Amplifier Construction(Load Invariant)

Not open for further replies.
Guys I need your valuable help. I'm planning to build an amplifier designed by Douglas Self(Load Invariant Amplifier). I just wanted to know how it performs and what is the maximum power output into an 8 ohms load. I can't get my hands on any of those articles which he published in Electronics + Wireless World( of course for free) so, if any of u guys have those articles it would be of great help.

I am looking at a power range of 120-200 Wrms. and while browsing i found this site( )may be u'll want to check it out, but I haven't built any of those kits anybody who has already tested it please let me know.



The load-invariant amp was designed to run off supply rails of +/-20V to +/- 40V. This gives a range of roughly 20Wrms to 85Wrms. Details of this amp are given in Doug Self's 'Audio Power Amplifier Design Handbook' which you might be able to obtain from your local library. Construction notes are available on Doug's website:

pcbs can be obtained from The Signal Transfer Company:


Your observation is relevant, but does not preclude the use of supply rail voltages above +/-24V. To quote Doug Self's own book, "The supply voltage can be from +/20 to +/-40V; checking power capability for a given output device fit must be left to the constructor".

The VI limiting must obviously be tailored to suit the output device charateristics, the supply rail voltages and the design load current. It is assumed that constructors attempting this design have sufficient knowledge to determine the changes required to suit different circumstances, for example changing the voltage sense resistors R30 and R31 in the VI limiting circuit to provide the correct protection locus.

Generally speaking, articles in Electronics World where this design was first published are not intended to be full construction projects. The 'load invariant' amp was presented as a theoretical basis of the design with a number of output arrangement options.

First I want to thank each one of you guys for pouring me with more info on this subject. I have found out something which I'd like to share...

Since I can't add images here, I'd like to proceed with the following assumptions for the LOAD INVARIANT AMP.

Assumptions made for only one supply rail(+).
Supply Voltage = Vcc = 40V.
Load Resistance = 8 Ohms.
Transistor Saturation Voltage = Vcesat = 1.5V Typical(mfg data).
Re Voltage Drop = Vre.
RMS Volatge at load = Vrms.

Vrms = (Vcc - Vsat - Vre)/ 2^(1/2) = 26.89V
Power Output into 8 ohms = 90 Wrms.
Load Current at 8 Ohms = 3.35A.
Load Current at 4 Ohms = 6.7A.
Power output at 4 ohms being 170Wrms approximate(Self doesn't recommend 4 ohms loading).

Two devices are connected in parallel and each should handle the current without any protection activation for the following reasons..

The protection is a Single Slope VI limiting.
If u carefully observe the schematic its clear that R30 and R34 form a voltage divider network connected across the output transistor(Q16 & Q18) collector & emitter and the center of the dividing network to the protection transistor Q12's base. The base of Q12 receives a continuous dc offset in accordance with Vce voltage of Q16 & Q18, therefore the circuit should perform without any problem at +/- 40V and may be even more for high power o/p with extra o/p transistor, if the small signal stages allows for it.


[Edited by Xavier on 09-28-2001 at 11:07 AM]

You appear to have deleted your postings for some reason, so I will not comment further on them as had been my original intention.


Let me say from the outset that I am not a fan of current limiting circuits on audio amps, so if I were building this design I would omit Q12/Q13 and the associated components and would take extra care not to short the output terminals.

With the component values as shown of the schematic, the output current will be limited to a little under 5A. Q12 will start to conduct and steal drive current from Q14 when its base is about 0.6V above the output rail. Due to R30/R34, this occurs when the voltage developed across the 0R1 resistor is roughly 480mV.

The current limit can be raised by adding a resistor between the base of Q12 and the output rail, thus forming a potential divider across the sense resistor. R30 can then be adjusted to vary the slope of the VI protection locus.
The paralleled output devices will make it easier to select a protection locus that will keep the devices within their SOA, but some calculation/simulation will be required to confirm this. A good starting point would be to use a value of 330R for the additional resistor whilst reducing R30 to 10k. This should give a peak output current limit approaching 10A. The foregoing assumes +/-40V supply rails.

Not open for further replies.