Unilab 10m FM conversion Transmitter PA mods - Final Stage

So, we now turn our attention to the final part of the modification. I suspect this will be the hardest part of the whole project. The existing PA needs a fair amount of work to get it running at 10m.

The challenges arises because:

• The BGY32 module is designed to work between 66MHz and 88MHz. This has to be replaced with a home brew alternative, and
• The MRF247 was never specified outside the 136-175MHz band, and
•  I will try to use transistors that are readily available.

That last point needs some clarification. I have a reasonable junk-box so I have plenty of devices I can recover.  However, I always try to build things that can be replicated. Which usually means I have more than 1 of whatever I used so someone else who is stuck can be helped. In this instance I don't know just what I have that could be pressed into service here so it will be a challenge.

If the MRF247 cannot be coaxed into service it will effectively mean a whole new PA strip.  Perhaps Murphy is elsewhere today.

Since the MRF247 is specified as a 75W transistor I decided that would be the target output power. However, I'd design the matching for each stage for a 30% margin for the output power i.e. almost 100W. It may transpire that I have more gain than the 10dB per stage I am allowing for and since the MRF247 could be driven towards 100W I wasn't going to say no to the extra 25W if it eventuated. There is no data that I could find on the series equivalent impedances outside the specified range of 136MHz to 175MHz so I had to make some assumptions for the matching networks. I assumed the following:

In an attempt to understand how the Unilab designers got the MRF247 to work outside it's specified range I looked closely at the input networks. I knew the BGY32 was a 50ohm output device. And the MRF247 was likely to have a series input impedance of 0.57+j0.3 at 77MHz. To test this assumption I tried a few software packages until I found one that helped me understand this: Iowa Hills Smith Chart software.

I found this software intuitively easy to use. It wasn't the first Smith Chart software that I tried but like the filter software from Iowa Hills this is clearly written by someone who uses the software. I do recommend you try this software.

With the values shown on the circuit I was able to confirm a match was possible at 77MHz between 50ohmns and the PA transistors input impedance of 0.57+j0.3. This was true whether the stripline was modelled and an inductor or as a stripline. In the end I left is as an inductor:

The next step was to see what I would have to change to get this existing input matching section to match a source of 9.4ohms. This would be the collector load required for a driver to deliver almost 10W. For a class C amplifier this is derived from the formula Po = Vcc * Vcc / ( 2 * R ).

With just an extra capacitor and a different inductor we can achieve a good match. An additional 180p capacitor at the BGY32 end of the stripline and an inductor of 86nH replacing the coil that is there is all that might be required.

So far, so good. I repeated the exercise for the output stage and decided that two 34.5nH coils and a 510pF capacitor to pad C603, or a 220p and 270p in parallel, would be a good starting point.

I'm going to make the changes outlined above and check my junk box for driver transistors before I progress any further.

73's
Richard