F5 power amplifier

Re: Re: Heat sink thermal resistance

Coulomb said:


One thing I have had trouble with when coming up with a good heatsink is not just the Tsink-avg, but the Tjunction. The Tj is probably your most critical parameter for longevity. The TO-247 Device max operating Tj is 150Deg C. Nominal operating temperature is 74Deg C. I think Mr pass has indicated a good comprimise is about 100Deg C.

It takes a lot of effort to get a TO-247 dissipating 32 Watts to operate under 100Deg C. Bodging the device to the heatsink will not work for a device operating in this range.

Regards

Anthony

So we shouldn't use too much heatsink because the MOSFETs will run too cool ?
 
Hi,
when designing your heatsink, you have choices.
Either set the maximum junction temperature (Tj) or set the maximum case temperature (Tc), or set the maximum heatsink temperature (Ts).

Let's take an irf240 which can dissipate 150W if Tc<=25degC which also brings Tj=Tjmax=150degC.
This gives Rth j-c = [150 - 25] / 150 =0.83C/W. agrees with the datasheet.
Rth c-s =0.4C/W, direct mounting using the best interface compound is ~ 0.2C/W

Let's also assume that each irf240 has to dissipate 35W in a worst case ambient temperature (Ta) of 30degC.

Method 1.)
Set Tj<=100degC
Tc=Tj - [35W * Rth j-c] = 70.95degC
Ts=Tc - [35W * Rth c-s] = 56.95degC
Ta=30degC
Rth s-a = [Ts-Ta] / Pd = 56.95 - 30 / 35 =0.77

From the manufacturers graph we can find the de-rating factor for Ts-Ta (deltaT) = ~27Cdegrees @ ~ 1.3
The required heatsink for one 35W device is 0.77/1.3 ~ 0.592C/W at sea level.

Method 2.)
Set Tc=70degC
Ts=Tc - [35W * Rth c-s] = 56degC
Ta=30degC
Rth s-a = [Ts-Ta] / Pd = 56 - 30 / 35 =0.74

From the manufacturers graph we can find the de-rating factor for Ts-Ta (deltaT) = ~26Cdegrees @ ~ 1.3
The required heatsink for one 35W device is 0.77/1.3 ~ 0.571C/W at sea level.


Method 3.)
Set Ts<=50degC
Rth s-a = DeltaT / Pd = 20 / 35 =0.57.
Use the same de-rating graph for DeltaT= 20Cdegrees gives ~ 1.5
the required heatsink for one 35W device at sea level with a maximum heatsink temperature of 50degC is ~ 0.38C/W.

Note:
for the conditions in method 1. the device is running at 55% of maximum rating.
For the more onerous conditions used in method 3. the device is running at 34% of maximum rating.

Also each method uses exactly the same formulae. Methods 2 & 3 just miss out the first one or two lines of method 1.

If you want to fit two devices on a common heatsink then the minimum heatsink dissipation will be double the single device size only if the two devices can be located at the optimum location for best dissipation of the combined sink and device.

Two devices operating @ Tc=70degC require a heatsink better than 0.28C/W at sea level and Ta max=30degC.

Each power amp and each ambient condition and each operational circumstance can and should be used to determine one's heatsink requirement.

Do you want a copy of these heatsink requirement calculations in the F5 construction thread?
 
Re: Re: Re: Re: Re: Re: Heat sink thermal resistance

labjr said:



I misunderstood what you were trying to explain. I thought you meant that the device also had to run at a minimum temperature.

My bad, I am having one of those irritable days. :(

These Devices will operate as follows.

Operating Junction and Storage Temperature Range TJ, Tstg - 55 to + 150
°C

Anthony
 
AndrewT said:
Hi,
when designing your heatsink, you have choices.
Either set the maximum junction temperature (Tj) or set the maximum case temperature (Tc), or set the maximum heatsink temperature (Ts).

Let's take an irf240 which can dissipate 150W if Tc<=25degC which also brings Tj=Tjmax=150degC.
This gives Rth j-c = [150 - 25] / 150 =0.83C/W. agrees with the datasheet.
Rth c-s =0.4C/W, direct mounting using the best interface compound is ~ 0.2C/W

Let's also assume that each irf240 has to dissipate 35W in a worst case ambient temperature (Ta) of 30degC.

Method 1.)
Set Tj<=100degC
Tc=Tj - [35W * Rth j-c] = 70.95degC
Ts=Tc - [35W * Rth c-s] = 56.95degC
Ta=30degC
Rth s-a = [Ts-Ta] / Pd = 56.95 - 30 / 35 =0.77

From the manufacturers graph we can find the de-rating factor for Ts-Ta (deltaT) = ~27Cdegrees @ ~ 1.3
The required heatsink for one 35W device is 0.77/1.3 ~ 0.592C/W at sea level.

Method 2.)
Set Tc=70degC
Ts=Tc - [35W * Rth c-s] = 56degC
Ta=30degC
Rth s-a = [Ts-Ta] / Pd = 56 - 30 / 35 =0.74

From the manufacturers graph we can find the de-rating factor for Ts-Ta (deltaT) = ~26Cdegrees @ ~ 1.3
The required heatsink for one 35W device is 0.77/1.3 ~ 0.571C/W at sea level.


Method 3.)
Set Ts<=50degC
Rth s-a = DeltaT / Pd = 20 / 35 =0.57.
Use the same de-rating graph for DeltaT= 20Cdegrees gives ~ 1.5
the required heatsink for one 35W device at sea level with a maximum heatsink temperature of 50degC is ~ 0.38C/W.

Note:
for the conditions in method 1. the device is running at 55% of maximum rating.
For the more onerous conditions used in method 3. the device is running at 34% of maximum rating.

Also each method uses exactly the same formulae. Methods 2 & 3 just miss out the first one or two lines of method 1.

If you want to fit two devices on a common heatsink then the minimum heatsink dissipation will be double the single device size only if the two devices can be located at the optimum location for best dissipation of the combined sink and device.

Two devices operating @ Tc=70degC require a heatsink better than 0.28C/W at sea level and Ta max=30degC.

Each power amp and each ambient condition and each operational circumstance can and should be used to determine one's heatsink requirement.

Do you want a copy of these heatsink requirement calculations in the F5 construction thread?

I live at 415 ft above sea level. How much more do I need? :D
 
Hello Andrew, I beleive the Newer Vishay/IR IRFP250N is a good device to use in the F5. It has a RthJC of only .65 and the input capacitance has dropped from 2900Pf on the old model to 2190Pf on the new model. As this device is not used in Parallel on the F5 the slightly higher Ic over the IRFP240N should not make an appreciable difference in the circuit.

The advantage is if you run the same thermal calculations, the lower RthJC significantly reduces the Tj under similar condions to the IRFP240 in your examples.

Anthony
 
labjr said:
So it's virtually impossible to keep the temperature of the device under 100 C while dissipating that much power. Not even with more heatsinking or a copper heat spreader to suck away heat from output devices faster?


As you can see from Andrew's calculations it is difficult but not impossible, given the right selection of parts.

Anthony
 
Coulomb said:



The FQA19N20C will run at a cooler Tjunction temperature than the FQA12P20.

The FQA19N20C has a RthJC of .69 a little higher than an IRFP250N but a lot lower than an IRF240N. It also has a much lower input and output capacitance range that the two IR/Vishay devices.

Anthony


I thought the the two Fairchild parts were complimentary devices with pretty much the same specs. I guess they aren't
 
In choosing a heatsink lets not forget that we dont want the device to be as cool as possible. I read the manual again and Nelson makes no note of device temp but does say it is good to have the sinks be 20C above an ambient of 25C. Remember that he says if you remove the thermistor that it will take longer for the device to get up to temp? I think it supposed to be at 58-68C but not sure. Its way back in the thread somewhere.
Uriah
 
udailey said:
In choosing a heatsink lets not forget that we dont want the device to be as cool as possible. I read the manual again and Nelson makes no note of device temp but does say it is good to have the sinks be 20C above an ambient of 25C. Remember that he says if you remove the thermistor that it will take longer for the device to get up to temp? I think it supposed to be at 58-68C but not sure. Its way back in the thread somewhere.
Uriah


Not so. It's all about bias.


Magura :)