[Amps] Power Transistor Question -- 2SC5125

[Manfred Mornhinweg]

Hi Steve!

> Spec sheet says they are rated for 80W output each or 160W per pair
> up to 175Mhz.
> 
> Wondering why ICOM only runs them at 100W?

Good question. Many people ask themselves the same thing. I will try to 
explain:

The RF power output rating of a transistor is a flexible thing. It 
depends very much on the class of operation, on the quality of 
heatsinking used, and also on the lifetime you want to obtain, the 
distortion level allowable, etc.

Let's see the specs of this transistor: Its collector to base breakdown 
voltage is just 35V. So, you cannot use it with a supply of more than 
about 15V. Clearly, it's for radios fed from 13.8V with a little 
plus/minus tolerance. Actually it's rated at 12.5V, which is a typical 
working voltage in a 13.8V radio, after the voltage drops in the feed 
cable, fuse, etc.

The current is rated as 25A maximum DC average. The current gain 
actually holds up to even higher currents, so from this point of view 
the transistor could carry more than 25A on RF peaks. But there is the 
problem of saturation, and of ohmic resistance! As the RF peaks go 
higher and higher, the losses due to saturation and resistance rise even 
more! Not only does that cause a dissipation problem, but also it 
compresses the waveform (distortion) and reduces efficiency. This 
transistor CAN produce in excess of 100W output per piece, but at 
reduced efficiency, high distortion and very high dissipation! That's 
basically why Mitsubishi rates it at 80 Watt maximum, and recommends it 
for 70 Watt operation.

In practice, even that 70W rating is a bit overblown, when running class 
AB, as is required for linear service. In class C, the efficiency could 
be as high as 70%, and there it's reasonable to run this transistor at 
70 Watt output. In class B in a tuned amplifier, it's a tad above 60%. 
But in an untuned class AB amplifier, as is required for broadband HF 
service, it's unreasonable to expect more than 50% efficiency. So, for 
every Watt of RF you get another Watt of heat.

The transistor is rated at 170 Watt dissipation. Sounds hugely enough, 
right?  Well, it's not that rosy: This 170 Watt rating is based on two 
conditions: First, that you allow the junction to heat up to its 
absolute maximum momentary temperature, 200 degrees Celsius. Running the 
transistor that hot will allow it to last only for a very short time, 
maybe just hours! You should run it at no more than 150 degrees to 
obtain a good (almost eternal) lifespan. Second, the 170 Watt figure is 
based on the requirement that you keep the transistor's mounting surface 
below 25 degrees Celsius! That's of course totally impossible in a 
practical radio! If you can keep the transistor mounting surface at 80 
degrees, call yourself lucky.

So, instead of having the whole span of 200-25 = 175 degrees between the 
junction and the case, in practice you have only 150-80 = 70 degrees! 
That reduces the real world dissipation capability of this transistor to 
  about 68 Watt, instead of the rated 170 Watt!

The exact value depends of course on the kind of heatsinking you use. 
With a water cooled copper block, ultra-good heat conducting paste, 
lapped mounting surfaces, etc, you might be able to push this transistor 
to 100 Watt practical dissipation. But why? It's cheaper to use two 
transistors, or a larger one, coupled with a plain and normal aluminum 
heatsink and a little fan on it!

So that's what any radio manufacturer has to do, if he wants to stay 
competitive: Start from the fact that he wants 100 Watt output, that 
this produces another 100 Watt of dissipation; and then choose the 
combination of transistors and heatsink that will give this dissipation 
at the lowest possible cost. So we end up with a radio that has two 
transistors rated for a total dissipation of 340 Watt, but in practice 
is run at only 100 Watt maximum! This allows to run the (relatively 
small) heatsink at 60, 70 or even 80 degrees Celsius, and the transistor 
mounting surface at up to 100 degrees, while still obtaining good 
reliability.

As a byproduct of running the transistor below its maximum RF power 
spec, you are running it in a current range where the linearity is 
better, due to less gain variation and less effect from saturation. This 
results in lower distortion, allowing the radio to meet FCC specs and 
allowing other local hams to use the same band while you are on the air!

> There is an adjust pot that will allow the output to be turned up 
> somewhat -- up to around 150W.

Yes. Don't do it! The purpose of this pot is setting the radio to a nice 
and precise 100 Watt, to keep both the distortion and the dissipation in 
their acceptable ranges.

> What is the undesirable effect of turning up the power to say... 
> 125W, if any?

Severe distortion, resulting in splatter all over the band. In addition, 
in high duty cycle modes (RTTY, or even speech-processed SSB) you will 
exceed the safe dissipation, and shorten the lives of the transistors. 
And then, depending on the design of the radio (I haven't checked the 
Icom for this), you might also loose the SWR protection by misadjusting 
this pot! It would be a darn stupid thing to do. Despite being such a 
stupid thing, some hams just do it, giving clear on-air evidence of 
their technical ignorance. Others ask first about the implications, and 
that's much more clever.

The fact is that your radio is designed for 100 Watt, and not a tad 
more! If you feel so, take off the covers and look at the final stage. 
You will see that it has an output matching transformer with one primary 
turn and four secondary turns. So it has a voltage transforming ratio of 
1:4. The primary is connected between the two collectors. Assuming the 
transistors are getting 13.0V DC (after 0.8V drop in the cables and 
fuse), and assuming that the transistors run into hard saturation at 
0.5V, in this push pull circuit you get a maximum peak voltage of 25V 
across the primary. That means a maximum RMS voltage of 17.7V. On the 
secondary of the transformer you would then get 71V. Across 50 Ohm, that 
gives just 100 Watt! If you drive the transistors any harder, you will 
indeed get more RF power, but instead of a clean sine wave it will be a 
clipped wave, tending all the way toward a square wave at a level of 
about 150 Watt output. The RF harmonics are not the real problem, as 
they are rejected by the low pass filter following the final amplifier. 
The problem is that this signal clipping also causes severe 
intermodulation distortion! All the large odd-order IMD products fall 
inside the band, and will be radiated, causing severe splatter over a 
bandwidth of 100kHz or even more!

So, if you would like to increase the power of your radio, by milking 
the transistors for all they are worth, and risking their destruction, 
at the very least you would have to add a fifth turn to the output 
transformer secondary winding. Then you could adjust that pot to 145 
Watt output before running the transistors into saturation. But the 
whole risk of burning the transistors, and perhaps some other parts, is 
yours. You would need a power supply capable of delivering about 30A 
instead of 22.

You would be increasing your transmit signal by a stunning 1.6 dB, or 
about one quarter of an S-unit... Do you think that's worthwhile? I 
don't. I would rather suggest to build a linear amplifier to raise your 
power by 10 or almost 12dB! That's certainly more worthwhile.

73,
Manfred.

----------------------------
Visit my hobby website!
http://ludens.cl
----------------------------