This is a continuation of a post at: Portable preamps for 6/2/70 part 1.
Part 1 showed a schematic for the preamp component. This would live on the antenna, near the feedpoint. Given the bias tee, it will only need to be fed by coax. It does not need any control lines or separate DC cables going up there.
Here is a picture of the PCB design:
I’ve got this under 2 inches by 1.5 inches, so I am happy with the size. I have two BNC connectors on board, these can be obtained from rfsupplier. I used microwave design techniques on this board with as much as possible of the RF being microstrip with 50 ohms characteristic impedance. Of course, with this approach, there is no using vias for anything apart from ground connections. In line with Minicircuits recommendations, the PGA-103 is surrounded with grounding vias. I also ensured that there were grounding vias near each of the relay RF inputs and outputs.
The sequencer needs to do several things:
- Provide the Vdd supply for the preamp and put this on the coax
- Only put Vdd on the coax when the radio is Rxing – otherwise the preamp will be exposed to RF far beyond spec, and magic smoke will not be far behind. As the preamp uses non latching relays, any removal of DC from the coax will cause the relays of the preamp to switch to bypass, allowing TX RF to safely pass.
- Allow for a delay between the radio commencing TX and high power RF going out the cable. Modern full mode rigs have TX inhibiter inputs, and the FT-817, 57 and 97 are no exceptions. The microprocessor can respond in a few microseconds, but relays need a few milliseconds to switch. The plan is to allow 25 milliseconds to be safe.
- Also provide for a short delay when TX stops before switching back to RX. This would be about 10ms.
- The local bias tee to put the DC on the cable needs to withstand 100 watts ssb. This will mean this bias tee will not need any relays to switch the RF. Of course there are relays in the preamp at the antenna end of the coax, where they are needed.
Microchip manufacture the PIC range of controllers. The task needed here is basically a timer and a state machine to deal with the various scenarios. Others who have built sequencers have used PIC controllers. In a previous life, I wrote real time control systems on PCs. The requirements here are far more simple! Looking at the range, the entry level PIC10 series will do the job. While not in the data sheet for the device, the Microchip web site also details that this device has current source/sink capability of 25mA. This is quite nice, and with a 5V supply, should give lots of drive and sink options.
On the FT-817 and 97 there is an ACC port. This has:
- TX GND (a sink when the radio wants to TX, otherwise floating – i.e. open collector),
- TX INH signal. This input feeds the base of a BJT through a 47k resistor in the radio. The transistor needs to sink 8V through a 3.3k resistor to prevent TX. This is 2.4mA. We need to provide 1/Beta of this current, and Beta should be 40+, so only 53uA or so is needed.
- A 13.8V (or actually whatever the voltage that is feeding the radio is – so on a FT-817 this could be 9V for instance. There is a 10ohm 1/10th watt resistor on the radio in series with this source, so this source is practically limited to about 80mA. We will not want to power things from this, but we can present high impedance to it and use it as a signal to indicate the presence of the radio.
- A GND
Here’s a look at the circuit:
EDIT: This diagram is updated to correct a small drawing bug.
I thought about whether I would use a transistor to switch the 5V supply to the bias tee and onto the coax, but there could be 300 to 400mA being pulled when on, and this is a bit to sink when it needs to be off. Instead, I introduced a relay to switch the supply voltage to the bias tee.
There are actually two bias tees. One for 6m and the other a compromise between 2 and 70. Both the FT-897 and the FT-817 have two antenna connections. The 897 is not selectable, one is HF/6m and the other is 2/70. I have largely gone with this configuration, however the 6m optimised tee could drive 2m. The FT-817 is fully switch-able, so either of its antenna connections can be used for any band. I have selected bias tee components that can be switched, so one of these circuits could have the left optimised for 2, the other for 70, or have both bias tees the same configuration. Again, I selected bias tee components to sink the fraction of power that will go through the tee from TX power at 100 w ssb, CW or 50w PSK. This is actually one of the key reasons for two inductors on each tee. One is a ferrite coil, the other is a chip inductor. Both are from Coilcraft and the part numbers are going to be quite specific. It will not be possible to use any inductor of a similar inductance value, because the actual inductance at the target frequency can be quite different to expectations. There is also the need to actually sink some of the power that comes through here as inductors are not known for having high Q – Q effectively being a ratio between inductor resistance and reactance.
The relay needs around 12mA to drive the coil. I am using two 4401’s with the coil on the collector of the first. Q1 has a PIC output on the base. If this is taken high, then Q1 will conduct. 2mA will be drawn from GP0 on the PIC, well within the 25mA limit. Q2 has the radio 13.8V output on its base. The resistor is sized again to provide for a 2mA draw. If the radio is connected, then Q2 will conduct, otherwise not. This setup is effectively and AND operation. Both the PIC output and the radio 13.8V signal (which just needs to be any meaningful non-zero voltage) are required for the coil to switch from NC to NO and provide the bias voltage onto the bias tees. No radio connected to this sequencer, then no Q2 conducting, and therefore no DC bias on the coax. This results in the preamps at the other end being safely deactivated and bypassed.
GP3 is an input only pin and it is connected to TX GND. This is an open collector pin, acting as a current sink when activated, with very low resistance. About 1mA will be pulled out of the GP3 chip when low. The PIC has TTL inputs, and also provides weak pull-ups, so no external pull-up resistor between TX GND and Vdd is required.
GP1 has a 1k resistor, which is effectively in series with the 47k resister in the radio. This should provide about 110uA of drive, again a tiny fraction of the 25mA PIC capability, and about double what is needed by the radio for TX INH. The 1k resistor is there mainly to protect the PIC in-case of an interface cable short – only a little less than 5mA would flow in case of a short.
The circuit has a number of 1nF capacitors. For the ones not associated with the bias tee, I can use some cheap X7R dialectic caps, but little compromise can be made with the 1nF cap in the bias tee. ATCeramics make some good high RF power rated caps for that job, I can use cheaper Ebay caps for the rest.
Finally, the code in the pic, which I will need to write, basically needs to implement the state machine. This code should be pretty straight-forward.
Next step is to design the board, and then it will be time to get some boards made. I have also designed a FT-817 to phone interface for PSK – see post here – and so I will get a batch job done with a few of each of these boards.
73 de Wayne VK3WAM
Part 3 can be found here.