What is the best quiescent current for this amplifier? Earlier you wrote about 20 mA. But hou do you measure it? Is it collector current common for four pairs of output transistors or is it emitter current of each transistor?

About the static current. I think 1-50 ma don't have too big effect.
The SELF is introduced in the book. Not suitable for large static current.

Because it can increase the distortion.

But I found in practical measurement, the influence of the static current did not too big.

Whether it is a 0 or 100 ma. There is almost no difference.

So I set about 20 ma. Is about 3 * 8 = 24 ma ma
May have some different actually. Even different temperatures will have some changes.

But in fact will not bring any influence.

I think my test is also suitable for other amplifier.

Of course L20.5 adopted CFP output. It of the distortion, voltage is the smallest.
LJM L20.5 and very simple MODification (feedback)

Review and Measurements of L20.5 Loudspeaker Amp

Review and Measurements of L20.5 Loudspeaker Amp | Audio Science Review (ASR) Forum

by YMY

SO..the mother board must be MODification, finally can be use, Warning!! Must remember don't use switching power supply at any time& Please don't use pre-amplifier

“science datasheel”

8Ω 200 W 4Ω 400 W CLASS-AB
frequency response 20-25 Khz -0.15 DB 20 Khz-0.05 DB 10 Hz -0.7 DB
AP TEST (8Ω)1-50 W
THD+N 50 HZ 10 W(9 Vrms) <0.0007
THD+N 1 Khz 10 W(9 Vrms) <0.001
THD+N 10 Khz&20 Khz 10 W(9 Vrms) <0.0008
THD+N 50 W(20 V rms) 1 Khz <0.0015
Unfortunately, the seller himself demostrated a very wrong way to use AP.. which led to the master design error and insisted on the wrong AP test method, we have different views, so we could not communicate with each other, fortunately, my friend is an experienced university professor who is very interested in this product and makes a simple MODification (feedback) I'll show you how to MOD the motherboard later.
Last edited:

Don't change anything.
There is no point in doing so.
It's just a wire. From top to bottom. There is no difference.
There are many people who think they will be better. As your neighbor told you, he found that chicken legs were delicious.

Take a look at the feedback from L20.5. It's very direct. Effective.
And it's almost the same as in the photo.
This change is like writing a word on it. Then tell you. Performance has improved.
I don't think it's necessary to waste time here.
There was no change at all.


  • 205FOOK.jpg
    119.3 KB · Views: 672
Last edited:
Здравствуйте. Собрал этот усилитель на выходе оказалось - напряжения питания, не могу разобраться в чём причина, нет ли у кого схемы?

Hello. I collected this amplifier at the output, it turned out - the supply voltage, I can not figure out what the reason is, does anyone have a circuit?
:cop: google translation added. please post an english translation when posting in another language.

:cop: добавлен гугл перевод. пожалуйста, разместите английский перевод при публикации на другом языке.
Last edited by a moderator:
LJM L20 V9.2 Lockdown Project Jan 2021

I tried to make the amp as compact as possible with much care over layout. There is no detectable noise or crosstalk and the sound is very clean. One amp total quiescent was 95 mA, the other down at 30mA - so I set both at 95mA which seemed to make an improvement. Sound is very detailed at all levels. I have to be a bit careful as the output is rather high for my Mezzo 6 speakers. The ultra quiet 2 speed fan keeps the whole system very cool.

Hi all I have purchased 10 of these boards from several different sellers on eBay all regular sellers with good feedback. I now have two of the boards left which I salvaged from the amplifiers after they burned up when in use. I have never pushed them hard and they were not fed with too much power. I queried one of the sellers and he said there was a design fault in the board but he is still selling the kits only. All of the boards I purchased were ready built and I only had to add the power supplies, something I am well capable of. The two boards I still have that have not been damaged I am afraid to use since when the amplifiers burned up they took out the bass units of the Kef 104ab speakers I had at the time. Anybody else have this problem and just to be sure you understand the boards blew when being played as background music after about two weeks of use. I am now using L12.2 boards with no problems.
  • Like
Reactions: 1 user