Question regarding Installing a Transistor wrong

Started by 80k, May 10, 2009, 04:35:25 PM

Previous topic - Next topic

80k

So thanks to confusion of pinout between two different manufacturers (Zetex and Fairchild), I had a couple Zetex MOSFETs (BS170) installed backwards on my distortion circuit with power running through it.

So I reversed it and it works now. I'm curious what the best practice is for a situation like this. Should I throw out those MOSFETs and install fresh ones, or are those still going to be in good shape to be used? I'm just curious if a reversed transistor generally causes undue stress on the component?

brett

Hi
The classic problem of reverse installation is in bipolar transistors, where the base-emitter junction is reverse biased and "broken over".  This makes the transistor noisey in subsequent use.
JFETs and MOSFETs are much more forgiving.
cheers
Brett Robinson
Let a hundred flowers bloom, let a hundred schools of thought contend. (Mao Zedong)

80k

Thanks, that's what I'd heard before, regarding MOSFETS and JFETS. I just played with the circuit a bit more, and it seems to be working very well.

80k

Quote from: brett on May 10, 2009, 10:20:58 PM
Hi
The classic problem of reverse installation is in bipolar transistors, where the base-emitter junction is reverse biased and "broken over".  This makes the transistor noisey in subsequent use.
JFETs and MOSFETs are much more forgiving.
cheers

Actually, to add to the question, if it's a bipolar transistor that was reversed, and makes it noisy, how noisy is it? Is it subtle, or does it make it immediately unuseable for the circuit? I'll probably do a breadboard and test this myself one of these days, but was just curious. Thanks again.

MikeH

As I recently discovered, sometimes they will work just fine backwards, just with low output.  A very frustrating, non-specific symptom.  And in my experience, I've never had any transistor- FET, MOSFET, PNP, NPN or otherwise- be damaged by being inserted backwards.  And I have put a LOT of them in backwards.
"Sounds like a Fab Metal to me." -DougH

Sir H C

It depends on the circuit as to how overstressed (if at all) the transistor will be.  In general bipolars hate reverse bias the most, Robert Pease has a section of his troubleshooting book about how this degrades them (can be unnoticeable to horrid), JFETs drain and source often are interchangable, on MOSFETs if you flip drain and source, the body diode will conduct, but currrent limited by the load resistor, gate put to the wrong place, it could overstress the gate, but that depends on the device.

kristoffereide

how would a faulty biased transistor sound? With equal voltages on EBC? I'm getting a synthy sound on a build, only responsive to high signal and clips fast and wonder if this is the problem. I don't remember enough about trannies anymore...

I haven't measured this yet
Quote from: biggy boy on April 12, 2009, 06:22:33 PM
I find it funny how I can have close to 1000 components, yet I never seem to have enough parts to make a project. :icon_eek:

80k

Quote from: Sir H C on May 11, 2009, 04:55:18 PM
It depends on the circuit as to how overstressed (if at all) the transistor will be.  In general bipolars hate reverse bias the most, Robert Pease has a section of his troubleshooting book about how this degrades them (can be unnoticeable to horrid), JFETs drain and source often are interchangable, on MOSFETs if you flip drain and source, the body diode will conduct, but currrent limited by the load resistor, gate put to the wrong place, it could overstress the gate, but that depends on the device.

interesting, thanks for the information. The circuit in question is a crossover distortion (Zvex Machine) device, and both MOSFET stages are essentially SHO circuits. There are also bipolars in the circuit, but luckily I did not flip those.

R.G.

QuoteThe classic problem of reverse installation is in bipolar transistors, where the base-emitter junction is reverse biased and "broken over".  This makes the transistor noisey in subsequent use.
Actually, this *can* happen, but quite often does not in 9V powered pedals. The base-emitter junction on a bipolar is indeed sensitive to reverse breakover or zenering. They get noisy, and remain that way. However, if the transistor would normally be biased on somewhat in a 9V circuit, there is a horse race between the operation of the transistor in the reverse mode and the collector-base voltage. If  the base is held up within maybe 5V of the collector by reverse mode operation, no zenering and no noise-damage happens. Still not a good thing to do, and a crap-shoot to be avoided, but sometimes it's not fatal. Bipolars are commonly used in reverse mode under low voltage situations as analog switches. The devil is in the details.

QuoteAs I recently discovered, sometimes they will work just fine backwards, just with low output.
True. The reverse-mode gain is often dramatically lower than the normal gain, as reflects that the specialized doping on the junctions are now in the worst places, not the best. In circuit, the base-emitter junction voltage is often several volts lower than the 9V supply, so they survive, and work, but with low gain.

Quoteon MOSFETs if you flip drain and source, the body diode will conduct, but currrent limited by the load resistor, gate put to the wrong place, it could overstress the gate, but that depends on the device.
Gate oxide on most MOSFETs is 15V to 20V thick. This is one reason a 12V zener is the commonest gate protection device.

Quotehow would a faulty biased transistor sound?
There is literally no way to say; it depends entirely on the circuit it's in. Could be literally anything.

QuoteWith equal voltages on EBC?
Equal voltages on EBC says you have either shorts holding them that way or two (or more!) of them are open.

QuoteI'm getting a synthy sound on a build, only responsive to high signal and clips fast and wonder if this is the problem.
It is certainly *A* problem. Put it this way: no bipolar transistor works as an amplifier or a switch with equal voltages on all its pins.
R.G.

In response to the questions in the forum - PCB Layout for Musical Effects is available from The Book Patch. Search "PCB Layout" and it ought to appear.

kristoffereide

Could you direct me to some good literature on the subject? 
Quote from: biggy boy on April 12, 2009, 06:22:33 PM
I find it funny how I can have close to 1000 components, yet I never seem to have enough parts to make a project. :icon_eek:

80k

R.G.
Thanks for the information. Very much appreciated, as usual.

R.G.

Quote from: kristoffereide on May 12, 2009, 05:06:21 AM
Could you direct me to some good literature on the subject? 
If by that you mean how transistor voltages all get to be zero, no, I can't. I may be able to look up some general bipolar transistor info that would tell you as a side note.

I can synopsize some things for you.

1) Bipolar transistors operate like two diodes linked together (which they are). One diode is the collector-base diode, the other is the base-emitter diode.
2) The collector-base diode is always run reverse biased, and which is responsible for NOT conducting the power supply current when it's left alone by the base-emitter.
3) The base emitter is always used as a controller on the collector-base. Current is injected into the base and passes out the emitter. As it goes into the base, some of the current gets into the reverse-biased depletion region of the collector-base junction. This "poisons" the ability of the collector-base to hold off current flow, and so it lets current through.
4) How much current goes through the collector-base depends on how much current is injected into the base region. In all normal transistor, MORE current goes through the collector-base than is injected into the base. It only takes a little poison to have a big effect...  :icon_lol: The amount of current let drop through the collector by the base-region current is called the current gain, beta, or HFE of the transistor.
5) With that as background, now we get to the things you need to know. If the transistor is working correctly, not damaged in some way, then it does not conduct current at all unless the base-emitter diode is forward biased. That means that the base must be between 0.45 and 0.7V higher than the emitter (for NPN transistors; PNPs will have the sign of the voltage reversed, but are otherwise identical). 

So if you measure 0V difference between the base and emitter (and are not making a measurement error, and your meter is working properly, and is calibrated, and..., and...) then the base-emitter junction cannot be forward biased, so the transistor cannot be conducting properly. There are two ways to get 0V between the base and emitter, shorted and open. If the base and emitter are shorted inside or outside the transistor package, then you can put as much current in as you like, but the current all flows through the short and cannot raise the base voltage. The transistor does not conduct. If the base is open, and there is no external bias voltage being applied to it, then the voltage between base and emitter is 0V, but it's 0V just like your meter reads 0V when you wave the probes around in the air.

Likewise, if there is 0V on the collector, there are three possibilities; one is that the collector is open, including open to the rest of the circuit, and there is no current available for it to conduct. Or it is shorted, either internal to the package or external to the package by somethign like a solder blob, and cannot raise the voltage. Or there is very little current available to it from a very high resistance collector load resistance, and the base current is forcing it to let through almost 100% of the available current. This last does not produce truly 0V, but it may get the collector-emitter voltage down to maybe 25-50mV, which some meters would say is "0V".

If the base-emitter voltage is 0V, then the collector should be open circuit, and its voltage should read the full power supply voltage through its load resistor. Since that is not the case, either there is no power supply voltage, or the collector load resistor is open, or the collector and emitter are shorted together.

As questions where this is confusing.

I don't know any place where there is literature explaining this kind of reasoning. There are bits and pieces of it scattered through semiconductor physics and circuit design.
R.G.

In response to the questions in the forum - PCB Layout for Musical Effects is available from The Book Patch. Search "PCB Layout" and it ought to appear.