Multi-loop antenna selector
Now the problem I have is that I have only a coaxial which I will use for the signal and as a phantom power to the loop preamplifier. No other cables are allowed.
So I am trying to find a simple way to energize each of the relays and de-energize the others, so that only one loop is selected each time.
I was thinking of a higher variable input voltage to the coaxial and a way to detect the different voltage levels and energize the appropriate relays.
Can you think of any voltage level detector circuit that can do this?
I think I will use this one https://circuitswiring.com/wiring-di...robe-model.jpg it seems promising. But how can I replace the LEDs with transistors, so as to switch relays instead of lit LEDs?
Or can I connect the relay coils in place of the LEDs directly? How much current can each output of the chip sink in DOT mode?
The maximum load current is programmable but can't exceed 30mA. I strongly suggest that you keep the current much lower than that and use a pull-up resistor instead of the LEDs then use the outputs to drive transistors, with suitable base resistors.
If you draw too much current or introduce spikes from the relays it will probably upset the voltage references and could even cause instability with several relays trying to operate at once, even in 'dot' mode.
If this is for receiving only, it might be possible to use the outputs directly to operate diode switches.
Brian.
It's for RX only. The preamplifier is attached.
I have thought of one dual contact relay to enable each loop antenna (connect it to the preamplifier). How can this be done using diodes, this would be interesting?
Ideally you would use PIN diodes because they are optimized for good conduction and good isolation but there might, stressing that this would be very experimental, be a quick and dirty way to do this by utilizing a property of the schematic in the article.
The outputs of the LM3914 are 'active low', they sink current to ground to turn the LED on. If you built one input stage T1 & T2 for each loop but combined the collectors of T2 from them all to the base of a single T3 stage, then lifted R2 from ground and connected it to the LED outputs instead, it would allow the bias to be switched to one loop amplifier at a time. It might help to wire an extra resistor of say 1M from the base of T2 to the positive supply to ensure it was completely turned off when not wanted.
If you try it, keep T1 & T2 close to the loops, it isn't too important, at least for HF if the switching signal and T3 are wired a short distance away.
Brian.
If it is to repeat the input circuit for each loop, (so as also to practically keep the T1/T2 close to each loop), I am thinking that maybe I could keep all the T1/T2 stages operative, and just switching the T3. This would also ensure that I do not mess up with the bias of T2. Repeating 3 more components for R3 isn't a big deal.
How about lifting R4 from the ground and connect it in place of the LEDs in the Lm3914? When R4 is lifted, there should be no RF flowing to the C3, is this the case?
Lifting R4 from ground will also ensure that the other preamps connected to K1, won't leak their signal to the other preamps grounds through their R4s?
The thing that worries me in all the previous solutions that leave all the loops connected to preamplifiers and just select the output signals, is that because the loops will be closely spaced (a plastic tube supports all of them and they are separated only in angle) the unselected loops may dissipate RF energy from the selected one.
Whereas if I use relays and completely disconnect the unselected loops from the preamps, these loops will be just wires not connected to anything. I do not think that these unconnected wires will have a significant effect in the directivity or notch of each loop, as parasitic elements in such low frequencies (HF).
You can try but I suspect the impedance matching of T3 stages in parallel might be a problem. 'Disconnecting' the bottom of R4 will still leave the B-E junctions of several T3's in parallel, possibly working as unbiased diodes. The reason I suggested R2 and adding the 1M resistor is because it not only stops the amplifier, it also reverse biases T2 to better isolate it.
It is true that allowing some RF currents to flow through the loops will make them interact more than if they were disconnected but I suspect the amount of coupling to be quite small and more dependent on the loop construction than the electronics.
For further experimentation - consider making the loops reversible so they can be used additively or subtractively, it might help to boost or reject from certain directions. You could also consider a single loop with its rotation controlled electronically, either with a stepping motor or a normal motor with gearing to slow it down. It is mechanically simpler but electronically a little more complicated but it offers true directivity and the option to automatically peak the incoming signal.
Yet another experiment would be to use a single loop with a short center pivot as an active rod antenna. That gives one antenna with omni-directional properties and one with directivity. Adding the signals together gives you the possibility to peak or null by rotating the loop alone.
Brian.
That would be a plus in repeating the amplifier stages. Switch two loops at once and add their signals together.
Now this is a bit tricky, how to add the signals together? Use a combiner like a Wheatstone bridge? What about if the signals are comming in from different phase, they would be canceled out.
A simple summing amplifier is all you need, at it's simplest just join all the signals together!
The phasing is exactly what you want, it gives the boost or cancellation essential for adding directivity. Don't forget that turning the same loop through 180 degrees will invert the signal it produces. The directivity pattern may be the same but if used in conjunction with another signal source it can add to or subtract from the other signal.
Brian.
All right, so instead of selecting just one loop at a time, if the system has the ability to discretely control which loops will be active (one or more) you can do all shorts of pattern experimentation. I like this.
I won't turn any loops physically, so the 0deg one will show the same signal phase as the 180deg one (the same antenna, the same electrical connections).
So this will not allow the use of LM3914 any more. Can you think of a simple scheme (I like discrete if possible) this individual loop enabling can be done remotely, given the fact that only a single coaxial cable is connected between the loops and the shack?
Its really easy using MCUs but I know you don't like programming them.
The simplest way is to get hold of two shift registers, one 'parallel in/serial out' and the other 'serial in/parallel out', you need some way to synchronize them together and carry the power to them. There are devices for doing this automatically called UARTs but they are mostly built in to MCU these days, if you can find a pair of the older 'stand alone' UARTS they should work fine. UARTs have the advantage of having the framing logic built in to them, you basically set the bits on pins at one end of the link and the same bits come out of pins at the far end. As they are almost all 8-bit devices, you get 8 individual switched lines straight away and they can be used in any combination.
Carrying power is fairly easy. Use chokes and filters to keep the wanted RF out of the power/logic circuits, use say 12V for the power line but at the sending end lift and lower the supply voltage with the logic signal from the sending UART. This could be a simple as connecting the ground pin of a 5V regulator to the UART output so a logic low produces 5V and a high produces about 10V. At the receiving end, use a Zener diode to drop 5V from the incoming voltage and you have your logic signal back again.
Personally, I would use two PICs, the code in them would be very simple: read an 8-bit port and pass its value to its UART at one end and read the UART and pass it's value to an 8-bit port at the other end. Just a few lines of code and they have built in framing and clock generation already.
If the distance is short, you could consider other methods of control, IR for example.
Brian.
Thanks for the great advice.
If I want to use one loop each time, can I use this simple level detector circuit? This will activate one or the other loop based on the voltage threshold of the zener and the voltage applied to the circuit. However, there will be some overlap.
This might do the trick. It splits the initial LOW/HIGH selection into two subsequent sections. The middle zener must be of a middle voltage, say 15v.
The right zener must be of a lower voltage and the left zener of a higher voltage.
The loads are shown as LEDs and these loads can be the power feeds to the preamplifiers instead, so no relays may be used.
This will gradually enable or disable the amplifiers (vary their voltage up and down until completely switched on/off) so this might "emulate" the behaviour of a continuously rotating loop.
And here is the circuit using 78L09 regulators (white boxes) to feed the preamplifiers.
And here is the complete schematic of the antenna.
Different voltage levels select each one of the 4 loops. Obviously the minimum voltage level should be >12v, so as to feed the preamps with 9v.
I wouldn't recommend that as a solution, it will work sequentially but not exclusively. It will add more loops as the voltage increases rather than allow individual selection. There would also be no abrupt selection, with some amplifiers possibly running but with reduced supply, it might cause problems with overload from strong signals.
Brian.
Are you sure?
I have designed it not to work that way but instead select only one output each time and disable the others.
The central zener and nearby complementary BJTs will select either the left or the right pairs of the BJTs, by providing power to either of them depended on the voltage. The right and left zeners and nearby complementary BJTs will select either the left or the right transistors in each pair, allowing power to flow through them.
I didn't check accurately and you may be right but it is rather complicated and don't forget that regulators usually pass lower input voltages to some degree even if it is below the voltage needed to reach regulation.
Personally, I would still go for a digital approach, it will give more reliable control and greater flexibility for very little cost.
Brian.