So let me make sure I understand. The driver transistors should be thermally coupled with the output transistors. To do that, they must be in the same heatsink.
So in my PCB design, I made it like that, as shown in the picture above. The output transistors are those very large ones, 2SC5200 and 2SA1943, being driven my MJE253 and MJE243 in between them. Is thermally coupling them exactly like that? I don't have a 3d model for a heatsink but just imagine those 4 transistors screwed in to a single heatsink.
So in my PCB design, I made it like that, as shown in the picture above. The output transistors are those very large ones, 2SC5200 and 2SA1943, being driven my MJE253 and MJE243 in between them. Is thermally coupling them exactly like that? I don't have a 3d model for a heatsink but just imagine those 4 transistors screwed in to a single heatsink.
CFP the drivers are isolated from heatsink.
Darlington stage it is more common.
for simplicity CFP one drop is justified by the driver
The VBE multiplier and drivers thermal track.
You will be thermal tracking T0-92 device with T0-126 drivers only.
they are isolated from heatsink
Darlington stage it is more common.
for simplicity CFP one drop is justified by the driver
The VBE multiplier and drivers thermal track.
You will be thermal tracking T0-92 device with T0-126 drivers only.
they are isolated from heatsink
As you can see in the P3A
Using T0-92 Vbe multiplier is mounted next to one of the T0-126 driver transistors
and zip tie to hold them in place.
Another possible better solution is use T0-126 for VBE and Driver.
Both could be mounted on a heatsink. Otherwise isolated from main heatsink
in either example
Using T0-92 Vbe multiplier is mounted next to one of the T0-126 driver transistors
and zip tie to hold them in place.
Another possible better solution is use T0-126 for VBE and Driver.
Both could be mounted on a heatsink. Otherwise isolated from main heatsink
in either example
Last edited:
Cool, but if VBE multiplier is only connected to one of the driver transistors, wouldn't it have the chance to cause imbalance for the two drivers?
Not really imbalance.
But yes good point.
I just found the quickest example I know about.
It just showing that you thermal track the drivers only.
I better design as mentioned would be to use a more common thermal tracking
device such as a TO-126 package for VBE. And if it is actual same device as
drivers even better.
Then yes the VBE and both drivers could be thermal tracked.
Or 3 devices mounted on a heatsink.
Which is a separate heatsink not mounted on main heatsink.
But yes good point.
I just found the quickest example I know about.
It just showing that you thermal track the drivers only.
I better design as mentioned would be to use a more common thermal tracking
device such as a TO-126 package for VBE. And if it is actual same device as
drivers even better.
Then yes the VBE and both drivers could be thermal tracked.
Or 3 devices mounted on a heatsink.
Which is a separate heatsink not mounted on main heatsink.
I guess it is about the stability of adjusted quiescent current resp. idle through the output stage.
Then check out this threads:
https://www.diyaudio.com/community/threads/differences-between-various-vbe-multipliers.117047/
https://www.diyaudio.com/community/threads/how-do-i-get-a-stable-bias-current.308041/
https://www.diyaudio.com/community/threads/vbe-multiplier-current-compensation-resistor.365670/
https://sound-au.com/amp_design.htm#s32
https://www.diyaudio.com/community/threads/vbe-thermal-coupling-issue.48971/
https://www.diyaudio.com/community/threads/optimizing-the-vbe-multiplier.216385/
https://www.diyaudio.com/community/threads/diodes-and-vbe-multipliers.385184/
For the most part.
It is simple, because of the topology and only one drop from
the driver.
Thermal behavior is justified by the drivers.
Only the drivers are thermal tracked and your interested in
not letting the drivers get to hot.
Hence mounting both drivers to a seperate heatsink and thermal tracking
with a similar device. Is better approach to keep CFP amplifier
from thermal runaway / bias thermal tracking.
if you attached the drivers to the hotter output transistors
it would heat them up more and make it thermally runaway
easier.
It is simple, because of the topology and only one drop from
the driver.
Thermal behavior is justified by the drivers.
Only the drivers are thermal tracked and your interested in
not letting the drivers get to hot.
Hence mounting both drivers to a seperate heatsink and thermal tracking
with a similar device. Is better approach to keep CFP amplifier
from thermal runaway / bias thermal tracking.
if you attached the drivers to the hotter output transistors
it would heat them up more and make it thermally runaway
easier.
I see, thanks a lot for the detailed explanation and even an example case. Many thanks!
I assume that if I am going to use both TO-126 package for VBE multiplier and driver, I would need to use a thermal pad, like the grey one behind the devices below,
I am indeed thinking of using both TO-126 package, I have yet to decide if I want the three devices to be coupled hence a heatsink, or just couple the multiplier with just one of the drivers. But still, thank you for you answers.
