Hello everyone!
Currently I am trying to build a power supply unit using LM723 chip. I have a transformer that is rated at 18 V and 30 Amps and after connecting a bridge rectifier and a capacitor I got around 25 volts DC. While the transformer is rated at 30 amps, in real life it would give out arond 20 amps for sure. So, the input numbers are 25 volts and 20 amps DC. To handle that much current I need to use a number of pass transistors and I am planning to use TIP2955 for that. However, they're PNP transistors and all the circuits I was able to find on the internet and even in manufacturers' datasheets have NPN TIP3055 transistors. I have attached a picture of such a circuit that best suits my needs and so that, you guys could understand my question.

Will the power supply unit work if I replace TIP3055 transistors with TIP2955 transistors? Since these two types of transistors are of opposite polarity, is it going to be enough if I just swap the polarity, i.e. connect the emitters of TIP2955 to the input via balancing resistors and keep the rest of the circuit unchanged? In the attached picture you can see that I've replaced first of the transistors VT1 to TIP2955 and driver transistor VT5 to TIP42C as an example.

 

PNP and NPN transistors are not interchangeable in a circuit. The circuit designer's choice for one over the other depends on the circuit topology.

The high current pass transistor in a linear power supply design can be either PNP or NPN depending on the design. Find a design that suits your needs. If it shows NPN transistors then use NPN transistors and vice versa.

 

Hello,

The transistors can not be mixed.
20 Amp is a huge current.
The 723 powersupplies I have seen are 10 Amp max.
Here is an example with a full PDF:


Bertus

 

Hello,

The transistors can not be mixed.
20 Amp is a huge current.
The 723 powersupplies I have seen are 10 Amp max.
Here is an example with a full PDF:
View attachment 359345

Bertus

I am not mixing the transistors. I am replacing all NPN ones with PNP variants. In that picture I have intentionally replaced the first transistor on the top to demonstrate how I was going to implement my plan. Remaining three will be replaced with TIP2955 too. The driver transistor that is connected to pin 10 of LM723 was also replaced with PNP version as well.

 

am not mixing the transistors. I am replacing all NPN ones with PNP variant

Doesn't work like that. The polarity is important. The emitter of the transistor has to be on the outgoing side with the balancing resistor for the circuit to work.

If you want to use the TIP2955 you must invert the whole approach so the transistors are in the -ve rail not the +ve rail - this is a non-trivial change if using the 723.controller, or change the way the 723 is connected as per the datasheet - but this is a complete redesign...


Why do you want to use TIP2955 (the replacement for the 2N3055 would be TIP3055)?

 

maybe not complete redesign. something like this may do:

 

Might work...

 

Doesn't work like that. The polarity is important. The emitter of the transistor has to be on the outgoing side with the balancing resistor for the circuit to work.

If you want to use the TIP2955 you must invert the whole approach so the transistors are in the -ve rail not the +ve rail - this is a non-trivial change if using the 723.controller, or change the way the 723 is connected as per the datasheet - but this is a complete redesign...
View attachment 359383

Why do you want to use TIP2955 (the replacement for the 2N3055 would be TIP3055)?

I just have a bunch of TIP2955 lying around and have pretty much everything else to work on this project except for TIP3055 transistors.
I'm wondering why Vout in this schematic is connected to CL and not the base of a driver transistor.

 

maybe not complete redesign. something like this may do:
View attachment 359384

Thank you for your suggestion. I may try this out under much lower output current to see if this works.

 

Might work...

View attachment 359386

Thank you. I will test this option too.

 

Thank you. I will test this option too.

The lower example In my post is the same as @panic mode suggested. It's a simulation of the output stage to see if it produces similar results; seems close enough to make it worth trying. I only put 2 transistors in for demo, but you may need more depending on output current and voltage drop.

 

The lower example In my post is the same as @panic mode suggested. It's a simulation of the output stage to see if it produces similar results; seems close enough to make it worth trying. I only put 2 transistors in for demo, but you may need more depending on output current and voltage drop.

Ok, got it!

 

The "cheap trick" to use when you only have PNP transistors is to regulate the negative side using a negative regulator circuit. I saw this in an old ARRL publication Then you simply connect the load with the required polarity. The big inconvenience is not having a traditional negative ground.

"Doug DeMaw claimed it was a brilliant scheme to use surplus parts that others did not want".

My own evaluation is that it can work very well but ONLY if you use a circuit for a negative voltage supply.