Bipolar pass transistor to Mosfet conversion

Thread Starter

yourownfree

Joined Jul 16, 2008
99
I have a power supply a Kenwood PS-30. It usesd two 2n5885 transistors. I have seen other power supplies use 2n3772 transistors. In any case I would like to use power mosfets instead of bipolar transistors if it would be better. I know the mosfet wont be turned on all the way so cant take advantage of the low ohms there, so I guess my question is: Will a mosfet be better than a bipolar, running cooler and dissipating less heat? If so what kind of modification would I need to make? I am thinking of adding a resistor from the gate to ground and adjusting the gate voltage accordingly to obtain the proper output. But I could be way off here. What is your take on this? How can I use a mosfet as a pass transistor in a power supply or is it a bad idea?
 

Jony130

Joined Feb 17, 2009
5,487
In linear regulator MOSFET will dissipate the same amount of heat as a bipolar transistor. Because power dissipation is equal to P = (Vin - Vout) * Iload. And it does not depend on the type of transistor we use.
 

AnalogKid

Joined Aug 1, 2013
10,987
Without a schematic it's hard to say, but three things:.

1. This is a linear power supply, so the power dissipation in the pass elements will be *exactly* the same no matter what the pass elements are.

2. The drive requirements for a power MOSFET are significantly different from those of bipolar transistors.

3. There are significant differences between transistors designed to be used as high power switches (such as low cost power MOSFETs) and transistors designed to be used as linear amplifiers - gain flatness, safe operating area, etc. These are more factors complicating the substitution you are proposing.

ak
 

Thread Starter

yourownfree

Joined Jul 16, 2008
99
I thought about that being in the linear region after the fact. Almost not worth the effort. Of course if I did go that way I would find a mosfet that could handle the current in the linear region and stay within temp and so on.For now I think I will just pay the $3.00 for a pair of 2n3772. Thanks for the reply it's been helpfull
 

ian field

Joined Oct 27, 2012
6,536
I have a power supply a Kenwood PS-30. It usesd two 2n5885 transistors. I have seen other power supplies use 2n3772 transistors. In any case I would like to use power mosfets instead of bipolar transistors if it would be better. I know the mosfet wont be turned on all the way so cant take advantage of the low ohms there, so I guess my question is: Will a mosfet be better than a bipolar, running cooler and dissipating less heat? If so what kind of modification would I need to make? I am thinking of adding a resistor from the gate to ground and adjusting the gate voltage accordingly to obtain the proper output. But I could be way off here. What is your take on this? How can I use a mosfet as a pass transistor in a power supply or is it a bad idea?
Your bipolar pass transistor needs around 0.7 - 0.8V to forward bias the B/E junction and make it conduct - your MOSFET will need somewhere in the region of 8 - 10V to meet the VGSthr.

If you need to output close to the available rectified voltage, VGSthr can knock a few volts off your headroom.

A common way round this is to capacitively couple a charge pump voltage doubler to the end of a convenient secondary - you only need the extra headroom, and only at low current, for driving the gate.
 

#12

Joined Nov 30, 2010
18,224
Summary:
It can be done, and most of us can do it, but we generally wouldn't bother to change an existing design because you aren't saving a milliwatt and it takes more parts to drive a MOSFET correctly.
 

Thread Starter

yourownfree

Joined Jul 16, 2008
99
Nice.Thanks. I am going to substitute the 2n5885's with 2n3772. A little less current handling by 5 amps per transistor but I think it will work. The fuse is a 20 amp originally in the output circuitry. I am thinking of dropping that to 15 amp. I am fixing this for a person. I can wait 10-15 days for the 2n5885 or three days for the 2n3772 type. The guy is not very patient. So I found a place that has a closeout on these. All said and done I bought 10 of them for a total of about $14 USD which included shipping. This the second one of his power supplies I have fixed. The last one used 4 2n3772's. I am kinda Leary about using these but with proper heat grease and a lower amperage fuse maybe he won't push it. If he does I am only out by $3 and my time. I will most likely sell the other transistors since I had to buy them in order to save money. Otherwise it would have been about $10 just for 2 cause they add shipping and handling fees under $10 purchase
 

studiot

Joined Nov 9, 2007
4,998
Perhaps you could improve the cooling?

Are the failures SOA or heat ? They are unlikely to be current or voltage directly with either transistor.
 
Last edited:

ian field

Joined Oct 27, 2012
6,536
Nice.Thanks. I am going to substitute the 2n5885's with 2n3772. A little less current handling by 5 amps per transistor but I think it will work. The fuse is a 20 amp originally in the output circuitry. I am thinking of dropping that to 15 amp. I am fixing this for a person. I can wait 10-15 days for the 2n5885 or three days for the 2n3772 type. The guy is not very patient. So I found a place that has a closeout on these. All said and done I bought 10 of them for a total of about $14 USD which included shipping. This the second one of his power supplies I have fixed. The last one used 4 2n3772's. I am kinda Leary about using these but with proper heat grease and a lower amperage fuse maybe he won't push it. If he does I am only out by $3 and my time. I will most likely sell the other transistors since I had to buy them in order to save money. Otherwise it would have been about $10 just for 2 cause they add shipping and handling fees under $10 purchase
The 3772 was a pretty good power transistor in its day, it might be worth looking at more modern TO3P devices, they will fit the metal TO3 holes and you save a mounting screw. Probably a newer better device for less money too.

As someone else suggested, a cooling improvement might help - you can salvage small box fans from all sorts of things these days.
 

KL7AJ

Joined Nov 4, 2008
2,229
Summary:
It can be done, and most of us can do it, but we generally wouldn't bother to change an existing design because you aren't saving a milliwatt and it takes more parts to drive a MOSFET correctly.
One viable advantage is that the mosfet is less likely to go into thermal runaway.
 

AnalogKid

Joined Aug 1, 2013
10,987
True, but if you think about the circuit conditions that lead to thermal runaway (in audio amplifiers, for example), they don't apply to a push-only output stage such as a power supply regulator.

ak
 

ian field

Joined Oct 27, 2012
6,536
One viable advantage is that the mosfet is less likely to go into thermal runaway.
And they're much easier to parallel than bipolar.

Some PSU manufacturers use a P-channel pass MOSFET as it doesn't need an extra gate drive overhead, but they're generally not known to perform as well as N-channel - and are more expensive.

Another alternative is to design with floating secondary/rectifier/reservoir so its easy to use N-channel on the negative side and ground at the output terminal.
 

Thread Starter

yourownfree

Joined Jul 16, 2008
99
Good thought. Maybe a fan just to help. I think this one was a wire or something fell on the parts. Not sure why the cover was open. The guy likes to think he can fix stuff. The end result I end up with it. He could have just picked it up from another person as he likes to constantly get new things try it decide it's not for him and then move on. I did get those transistors actually for $1.647 each. This is including shipping. The circuit is basic for this power supply. One transistor that has a trim pot attached which in turn drives an nte 292 pnp which supplies the base of the pass transistors with the proper voltage. The circuit has no protection other than a fuse. Not even a crowbar circuit which I hate. I was thinking about adding a circuit to give a yellow warning light if it starts to approach a current level and then of course a red one that says you have overdrawn and then just latch it and shut it down until you press the reset on the new circuit. I am not a fan of a crowbar circuit.
 

Thread Starter

yourownfree

Joined Jul 16, 2008
99
Ok I just uploaded the schematic for you all to ponder over if you want.
I like the discussion, the fun things about electronics always a discussion about how to come up with an end result and there are so many ways. That's one of the reasons I like electronics is there are different ways to come up with an end result and you can even design your own way or follow a proven design or maybe improve on a proven design. It's all fun, one of the few things you can build in your mind while you work,drive, maybe sleep.
The writing on the diagram is not mine.
 

Attachments

Top