my homemade PLL transmitter

RADIODIRTYSPRIN your circuit board JPEG goes way out of focus on my monitor...  maybe it's time I got a better monitor... but from what I could tell the circuit looks neat but I can't tell if it's running Class E.

It is great to see the results of someone homebrewing a transmitter and thanks for sharing. The picture is out of focus but shows the gist of your build.

Even with a scope it is difficult to confirm Class E operation but you can tell if you are close by measuring the input power to the final and the output power, usually into a 30 pF and 30 ohm resistor in series, and calculate the efficiency. If this is above 80% then you are doing well.

The LM386 is a neat chip but it can do some nasty things such as going unstable under load. If you used the R C network on the output as recommended in the data sheet then it probably will be OK and I only mention this so you can watch out for this problem.

Neil

Hi Horseman,

I hvae Tom Polk's "12 dollar AM band Frequency Synthesizer" page bookmarked in my browser, bu tit just has the synth part only, intended to add to a transmitter you may have, it doesn't show a final amp or modulator section. (Polk's page not loading right now, but search for it, it was probably moved).

Did you build a the final and modulator yourself, or find the design somewhere? It would be nice to see a schematic. I like the idea of using a 2N7000! Other designs use a high power MOSFET like an IRF510 which needs a buffer amp, driving the 510 with up to half a watt, which then has to be cut back down again to 100 mw on the antenna side, which seems wasteful.

The 2N7000 should be able to be driven directly, no buffer amp, and put out 100 mw easy. The 2N7000 is small, but it's a real power mosfet, just a lower power one. Its advantages are low gate capacitance for low drive current needed, smaller size, and it's real cheap, and costs ten cents for one, plus I have like a hundred in my shop!

My ideas on the LM386 modulator, it can be good, especially used in the lowest gain mode without the bypass electrolytic to keep noise down and the chance for feedback. Put a resistor from the output pin of the 386 to the modulated amp, since I found that the 386 doesn't like high capacitance at is outpup, like the bypass caps on the RF stage. The resistor can be 10Ώ  to 1kΏ or more, and double as the power input control resistor.

Keep experimenting, this is cool, and let us in on the results.

I'm not an engineer, so everytime this

sort of thing comes up, I learn a little

bit more.

Brooce, Part 15 experiments and other stuff

RADIODIRTYSPRIN.. I have a question when you get back, do you have 10 khz reference frequency leakage through your PLL's loop filter?

Nice to see tech talk on a Part 15 page. LOL!

I agree, sometimes politics are the majority of the discussion on the boards, but I like the tech, learning about circuits like kids did in the past, and hoping to make new discoveries.

Another thing Horseman, did you heat sink your 2N7000 final?

i'm assuming this is in class E because i was able to run the RF section at 500mw with no heat on the output FET on a dummyload. there is a resistor throttling back the juice it sees upstream from the choke so i don't fry the LM386. it runs stable with the output set to about 80mW

i do not notice any ringing in my signal. only my town's dirty grid system when i get farther away from the signal in the car at 60hz

i do not notice any ringing in my signal. only my town's dirty grid system when i get farther away from the signal in the car at 60hz

Looking Cool!

FWIW his name is so long it gets cutoff. The Fullname is Radio Dirty Springs 1700.

The ringing would be due to the 10 khz reference of the PLL, getting through the loop filter. On a receiver it would sound like tiny carriers every 10 kc on either side of the main frequency you've set. I can't think it's much of an issue with a Part15 transmitter though.

I'd think you'd have the makings of class E, even if you can't tell for sure, since it seems you'd be feeding square waves right into the FET, and that's one part of it, then it's how the output is tuned.

A MOSFET as a final is great, it can be a nearly perfect modulated device. If you could share a schematic of your design, that would be cool.

It's actually MY PLL from the early 80s.  We build them here in Europe (9kHz channels) with 4.608MHz crystals.  Tom Polk has got the loop filter completely wrong - it needs a much longer time constant than he's given it.  I'll put the right values up here later.  His version will produce a rough-sounding carrier!

My little MW transmitter uses this type of PLL, driving into a 4049 hex buffer for a bit of drive, into a drain-modulated IRF510.  If you keep the supply voltage to the final way down, you could make it easily comply with Part 15, but if you raise the drain voltage, you can get several Watts of carrier power!

Incidentally, if you invert the output of the 4046 oscillator (as I do with the 4049), you can switch the carrier by switching pin 5 of the 4046 - low for on, high for off.  The 4049 means that the "off" state gives a logic "low" into the gate of the final FET, so there's no risk of it drawing any current.

Hi Mictester, it's very cool that it was you who came up with that PLL design. Tom Polk's version of it (crystaled for 10 khz channels for the Americas) has been built by many experimenters in the US and it's a relatively simple circuit using common discrete components and no microprocessors to program.

Tom does write that his loop filter might not be optimum, and encourages experimentation to make it better. His filter is fast, it locks right away at power on.

I have noise issues with the Tom-PLL, the main problem being lots of the 10 khz reference is getting through to the output, and a warbling sound is heard on the carrier below 1 khz. It all seems to be a form of phase modulation too, tuning to dead center of the carrier on a receiver almost nulls it. Tuning one of the PLL's harmonics on a FM receiver and you'll hear a strong 10 khz whistle in the audio, plus the warble.

The hex buffer sounds good, and what I've seen some use is an NPN driver transistor feeding complimentary current amp transistors, then a power FET final, based on Larry Lythall's circuit. I like the idea of inverting the output, I've wondered about the output being caught in a logic high, and needing current limiting in the final because of that.

I'd like to see your filter values, and it might help others to put out cleaner signals.