PDA

View Full Version : RF transistor troubleshooting help



W8KT
11-03-2007, 09:54 PM
Hi, I have a solid state amp that has had a transistor fail. This is a 6 transistor amp and I am looking for advice on the best way to test the transistors to see what one (or ones) has failed.

This is an NPN device

Can this be done in circuit?

What is the best way to test?

Thanks,

Bob

W4HAY
11-03-2007, 10:04 PM
Here ya go! (http://www.electronics-radio.com/articles/test-methods/meters/multimeter-diode-transistor-test.php) Since you didn't specify the type of transistor, I'm assuming junction instead of FET.

This is an out-of-circuit DC test for shorts/opens, but that's the way most transistors fail. If they all test good with this method, you'll have to delve deeper into the problem

VK2TIL
11-04-2007, 12:13 AM
HAY has given good advice.

As the link says, a transistor tests as two back-to-back diodes (but two diodes don't make a transistor http://www.qrz.com/iB_html/non-cgi/emoticons/smile.gif ).

Check the polarity of your multimeter on the ohms range; some have the red testlead positive whilst others have the red lead negative.

You can check this with an ordinary diode (1N400X or similar).

Some power transistors have internal protective diodes that can confuse the simple test.

Here is the data sheet for your device;

http://www.datasheetarchive.com/preview/272792.html

It doesn't seem to have diodes (if it had, the data sheet would probably list that as a feature).

W9GB
11-04-2007, 02:23 AM
Bob -

According to the datasheet link this is an NPN bipolar by Toshiba. TIL is correct about the testing (and HAY).
http://www.electronics-radio.com/article....ing.php (http://www.electronics-radio.com/articles/test-methods/meters/transistor-circuit-fault-finding.php)

IF you go to this page on the KBT DC Supplies web site,
it describes how to test an NPN transistors with a VOM or with a DVM (please note the different appraoches).
http://www.kbt-dc-supplies.com/fixem1.php

While the 2N3055 and 2N3771/3772 are used a pass transistors in linear power supplies --
they were also used as final power transistors in audio amplifers (Dynaco for example).

IF the damage to the RF transistor is severe -- please NOTE (in the Toshiba datasheet):
CAUTION
Beryllia Ceramics are used in this product.
The dust or vapor can be dangerous to humans. Do not break, cut, crush or dissolve chemically.
Dispose of this properly according to law. Do not intermingle with normal industrial or domestic waste.

w9gb

VK2TIL
11-04-2007, 03:38 AM
HAY was also correct!

http://www.qrz.com/iB_html/non-cgi/emoticons/smile.gif

WE6A
12-30-2007, 04:40 AM
Simple way: Fire it up. Then put your finger on each one of them in turn; the cool one is dead.

73's

we6a@wt.net

K7FE
12-30-2007, 05:16 AM
Quote[/b] (we6a @ Dec. 29 2007,20:40)]Simple way: Fire it up. Then put your finger on each one of them in turn; the cool one is dead.

73's

we6a@wt.net
I have used many 2SC2879 transistors. The most common failure mode of a bipolar RF power transistor is a "short", so they get HOT.......not cold.

It could be difficult to tell which one is hotter than the others when mounted on a heat sink and a fan is blowing.

The best way to test the transistors is to remove the base and collector leads. Leave the two emitter leads soldered to the ground plane on the PC board. Use the VOM method described in a previously post or use my method described below.

I use an analog VOM plus my hand to act as a bias resistor.
Set the VOM for a 10K or 100K scale.
Connect the positive lead of the VOM to the collector. (for an NPN, reverse it for a PNP)
Connect the negative lead of the VOM to the emitter.

If the meter shows conduction, the transistor is bad.

If not proceed to this test. Keep the meter leads as in the above test.

Use different fingers on one hand to touch the collector and the base leads at the same time.

When you do that, the meter should show a decrease in resistance as the transistor conducts. If it does, the transistor is good.

When you solve the transistor issue, tell us how your amplifier is biased and how the power combining/splitting was accomplished.
73,
Terry, K7FE

VK6ZGO
12-30-2007, 12:27 PM
Bravo #K7FE!
I was waiting for someone to mention this way of testing transistors. I usually don't bother with the "back-to-back diode " test,as your method is so simple,& gives so much info about the device.

73 VK6ZGO

VE2NSM
12-30-2007, 05:50 PM
Also, if one is bad, you should replace his companion as well with a matched pair, since I can bet this is from a push pull amp.

AG3Y
12-30-2007, 06:33 PM
Quote[/b] (we6a @ Dec. 30 2007,01:40)]Simple way: Fire it up. Then put your finger on each one of them in turn; the cool one is dead.

73's

we6a@wt.net
The "finger on the transistor" test can be very uncomfortable! I have seen transistors get so hot that the cases will melt solder ! Same thing with ICs. You should NOT go indescriminantly placing your fingertip on the cases of suspected defective devices unless you don't mind having a fingerprint replaced with a scar !

Just a word to the wise ! 73, Jim

K7FE
12-30-2007, 08:49 PM
Quote[/b] (vk6zgo @ Dec. 30 2007,04:27)]Bravo K7FE!
I was waiting for someone to mention this way of testing transistors. I usually don't bother with the "back-to-back diode " test,as your method is so simple,& gives so much info about the device.

73 VK6ZGO
It is a transistor testing method that I have used since the early 1960's. I do not remember how I came up with it. The test gives an indication of the following:

1. Leakage current between the collector and emitter.
2. Beta or gain of the device (approximate).
3. PNP or NPN determination.
4. Short test.
5. Proof that it is a working device.

You can "Beta match" transistors for multi device amplifier applications. Just select ones that have a similar deflection on the ohm meter.

I often use two hands to provide the base/collector resistance, not just one. The low current from the ohm meter's internal battery flows through your body. Your body acts like a bias resistor between the base and collector, which turns a "working" transistor "on."

You could substitute a fixed resistor for your body, but that would mean that you need to scrounge for a resistor in order to test a transistor............why bother.

Small signal transistors like 2N2222 have a high DC gain (beta) and will deflect the ohm meter well.

RF power transistors like 2SC2879 have a larger geometry chip, higher leakages, lower DC gain and will not deflect the ohm meter as far. You need to use higher ranges on your ohm meter.

73,
Terry, K7FE

KA8DKT
12-31-2007, 01:33 AM
This last is a good test. I can remember finding transistors that passed the junction test but had no beta. The junctions were good but the transistor had no gain (beta of 1). This seems to happen a lot to 2N2222As.

Of course I was using the beta tester on my digital mulitmeter.

-gary

ad: giga-rw