Wideband transistor overheat

[Alessio]

Please consider the TV aerial system in the smallest pic.

VHF/UHF amplifier is a circuit projected around a wideband transistor which uses all the power from the PSU.

The wideband transistor is a BFG135, its maximum working frequency is 7GHz and the maximum allowable
dissipated power is 1W. Highest frequency for UHF is 900MHz so BFG135 must be "sleeping", in sense it
can work so much faster than the actual UHF frequency.

But if I put a finger on the top of BFG135 I can keep it in place for MAX 3 seconds, then my finger
gets very hot, so BFG135 is not working properly.
At the bottom layer of PCB there is a black (overheated) area under the transistor.

The entire VHF/UHF amplifier is supplied with 12V, 200mA, so there is a power consumption of 2.4Watt.
According to datasheet specification, Rthj-s (similar to Rthj-c) is 30K/W, so 1W result 30K, 2.4W result 72K.

Assuming the entire power demand of the amplifier must be drawn by BFG135, I'm suspecting the transistor
is dissipating 2.4W, so this could be the reason why it's "overheating".

But there are other reasons that could overheat this 7GHz transistor, so I've inspected PSU:
It's not the original PSU required for this VHF/UHF amplifier, I'll tell you why.

First look at the original PSU figured in the biggest pic.

Due to high tendence of PRIMARY IC & RELATED CIRCUIT to reverse in PROTECTION MODE the entire original power
supply has been substituted FOR EMERGENCY with a generic SMPS with this feature: 12V voltage regulated, 1A.
Obviously a generic SMPS has bad filter circuit, in any case VHF/UHF amplifier works fine untill now.

Please note VHF/UHF amplifier is working with this GENERIC (scrappy) SMPS since 2006.

Now I'm using the original CLCC filter in a linear PSU, 12V regulated.
The actual current consumption is unchanged: 200mA and the wideband transistor continue to overheat.

I'll post clear picture of VHF/UHF amplifier if this could help.

My big questions are: Could the generic SMPS have ruined the wideband transistor during this 5 years?
Does BFG135 overheat like this because it's too small for substaing the job?

[sfiga69]

you are sure that with the aging of the CS or infiltration of humidity, the transistor, due to stray capacitance generated by oxidation, do not auto-oscillation?

[solutions]

1) Sounds like a class A amplifier.  It will run hot
2) Sounds like you may have a bad thermal connection between the board and the transistor...doesn't take much degradation here to cook a part
3) You may be overdriving the input of the amp with a near in signal or too high a gain of an antenna...the design sounds too cheap to have an AGC in it
4) 7GHz has nothing to do with it "sleeping", but a lot to do with the gain you'll get out of it at 900MHz
5) A big SMPS like 1 Amp will suck at low current voltage regulation, so that may also explain why the transistor runs excessively hot...it may be running 15V for instance at light load.
6) 5 years is great..replace the transistor, grease and bolt it down really well, wait 5 more years, rinse, repeat.  If you are totally paranoid, you can see what the light load regulation is and put a 12V linear LDO in behind the SMPS, but for 5 years of it working, why bother?

[Alessio]

Ciao sfiga69.

Aging is the most common and costly source of electronic problem.
The amplifier take place where no infiltration of humidity occurs.
I can't say (and measure) if the transistor auto-oscillate.

1) I didn't consider this.
2) Surely there's a bad thermal connection: you could see when I'll post a pic next week.
3) No built in AGC here, but the gain of the antenna is far to overdrive the input of the amp: without amp I can receive few signals.
4)  
5) The SMPS is voltage regulated, 12V measured. I also tried a linear PSU with the original CLCC filter but the transistor is always hot.
6) I'll buy a new BFG135.

Take a look of the picture. This is the solder layer, in the component layer there is the "burnt area"

Thanks you, I'll leave the amplifier working for the rest of its days.