One project that I have been working for a little while is the design and eventual build of preamps for the 6m / 2m / 70cm bands.
My requirements are:
- Relay switching for TX, by the presence of DC supply
- Preamp activated by DC power, otherwise off and bypassed
- 100 watt TX capability, but only at 50% duty cycle
- DC supply for the preamps to be on the coax, so bias tees will be needed
- Said bias tees must also work in the presence of 100w TX power
The question needs to be asked. These bands are often quite low noise, certainly on many SOTA (Summits on the Air) mountain tops. Most of my SOTA activations have been HF, however there have been a few that have been on 2m. During the John Moyle Field Day contest this year, I operated from Mt Torbreck, VK3/VN-001. While I got a good score, there were a few QSO’s going missing who could hear me, but I could not hear them. Some of them were using more power than me. Some have reported that they hear much better with preamps. So, there seems to be enough of a reason to do it.
DC power supply from the coax
It takes some design work, but clearly it is going to be convenient to have DC power come down the same cable as the coax from the radio. Near the radio, I will need a bias tee to put the DC on the cable, and it will only do this when not TXing. When it gets to the preamp, there are three possible states:
- The radio is TXing and there could be lots of RF energy on the coax, no DC
- The radio is RXing and there is DC on the coax
- The radio is RXing and there is no DC on the coax
State 1 needs to ensure that there is not large amounts of energy that gets on the Vcc rail in the preamp. An inductor blocks (in part) AC, including RF. It does it through reactance, and because its Q is not infinite, some resistance as well. Unless the Q is very low, it is by far through reactance.
State 2 needs to get the DC to the Vcc rail of the preamp. This will flow happily through the inductor, but the inductor does have some DC resistance – so there is going to be a voltage drop. The DC should not continue further down the coax towards the antenna, so a DC blocking cap is needed – this needs to have very low effective series resistance and very low capacitive reactance at the RF frequency. It also needs to be able to handle the RF current during State 1.
State 3 has no special circuit requirements for state 3 as no DC power is on the coax, and the inductor is going to block the RX RF energy from going up into the preamp, instead it needs to go to the radio where hopefully it results in received intelligence 🙂
The impedance of the inductor needs to be high, much higher than 50 ohms at the RF frequency. It needs to have low DC resistance. It also needs to take into account the fact that inductors are imperfect beasts at RF frequencies – they have resonances and act like capacitors above the resonant frequency. I was able to use some design formulae from RF Circuit Design by Chris Bowick. These give a whole range of impedance matching values derived from S parameters. It can also be used to find actual inductance and capacitance values for a given component at a given frequency. The nominal inductance of an inductor is not a fixed value, but changes with frequency. Different inductors change in different ways, based on their construction.
Getting the right inductors has been a major challenge. Even if I have an insertion loss of only 0.2db, there is still about 3 watts getting into the preamp circuit if there is 100 watts of RF outside, and some of this will have to be dissipated by the inductor’s resistance. Also, some of these inductors are quite large, but if they are too big, their resonances will be at far too low a frequency to be useful for the targeted 6m, 2m and 70cm bands.
In the end, to deal with power dissipation requirements, I used a specific bias tee inductor and some chip inductors. If I used just one inductor, I would lose a little too much DC across the tee (x2 of course for the fact that I need to put this DC on the coax back near the radio!)
Ripple on the Vcc rail
So now, I have got my DC when I want it, and are not likely to release magic smoke when at 100 watts 50% duty cycle modes. I still have some ripple on the supply, and I need a bypass cap to sink it to earth. The inductors suck up most of the heat dissipation. If I use a single cap, this leaves a lower frequency pattern, so I have found that having two caps, one around 1nF to 5nF based on the band (higher band, lower cap), and a second 10uF cap seems to get the ripple down to less than a millivolt. In (a different) RF Circuit Design by Richard Chi Hsi Li, the point is made than rather than having a capacitor forest of low to high values, a single cap should be selected around self resonant frequency (SRF). This exact frequency is going to vary, and the cap cannot be ordered with some precise, eg 2.2235nF value, so I target a little below SRF for the small cap. Some simulation tests show that these caps are only dissipating a few milliwatts, so it is looking good.
The next thing to discuss is getting the amplifier that I am going to use out of the way during TX, or I will release the magic smoke. I have found a suitable RF surface mount relay that looks suitable for the power rating required from Axicom. It has a continuous 60 watt rating, so 100 watts SSB or even 100 watts CW should be a piece of cake. Lets not try 100 watts PSK though! I will need two of these relays, one each side of the amplifier. Normally closed is no DC voltage on the coax, state 2 and 3. This would bypass the amplifier with the signal going directly from one relay to the next. Normally open will connect the amplifier.
One issue is that using the same relay, the RF paths have to cross. If I use a two sided board with a ground plane with the RF signal going on microstrip, I cannot cross the signals on the board. I am going to have to have a short jumper on the board for one of the signals – I have chosen the normally open (preamped signal).
From the Antenna side relay normally open the RF comes into the amplifier. Now, I could just use a JFET or MOSFET, or even a MMIC, but Minicircuits have a device called PGA-103+ which should make this part real easy. It has three (well four, but ground shares two) pads. One for RF in, one (two) for ground, and one for RF out and DC bias. This device works on either 3V or 5V. The device needs a DC blocker cap on the input, and one on the output. The bias voltage needs to get in there somehow, and so this calls for another bias tee on the output pin, before the output DC blocker cap. Another bias tee (that’s three now), but at least this one does not need to dissipate large amounts of current – it’s I^2R that causes the heat – and current is squared – damn!!!
Having three tees with the DC voltage drop causes quite a lot of design issues, and it took some time to identify the right components to do the job. There are chip inductors, air coil surface mount inductors. Coilcraft even make some core based smt inductors. The challenge is to get something with a high enough inductive reactance that does not take up too much space, and has not effectively become a capacitor because it is above SRF. Some devices are designed to be used above SRF and that’s ok, but most inductors are not, so this does provide a limit. The air coil ones just don’t have the reactance, even though they have low DC resistance.
The amp provides 26dB gain on 50m and 25dB on 2 and 70. It has a noise figure of 0.5dB. That is really nice. This should be a good quality preamp.
The 5V supply requirement of the amp meant that I selected 5V relays. I did not want to add some BJT device with a zener regulator to bring down 12V (or 13.8V or so) to 5V. Extra complexity for the preamp, duplicated on each one. It will be far easier to get a 5V DC-DC step down converter off ebay, plug some Anderson Powerpoles on it and run that from a 12V-15V or even wider supply. So this is the plan.
The DC current requirements of the amp and the two non-latching relays are 160mA. I needed to ensure that across all the bias tees to the amp (thats three remember!) that I was getting at least 4.75V there. I did contemplate using the amp in 3V mode, but that needs a zener plus a BJT, so I tried to avoid that. In the end, I can get 4.87V to the amp, and that is well within spec.
Here’s a look at a Schematic for the preamp.
It should be noted that component values will change for each band. It does not seem possible to retain simplicity and have a preamp for all three bands, but at least I should have a common PCB board design.
I will need a separate schematic for the radio side bias tee. I will talk about this more on a future post, but it is going to need to take the TX indication out of the radio, use a PIC chip to TX inhibit for 30ms or so, and switch the DC – through a BJT and then onto the bias tee onto the coax.through two)
Part 2 can be found here.