EMRFD Message Archive 2671
Message Date From Subject 2671 2009-02-04 13:36:27 trabem mebart BS-170 PA heatsink Hi All,
I am looking for an alternative to the BS-170 mosfet used in a power amplifier for a QRP rig. The designer chose the BS-170 because it was cheap and very available. But, it is not available in a TO-323 package (which is a miniature version of the TO-220). My hope was to find a BS-170 in a small SMT package that would have a heat sink tab on it, allowing it to use the pcb as a heat sink.
The design currenty calls for a TO-92 version, which uses an outboard cheap outboard heat sink, with one side pressed against the board and the other side pressed against the heat sink (with thermal compound on both sides).
I'm not sure if the heat sink actually does any good, but I guess it's been used in this manner for quite awhile now-despite the inability of the TO-92 case to dissipate power well.
However, the BS-170 is also available in an SOT-23 and an SC70 package for the same cost. Since these packages are smaller, the active portion of the device would be much closer to the heat sink if a heat sink was used. I think that using an SOT-23 or an SC70 case (instead of the TO-92 case) would result in better cooling with the same size heat sink.
Has anyone on the list ever tried this before or have comments?
Thanks,
Art, KY1K
[Non-text portions of this message have been removed]2672 2009-02-05 06:59:16 Vojtech Bubnik Re: BS-170 PA heatsink Hi Art.
I was asking myself the same question. I spent couple of hours by
searching in Mouser catalog for BS170 replacement. I was looking for a
MOSFET with low gate capacity. These are the two parts from Mouser
catalog, that seemed promising to me: FDD6N20TM and FQPF3N25. They
have higher gate capacity than three BS170 in parallel, but lower than
venerable IRF510.
Regarding BS170, according to Fairchild datasheet, BS170 and 2N7000
are really the same chips, only the package has much better thermal
conductivity. The BS170 in SMD package has lower power rating than the
TO220 part. Until now there seems to be nothing beating three BS170
with simple heatsink for Class-C or Class-E HF PA in price and
performance.
73, Vojtech OK1IAK2673 2009-02-05 07:48:01 cwfingertalker Re: BS-170 PA heatsink Art,
These devices interest me. I intend to use them in future QRP
transmitter projects. Maybe what you are looking for is here too.
Take a look.
http://www.rfparts.com/transistors_RD.html
Bill N7EU2676 2009-02-05 12:28:37 Glen Leinweber BS-170 PA heatsink Art,
Spec sheet gives thermal resistance from the junction
to outside "air" for the two packages:
SOT-23: 417 degreesC-per-watt
TO-92: 150 degreesC-per-watt
You should try to keep the chip temperature below
100 C.
Clearly, with no heatsink, the TO-92 can dissipate heat
better. Adding a heatsink to these epoxy-case devices
doesn't help much - a good deal of heat is actually
conducted down the lead to which the chip is bonded.
This is also true for the SOT-23.
I've asked this question before, with no answer:
which of the three leads is the the chip bonded to?
I suspect its the drain. Adding some heat dissipation to
this lead should assist in keeping chip temperature down.
But the spec sheet shows no thermal resistance from
chip-to-drain lead.
For the SOT-23, the thermal path from chip to the
outside world is pretty short - i'd guess that if you choose
the right pin, some printed circuit board area to dissipate
heat could allow the SOT-23 to do nearly as well as
the TO-92.2677 2009-02-05 15:38:23 Russell Shaw Re: BS-170 PA heatsink Glen Leinweber wrote:
> Art,It's usually the drain. It is the largest base-area of silicon
> Spec sheet gives thermal resistance from the junction
> to outside "air" for the two packages:
> SOT-23: 417 degreesC-per-watt
> TO-92: 150 degreesC-per-watt
> You should try to keep the chip temperature below
> 100 C.
> Clearly, with no heatsink, the TO-92 can dissipate heat
> better. Adding a heatsink to these epoxy-case devices
> doesn't help much - a good deal of heat is actually
> conducted down the lead to which the chip is bonded.
> This is also true for the SOT-23.
> I've asked this question before, with no answer:
> which of the three leads is the the chip bonded to?
> I suspect its the drain. Adding some heat dissipation to
> this lead should assist in keeping chip temperature down.
> But the spec sheet shows no thermal resistance from
> chip-to-drain lead.
upon which everything else is built.
> For the SOT-23, the thermal path from chip to the
> outside world is pretty short - i'd guess that if you choose
> the right pin, some printed circuit board area to dissipate
> heat could allow the SOT-23 to do nearly as well as
> the TO-92.