Choosing a replacement transistor

[PoDuck]

I am working with an old power supply, and I have a couple transistors I am not able to find reasonably priced replacements for, but I'm not sure how to make sure I get the right specs for a replacement.

The first transistor is an MPSU51A, and the datasheet is here:  http://www.datasheetspdf.com/pdf/542724/MotorolaSemiconductor/MPSU51A/1

The second transistor is MPSU01A, and the datasheet is here:  http://www.alldatasheet.com/datasheet-pdf/pdf/113390/MOTOROLA/MPSU01A.html

These are both in a power supply, apparently in a regulator circuit, and both TO-202.

While it would be nice to just find a replacement, it would also be nice to know what I need to look for in order to find my own replacements in the future.

Thanks

[amspire]

Probably need more information for a good answer. This is where a photo can really help as we do not know what kind of power supply you are talking about and how the transistors have to be mounted.

Is this a power supply in a device, or is this a bench power supply? Can you find a circuit - is this a brand name product?

If they are transistors in a linear power supply that are driving the regulator power transistor, their characteristics can be significant in the stability of the feedback loop. If they are the actual supply regulator transistors, it is probably a matter of just finding something a bit better. Do you have to have a TO202 package?

[PoDuck]

The schematic is here at page 164:  https://ia600901.us.archive.org/14/items/TRS80Model4PServiceManual/TRS80%20Model%204P%20Service%20Manual.pdf

It is a switch mode power supply, and they are the ones used at Q1, Q3, Q4, and Q5.

As for whether or not they have to be TO202, I guess it's not absolutely necessary, but I would have to figure out how to mount something else.

[amspire]

So Q1 and Q3 are just used as switching transistors driving a power transistor on the mains side of a switching power supply. Shouldn't be too hard to replace.

Q4 and Q5 are even easier - they are connected to work as an SCR/Thyristor. The are off until R26 charges C14 to 30V and then they turn on so that there is less then 1V across them. They are just delaying the start of the power supply for 1 second after applying mains. C14 then has to provide the initial power for the switching regulator until there is power available from pins 2 and 3 of the transformer.

You will find more transistors commonly available in the TO126 or the bigger TO220 package.

Are they all blown? Are they all on heatsinks?

If they are all blown, you want to check CR4, CR1, R23, R19, C6 or C8? (cannot tell - going to the power transistor), Q7.

If the zener VR1 went open circuit, then the voltage on C14 would have risen to the point that Q4 and Q5 had avalanche breakdown and if this is the case, you need to replace C14. Even if C14 appears to work, it would have had overvoltage. I suspect zener VR1 is probably fine.

If Q1 was the transistor that shorted which then caused the other three to fail, then the regulator IC could easily be blown. If Q3 shorted first, the regulator IC is probably OK.

Do you have any transistors lying around that you might be able to use?

[PoDuck]

The zener has blown twice.  The first time, I'm not sure what caused it, and the second time I suspect it was because I powered it up without a load on the outputs.  At that point though, I only had a 1 watt zener to replace it with, so I'm not 100% sure that it blew because of not having a load on it, or if it didn't blow subsequently because I replaced it with a 1 watt zener.

To be completely honest, I am essentially shooting in the dark at this point, trying to figure out why I am getting exactly double the voltage I should be on all the output rails.  I asked in another topic for help figuring it out, and nobody responded, so I am just trying to muddle through it on my own.

As far as I know, the transistors are actually fine, but they are more a mystery to me than passives, given all the specs that I have no idea if they can change without testing bad, so I figured I might just as well replace them if I can find replacements, even though they test out okay.  Nothing I am testing seems to be bad, and I have tested everything that seems to even be close to the primary side of the transformer, although admittedly not everything on the entire board.  I've also replaced all the electrolytic caps (including C14), just because this machine is 35 years old, and a couple were bulging slightly.  Come to find out, about half of the electrolytics were at least 25% out, and 1 was basically just a resistor.  I also replaced the two x-caps, since they are commonly known to go bad, although they wouldn't really harm anything.

I put my scope on the test points, and they all have what appears to be the correct pattern, albeit they are all over voltage.  I have to keep from running the board too long though, because two of the resistors get extremely hot very quickly at the voltage it's running at.

In other words, it's come down to me randomly replacing things, because I don't know what to do next.  lol

[PoDuck]

I forgot to answer about what I have laying around in terms of transistors.  I have a few.  I have some BC547 and some 2N5401, both TO-92.  I have no idea whether or not they will work.  It may help to know that those transistors that are on the board do not have heat sinks on them, and they don't seem to even get warm, even though the voltage on the board, almost everywhere, is way too high.

[amspire]

The zener has blown twice.  The first time, I'm not sure what caused it, and the second time I suspect it was because I powered it up without a load on the outputs.  At that point though, I only had a 1 watt zener to replace it with, so I'm not 100% sure that it blew because of not having a load on it, or if it didn't blow subsequently because I replaced it with a 1 watt zener.

The main that way the zener will blow is if the output is higher then intended. This is a flyback transformer switching setup. If the output gets voltage too high, the transformer can directly dump current into the zener and the zener will quickly blow. When Q7 turns on, it puts power into the transformer. When Q7 turns off, energy in the transformer is dumped onto the load circuits including C14. The only thing stopping the output rising way above the intended level is the regulation. The lower the load, the shorter time the regulator IC turns on Q7 and so the less energy is put into the transformer.

To be completely honest, I am essentially shooting in the dark at this point, trying to figure out why I am getting exactly double the voltage I should be on all the output rails.  I asked in another topic for help figuring it out, and nobody responded, so I am just trying to muddle through it on my own.

As far as I know, the transistors are actually fine, but they are more a mystery to me than passives, given all the specs that I have no idea if they can change without testing bad, so I figured I might just as well replace them if I can find replacements, even though they test out okay. 

If the transistors check out on a multimeter, then they are fine. That is not your problem

I put my scope on the test points, and they all have what appears to be the correct pattern, albeit they are all over voltage.  I have to keep from running the board too long though, because two of the resistors get extremely hot very quickly at the voltage it's running at.

In other words, it's come down to me randomly replacing things, because I don't know what to do next.  lol

Can you describe exactly where you saw double the output voltage? It should not be possible for the 5V rail to get above about 6.5V - it has a SCR crowbar circuit to stop this. If you saw 10V on the 5V rail, then there is more to fix then just the original fault. If you got 0V on the 5V rail, then you can easily get overvoltage on the other rails.

Now the actual voltage regulator device is in the 5V rail. It is a TL431 IC and these are cheap and very handy to have as programmable regulators - It is a good idea to get some for you stock.

This then works through an optocoupler (4N50) to control the switching regulator IC. Need to check it works and it should be easy getting replacements. Best not to replace it if it still works.  The switching IC could also be damaged, but if it is not getting feedback through the Optocoupler, it will definitely generator overvoltage even if it is working. Do not switch this circuit back on until you have narrowed down the fault. After it is all working again, you should replace C14 again with a new one.

[PoDuck]

There are some things that might help that I didn't mention.  Initially, I was having a shorting problem caused by a bad IC on the RS232 board.  It was causing the power to cycle on and off.  As I was trying to find that problem though, the SCR, which is Q6 started releasing its magic smoke.  I replaced that, and while I had the board out, I noticed that many of the pins on the output header had cold solder joints, so I reflowed them.  When I plugged the board back in, it wouldn't power up.  It was this point where I found that the 30V zener had blown.  When I replaced it, I got no power still, but I had given power to the board without a load on it, and found that the zener had blown again.  I replaced it the second time with a 1 watt zener, and it was then that I got double voltage on all the rails.  That is the 5V, +12V, and -12V rails.  I got 10V on the 5V rail, and +24V and -24V on their respective 12V rails.  I thought maybe something by the crowbar circuit had blown, since I was assuming that I shouldn't get more than 5 volts if that circuit worked.  All the components tested out okay, although I only checked the 5.6V zener to see that it was still acting like a diode, but not that it was working at 5.6V.

Last night, I was looking through my zener collection, and I came across some 1 watt 5.6V zeners, and replaced the zener in the crowbar circuit, just because it seemed like a possibility to be the problem, but I have not tested the board since.  In my mind, that wasn't what would regulate the voltage, but simply open up the SCR, and wouldn't be my problem, but it may be why I could see 10V on the +5V rail.

I put in an order for a bunch of capacitors I am getting short on last night, and decided to order a handful of optocouplers, although I haven't tried testing the one on the board, so that will be in the mail in the next few days.  I'm going to get hold of Arrow right now and see if they can't add some TL431s to the order before it ships.

Anyway, you are now up to speed, and I think you could be right about the voltage regulator.  I didn't look on that side of the board for the problem, since I was getting too high of voltage on the primary side of the transformer.  I will test it when I get time later today.

Thank you for all your time helping me.  I really appreciate it.

[amspire]

The crowbar should never trip. The TL431 is very easy to check. If you connect pins 1 and 3 together, apply 0V to pin 2 and 5V to pins 1 and 3 via a 1K resistor, you should get 2.495V +/- 2%  on pins 1 and 3.

The Optocoupler can also be tested with a multimeter. The LED diode attached to the TL431 should have a forward voltage of about 1V. The reverse breakdown voltage may be 3V.

The other end is a NPN transistor and so you can test it just like any other NPN transistor. If both sides appear OK, it is almost certain it is OK.

[PoDuck]

Okay, looks like the TL431 is bad.  I did as you said.  I connected pins 1 and 3 together.  I put a 1K resistor on them and connected them to +5V, and connected the center pin to negative.  At the resistor, the reading started at 5 volts, and slowly started going down.  It got down to about 2.9V, but the TL431 got so hot that it melted the plastic on the wires connected to it, so I stopped there.

So, it looks like I need to make another order of parts.

I'll cross my fingers that this fixes it.  I have been working on this board way too long.

Thanks again.  I'll post whether or not that fixes it when I get the parts.

[amspire]

Okay, looks like the TL431 is bad.  I did as you said.  I connected pins 1 and 3 together.  I put a 1K resistor on them and connected them to +5V, and connected the center pin to negative.  At the resistor, the reading started at 5 volts, and slowly started going down.  It got down to about 2.9V, but the TL431 got so hot that it melted the plastic on the wires connected to it, so I stopped there.

So, it looks like I need to make another order of parts.

I'll cross my fingers that this fixes it.  I have been working on this board way too long.

Thanks again.  I'll post whether or not that fixes it when I get the parts.

It probably was working but not now. The 1K resistor is meant to go from pins 1 and 3 to the 5V. You do not have to change the 5V. If you connected pins 1 and 3 to 5V, the device will turn hard on and cook.

[PoDuck]

I connected pins 1 and 3 together, and then to the 1K resistor, and then I connected the 1K resistor to the positive lead from my power supply.  I then connected pin 2 to the negative lead.  I didn't put the resistor between the positive leads.

I did it the way it is done in this video:  https://youtu.be/jbbl3vQ2E7A

[amspire]

I connected pins 1 and 3 together, and then to the 1K resistor, and then I connected the 1K resistor to the positive lead from my power supply.  I then connected pin 2 to the negative lead.  I didn't put the resistor between the positive leads.

If all the current was going through the 1K resistor, then the TL431 couldn't melt anything. It can only get a maximum power of 25mW through a 1K at 5V and that is not enough to make the TL431 even slightly warm.

If things were melting, then you either had the 5V going straight across the TL431 or the resistor was not 1K.

[PoDuck]

That is really strange, since I checked the resistor and the voltage coming from my power supply with my meter before hooking things up.  My power supply is not very accurate, so if I need a specific voltage, I always check it first, because I trust my fluke more than the analog voltage meter on the power supply.  The resistor still checks out at 1k.

Anyway, either way, the thing is not working as it should.  I've ordered some new ones.