I bought a few from these folks:
http://www.surplussales.com/Crystals/Crystals-2.html

That's seriously kewl, I like it...


well little Johnny, I bet you won't touch that again...

As far as coupling to the antenna, I can promise you the Pi net single coil and variable cap I use in mine will work just fine, your output needs to be tuned for best operation... At 1110Kc you'll probably need around a 800-900uh coil to match a 10ft antenna...

Normally a crystal should not rattle but if it works OK, I wouldn't worry about it...

BTW since Bob didn't mention his custom wound output toroids, I'll link to one of his Tx that uses a switchable coil to tune the output(along with a VC)... I have one in a VFO Tx of mine and it works like a charm...

http://www.antiqueradios.com/forums/vie ... highlight=

Tom

I decided to do some some rewiring. Now I have a Hartley oscillator with a 56. Four turns of wire around the tank coil couple the oscillator to an untuned buffer stage using a 47. The buffer stage is then connected to a plate modulated 47. Although the rf output from the buffer stage varies from 80 to 180 volts across the broadcast band (B+ is 100V), the output amplitude from the final 47 stays completely consistent.
Now the transmitter uses more tubes and less coils, capacitors. Both good things!
My next task is to rewire the final stage to match the 6F6G tx schematic. Is the only change needed for variable frequency work replacing the 390 uH antenna coils with one of those tapped coils?

I also built the crystal oscillator today, also with a 56. It works fine but it goes wild if its jarred. I think I am going to open the crystal and try to refasten whatever is rattling.

Now that is pretty neat! Curt would of been right proud of you.

Can't wait to see pictures of your revised version.

Kevin

As long as your stages are untuned(other than the VFO of course), there should be no other changes necessary... Bob uses a small variable cap on the on the antenna output(which is the proper circuit) but unless you are going to use a very short antenna, I've not found it necessary... With lengths of more than approx 8-9ft, the antenna has enough capacitance that it also acts as the output cap...

The coil I have that ranges from approx 250uh to 900uh and will allow antenna tune from above the BCB down to around 750Kc... I don't remember if Bob has specifications for the ones he did but it should be easy enough to add additional windings if you'd like to tune lower(maybe he'll chime in again)... Per his recommendation, I mounted mine on a six position switch available at Radio Shack...

Tom

Gee, where were you guys when I was giving away crystals for $5.00 shipped? I sent them priority mail which pretty much took care of the $5.

I still have some left if anyone is interested. I'd have to dig them out to get the freqs but they are all above 1khz.

John

Here's the transmitter as it currently stands. With a 500uH antenna coil I was only able to get the final amplifier to tune in the neighborhood of 1200-1400kc. I assume that is why a tapped coil is needed? Otherwise it works O.K. I will probably still be using 6v auto tubes for the final version. I just wanted to know if the 2.5 volt tubes would work too.

Gee, where were you guys when I was giving away crystals for $5.00 shipped?
--------------------------------------------------------
I sent mine to Wyatt

Darn, I guess I missed that John. Though, when I got mine from Surplus Sales, it was a few years ago, and they weren't nearly as expensive as they are now.

BTW, If anyone wants to use a crystal oscillator in the lower part of the band, a ceramic resonator is pretty much a direct substitute for a crystal; you can get them down to 540 kHz (or even lower if you want to go longwave), and they are very cheap.
1000 kHz: http://search.digikey.com/scripts/DkSea ... me=X926-ND
800 kHz: http://search.digikey.com/scripts/DkSea ... me=X925-ND
640 kHz: http://search.digikey.com/scripts/DkSea ... me=X924-ND
540 kHz: http://search.digikey.com/scripts/DkSea ... me=X923-ND
You won't transmit as far at these lower frequencies, which may be a good thing if you're worried about having the neighbors complain.

I think it's great!!! Are you planning on including the buffer stage in the final build??? Reason I ask is I haven't seen it as necessary, but then I've only used a single frequency...

Yes the reason for the tapped coil is to increase antenna tuning range... If you'd rather not use the switchable coil, you can always wind some different value plug in's(defiantly more period)...

At first the idea of a tunable Tx is enticing but I rarely move mine, in fact the one I use most has been on the same freq 99.999% of the time since I built it back in '07... Of course I have to pull it out of the case and adj the coil, so retuning isn't exactly convenient... The other VFO Tx is adjustable from the front panel, but the antenna tuning is on the back and of course a method of verifying a peak is necessary when retuning... For that, one can use a 50ua meter with a germanium rectifer to build a meter(either portable or Tx mounted)...

Tom

It was an FT114A-61 ferrite toroid with 75 turns of #24 (or thereabouts) enameled wire. It was tapped at 45, 52, 58, 64 and 70 turns. The reason for all the taps, was that I had no clue what inductance value I actually needed, so I made it adjustable. It covers the range reasonably well, but if you're aiming to transmit at a frequency below 1000 kHz you may want to add a few more turns. Of course, it also depends on your antenna.

In operation, I found that there would be two or three different tap positions that would work for any given frequency, and it was a matter of adjusting the variable cap(s) to get a peak in the output power. However, there would always be one tap position that gave the best peak, so there was some minor trial and error to get the optimum tap position. The RF signal strength meter was a great help.

I thought it might be nessessary to prevent the transmitter from putting out FM, but it seems it was more trouble than it was worth. I am coming to the conclusion that it is impossible to expect any level of linearity from an untuned stage.

Here is the updated schematic. With B+ =100v, the '38 is receiving about 100 v p-p of rf and the grid is drawing enough current to drop about 40v through the grid leak resistor. Therefore, it is definitely running in class C and the cathode resistor is unnecessary but I left it in to protect the tube should excitation be removed. The sound quality is excellent. By reducing the screen voltage I am nearly down to the FCC max input power of 100mW, I hope I can make it.

About the antenna coil:
I wound an enormous antenna coil, about 1.7mH. It allows me to tune from 550 to 1700 kc. However I notice that the signal from my transmitter is spread out over a somewhat wide area on the radio dial and is also releasing fairly considerable harmonics. Would winding a higher Q antenna coil, such as the ferrite toroid one, fix these problems?

I also have a question about tuning the antenna. Do I want to tune it to the highest rf voltage on the antenna? This point corresponds pretty closely to a substantial drop in plate current. For a B+ of 100v, the unmodulated carrier has an amplitude of between 400 and 500 volts p-p. The amplitude doubles when fully modulated and I can no longer view the entire waveform on my scope with 10x probes. If I turn the B+ up to 250v and fully modulate, purple glow zaps between the antenna capacitor plates.

Ahhh the circuit is looking approx like I thought you'd wind up with...

I can't answer about the Q of the coil but I'd think it would help, the ones I've used have been fairly high... I do suspect most of the harmonics are directly out of the Tx and not necessarily from the antenna... as long as they are not strong enough to be received more than 10ft away I wouldn't be to concerned...

On the tuning, yes the peak in antenna voltage should occur at minimum cathode current... I would wrap a turns few turns of wire around antenna insulation and use that as scope pickup, having it coupled direct is likely causing a detuned condition once removed...

Tom

How about a PI filter on the output to filter out the harmonics and match the antenna impedance?

http://www.smeter.net/amplifiers/pi-net ... output.php

It already has the pi filter (minus the second var. cap.). It's just drawn a bit different than normal. Maybe the second var. cap. would help in this situation. Not Sure.

When you say that "the signal from the transmitter is spread out over a somewhat wide area on the radio dial" if these are not at exact multiples of your transmitter frequency, then they are not harmonics, and increasing the coil Q won't help. In fact it's not likely a problem. It may simply be that your receiver is getting such a strong signal that it's overloading. It may also be due to the fact that you're not limiting the bandwidth of your audio signal. If you feed the transmitter with a signal that has a 20 kHz audio bandwidth, you'll be transmitting hi-fi audio (a good thing), but it will be spread out a bit more than normal, but not a lot.

If you are getting significant harmonics, then you may want to play around with the pi network a bit. High Q should help in this situation.

I determined that reducing screen voltage was not a viable option to achieve Part 15 compliance. In order to get the PA tube to draw only 1mA at 100 volts, I needed to reduce screen voltage to only 20 volts. The result was that it required just as much modulation power as if the screen was at 100 volts but at much less power output. Bypassing the screen resistor with a capacitor greatly reduced the amount of power necessary for modulation, but some nasty looking distortion resulted. Either way maximum modulation percentage was only about 50%.

So then I tried connecting the screen grid directly to the modulated B+ again and just reduced the B+ voltage until the PA tube drew 100mW, This occurred at 50 volts and about 2mA. Now I can modulate it with only a little power and achieve nearly 95% modulation. The modulated wave looks really good from 20 to 20k c/s. I am pretty happy with the design now and am going to start the construction of the final version soon

Some observations at the antenna with the plate and screen at 50 volts and 2mA:
Peak to peak unmodulated carrier voltage: 250v
Peak to peak modulated carrier voltage at 1000 c/s: 470v

With a 3m antenna located in a second story room in a brick building, and the transmitter operating at 900kc/s, the usable range is a bit more than 1 building from the originating building in all directions when using a walkman.

About the harmonics: adding a second capacitor to make it a true pi filter did not really help. My toroids haven't come yet so I do not yet know their effect.

The picture below is the carrier being modulated with a 1000 c/s wave, with the 'scope set at 100 volts per division.


Thanks for all the help. It was much appreciated.

Nice Jon, thanks for the report...

If I locate a 2.5v transformer, think I'm going to build one using the 47 I have... There are several options building with lower voltage tubes that I already have on hand... I'd just buy a S bulb 42 but those are like gold...

Maybe I'll try driving the 47 with a pair of 45s in PP... OK I'm kiddin' on that one, I do want to try modulating a S bulb 112(and maybe a 45) I have...

Tom

That transmitter is far out Jon! Simply far out! Nice job, and with globe tubes even! Who would have imagined?

Thanks!


I tried it out with a nice big globe '45, but I had to drop plate voltage down to 38 volts in order to achieve Part 15 compliance. If I ever build an all 1920's breadboard transmitter I will probably use a 171A.
---------------------
I got my toroid today. I bought the FT-140-61 toroid to allow a little more winding room.
The result? HUGE decrease in harmonics . Substantial increase in range and antenna voltage. I can also easily modulate 100% now too. I am not sure if that is from the new toroid or from something else that might have happened while I changed over the coils.

I breadboarded the modulator and it worked fine. I also tried using a single UY236 instead of the pair of UY237's, with 180 volts B+, 18 volts on the screen, and -1 volts on the grid. It had plenty of gain and worked just great. However I think I will stick to the '37s since the '36 isn't actually listed for audio duty and since I want to pad tube count a bit.

I also need to find a suitable setup for a modulation transformer. The load impedance of the '38 in the modulator is about 12,000 and the input impedance of the PA is 25,000 ohms. I was originally going to put two radio output transformers back to back, but the ones I tried had NO bass, probably since their inductance was too low. For testing I have been "borrowing" the Edcor GXSE10-8-5k's from my '45 amplifier project. They have plenty of bass despite being only 5000 ohms impedance. I might try a 5k connected to a 10k from their cheapest line of transformers but that would be a bit pricey at $36 dollars for the pair.

Maybe I'll try the 1623 I have, not a globe but the ST-19 bulb is impressive enough...

http://www.nj7p.org/Tube4.php?tube=1623

Actually I went through my stash today and found a globe 71A I'd forgotten about... Found it and and a globe 80 in a falling to pieces Radiola 17, at the local surplus whse...

I'm thinking I'm going to use a power transformer from Precision 612 I have with a bad meter... It has a dedicated 5v rect winding and the other secondary is tapped from 1v to 300v... Thinking I may be able cut the tap at maybe 12v and use the HV winding separate from the heaters...

Tom