|
Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets. |
|
Thread Tools |
13th May 2014, 3:30 am | #1 |
Hexode
Join Date: Jan 2010
Location: Blackburn, Lancashire, UK.
Posts: 399
|
Transistor junction temperature query
Is there a way to calculate the junction temperature of a Transistor by the temperature of the case?
I have a SD1127 Transistor operating in class C as a final operating at 13.8V drawing approx 500mA, and the temperature of the casing after several minutes reaches 59oC according to an IR Thermometer - I can hold my finger on the Transistor for around 12 Seconds before it becomes uncomfortable. I've been reading various Heatsink, Thermal resistance, and other associated information all day and in the process have overloaded my brain and am now truly baffled. With the aid of the said Transistor's datasheet in addition to the information provided above, please would someone be so kind as to tell me what the rough junction temperature is? Here is the datasheet for the SD1127 - http://html.alldatasheet.com/html-pd.../1/SD1127.html |
13th May 2014, 7:59 am | #2 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
|
Re: Transistor junction temperature query
Only if you know the performance of the heatsink or the actual power dissipation of the transistor. Then it's just like Ohm's law.
The data sheet should give the thermal resistance from junction to the mounting face of the transistor in degrees C. Multiply by the power flow and you get the difference in temperature, add to the mounting flange temp and there you are. If the heatsink is a commercial one with specified degrees C per watt, you can estimate the power flow from its temperature rise over ambient with it oriented as per its data in free air. Getting the power flow is the difficult bit, and measuring the temp at the mounting is awkward. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
13th May 2014, 8:00 am | #3 |
Nonode
Join Date: Aug 2007
Location: South Bradford, West Yorkshire, UK.
Posts: 2,571
|
Re: Transistor junction temperature query
The junction to case thermal resistance is 21.9degrees C /W. From your measurements the transistor is dissipating 6.9W. Multiply this by the thermal resistance and you get 151 degrees C. This is the temperature rise between the case and the junction. If the case is at 59 degrees C then the junction is at 210degrees C which would be out of spec.
This assumes all the power is being dissipated in the transistor. I assume this is being used as a PA transistor so the actual dissipation of the transistor will be approximately 6.9W minus the measured output power. E.g. if you are getting 4W out the transistor will be dissipating approximately 2.9W leading to a temperature rise of 63.5 degrees C and a junction temperature of approx 122degrees C. Hope that helps Keith |
13th May 2014, 10:10 am | #4 |
Octode
Join Date: Dec 2008
Location: Wincanton, Somerset, UK.
Posts: 1,757
|
Re: Transistor junction temperature query
There are two pieces of information missing which are needed to answer your question:
1. What's the RF output power? 2. is the transistor operating with any kind of heatsink, and if so, what is it / what does it look like? |
13th May 2014, 11:29 am | #5 |
Hexode
Join Date: Jan 2010
Location: Blackburn, Lancashire, UK.
Posts: 399
|
Re: Transistor junction temperature query
Thanks for the replies, Guys, as it's really cleared things up.
John, the attached heatsink specification states 48oC/W, and the output power of the said transistor is 1W into a dummy load. Here is an image of the type of heatsink attached - |
13th May 2014, 8:00 pm | #6 |
Octode
Join Date: Dec 2008
Location: Wincanton, Somerset, UK.
Posts: 1,757
|
Re: Transistor junction temperature query
OK, well, somethig doesn't add up here.
If your crinkly is about 48C/W (which sounds about right), and the transistor + heatsink are at about 60C (finger test agrees), and your ambient is 18C? (and no fan?), the thermal equation says you're only disspating [60-18]/48 watts, which is a lot less than the power (DCin - RF out)? Anyway, with the figures you have already, we can say that the temperature of the case will sit at around two-thirds of the difference between the ambient air and the junction temp., as the resistance from junction-to-case is about half that of the case-to-ambient, (22 & 48 C/W). Therefore, if you add half the temperature rise of the case (say 40 degrees = 20C) to the temperature of the case (say 60C) you have the junction temperature (about 80C). |
13th May 2014, 11:04 pm | #7 |
Hexode
Join Date: Jan 2010
Location: Blackburn, Lancashire, UK.
Posts: 399
|
Re: Transistor junction temperature query
Hi John,
Thanks a lot for the calculation, I really appreciate it. When I say the SD1127 final was drawing 500mA at 13.8V it was an estimation, well, more of a wild guess really; The whole circuit draws around 750mA at 13.8V, and it consists of a 2N4427 driving the SD1127, with a PLL circuit driving the 2N4427 with 3 LED's. My wild guess is that the PLL with LED's probably draws around 100mA, the 2N4427 (SD1127 driver) around 150mA, with the remaining 500mA or thereabouts for the final which will draw the most. There's no way of determining the Collector current or Vce at present as my multimeter is broken, so I've just quoted the current draw as shown on my PSU. Do you think my estimation of the Collector current could be well out? Again, many thanks for your contributions. |
19th May 2014, 7:36 pm | #8 | |
Octode
Join Date: Dec 2008
Location: Wincanton, Somerset, UK.
Posts: 1,757
|
Re: Transistor junction temperature query
Quote:
You'd normally expect a class C stage to operate at >50% efficiency if correctly aligned, so for 1W out, perhaps 1W dissipation, whch is in the right ball-park. That would put the SD output-stage current at about 150mA. Your "thermal" equation puts the dissipation at around 870mW, equivalent to 53% efficiency. John |
|