Comments? email to
There are lots of different 2 meter low-power rigs in use out there, ranging from hand-helds for FM, older multi-mode transceivers, and even the newer all-purpose types like the FT-817 or the Elecraft 2M transverters. QRP operation can be fun, but if you're like me, you probably have the occasional need for a bit more power.
Good news...if you have a couple of afternoons to spend on a project, you can build this 80 watt multi-mode amplifier with ease. Easy because it uses one of the newer S-AV36 Toshiba modules, 50 ohms in and 50 ohms out, with gain galore. So much gain that less than 50 milliwatts will drive it to full output in any mode.
My original intent in making this was to have an amplifier capable of boosting an older 10w multi-mode radio up to 80 or 100w. I wanted to keep it low-cost and simple (no preamp or power meters), yet capable of mobile operation in any mode from 12v power.
After absorbing the specs in the data sheet, it was clear this module could be driven by almost any low-power rig; thinking about it a bit more, and keeping in mind the low cost and simplicity requirements, a few more useful features came to mind, such as:
The devil is in the details for the designer, though, and it did take a little planning, but the end result was a small PC board and just a few interconnecting wires. Add a heat sink, connectors, switches, a couple of sheet metal parts for the enclosure, and that's about it.
The Photo to the left is the first prototype, using a board made with common
Designing the amplifier
The S-AV36 module is pretty easy to use; aside from RF in and RF out, there are two power connections; one is for bias (this turns the module on and off), and the other for main DC power, 13.5v nominal at up to 15 amps. Since the input power required to drive it is only about 50 milliwatts, the first thing to do is design an input attenuator to match the output of the driver to the S-AV36.
The chart shown on the right lists the resistor values for 50 ohm attenuators at drive levels
ranging from 1 to 10w. There are some strange values there, but these
are not terribly critical, you just have to get within a few ohms to get the job
done. For example, a 23db attenuator is needed for a 10w radio; the resistors
chosen were those readily available from major distributors, so 58 ohms became
56, and 351 became 360 (close enough). L5 is not really necessary, it's purpose
is to compensate for the stray capacitance of R7 at 2 meters (a 35w tab-mounted
resistor). The input SWR was acceptable without it, but it does make the input
match almost perfect.
The low pass filter
Now that the input is taken care of, let's deal with the output; the data sheet says the second harmonic will only be down about 25db, and the third about 30. Not good enough for the FCC, so we need an output filter that will put us in good graces with at least 60db total suppression. For that 25db second harmonic, we need another 35db.
The filter shown is a standard pi-type Chebyshev, 7 poles, and provides the required suppression with very little insertion loss at the operating frequency.
The last analyzer display shows the sampled harmonic content of the amplifier
tested at 90 watts output. Good filter.
The Antenna Relay and switching controls
In the sprit of keeping costs low, a PCB-mount type of DPDT general-purpose relay was chosen. Less than $5 in cost, the contacts are rated at 8 amps. At 2 meters, a bit of reactance is introduced by this part, but compensated for by a small capacitor (C15) in series with it's input.
The best way to tell the amplifier to switch on is to use a control line back to the driving radio (PTT). If this is unavailable or inconvenient, the amplifier has an RF-sensing circuit that samples a bit of drive from the input connector to provide the transmit trigger.
Another little twist; switching from receive to transmit should be sequenced for two reasons; first, the S-AV36 is tough, but no self-respecting amplifier module likes seeing an open circuit while those lazy relay contacts are moving, even if it only takes 20 milliseconds to happen; it's not good for the module, and just plain rude. For this reason, the module has to be kept off while the relay contacts are settling. The other reason is to protect the relay contacts from that 80 to 100w the amplifier will generate before they finally settle; it tends to shorten the life of the relay.
Showing the full schematic now, C4, D1, and D2 sample the input, and C5, C6, R1 and R2 provide filtering and some timing, depending on the position of S1. In SSB mode, the circuit acts like a vox, providing a second or so of delay before reverting back to RCV mode. In FM mode, the switch back to receive is much quicker, as the delay is not necessary for FM operation. The circuit is sensitive, and will trigger with less than 1/2w drive.
Q1 is the switch that operates the relay. When the relay is turned on by Q1, it also turns on Q2 (the bias switch) after a short delay. This delay is provided by C9 and R4, and is about 50 milliseconds in duration, allowing those relay contacts to settle before the module becomes active.
When switching back to RCV, the bias to the module is cut off before the relay contacts open. This fast cutoff is timed by C9 and R5, and is only about 5 milliseconds in duration.
Another noteworthy component is D6, the reverse-polarity protection diode. This diode's purpose in life is to blow an in-line fuse in case you accidentally connect the power cable up backwards (come on, we've all done it).
The extra contacts on power connector J3, pins 3 and 4, provide a means to disable the RF sensing and connect PTT directly to the driver should the RF sensing be deemed unnecessary.
L1 and L4 are 4 turns #18, 4mm ID and 8mm long. L2 and L3 are 7 turns #18,
4mm ID and 10mm long.
Building the amplifier
A bill of materials (BOM) is provided for gathering the parts you don't have on hand, and is grouped by the recommended supplier. For parts supplied by Mouser, you can order all of these by ordering the project list from their web site at: http://www.mouser.com/tools/projectcartsharing.aspx. The access I.D. code for the project is c3ad150d1a . At the time of publication of the QST article, Mouser was temporarily out of stock on the 56 ohm 35w resistor (R7). A 20w resistor can be substituted there, part # PWR263S-20-56R0J. If you use the 20w part, also substitute a 270nh inductor for L5, part # imc1210err27k.
RFPARTS (www.rfparts.com) is the supplier for the Toshiba module and coax connectors; artwork for the PC board is provided here, and for those not wishing to make their own board, commercially made boards will be available on the parts page here. Drawings for making the chassis parts are here: front panel rear panel bottom cover. For heat sink material, a heat sink with a base thickness of .300 or better should be used. The size of the one I used was 8" long x 5.375" wide.
Here are the recommended steps, in order, for constructing the amplifier:
Testing the amplifier
Once everything is wired and in place, you can test the amplifier using the following procedure:
I experimented some with various IDQ settings, and concluded that Toshiba must have designed the module to operate close to class A; setting IDQ lower tended to introduce lower overall gain and crossover distortion in SSB, and setting it higher resulted in higher gain and saturated output power. At 10 amps IDQ, for example, the amplifier could be driven to over 100w output with about half the drive required at 8 amps IDQ. This does exceed the manufacturers ratings for the device, and really doesn't make any difference on the air, so I resisted the temptation to leave it that way.
Knowing how bias levels work could be useful in the non-linear modes (FM, CW); you could set the bias level to control output power, and perhaps have a front panel control knob in place of VR1 for this purpose. For all-mode versatility, leaving IDQ set at 8 amps is best.
last note; at 10w drive level, I noticed R7 (the 35w attenuator input resistor)
ran hot; this was due to the inadequate heat transfer of the PC board I made for
the original prototype, which has just a few rivets where there should have been
multiple plated-through holes surrounding this part. My solution was to
use a piece of .040 copper strip soldered to the ground tab of the resistor,
using it to sink heat off to the heat sink by fastening the other end to it with
a #4 screw. Most of us making our own prototype boards at home don't have the
ability to make plated-through holes like the commercial board houses, so if you
make your own board for this project, you'll probably need to implement a
similar solution. If you purchased the commercial board, you won't need the
copper strip; but remember to fill the plated-through holes surrounding (and
underneath) R7 with solder to help with heat sinking.