Maker Pro
Maker Pro

Testing FET transistors ?

P

phil

Is it possible to test if a FET transistor is o.k. with an ohmmeter,
i.e. with bipolar we simply check that the two diodes are o.k.
but with FET we do not have this npn structure, so what do we
do?
 
F

Fritz Schlunder

phil said:
Is it possible to test if a FET transistor is o.k. with an ohmmeter,
i.e. with bipolar we simply check that the two diodes are o.k.
but with FET we do not have this npn structure, so what do we
do?


By FET I assume you mean MOSFET, rather than JFET or MESFET or anything
else. This is an important distinction in this case since testing a JFET or
MESFET for instance isn't the same as testing a MOSFET.

Usually simply doing an ohm test between gate and source (or drain for that
matter) is sufficient to reveal if the device is any good or not. A good
device will normally read many meg-ohms or over range on your digital ohm
meter while a bad device will typically measure 0 ohms to say a few hundred
ohms. You can also (if you want) do an ohm test between drain and source
(red lead to drain, black to source). A bad device would typically read
zero ohms to a few hundred ohms as well even if you short the gate to the
source. A good device should read many meg-ohms or over range if the gate
is shorted to the source.

This isn't a 100% foolproof test however. Sometimes power MOSFETs explode
(same goes with BJTs and other devices though) sometimes breaking the
plastic case. Sometimes the whole case is blown apart or sometimes just a
little hole is made in the case and some high pressure gas escapes
accompanied by a visible and audible snap. This is much more likely to
occur in directly mains powered circuits than from other power sources. In
these cases sometimes the MOSFET is so horribly damaged that the pins don't
always remain connected to the die (sometimes the whole pins with some of
the plastic case come flying off the rest of the package). In these cases
sometimes the above indicated gate-source ohm test will suggest an infinite
resistance result even though the device is clearly destroyed. Typically it
is hard to misdiagnose this situation however since the MOSFET is visibly
damaged and/or when it failed it made a very memorable snapping sound.
 
S

Seth Koster

phil said:
Is it possible to test if a FET transistor is o.k. with an ohmmeter,
i.e. with bipolar we simply check that the two diodes are o.k.
but with FET we do not have this npn structure, so what do we
do?

If its a n-channel JFET pos lead of ohmmeter connects to gate and a
diode junction should be indicated from gate to source and gate to
drain, reversing the leads (neg to gate) should produce no connection
from gate to drain or gate to source. When leads are connected to
source and drain there should be some resistance, amount dependant
upon specific device. To test a p-channel JFET reverse the procedure
(neg to gate).
MOSFETS are difficult to test at best. Use a LOW POWER ohmmeter
set at HIGHEST resistance. If a DE-MOSFET there should be some
continuity between source and drain but none from gate to either
source or drain. A E-MOSFET should show
no continuity between any of the pins. You can sometimes destroy
MOSFETS just by touching them with a finger, though.
 
N

nihoa

phil said:
Is it possible to test if a FET transistor is o.k. with an ohmmeter,
i.e. with bipolar we simply check that the two diodes are o.k.
but with FET we do not have this npn structure, so what do we
do?
Get yourself a Transistor/FET tester. If money is a problem, you can get
one for cheap on Ebay. Search for Sencore. I have one and it works great.
Having a checker sure helps to eliminate the guesswork.

Nihoa
 
R

Rich Grise

nihoa said:
Get yourself a Transistor/FET tester. If money is a problem, you can get
one for cheap on Ebay. Search for Sencore. I have one and it works great.
Having a checker sure helps to eliminate the guesswork.

Nihoa

I've done a really, really crude test with an ohmmeter. For an N-channel
MOSFET, for example, I connect the meter S-D, with the gate floating.
I get an "off" resistance. Then, I touch the drain lead and gate lead
with my finger(s). If the FET is working, that little leakage through
my skin _should_ make it conduct.

Interestingly, this can work with darlingtons as well.

I don't know about a JFET - it probably looks like a diode S-G and D-G,
but for a transconductance test you'd have to come up with some
reverse bias.

Have Fun!
Rich
 
D

ddwyer

Rich Grise said:
I've done a really, really crude test with an ohmmeter. For an N-channel
MOSFET, for example, I connect the meter S-D, with the gate floating.
I get an "off" resistance. Then, I touch the drain lead and gate lead
with my finger(s). If the FET is working, that little leakage through
my skin _should_ make it conduct.

Interestingly, this can work with darlingtons as well.

I don't know about a JFET - it probably looks like a diode S-G and D-G,
but for a transconductance test you'd have to come up with some
reverse bias.

Have Fun!
Rich
Same thing also works with jfets my normal means of testing all fets
 
R

Robert C Monsen

phil said:
Is it possible to test if a FET transistor is o.k. with an ohmmeter,
i.e. with bipolar we simply check that the two diodes are o.k.
but with FET we do not have this npn structure, so what do we
do?

Using your finger on MOSFET devices can blow the gate due to static buildup.
Here are some simple tests:

N-MOSFET enhancement, get two 1k ohm resistors and a normally open push
button. Hook the gate and drain up to 10V through separate 1k ohm resistors,
and the source to ground. The source should be close to 0. Now, hook up the
push button between the gate and ground. When you push the button, the
source pin should shoot up to 10V. When you release it, it should go back
down to near 0. If you can measure the voltage accurately (call it V), when
the button isn't pushed, the Rds should be around 1000*V/(10-V).

10V
|
+---+----+
| |
.-. .-.
1k | | | | 1k
| | | |
'-' '-'
| | Vdrain
| +--------------o
| |
| ||-+D
| ||<- N-MOSFET (enhancement)
+----G||-+S
| |
| o |
|=|> |
| o |
| |
+--------+
|
===
GND
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

P-MOSFET enhancement, same circuit, except exchange 10V and GND everywhere
in the description.

N-JFET, its a depletion device, which contains diodes. A diode tester will
give you a Vf if you hook the gate to the + and either source or drain to
the - terminal. However, depending on the JFET and tester, this might not be
good for the JFET.

Another way to test it is by running it. You can do this by hooking up the
source to the gate should let current flow. If you lower the gate voltage
below the source, it should turn off. So, hook up a 1k resistor between the
source and ground. Hook the drain to 10V, and the gate to the source. The
voltage at the source should be fairly close to 10V. Call this Vsource. If
you then hook up the gate to gnd, the voltage at the gate should drop to
near 0. Assuming you measure Vsource, then the Idss (which is the 0V
current) should be Vsource/1000.

10V
+
|
|
|
SPDT |
|--------o |-+D
| __--oG->| N-JFET
| |---o |-+S
| | | Vsource
| --------------+-------------o
| |
| .-.
| | | 1k
| | |
| '-'
| |
+---------+--------+
GND
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

Again, reverse the GND and 10V to test a P-JFET.

Note that you can destroy a JFET by forward biasing the gate and channel, so
make sure you never make the gate of a N-JFET greater than 0.6V than either
source or drain.

Regards,
Bob Monsen
 
P

phil

Robert C Monsen said:
Using your finger on MOSFET devices can blow the gate due to static buildup.
Here are some simple tests:

N-MOSFET enhancement, get two 1k ohm resistors and a normally open push
button. Hook the gate and drain up to 10V through separate 1k ohm resistors,
and the source to ground. The source should be close to 0. Now, hook up the
push button between the gate and ground. When you push the button, the
source pin should shoot up to 10V. When you release it, it should go back


How exactly will source shoot up to 10V if it is connected to ground ?
 
T

Terry Pinnell

phil said:
Is it possible to test if a FET transistor is o.k. with an ohmmeter,
i.e. with bipolar we simply check that the two diodes are o.k.
but with FET we do not have this npn structure, so what do we
do?

If it's a tester you want, rather than an answer to your question, I
built the attached into a 2 oz tobacco tin fitted with three short
miniature croc-clip connections about 20 years ago. Seems to work OK.

http://www.terrypin.dial.pipex.com/Images/FET-Tester.gif
 
P

phil

Terry Pinnell said:
If it's a tester you want, rather than an answer to your question, I
built the attached into a 2 oz tobacco tin fitted with three short
miniature croc-clip connections about 20 years ago. Seems to work OK.

http://www.terrypin.dial.pipex.com/Images/FET-Tester.gif

Well, I was hoping for a simple procedure like in bipolars. Since I was
actually refering to jfet, someone has pointed out that it is the same as
with bipolar and also there is no need to be worried about static.
With mosfet, I gather from replies here that 1. there is no simple way,
i.e. it can only be tested by testing an application (some circuit which
incorporates it) and seeing if the application is working and 2. I need
to be carefull not to touch the legs of the transistor because they are
sensitive like pc memory. I gather your circuit is some sort of buzzer,
I will have a look at it and let you know my (hobbyist) opinion.
 
R

Robert C Monsen

phil said:
back


How exactly will source shoot up to 10V if it is connected to ground ?

typo, its the drain pin that will go up. thanks for pointing that out. When
you
push the button, the gate will be grounded, so the mosfet won't conduct.
That
allows the drain resistor to pull up the Vdrain output.
 
Top