The AM Forum

Good day,

This is an update on the work from an old post on the subject I started several months ago.

Thanks for the emails and prior posts. The unit is packaged as shown here and the updated schematic and some results noted. I guess the most significant facts found are:

1. Yes, it is possible to apply UHF or possibly higher operating frequency bipolar or MOS devices to HF. In the current case 500 MHz device operating at 3-30 MHz. Next step, try some W-CDMA 125 W S band (2 GHz) LDMOS down at HF in a PP arrangement. All this requires care about negative feedback

2. You can build this stuff on FR-4 using copper tape as the pad lands. If the current is high, in my case, about 6-7 amp peak, then add bus bar wire as required.

3. The PP balance from 2nd harmonic measurements, places it at 30 dBc or better. Need to measure 2-tone IMD to further verify how well balanced is this arrangement.

4. The addition of the LPF with diplexer is essential. Any screwball load Z will not affect stability with the HP diplexer in place.

I have updated the schematic to include this item. On the SA pix the screen top is 60 dBm, a 30 dB attenuator in place. Accurate measure done with a Booton power meter. The LPF is a 4-section, good for 24 dB and the PP balance providing 30 dB.
Alan



This was a FUN project.

Here is a pix of the inside with power supply, amp and filter.

Alan

Very cool. What is the price of the transistors?

Very nice amp.  Looks like your experiment paid out.  Looks pretty clean too.
Couldn't immediately find the MRF-327 at Mouser but here's one similar and probably tamer that Mouser carries for $52.21 each, single.

80w, 28v, N chan. mosfet, enhanced mode, etc.

http://www.macomtech.com/datasheets/MRF173.pdf (http://www.macomtech.com/datasheets/MRF173.pdf)

Gain at 20Mhz is 20 db.
Never seem to see an IM spec for these.  I guess you have to roll all those S parameters into a model and see.  ;)

Hi and thanks for the feedback !

The correct schematic showing the diplexer is attached here. The previous output match on the above schematic to interface to a LPF, was found unnecessary. A 4:1 output xmfr connection to the filter is an adequate interface. The 50 ohm termination formed by 4 - // 220 ohm 2 W is sufficient to handle 8-10 W CW at the 100 W power level since the HPF corner is set to 10 MHz.

These devices and others like it were in the junk box. Not sure of their current cost or availability Steve. You might run across UHF mobile radio parts as well as WCDMA 2.1 GHz base station devices that are surplus to the users and being "re-cycled". In many cases the package costs of these devices, the gold and ceramic, are greater than the silicon ! I have seen the LDMOS 21125, 125 W 2.1 GHz device going for 10-15 $. Class D and E power FETS are probably a much better deal on a dollar/watt basis ! In any case,  my intention was to address the notion that HF amplifiers can or cannot be built with UHF or higher frequency devices. Consider the MRF327, the gain at 450 MHz is 12-13 dB,  so as we approach 4.5 MHz with the gain rising at 20 dB/decade, you are looking at over 40 dB of small signal gain ! Clearly, you better be careful about feedback and terminations. Otherwise you will be seeing a power oscillator for sure.

The internal supply is 24 V (28 V would have been nice but the xmfr is stressed) and the Icc at ~ 6.8 A. The supply VA is only 100 W but I am easily geeting away with using it for CW. The PA efficiency is 58 %.