Hi Hans,
Sorry, I got off on a "maximum DC load you can control with a BS170" tangent, which is 60V minus Vds(on) times the current that causes the transistor to dissipate 830mW.
In _that_ case, the supply looks more like a constant current, and increasing temperature will raise Vds(on) and increase dissipation, hence positive feedback.
We're both right! #PerspectiveIsEverything
73, Willie N1JBJ
toggle quoted message
Show quoted text
On May 1, 2023, at 1:22 PM, Hans Summers <hans.summers@...> wrote:
Hello Willie
Also note that RDS (on) rises as temperature rises, which
causes more heating in the device as it gets hotter, so
positive feedback for dissipation. Bipolar transistors have an effective resistance which decreases as temperature rises. They have positive feedback for dissipation, also known as thermal runaway.
MOSFETs do not. As the resistance rises, the current falls (assuming constant voltage). Put another way since
Power = V^2 / R
If V remains constant and R increases, then power DECREASES. So it's a negative feedback, not positive. No thermal runaway.
Which of course isn't to say that you can't destroy MOSFETs by too much heat. But at least it is not such a disastrous thing as in bipolar transistors. This is also why we can safely use MOSFETs directly in parallel. If one transistor does more work than the other due to variations in characteristics, its resistance will increase, causing the power dissipated in that device to fall and so it is a negative feedback that stabilizes the load share between the devices. If it was bipolar transistors you can't put them in parallel because one transistor does more than the other, its resistance decreases, current increases, power dissipation increases, and being a positive feedback everything gets worse and worse until it melts.
73 Hans G0UPL
On Mon, May 1, 2023 at 8:10?PM William Smith <w_smith@...> wrote:
Also note that RDS (on) rises as temperature rises, which causes more heating in the device as it gets hotter, so positive feedback for dissipation.
73, Willie N1JBJ
On May 1, 2023, at 1:03 PM, Tony Scaminaci <tonyscam@...> wrote:
?Interesting results John¡ plugging in 25C for the ambient temp, 150C for Tja, and 830 mW for Pd, gives a junction temp of 150C. This data is 100% consistent with the maximum ratings.
Has anyone measured the air temperature inside the case during normal operation? Also, during transmit, has anyone measured the actual current draw through the BS170? 400 mA may be too high of an estimate when operating with a decent SWR. As Allison stated, operating with a poor SWR will make things considerably worse as will attempting to squeeze more RF power out of the system.
Tony AC9QY
On Mon, May 1, 2023 at 5:51 AM John Z <jdzbrozek@...> wrote:
For those of you who are interested in the thermal aspects of the BS170 application in QDX, here is an interesting website. It is equipped with a table of thermal characteristics for various component packages, and some calculators for designing thermal solutions or estimating performance.
The table of characteristics is especially interesting as it attempts to separate junction-to-case and case-to-air thermal resistances from the aggregate junction-to-air resistance for the TO-92 package. In some cases the device manufacturer will provide such data. The Fairchild PN2222 NPN transistor datasheet is a good example.
In some cases the manufacturer will not, as is the case for the BS170's that we use. I believe the data presented on this web site is a mix of spec sheet data and empirical data. We can argue about that provenance, but I think it is better than having no data.
The first calculator can be loaded with numbers representing the BS170 operating without a heat sink, dissipating power consistent with our application. We can load the manufacturer's junction-to-air resistance here, along with our ambient temperature.
The calculator shows that in this case, with 800mW dissipated, a 25 deg C ambient, and a 150 deg C / watt resistance, the BS170 is in trouble with a junction temperature that exceeds the manufacturer's maximum rating. It gets worse with a 40 deg C temperature assumed in the box.
The next calculator allows you to do what-if calculations with various inputs concerning the package thermal properties, the environment, and the heat sink solution. It can be used to demonstrate the degree of performance leverage can be had by attacking the BS170 case-to-air resistance with a heat sink solution. Results are of course dependent on the values you assume and load. YMMV.
Have fun with this! I hope it takes some of the 'hot air' off the thread.
73 JZ KJ4A
On Mon, May 1, 2023 at 3:28?AM John Zbrozek <jdzbrozek@...> wrote:
Willie,
That looks right to me. It's an extreme design point that I would never use, but with the subtraction of the 2V Vds your math works out.
JZ
On Mon, May 1, 2023, 1:10 AM William Smith <w_smith@...> wrote:
Right, so 58 volts across the load at 400 mA is 23W in the load and 800mW in the transistor, no?
73, Willie N1JBJ
On May 1, 2023, at 12:51 AM, Tony Scaminaci <tonyscam@...> wrote:
?
Willie,
The voltage at the drain is never 60 volts when the device is on. With a 5 ohm on resistance and a drain current of 400 mA, Vds would be 2V. The 60V max spec would apply only when the transistor is off.
Tony AC9QY
|
