View Single Post
Old 20th May 2017, 2:14 pm   #8
Andrew2
Nonode
 
Andrew2's Avatar
 
Join Date: Mar 2010
Location: Dukinfield, Cheshire, UK.
Posts: 2,034
Default Re: My pantry transmitter.

All comments noted. Here's a more detailed description, but as it was built 'from head to board' with only minimal prototyping and measurement, the notes are still rather poor. Still, anyone with a basic understanding could knock one up I reckon.
Re the bipolar PA. Please no. Push pull ones are nice, but single-ended ones always make me (and my scope) shudder with the nasty-looking waveforms. The bias supply is a pain too. I'll let one of you brave souls do it!


Here's a better (?) description of the pantry TX.

Osc and divider.

The oscillator is a crystal-controlled Colpitts and it's about as basic as it can get. The xtal in this one is 14.169 MHz and is tweaked onto 14.175 MHz by the 20pf trimmer in series with the xtal. The transistor is a BC548, so probably anything will do. The only thing to fiddle with (apart from the xtal trimmer) is to set the base bias so that the DC level at the emitter sits at about 2.5 v. Adjust the bias to ensure reliable clocking of the HC4017 counter chip. This chip must be an HC part as the standard CMOS 4017 won't work at anything more than a couple of MHz.
There should be about 1v p-p of RF at the clock input (pin14). Dont forget to enable the chip by decking (pin 13).
(Pin 15) is the reset pin and is returned (in this case) to (pin 11) which will cause an f/9 division, outputting a nice 5v p-p square wave at 1575 KHz from the carry out (pin 12).
The diagram shows a crude potential divider to provide a 5v supply for the 4017, but since then I have reduced the 560 ohm resistor to 390 ohm and stuck a 4.7 v zener to deck to replace the 2.7k.


Buffer.

Not much to see here. Another BC548 is pressed into service, this time as an emitter follower. Input at 1575 KHz is directly applied to the base and the output is taken from a potential divider in the emitter. This divider is a hangover from the prototype, when I was driving the modulator from my sig gen. When building, I simply tried to match the signal levels I had used at the time. In my completed unit, I get 4v p-p at the buffer's emitter and about 1 v p-p at the junction of the resistors. If you are testing as you build, you may get higher readings than this as the loading effect of the next stage (the modulator) is absent.


Modulator.

The first transistor that came to hand from the mounting pile of scrap on my bench was a 2N2222, but I'm sure another BC548 would work a treat. This circuit was the result of thinking 'there must be a simple way of applying AM that is linear, without resorting to LTP's and stuff'. And this is what came out. The RF from the buffer goes to the emitter via a 10n cap, and the audio goes on the base via a 1uf cap and a 1k series resistor. There's a 1n to deck from the base to get rid of any RF.
OK, the first of the dreaded wound components appears here, in the collector. It uses a junk-box ferrite toroid about 1cm dia, and I wound about 6 turns each for primary and secondary. It may be worth experimenting with the turns ratio when you have finished the TX, but as mine worked 'well enough' I didn't bother. There's a 300pf cap to deck from the collector to remove some harmonic energy. My scope sees about 500mV p-p without modulation, but the exact amount will vary with the setting of the bias pot. A 4.7k preset was used here as I had a couple in the drawers.
At this stage you can try out the TX, getting a feel of what the 'set mod bias' pot does and how much output it can provide without squashing the audio.

Final amp.

I used a VN67 fet device here as I find them much nicer to work with than bipolars for this type of thing. I believe VN66/67 devices are now like hen's teeth, but perhaps an IRF540 or similar would do if the bias is turned up. Make sure you don't buy cheap Chinese copies - I did once and I was led a very merry dance.
There's nothing much to it. Input comes from the secondary of the modulator toroid via a 22 ohm gate stopper to the gate of the fet. The other end of the secondary goes to deck via a 1uF cap. Bias for the fet comes from another 4.7k preset to the top end of this cap. Another junk-box ferrite toroid goes in the drain, again with half-a-dozen turns pri & sec. I'm sure this could be substantially improved with a bit of fiddling with turns ratio.
Without RF drive, set the bias pot for about 60 to 80 mA drain current. A small heatsink might be advisable. Before applying drive, solder a 47 ohm 0.25 watt resistor across the secondary of the output toroid.


Testing...

On with the volts and the set up can begin. Trim your osc to frequency and check the various waveforms look something like sensible. Without modulation you should be able get about 3.6 v p-p into the 47 ohm resistor, and at 100% modulation the p-p voltage should double to over 7 v p-p. You will find that you can get a lot more than this by advancing the mod bias pot, but doing this will leave less room for the modulation. 7 v p-p into 47 ohms is 130 mW peak envelope power and I find it easily enough to cover our house when fed to my homebrew hula-hoop loop. Note, these junk box ferrites are pretty random things and if you cannot achieve the results I have, it may be worth trying different ones. The LPF is optional as the loop aerial (see later) is very sharp and should (maybe...) attenuate any harmonic content quite well. However, I'm a bit anal about such things so I included it.


Loop aerial.

Trying to radiate MW signals from a short wire aerial is a pretty depressing thing to do. It never travels far. I've had very good results with a 3 foot diameter frame aerial hung on the wall of my upstairs shack. It's just a normal plastic hula-hoop, the type with a join that can be pulled apart. I can't remember exactly how many turns I threaded through the hoop, but I've just measured the inductance of the winding and it is 23 uH. Leave a few inches of wire sticking out at each end and push the loop back together. The tuning/matching circuit can be made up on a scrap of Veroboard and left to dangle on the loop wires. My method of feeding the TX into this is shown in the diagram. Basically it is brought to resonance on the right frequency by the combined capacitance of the trimmer (a large 1000pf compression type) and the 8.5n capacitor in series. The fixed cap can be made up from several caps in parallel if necessary. The RF from the TX goes across the fixed cap. Note the tuning is pretty sharp. If you don't have a way of finding when the loop hits resonance, just leave your scope probe nearby and tune the trimmer for a sharp increase in the scope display.

So there it is. Sorry I've had to be a bit vague in parts and as usual many of the bits are un-specified junk-box components, especially the ferrites. Suck it and see is good when it comes to wound components.
Re the photos. I've started building the board into an old Icom case so I couldn't get the shots I wanted, but I think they show what's what. The smaller board next to the RF board is just a bit of audio amplification and HF filtering and there's a little circuit that flashes an LED when the modulation approaches 100 per cent.
Attached Thumbnails
Click image for larger version

Name:	new1.jpg
Views:	299
Size:	62.9 KB
ID:	142870   Click image for larger version

Name:	New2.jpg
Views:	243
Size:	99.0 KB
ID:	142871   Click image for larger version

Name:	Pantry osc.jpg
Views:	320
Size:	37.3 KB
ID:	142872   Click image for larger version

Name:	Pantry modPA.jpg
Views:	253
Size:	37.6 KB
ID:	142873   Click image for larger version

Name:	Loop match1.jpg
Views:	255
Size:	23.2 KB
ID:	142874  

__________________
Andy G1HBE.

Last edited by Andrew2; 20th May 2017 at 2:23 pm.
Andrew2 is offline