Leach SuperAmp with 2/4 extra pair of output devices?


I am going to build me a Leach superamp. Not a clone, or if my idea works, it will be a clone, or just a Superamp with more power as i wont change anyting just add more MJ15003/4 devices.

I would like to put two or four more pairs on the amp.

I know i have to do my own PCBs but my thought was tp use a program to create a Gerberfile with the exact layout of the superamp but with 4 more output devices.

Is it as simple as that? Do you think i will have to change the value on R52/54
and R53/55?

The 4 pairs more superamp has, lower the voltage across the devices the low tim leach has, what will happen when i add another 4 pairs of devices?

The idea is to upp the voltage a bit, past 100V so we get a monsteramp with superfidelity :D

I know about Jens clone and i am not interested in it. It would have trouble putting out the same power as the superamp even if he has 3 more devices.
Plus, its Slew Rating seems to be a bit worse, with overshooting on 1khz square waves, I wonder how it looks at 20 Khz.

I want to keep it as close as possible to the original, just with more devices to upp the voltage so that i get even more power.

I know already the superamp is a heafty bit of amp, but i want a monster.

I am building speakers that can play over 120db without xmax limitations and with low distortion.
To them, i want an amp that when the speakers are driven fully out, the amp should have plenty left for transients as i am a headroom liking guy.

My idol and the one and only person i completely ytrust in the word of hifi, Ingvar oehman says that for a speaker to be used at fully, you need 4-500w, thats just for the need of headroom.

I suspect other people might be interested in this project, so please chime in and say what you think.


1. This is no redesign thread. It is going to be as close as possible the original. Only acceptable changing is for increasing the Slewrate but it already has a good rating. Same as the Leach low tim amp as it is an low tim with more devices.

2. The PCBs, should have the same lines and traces as the superamp plus what is needed for the extra devices.

3. NO redesigning, just talk about the org superamp and this project.

4. Designing for 8R first hand, 4R second hand but it shall be stable even for speakers with 4R nominal rating.

Project name: Leach TMR Superamp.

Anyone interested in a superamp with 400+ watts?

Best Regards//Mathias AKA TMR
Last edited:
Current limiting is already in place as designed by Leach.

What would be nice would be changing the I limiting to IV limiting by incorporating the extra resistor to the supply rails.

The three quad output stage will give 50% more output current for the same limiting circuit values. The values will need to be checked, particularly if the voltage of the supply rails is changed from the +-58Vdc available from the recommended 40+40Vac transformer.
It is the Lo Tim that uses the 40+40Vac transformer.
What is the recommended transformer for the Double-Barreled Amp?

found it
T1 - The transformer should have either a center tapped secondary or two separate secondary windings which can be wired in series. With 120 V ac rms applied to the primary, the no load secondary voltage should be 120 to 130 V ac rms for a center tapped secondary or 60+60 (60x2) to 65+65 (65x2) V ac rms for two secondary windings. This should give a no load amplifier power supply voltage of plus and minus 85 to 93 V dc. Some transformers are rated at 115 V ac rms on the primary. With 120 V ac rms applied, the secondary voltage will be greater by a factor 120/115. If the transformer is rated at full load, its no load voltage will be 15% to 20% higher. I would recommend a transformer current rating of at least 6 A. The transformer I used in each of my two original Double Barreled Amplifiers was the Signal 230-6. It had two center tapped 115 V 6 A secondaries which I wired in parallel to obtain a secondary rating of 115 V at 12 A. The primary had three voltage taps: 105 V, 115 V, and 125 V. I wired the AC line input to the 115 V tap. With 120 V AC applied to the 115 V tap, I got plus and minus 85 V DC on the power supplies and 270 W into an 8 ohm load. If I had used the 105 V primary taps, the power supply voltage would have increased to about 93 V and the amplifiers would have put out over 300 W. The Signal transformer was definately an overkill. It weighed 38 pounds. But it would really kick you know what. To my knowledge, this transformer now is available only by special order.
60+60Vac or 65+65Vac for the Double-Barreled Amp.
Last edited:
Wtih 93V -+ it delivers 300w, 87V+- 270

I have all documents that is possible to find on this amp, even the original artice in audio from 1980. What is written on the professors page is fairly new documentation and is for the 2nd generation of this amp and thats the one i would like to a bit.

I am fairly new at this and i don+t know much but as i suspected there are more interested in the idea that is a lot more knowledgeable to me.

But for anyone to be able to help it is necessary to read what the professor has written on his page. Link is posted some posts further up.

I dont want to change the layout more so than whats needed to use 4 pair extra output devices.
Last edited:
300W into 8r0 needs 69Vpk and 8.6Apk.
That is very easily achieved from +-93Vdc.

The cascode arranged outputs and drivers will lose a bit more voltage than an ordinary output stage, but taking account of that I would expect a PSU and amp optimised for economic maximum power to maybe get to 80Vpk meaning a target of 400W into 8r0 may be reasonable.
700W into 4r0 might even be achievable from a 3qaud or 4quad output stage.
The proffessor put the 2 pair of extra out. devices on the lowtim amp, its a cascoded output stage. His idea was to put 2 more pair to be able to use higher voltage and in that way get more power output. That the way this should be done to. Using his methods to increase power and to change as little as possibe.

Here is his exact words:

"Transistors Q24 and Q25 connect in series with the pre-driver transistors Q14 and Q15. The base of Q24 floats half way between the output voltage and the positive rail. The base of Q25 floats half way between the output voltage and the negative rail. The addition of Q24 and Q25 cause the voltages across Q14 and Q15 to be approximately one-half of what they would be without Q24 and Q25. Similarly, transistors Q26 through Q31 cause the voltages across Q16 through Q21 to be approximately one-half of what they would be without Q26 through Q31. By connecting the transistors in series in this way, the rail voltages can be increased for higher output power."

Is it possible to do this again for another 4 pairs? What has to change?
This is the parts that is added to the LOW_TIM Leach amp to make it a superamp


C10, C11 - 15 pF mica
C13, C14 - 100 uFd 100 V radial electrolytic
C21, C22 - 47 uFd 100 V radial electrolytic
C26, C27 - 270 pF mica
C28 - 0.01 uFd 250 V film

Q1, Q2, Q5, Q7, Q9, Q10 - MPS8099 or MPSA06
Q3, Q4, Q6, Q8, Q11 - MPS8599 or MPSA56
Q23, Q24 - 2N3439
Q22, Q25 - 2N5415
Q26 - MJE15030
Q27 - MJE15031
Q28, Q30 - MJ15003
Q29, Q31 - MJ15004

D5, D6 - 1N4934 fast recovery rectifier
D13 through D16 - 1N5250B 20 volt zener diode

R13, R14 - 5.6 kohm 1 watt (This value is for 85 V power supplies. For other power supply voltages, the formula is on the Parts List page for the Leach Amp.)
R28, R29 - 200 ohm 1/4 watt
R30, R31 - 3.9 kohm 1 watt
R37 through R40 - 470 ohm 1/4 watt
R41 through R44 - 10 ohm 1/2 watt (changed 6/27/00)
R52 through R55 - 6.2 kohm 1 watt
R56 through R59 - 10 ohm 1/2 watt (changed 6/27/00)
R60 - 39 ohm 1/4 watt
R61 through R64 - 0.33 ohm 5 watt. These 4 resistors are mounted on the heat sinks between solder lugs on the power transistor sockets. The wires that connect the collectors of Q18 and Q20 and the collectors of Q19 and Q21 are also soldered between the lugs on the sockets. Keep all leads as short as possible and use insulation stripped from hookup wire around the bare leads of the resistors.
R65, R66 - 300 ohm 1/4 watt""
300W into 8r0 needs 69Vpk and 8.6Apk.
That is very easily achieved from +-93Vdc.

The cascode arranged outputs and drivers will lose a bit more voltage than an ordinary output stage, but taking account of that I would expect a PSU and amp optimised for economic maximum power to maybe get to 80Vpk meaning a target of 400W into 8r0 may be reasonable.
700W into 4r0 might even be achievable from a 3qaud or 4quad output stage.

That was what he got out from the original one, the old one with outpu devices that were 2SCsomething.

My goal is 400 @ 8R, don´t mind a tad more but thats what i need.

Would adding just 1 pair of devices make that possible or do i need 2 pairs? Maybe it is safer using 2 pairs as i will use it to 4R speakers and want it stable for 4R nominal speakers.
Anyone that knows a freeware PDC program that don´t have any size limitations?

I have found devices that is the equivale of what he used from the beginning and he explits he still would like to use these. They are available in europe to.

"The 2N3439/2N5415 pair is the one that I originally used for my prototype amplifiers. It is the complementary TO-5/TO-39 case pair recommended in the RCA Power Transistor Manual for audio amplifiers. For all practical purposes, the 2N3440 and 2N5416, respectively, are equivalent to the 2N3439 and the 2N5415 for the Low TIM amplifier."

I am about to check the differences between them.
English isn+t my native language as you might understans so i shall write it in another way.

Today, my speakers are an easy 8R load. BUT, my next paair of speakers might be 4R nominal.

So, i need the amp to handle 4R. That means heatsink requriments are higher and more expensive.

But, i am not a fan of 4R and usually dont use elements or designs in 4R so for me, it is enugh for now that the amp is designed for 8R.
Heatsinks willl be cheaper too.
I hope your not listening at 120 db for very long as hearing damage starts at 85 db and is time related.
This damage happens slowly and before you know it....well


ofcourse not... as i stated earlier this is for headroom, music has transients 10-20 times higher than normal level.

Thats why i ant a high power, very good sounding amp.
okey, mcd99uk for your kind help.

Iam a beginner and i am a fast learner, as i se it, the output stage is a stacked one. Is that right? I need to know that i have got it right, thats that is the right toplogy so that i van read more about that to figure out how the resistor values and so on.

I have never ever done a PCB design, so i will start with doing an exakt copy of the org superamp PCB, then post it here.

After that, i will ned your help to figure out what more neds to change and values of resistors and so on.

If i get this done, i am a bit confused that there isnt more interest in this.

Isn´t people interested in high power versions of a well regarded amp?

Jens thread got so much interest, uit that maybe was because he is a good pcb designer and a more knowledgeable than me. I am a novice a this, i am not trying to hide that.

The reason i want to do this design and not just use jens redesigned Leach is that his redesign is, in my eyes not good. I have the original article in audio mag from -80 and in that articke, there where no ringing, no overshooting no nothing. Perfect square waves at 1khz, 4khz, 10khz, nad at 23khz, they where changed because they drove the amp to clipping.

Jens redesign is not as good as an Leach amp or as good as the Double barrelled/superamp we discuss here.
And thats also the reason to why i want to use the exakt same parts as the Prof. has used for his own superamp. I am going to use his favorite Transistor in every place. The output devices does no longer get produced by ONsemi bu STMicroelectronics in EU makes them.

And thats the decices i am going ro use.

This project is a Leach SuperAmp with more power. IT will just have more output devices, and/if another value on the resistors thats there. NOTHING MORE!

PS. Sorry Jens for my harhs words, i dont mean to be a bitch or so, i just say what i think when i think it. I have no hate to you or so. Your design just dont suits my need.