Portable HF linear amplifier

 

Having got back on the air and started lots of portable operation with my FT-817, I have realised that quite often a little bit more power would not go amiss. I have started on a series of projects to get this power through the use of some simple, portable linear amplifiers. I was encouraged along this route by the publication of Eamon Skelton's "Homebrew" column in the July and August 2008 RSGB RadCom, which suggested an HF linear could be built using cheap MOSFET transistors - for me cheap is good.

Output transformer

The linear is a class AB push-pull design and needs a suitable broad band output transformer, which can easily be made using a couple of large ferrite chokes. The primary winding is a single turn, which is made as a tube using the braid from coax cable and copper clad board end pieces. Here are the bits:

It is easiest to assemble this by leaving the coax core in place whilst the braid is soldered to the end pieces; here the braid is pushed through ready for soldering:

In this next stage, one half of the transformer is complete and the other side is ready for soldering. I forgot I had a 25W soldering iron stashed away which would have made this much easier than with my 15W Antex!

Here is the transformer assesmbled, with the secondary winding threaded through the primary.

Heat sink

Even a relatively low power linear needs a heatsink; efficiency is not going to be much better than 50% and it may well get run at higher powers than originally anticipated for portable use when it gets used car or caravan portable! I had a number of old heatsinks in the junk boxes, but wanted to be able to get the transistors inside the case; this needed a conductor and I decided copper bar would be best. Blake's metal centre got hold of come lovely stuff for me; I now have enough to see me out.

I needed to tap a couple of holes to take screws to hold down the transistors. Having no handy taps available (must get one before I next need to do this) I made a couple from some steel screws by filing flats near the end. This works absolutely fine since copper is so soft, providing you use plenty oil for lubrication.

Here you can see how the copper bar is brought through a slot in the box, with the heatsink mounted on the outside. The screw standoffs are to mount copper clad board to take the components. The bar also has a shallow hole to take a thermistor for a bit of thermal feedback on the bias.

Testing the basic design

For the initial design I intended to use IRF740 MOSFETs since these seemed to have suitably huge power characteristics. Before putting in the transistors though, I assesmbled the output and input broadband transformers onto the copper clad boards. No PCBs for this project - the basic circuit is very simple and can mostly be put together with point to point wiring. The initial PA circuit was basically that given by Eamon Skelton in the Homebrew column in RadCom, July 2008, which used IRFP260 FETs.

The MOSFETs need DC bias, and this is set up on a subsidiary stripboard using a 5V regulator and trimpot. The stripboard also carries a relay to turn the bias on when PTT is active on the transceiver; this saves power during receive. Here is one transistor in place. However, extensive testing ended up with me being unable to cure instability with the IRF740; even with just one transistor in place, the thing generated RF at several unwanted MHz! Its characteristic is very steep, so this is not too surprising; rather than spending more time on it I decided to try some IRF510 MOSFETs instead, since these have been used successfully in a number of small linears like this.

The IRF510s were nice and stable, but it proved hard to get a lot of power out of them with the original design. Eventually I swiped some ideas from a circuit by Mike Cossor WA2EBY (QST March 1999) - also the Packer amplifier described by Virgil Stamps K5OOR (2007) which is essentially the same circuit. These are (a) a better input transformer arrangement, and (b) a better HF choke arrangement to feed the MOSFETs. The photo below shows the amp at this stage, ready for a rebuild. 

The redesign looked pretty good, and put out over 60W RF into the dummy load at 3.5MHz using two 12V SLABs to give a 24V supply and using 2.3W input, though some of that power is in the harmonics. I will probably not drive it quite so hard for portable use.

Rebuild

Here's the rebuild underway; the original input transformer has been removed but the bias circuitry is in place. The output transformer has been moved closer to the FETs.

Once the rebuild was done apart from the low pass filters (I had just ordered the toroids from Heros Technology - excellent service!) I wanted to try it out since we were off for the weekend. I thought I might get by with an 80m LPF made using ferrite toroids I happened to have instead, but unfortunately a sniff of full power RF from the amp fried the first toroid. Reassuring in some ways in that it proves there's a fair amount of power there!

The photo below shows everything in the box - at least, except for the topband and 10m bandpass filters. The filter board is in the top right, switched manually centre right, and the input and output relays are now in place. The new input transformer is much neater and the HF choke for the FETs is counter wound as in the Packer Amp. The TX control lead from the FT-817 is also made up and plugged in on the left just above the input BNC socket. The treacle tin dummy load is now filled with out of date veggie oil which should stop the resistors frying. The iron dust toroids don't even get warm under full power, thank goodness! Signals all look pretty clean on the scope with the linear driven using an FM carrier, and look fine with SSB as well, though I must build a two tone test oscillator. It appears to be able to produce 35W out for 2.5W in easily enough over the lower bands - can't go much higher than that for portable as I have to carry the batteries.

Here's the finished product (almost) - it still needs some rubber feet in case it spends time on the XYL's best table at the caravan, and the innards still need the LPFs for 160m and 12m/10m. On this end are the BNC connection to the transceiver, the mini DIN connector to the PTT signal on the data in/out on the transceiver, an LED to indicate transmit (i.e. FET bias voltage on and relays switched), and the 4mm power connectors. 

At the other end we see the band switch, and the BNC connector for the antenna. The copper bar connecting the FETs to the heatsink (thermally) can also be seen down the tunnel in the heatsink. I have another of these - they were bought as a pair for a Hi-Fi amp that never got completed - we bought one instead! The other one will probably be used for the 6m linear once I get that started.

As of March 15 2009 it has been on the air - I did a SOTA activation on Cairnpapple Hill with great success using an inverted 'V' with the apex at 6m (fishing pole) and the ends about 1m off the ground - I use garden canes to support the strings at the ends. Made about 25 HF contacts (the bulk of these after being "spotted" on the SOTA website) but included the De Havilland Museum special event station in Hertfordshire and my first European mainland HF contact in Belgium. The weekend after I operated at the caravan in south west Scotland on 40m with a quarter wave vertical (on the Spiderbeam 12m pole) - it was the Russian DX contest weekend and so I picked up several eastern European contacts. The following morning I set up near the shore operating on 80m with a full size inverted V on the Spiderbeam pole. It has received quite a few favourable reports as to signal quality so I'm very pleased with it and will play a lot more with it in the future - I can have some fun experimenting with HF aerials.

Circuits

This is the main part of the rig - basically the Packer amp circuit referred to above, which is online in the HF Projects website construction manual.

The input to the PA has an attenuator to match the FT-817's output.

Here's the relay TX switching and bias circuit which is driven off a 5V regulator and has some thermal compensation. Not likely to be much of a problem where I normally operate - in the cold wind on the top of a Scottish hill! I needed to add the 0.22µF capacitor on the TX input switch as the relays started chattering when in use on 20m - clearly detecting and being overpowered by RF.

The power input allows for a range of supply voltages, since there's a 12V regulator for the relays. It still works fine with 12V input, though the output power is then somewhat limited.