If so, you'll understand relying on "radiation across a 4mm air gap" may not be the best solution 😀
Use da grease.
You are 'right' about the "slow response to delta Tj". But the best strategies allow for this while still providing conditions where the amp neither releases the Holy Smoke or departs from low THD under all important scenarios.
You may like to list these 'important scenarios'.
There's an ~4mm gap.
How does the grease stay there when cold? And when it warms up where does it flow to?
That is why I thought about but have not implemented an aluminium spacer to fill the gap and conduct heat to the sensor transistor.
In the meantime I will use NJL for the second build and investigate whether the routing of the diode string wiring is a problem.
The sensor transistor relies on the fact that the output transistor collector lead follows the junction temperature quite closely and quickly. Run some thermal grease from the collector lead to the sense transistor (use a small amount - you simply want to couple the SMD sensor a bit better to the output transistor collector lead)
Remember heat travels up the device leads - not just radiation.
Adjust for the correct Iq when the heatsinks are warm (hot). That way, you will start with low Iq and as the amp warms, it will settle in the correct bias range.
Your heatsink size and configuration will of course also affect the system response.
Thermal grease used as I describe is quite ok
Last edited:
You are right, the collector lead is a solid lump of copper connected directly to the junction.
It follows Tj accurately and fairly quickly.
It's only it's length from the junction that introduces some delay and some temperature attenuation.
But the PCB has the sensor transistor on the PCB. It is not on and not near the collector lead out.
The gap between these is >3mm
It follows Tj accurately and fairly quickly.
It's only it's length from the junction that introduces some delay and some temperature attenuation.
But the PCB has the sensor transistor on the PCB. It is not on and not near the collector lead out.
The gap between these is >3mm
Last edited:
Conducting heat from Collector to Sensor ?
I wonder if I could recover the backplate of a damaged To126 and make a solid copper strip to span the gap from the collector lead to the sensor?
A 6mm to 7mm long strip about 1mm thick and 2.5mm wide could be soldered to the wide part of the collector lead as it exits the package and then bent slightly to "lean on" the top of the sot23 sensor.
A spot of Thermal Goop between the sot23 and the copper strip to help with heat transfer.
I wonder if I could recover the backplate of a damaged To126 and make a solid copper strip to span the gap from the collector lead to the sensor?
A 6mm to 7mm long strip about 1mm thick and 2.5mm wide could be soldered to the wide part of the collector lead as it exits the package and then bent slightly to "lean on" the top of the sot23 sensor.
A spot of Thermal Goop between the sot23 and the copper strip to help with heat transfer.
Andrew's, just couple the two devices with a small amount of thermal grease
The edge to edge distace between the two collector pads is 2.96 mm. I measured it. That's quite ok to bridge with thermal grease.
No need to agonize over it!
🙂
The edge to edge distace between the two collector pads is 2.96 mm. I measured it. That's quite ok to bridge with thermal grease.
No need to agonize over it!
🙂
Last edited:
The grease is just going to run out.
It will not fill a 3mm gap and stay put as the sink heats up.
I'll end up with a puddle on the floor.
It will not fill a 3mm gap and stay put as the sink heats up.
I'll end up with a puddle on the floor.
and the lead to lead distance is ?
The grease does not run Andrew.
You are not filling a 3mm gap. You are using a small amount to couple the power device collector lead to the SMD device to improve the tracking (performance of Iq comp circuits are dependent upon the heatsink size and a whole lot of other factors as you well know). The grease also holds the two devices pretty much iso-thermal.
You are not filling a 3mm gap. You are using a small amount to couple the power device collector lead to the SMD device to improve the tracking (performance of Iq comp circuits are dependent upon the heatsink size and a whole lot of other factors as you well know). The grease also holds the two devices pretty much iso-thermal.
Hi, having trouble sourcing these (KSA1381/KSC3503) in the same gain grade.
For the SX amp: Is it possible to use KSA1220A and KSC2690A?
Also difficult to find bc547c and bc557c actually. How important is the "c" gain grade here? I read that bc550c and bc560c would also be fine?
Using rs-components and mouser.
There's also Heatsink Compound, the white stuff which is impossible to get off your clothes. Much more likely to stay put compared to Silicon Grease.
Its the compound I am talking about - this is the stuff I use.
Sorry for the confusion.
(Andrew, you can get it from RS)
Sorry for the confusion.
(Andrew, you can get it from RS)
Hello Bambadoo,
for the TIS (i.e. VAS) transistors, you really need to use a low Cob device. The KSA1381 specifies this at c. 3.1pF while theKSA1220A specifies Cob at c. 26pF.
For the driver stage, the KSA 1220 and its complement are quite ok and I recommend that you use them instead of the 1381 and 3503. But for the TIS, you may run into rail sticking and potentially worse, instability. I have not tried the 1220/2690 in the nx or sx-Amp, but if there is any instability due to high Cob, it would require the compensation design to be looked at again and/or potential HF parasitics to be addressed.
The short of it is, you need to use the specified devices if you want to avoid potential problems.
Sorry I don't have a better answer for you at this stage.
For the small signal devices, I am really surprised you cannot get the 547C/557C.
I just looked on Mouser and Digikey - indeed it looks like they are getting harder to find. I think the mainstream suppliers are stopping with the TO-92 devices.
The SMD versions are available (857C/847C) - however, that is no comfort for us builders.
The SMD versions are available (857C/847C) - however, that is no comfort for us builders.
My tube of thermal compound is now over 30years old, maybe nearly 40years old.
It gets used for thin mica insulated power devices.
I am down to about 2.6ml left and will have to buy some more in the next 5years.
Took apart a damaged To126 yesterday, but the copper blank is the wrong shape and size to make a good heat conducting path to the sensor. It would make a poor one and might be worth experimenting with. It needs to be about 9mm long, not the 6mm to 7mm that I suggested in a previous post.
That's the path length you are suggesting I fill with grease/Thermal Compound !
Last edited:
Hi Bonsai I have kept the emitter to emitter voltage for a pair of transistor to 100mv instead of 82mv mentioned in the document.As my heat sink is huge I just want to know increasing the blase voltage have any issues on the stability of the amplifier.Please let me know.The Heat is very stable as I have big heat sink.
I'd like to build a pair of these and am etting ready to order parts. I searched the thread but can't find a BOM in excel tomakeorderin parts easier. Does one exist?
Thanks
Thanks