Tunable Harmonic Generator ideas?

[pwdrive]

Just throwing an idea around, as per title, it's for a regenerodyne project I hope to do.
The standard set up for shortwave would be a switched xtal osc to feed the mixer, the mixer output would be up around 3 mhz feeding a tunable regen detector which would have a tuning range of about 500 khz, this set up would help to reduce image problems.
The question is does anyone have any thoughts on the following: osc would be xtal say 1 mhz operating in a Pierce osc. configuration so therefore class C operating, it should be rich in harmonics so should in theory be able to tune the output of this osc. to the harmonic multiples and use that as the osc. feed to the mixer. Probably have to have say three switched ranges (coils) tuned by a common (to each coil) variable capacitor in order to tune all of the shortwave band, It's a bit similar to the idea used in the Racal RA17 etc.
The idea is to use the above as a band set osc. in harmonic incriments of the xtals fundamental frequency ie: 1, 2, 3, 5, 6 .....etc mhz. As mentioned above the difference frequency (the IF) needs to be up around 3 mhz in order to obtain reasonable image rejection, the IF will be tunable by a regen detector.

Lawrence.

[Skywave]

Hi Lawrence,
I've been sitting here with pencil and paper trying to produce a block diagram to figure out the overall idea: and I'm sorry, I'm stuck! One thing does cross in my mind: since your detector has a tunable bandwidth of 500 kHz, surely this means that your 3 MHz I.F. must have a bandwidth of from 2.75 MHz to 3.25 MHz, i.e. centred on 3 MHz ? Therefore, irrespective of your method of deriving the local osc. freq., surely that means, in turn, that the necessary pre-selection before the mixer needs to have a selection of tuned circuits similarly tuning in 500 kHz segments. And that is a lot of L/C ccts. to cover the whole S/W spectrum !

It's obvious I've misunderstood things So if you could send an attachment with a block diag. explaining your idea, I would be much obliged; thank you.

Cheers,
Al.

[pwdrive]

Hi Al, just got in and read your reply, I will try and answer and give a bit more info before the evening is out, thanks.

Lawrence.

[pwdrive]

http://www.qsl.net/wd4nka/TEXTS/REGENf~1.HTM

Hello again Al, above and below is a link to the design I want to base my Regenerodyne project on, within that link you can access the circuit and other relavent info. I must stress that the circuit is only a basis for what I have in mind ie: I will not be following it to the letter and probably will be using different valves.
The idea is to build a little SW Receiver that isn't a full blown comms. set but is better than the average one or two valve Regen. I have set a valve count to 4 or 5 envelopes max for the receiver, solid state signal diodes are allowed, the power supply will be a seperate unit all together. So that's the brief.
As can be seen from the circuit in the link the RF stage tunes from 7 to 14.5 mhz and there are three switched xtals in the osc. these could be chosen for say 7, 10 & 14 mhz coverage of the receiver (3 amature bands) taking into account that the xtal frequencies are chosen to give a center IF frequency in the receiver of approx 3.25 mhz relative to the band selected, note that in the circuit only the secondary of the IF transformer is tuned, the approx range of the tuned secondary is +/- 250 khz centered on 3.25 mhz.
My idea is to expand on the stability of using a xtal osc. but at the same time reduce the number of xtals needed for different bands, so I was thinking maybe just one xtal (maybe two) in the osc. but the osc. producing good harmonics, as it would say in a well designed xtal marker circuit.
I have not yet worked out the frequency of the xtal needed but to take an instance, if it was 1 mhz and the osc. circuit it was used in produced good usable harmonics up to say 33 mhz (the harmonic generator), then it might be possible to cover all the SW band with the one xtal. The idea would be to have another variable tuned circuit in the output from the osc (before it goes to the mixer) this tuned circuit would be able to "peak" the harmonic chosen for a particular part of the band that you would want to receive ie: if using a 1 mhz xtal the osc. the osc. feed to the mixer can be increased in i mhz increments, you could maybe go one stage further and have 2 selectable xtals in the osc. one 500 khz above the other, the harmonic "peak" tuned circuit as previously mentioned would still probably cover the new set of frequencies introduduced by adding the other xtal, the added advantage of a second (offset) xtal is that you can then have the IF either above or below Fs thus giving more options and control in reducing a particular strong station image problem. So in the above Idea...3 to 4 switched tunable RF ranges to cover say 1 to 30 mhz, 3 to 4 switched ranges to tune the harmonic required from the osc.
one tunable IF control, regen control, vol. control and maybe a 2 position xtal switch.
Can you see where I am coming from with this? I will try and do a block diagram and post it up tonight.

http://www.qsl.net/wd4nka/TEXTS/REGENf~1.HTM

Lawrence.

[pwdrive]

Hello again prelim block diagram for harmonic osc regenerodyne in link below.

Lawrence.

[pwdrive]

Hi Lawrence,
I've been sitting here with pencil and paper trying to produce a block diagram to figure out the overall idea: and I'm sorry, I'm stuck! One thing does cross in my mind: since your detector has a tunable bandwidth of 500 kHz, surely this means that your 3 MHz I.F. must have a bandwidth of from 2.75 MHz to 3.25 MHz, i.e. centred on 3 MHz ? Therefore, irrespective of your method of deriving the local osc. freq., surely that means, in turn, that the necessary pre-selection before the mixer needs to have a selection of tuned circuits similarly tuning in 500 kHz segments. And that is a lot of L/C ccts. to cover the whole S/W spectrum !

It's obvious I've misunderstood things So if you could send an attachment with a block diag. explaining your idea, I would be much obliged; thank you.

Cheers,
Al.

Hi Al, when tuning the tunable IF the RF tuning can be set for max signal for any point on the RF range selected, none of the tunable circuits are ganged. Any help/comments welcome.

Have also replied with better description etc and preli. block diagram.

Lawrence

[Skywave]

Hi Lawrence,
I now have a much better idea of your thinking; thank you.

I suggest you make your detector tunable over a 1 MHz range. That way, the osc. injection can be in 1 MHz steps with front-end tuning to match. By using 500 kHz range, the osc. needs to change at 0.5 MHz intervals - and that makes the design of the oscillator injection rather awkward.

Al.

[pwdrive]

Thanks Al, yes 1 mhz tunable range on the IF transformer would be a good idea, if so do you think I should increase the IF frequency?

Lawrence.

[Skywave]

Well it's certainly going to need to be tunable over 1 MHz, and in the range of about 3 MHz, that's not ideal; 500 kHz will give a sharper tuning; not a lot more, agreed, but it will help. As for determining what freq. that should actually be, the question of image responses arises. Obviously we do not want an IF and osc. injection freq. that gives us an image band that corresponds to broadcast bands which will have high-powered transmitters in them. At the same time, we do not want an IF that is that high that tuning becomes a problem. The calculation of possible IFs and osc. freqs. is just a matter of sitting down and doing a few sums, for which Excel is particularly handy.

That said, I've been thinking about your original idea of a receiver covering the three Amateur Bands, 7, 10 and 14 MHz, and came up with the nucleus of an idea. It doesn't use a harmonic generator as such, but does use two frequency multipliers and an additional mixer. (Ideally, that mixer really needs to be a double-balanced affair: easily accomplished with a twin triode or a 4-diode ring modulator. For this exercise, I'd go for the twin triode: greater output). The freq. multipliers and that mixer obviously need pre-set tuned L/C ccts. to select the required harmonic.

I think the attached drawing says the rest.

[attachment=4681]

Note: The 'R.F. pre-select' involves the use of a valve with tuned L/C ccts. in its grid and anode, so a twin-gang variable cap. is required here, in addition to another wafer on the 2-pole, 3-way switch , as shown. An R.F. gain control would be a useful addition, too. I'd also make the signal / osc. mixer another DBM too.

On that same theme, and with a tunable IF of 3.0 - 3.5 MHz, and previous to the above, I did come up with a scheme using two crystal oscs. at 3 and 4 MHz, and by the use of two mixers, produce osc. injection freqs. of 4, 7 and 11 MHz, (for RF input freqs. same as above). Unfortunately, upon an analysis of images, the 7.0 - 7.5 MHz signal input band has an image band of 1.0 to 0.5 MHz: the MW broadcast band! So that hit that idea on the head.

All-in-all, an interesting exercise though.

Al.

[pwdrive]

Thanks for your input Al and the time to do your block diagram, I have noted the frequency multiplier idea and the DBM's, I did a bit with DBM's years ago when messing around with direct conversion receivers using the old SBL1 device.
I agree, working out the osc freq. with respect to potential strong image transmissions would be the key to most of this, obviously there is a trade off and a limit with such relatively simple designs and fully realize the limitations. However I would like to stick to the 500 khz tuning if poss as when coupled to a suitable reduction drive the tuning should feel about right, having said that if I chose to go the 1 mhz route I can always add a low value vairiable capacitor across the main gang for fine tuning.

Lawrence.