|
|
LDMOS transistors screwed!! - RF Cafe Forums
|
Corvus
Demenus |
Post subject: LDMOS transistors screwed!!
Posted: Thu Jul 28, 2005 11:18 am
|
|
|
Hi everybody!!!
I am working with LDMOS power
transistors managing powers of about 40dBm and I
am having a lots of problems with all of them. They
are very sensitive not like bipolar transistors
and its very simple to get them screwed!!Anyone
knows the weak point of this transistors in order
to attack directly the problem?
Cheers
|
|
|
|
|
IR |
Post subject:
Posted: Thu Jul 28, 2005 1:13 pm
|
|
|
Site Admin |
|
Joined: Mon Jun 27, 2005
2:02 pm Posts: 392 Location: Germany
|
Hello Corvus Demenus,
Yes that is a known
problem with LDMOS transistors -
but not
all of them. The problem is due to a protection
diode in the Gate Source region, that protects the
transistor from negative voltages. When a negative
voltage reaches a certain level the diode clamps
it, and by that is being short-circuited. I know
that some people are fusing the diode by providing
a large enough spike of voltage to the Gate, but
I don't know how much this is effective.
A better solution to this problem would be to
limit the input power to the transistor. I can tell
you from my long experience with LDMOS (and I am
talking about Freescale's MRF series) that for UMTS
band the transistors don't suffer from this problem.
For lower frequencies e.g 470-860MHz this problem
exist. You can find another manufacturer (like Philips),
which doesn't use (so far) in this protection mechanism.
Of course you have to be very cautious and
use ESD wrist when handling the devices, and apply
the correct input and output matching to the transistors
(According to the data sheet).
Please let
me know what frequency band do you use?
_________________ Best regards,
-
IR
|
|
|
|
|
Corvus Demenus |
Post subject:
Posted: Fri Jul 29, 2005 10:03 am
|
|
|
Thanks IR,
I am using Motorola transistors,
within frecuencies between 470-860MHz. I was wondering
if you can said me how can I fuse the diode providing
voltatge to the Gate, sounds a great idea!! By this
moment I am trying to limit the input power and
certainly I improve the life of transistors but
is not enough, I need 99% security!!
Please
it would be wonderful if you can provide me information!!!
|
|
|
|
|
IR |
Post subject:
Posted: Fri Jul 29, 2005 12:57 pm
|
|
|
Site Admin |
|
Joined: Mon Jun 27, 2005
2:02 pm Posts: 392 Location: Germany
|
Hello Again,
I think that positive voltage
of between 10-20V applied between the Gate and Source
(Gate is the positive) should do the job. You should
see a current transient of few ampers, that means
the diode is fused.
I haven't
done that myself
The best thing would
be to check other manufacturer as I said in my previous
post.
Do you use the same kind of transistor
model in consecutive stages? Namely does a MRF372
drive the following stage, which is MRF372 as well?
If yes, this is the main cause of the problem, since
you should always use a stage with a Pout capability
that is at least 5dB lower than the following stage.
The driving transistor can output power levels which
can be higher than the input of the following stage
can handle, especially if you are using high peak/average
signals. I understand that with Motorola it is impossible
since they narrowed the 470-860MHz product family
to 2 transistor models. This is why I recommend
to shift to another manufacturer.
_________________ Best regards,
-
IR
|
|
|
|
|
Corvus Demenus |
Post subject:
Posted: Mon Aug 01, 2005 7:08 am
|
|
|
Well, in fact I am using BJT class A transistor
as a driver and a class AB pallet as a power amplifier,
no problem with consecutive stages. The pallet uses
two MRF 372 transistors from Motorola, so its difficult
to change the manufacturer, I should find an equivalent
pallete... I am looking at equivalent Phillips transistors
but I only find BLF647 and BLF861, both giving only
150W less than 300W from Motorola transistor.
|
|
|
|
|
IR |
Post subject:
Posted: Mon Aug 01, 2005 9:27 am
|
|
|
Site Admin |
|
Joined: Mon Jun 27, 2005
2:02 pm Posts: 392 Location: Germany
|
I think that Philips has a model that gives a power
level of 250W. Do you use the MRF372*2 in a
balanced amplifier configuration to output the 300W?
Do you use the same design as in the reference
design provided in the transistor's data sheet?
I am asking, because this transistor need thermal
compensation for Ids drift.
_________________ Best regards,
-
IR
|
|
|
|
|
Corvus Demenus |
Post subject:
Posted: Mon Aug 01, 2005 11:14 am
|
|
|
I am using an entire pallet named LDU300 from Richardson
E. providing me enough power output, so all the
configurations for transistors have done by manufacturer.
Another think is in which way the manufacturer has
done his work properly. Possibly I have to think
into design for myself the entire power stage avoiding
this product. If it was this case, can you recommend
me a good info source:books, webs...
|
|
|
|
|
IR |
Post subject:
Posted: Mon Aug 01, 2005 11:43 am
|
|
|
Site Admin |
|
Joined: Mon Jun 27, 2005
2:02 pm Posts: 392 Location: Germany
|
You can find useful information in Eudyna (Formerly
Fujitsu) website.
https://www.us.eudyna.comThere are good
app. notes there (With their specific transistors,
but still discussing the general idea of balanced
and push-pull power amplifiers). I recommend
that you stick to the suggested reference circuit
per given transistor: The substrate and the matching/bias
networks. Don't get into adventure by trying to
alter their design .
_________________ Best regards,
-
IR
|
|
|
|
|
Corvus Demenus |
Post subject:
Posted: Tue Aug 02, 2005 6:35 am
|
|
|
Thanx for the web address!!Very useful info!!!
I am using a balanced configuration thanks to
baluns and 3dB quadrature couplers. I think is a
good configuration for my requirements..
|
|
|
|
|
komm1 |
Post subject:
Posted: Mon May 22, 2006 5:18 pm
|
|
Joined: Mon May 22, 2006
5:02 pm Posts: 1 |
Hi guys,
I am having similar problems with
freescale ldmos. I am trying to test 2 different
versions for frequencies around 2.14 GHz and a keep
getting a short circuit between gate and source.
It should handle output of at least 30dBm. I have
a specific design based on microstrip, so no other
ldmos will do for now. The gate-source is shorted
before I even test for small signal s-parameters.
I’ve checked and rechecked the circuit and re-simulated
1000 times and still have no idea of what I am doing
wrong. Any help would be much appreciated. Thanks
Helias
|
|
Posted 11/12/2012
|
|
|