amplifier transistor bank

ntetlow

Joined Jul 12, 2019
63
Hello,
in the attached power amp schematic there is a bank of 8 transistors (4 banks of push-pull) where the base input is common to all four pairs. Each pair must therefore amplifier to the same degree. Can someone tell me what is going on here, what will be the end result? Could the output be four times that of ant one pair?

https://www.elcircuit.com/2018/02/1...RdXZ3UnJsZnAtRGd2MXRMdS1SUSZ0PUFBQUFBR0FhaWo0

AlbertHall

Joined Jun 4, 2014
12,326
The amplifier voltage gain is fixed by R8 and R9 negative feedback circuit.
The four transistors in parallel will increas the maximum current output capability by four times compared to a single transistor.

ntetlow

Joined Jul 12, 2019
63
The amplifier voltage gain is fixed by R8 and R9 negative feedback circuit.
The four transistors in parallel will increas the maximum current output capability by four times compared to a single transistor.
Thank you for your reply. I simulated a basic amp with ltspice then added a push/pull pair to the original (i attach both files).
The output current of around 10u amps seems to be the save value at the load resistor of each simulation, don't get it.

Attachments

• 2.9 KB Views: 6
• 3.3 KB Views: 6

AlbertHall

Joined Jun 4, 2014
12,326
In each case you have a 100k load so the current will be limited by the value of that resistor.
The parallel transistors increase the output current capability, that is the heat disipation will be spread across all the transistors.
This kind of circuit is generally used for powering low impedance loads e.g. a loudspeaker.
With your two simulations try with say an 8Ω load and use the simulator to calculate the disipation in a transistor in each simulation.

Audioguru again

Joined Oct 21, 2019
6,604
The audio amplifier has an AC output, not the DC output current you are measuring.
Both of your simulations have the same +1V and -1V peak output AC signals.

ntetlow

Joined Jul 12, 2019
63
The audio amplifier has an AC output, not the DC output current you are measuring.
Both of your simulations have the same +1V and -1V peak output AC signals.
I can see that the current into the base has been divided by the number of push/pull transistors used and the resulting load current is ultimately the same no matter how many are used, this means that the power dissipated by each transistor is reduced thus reducing the heat generated by each, as you say.
I cannot however see by what theory (ohms law?) this has been achieved?

Audioguru again

Joined Oct 21, 2019
6,604
Your 100k output resistor has 1V peak across it which is a peak current of 10uA and an AC current of 7.1uA RMS.
When you double the number of push-pull transistor pairs then each pair uses half the AC output current therefore each pair shares the total amount of heating.

I tried your circuit with a real 8 ohms speaker. Do you know why it cannot produce more than a puny 0.3W output without severe clipping?

Attachments

• 22.2 KB Views: 7

AnalogKid

Joined Aug 1, 2013
10,943
I can see that the current into the base has been divided by the number of push/pull transistors used and the resulting load current is ultimately the same no matter how many are used,
Yes, but the maximum possible load current increases with each added pair. The maximum output current increases with each additional push-pull transistor pair added in parallel. The max allowed current through any one pair can be a function of internal heat, the device's max. current rating on its datasheet, the device's max. power rating, ambient temperature, safe operating area restrictions, etc. Whatever the limiting factor is for a single pair, multiple pairs overcome it.

This is a common problem in LTSpice. If you put in a 0.1 ohm load, the simulation will happily show a gazillion amps going through it. Note that in the real world, a gazillion amps would go through it for only a few milliseconds before something flamed.

ak

AlbertHall

Joined Jun 4, 2014
12,326
Maybe LTspice should add a 'sanity check' function?