Lash up Direct Conversion Receiver

[pwdrive]

Cheers Al.

Lawrence.

[pwdrive]

Just another thought, would welcome any ideas/comments etc on this:
So far the tunable IF that has been selected is centered on 2.5 Mhz and tunes approx +- 500khz either side of that frequency, the anoe of the Mixer is connected in series with the primary winding of the tunable IF tank circuit (Regen det. tank), from a practical point of view the primary winding cannot be variably tuned because of loading effects on the high impedance secondary tank circuit, the point I am coming to is this:

So far as a harmonic generator for the local osc is concerned I originally thought of a tuned tank in that stage to select the harmonic of a 1 Mhz osc to set the band whilst rejecting the adjacent harmonics to a low enough level as not to be to noticible in any mixing process, I am also thinking now of another approach, it is nothing new, Racal use it but not entirely in the same context as the Regenerodyne. The idea is this, instead of selecting the required harmonic with a tank circuit in order to supply the mixer how about having no tunable tank and instead feed all the harmonics generated to the grid of the mixer ie a "comb" of frequencies ie: untuned and beef up the filtering in the Mixer anode circuit before the signal is coupled to the tunable regen detector by way of a bandpass filter?
Can anyone see any fundamental flaw in this (theory wise) The signal input to the mixer is preceeded at the moment by a double tuned RF tank (RF grid and Mixer grid) I realise that this would not prevent any spurious mixing by any of the harmonics presented to the mixer in combination with any signal that the double tuned RF tanks cannot reject down to a certain level, I also realise that this is more of a problem the higher up the shortwave spectrum I go. It is possible that I could use a triple ganged RF tank as triple gang second hand caps still seem to be available this would enable a double tuned tank to be used in the grid of the RF amp with top capacity coupling as well as a tank in the Mixer grid, I realise that the gain stakes are now going down but by using more modern valves than what I am presently using I feel sure it could be made up for.

Lawrence.

[Skywave]

I think that that approach is heading for problems. I haven't done the full sum & product evaluations, so I'll argue my case qualitatively.

At the anode of the mixer, before the resultant spectrum gets to the BPF, you will have products of the 1 MHz tuned spectrum (R.F. tuning) and all the harmonics from the 1 MHz oscillator (your 'comb'). Therefore, the BPF will need to be tunable to reject all the unwanted products. To produce a BPF that is tunable will require several tuned circuits, all ganged to the R.F. and mixer grid tuning capacitor: sounds like at least a 4-gang tuning capacitor will be required - maybe more. And then there is the little issue of the tracking: the tuning of the BPF will have different Cmax and Cmin ratios to that of the R.F. grid tuning and the mixer grid tuning.

In conclusion, I reckon that your best approach is to arrange for a 'clean' L.O. injection signal at the required freq. (according to the R.F. & mixer tuning freq.) and feed that into the mixer. It may be possible to arrange for a 3rd. section of a 3-gang tuning capacitor to perform the necessary filtering to produce just the required nth. harmonic of the 1 MHz master osc. source. The simplest route, of course, is to build a stable wide-range tunable osc. that covers the required freq. range for the L.O injection and simply select the required osc. freq. prior to tuning over the required signal range - in effect, what you have been doing by using a sig. gen. as the L.O. injection source! Of course, that moves well away from the H.G. approach that you prefer, so presumably that's a no-no.
Further that that, at the moment, I can't think of any other suggestions. What is quite evident is that your requirement to tune the whole R.F. range from 2 - 29 MHz (or something like that) does make the design difficult. Now if you were aiming at only the Amateur Radio bands (or any other specific R.F. freq. ranges) then the design becomes a lot easier. Whatever, I'll have a further think about your general freq. coverage requirement and see if I get any bright ideas, but I'm getting a sneaky feeling that if anything relatively simple can be done for general coverage, then Racal would have done it, instead of the design they produced for the RA-17! (Or Plessey with the PR155, and various other manufacturers and their designs, which are nearly all synthesizer or pseudo - synthesizer based).

We'll talk again. In the meanwhile . . . . keep at it!

Al.

[pwdrive]

As always Al. thanks for that you're keeping me going! your initial thoughts on that last idea I posted run similar to mine but I thought I would throw it open for comment.
It's been piddling down all day today so I spent the day at the Regenerodyne bench, I hooked up the output from that 500Khz Xtal marker in the B40D again and fed it to the grid of a triode, the triode does not have a particular high Gm, nevertheless I made up a tank circuit for the anode load the fed it's junction via a capacitor to the grid of the mixer, I have tried this before but I took a bit more care this time, the results were reasonably encouraging, although I could not get much out of it much above the 20th harmonic, at the 20th harmonic (10 Mhz) I peaked the tank and set the RF tune to the 40 meter slot, it worked quite well although the mixer conversion gain seemed to be down which was kind of expected. The reason I mention this is that if I can get that with a 1 Mhz xtal job then that might take the HG source up to 20 Mhz which will be two thirds of the way there....
I then substituted the Xtal marker input with the sig gen at 10 Mhz (for 40 meter coverage) and the gain shot up considerably once the triodes tank had been tuned, unfortunately I have no 'scope or means to measure RF which is a pain.
I also tried a short while ago setting the sig gen to 17 Mhz (for 20 meter coverage) using the same arrangement and proceedure, the results were equally as good. I also tried resetting the sig gen to half frequency in both instances but tuning the tank at double the half frequency this worked well, the gain as best as I can ascertain being only marginally down.
So in short, amplifying the output from the sig gen then tuning that amplifiers anode tank circuit to the sig gen frequency increases the Mixer gain, this is to be expected as normally mixer grids have several volts on them as opossed to the 200 mv I previously had (un amplified, I also noticed that certain unwanted hetrodynes disappear by using this method but I guess that is to be expected also as we are cleaning up the sig gens output..
Of course all this is fine but I need to try out the HG idea with some decent valves/ components before knowing which path to take..I have still yet to order them!
If I have to go some other route then one way could be a VFO say for general broadcast transmission tuning and a few switched Xtals for the Amatur bands for that extra stabilty.
I might even put it all together properly and use the sig gen, it would make a nice compact receiver.
I have also rewound the RF tuning coils which has resulted in improved tracking between them, the RF tune peaks up really well now on 20 meteres with a nice sharp peak, a reduction drive will not be required on the RF tune just a decent size control knob say 2 " diameter with a calibrated skirt or pointer.
If anyone is thinking of building a Regen for SWL then I would certainly recomend the RF amp-Mixer-Tunable Det- A/F Output route which is basically what the Regenerodyne is.

Lawrence.

[Skywave]

Quote, PW-Drive: "Unfortunately I have no 'scope or means to measure RF, which is a pain."

What? You're doing all that R.F. R & D work without a 'scope? You're a braver man that me, Lawrence! Without a 'scope (and a good one too) I wouldn't contemplate it within one micro-second! I take my hat of to you for your excellent demonstration of the necessary "three P's" in this type of work: Patience, Persistence, Perseverance.

Al.

[pwdrive]

Yes Al you read that correct...no 'scope...just intuition/hunches and a bit of experience. I might treat myself to a 'scope sooner than later.
Just a quick note to say that I had a concerted effort to get the "borrowed" Xtal marker (500Khz) to come through on the 37th Harmonic (17 Mhz) in order to tune the 20 meter band, I managed to resolve a couple of strong SSB transmissions this morning, I double checked with the sig gen and it was definately the 37th harmonic, my conclusion is that at that harmonic it's getting pretty weak at that 'n th. Seeing that the end game is to use a 1 Mhz Xtal then it offers some hope.
In order to improve things I am now thinking of say taking the output from a Xtal osc then symetrically clipping it with a couple of solid state switching diodes (these would be allowed as per my original brief) and feeding that to a to a higher gain valve and tuning the tank to select the harmonic as I have already been doing, so overall it's Xtal osc...clipped....amplified...tune output at the harmonic required, there might be problems however as the output from a clipper might not be high enough to drive a single amp enough, if not then another stage might be required, I think this could still be done with two valve envelopes (including Xtal osc) as was originally planned. If I can find a spare valve and a couple of signal diodes I might try that in a lash up form just to see.
Of course in doing all this the ideal tool would be a spectrum analyzer, I have never owned or used one but one day I might consider buying a good used one, an old HP or similar, do you or any members have any experience with these things?

Lawrence.

[Skywave]

Ah yes, spectrum analyzers: wonderful things, provided that you:
(a) get one that is of a reputable make,
(b) get to understand its limitations, since they can easily give erroneous results,
and
© be very careful not to damage it!

For (a), I am thinking of HP, Tek and Anritsu; that's based on my experience.
For (b), you need to be aware that there can be combinations of input signal levels and control settings that will give misleading results.
For ©, this remark is really limited to using an analyzer with high R.F. power sources - such as transmitters: always insert an attenuator pad (typically 30 dB) between that TX and the analyzer's input, to avoid burning out the input attenuator on account of sheer overload. (Please don't ask me how I know this! )

There are basically two distinct types of analyzer: ones with and ones without a tracking generator (T.G.). Those without are O.K. for things like EMC and field strength measurements, but if you need to look at the response of things like filters, amplifiers, etc., you do need to have a tracking generator. Early analyzers tended to have a T.G. as a separate box; modern ones have them integrated into the unit.

Probably one of the commonest S.A.s on the second-hand market these days is the HP 141T: they typically change hands in the region of £300 - £400. At the other end of the market, there are things like the Tek 494 (> £3000) and the superb Anritsu MS2687B (typically £14k - £15k). The Anritsu range with the prefix MS26xx are superb analyzers and a joy to use . . . but they don't come cheap!

Al.

[pwdrive]

Thanks Al...HP 141T it is then....

Lawrence.

[pwdrive]

Had another go at this today, first thing was to tidy it all up a bit to reduce all the stray C, that turned out ok.
It is clear that the signal from the donor Xtal marker in the B4O D does not have enough output so what's to do...
I switched the B40 to band one and pulled the mixer valve removed the screening plate covering the acsess to the turret tuner, I then slid a 100pf between contacts 2 and 3 on the oscillator section and screwed the osc coil down to 1 Mhz for use as a fundemental frequency generator, this was then fed from pin 3 of the mixer valve socket in the B40 into a harmonic amp and the anode tank was duly tuned, with the 10th harmonic tuned and fed to the mixer of the regenerodyne the reception on 40 meters was very good so I changed the tank coil and re tuned for the 17th harmonic to test the reception on 20 meters, the good news is that there was a contest on so plenty of transmissions and also I had plenty of output from the tank on the 17th harmonic to drive the mixer, reception was very good.
The bad news is the Q of the harmonic selector tank starts to fall off somewhat at the 17th and when tuning the tunable regen det, products of one of the lower harmonics were coming through even though the tank was tuned to the 17th, bit of a blow but kind of expected, back on a positive note, feeding the sig gen into the harmonic amp and tuning the harmonic selector tank to increase the sig gens output produced very good results. It might be posssible to have a double tuned tank to increase the overall tank selectivity for the harmonic selection, next time I have a spare day I might give that a go.

Lawrence.

[pwdrive]

Did a few more hours on the Regenerodyne project late this afternoon and this evening, made some progress, first I replaced the RF Amp (6AC7) with a 6SH7 as on the high frequencies the 6AC7 was starting to go into regeneration which I don't really want in the RF stage, the 6SH7 has an internal twin cathode connection lead out thus reducing the internal cathode lead inductance, the RF Amp now performs better with this valve fitted.
I decided to change the harmonic generator/amplifier from one half of the 6SC7 (triode) to the now spare pentode (6AC7) one of the reasons was that the other half of the 6SC7 is used as the first AF amp but unfortunately the valve has a common cathode so poss harmonics/RF getting into the audio line and maybe back through the HT or heater line.
I plumbed in the 6AC7 as a pentode connecting the control grid via a 100pf to the donor 1 Mhz osc in the B40D, I rewound the two anode tank coils (one for the low bands and one for the high bands) optimised the bias and screen voltage and fired it up, the results were a lot better than I expected, so much so that on tuning the harmonic selector tank to the 21st harmonic and peaking the RF tunining control for 18 Mhz I was rewarded by listening to a amature transmission from Venuzwalia (YV4EH) to a bloke in Bulgaria on 17 meters so really pleased with that, also now the mixing products from the harmonics either side of the tuned harmonic are somewhat reduced to what they were before.
Hear is are some rough unscientific test results:

Regenerodyne tuned to a medium strength European amature transmission around 18.1 Mhz using the 21st harmonic of the 1 Mhz donor osc. and manually tuned the sig gen which was set to full output on its attenuator (25 mv) and set to CW, its center insulated coax lead was wrapped round the aerial lead to give a bit of loose coupling, outer of the coax was left floating.
Tuned sig gen to received signal......blottoed as you would expect.

Lower harmonic products:
17.1 Mhz, quite noticable when tuned through the amature transmissionon on 18.1 Mhz but still able to resolve transmission ok.
16.1 Mhz, noticable but generally weak.
15.1, 14.1, 13.1 Mhz etc, nothing noticable.

Higher harmonic products:
19.1 Mhz, as 17.1 Mhz but weaker.
20.1 Mhz, as 16.1 Mhz but weaker still.
22.1, 23.1, 24.1 Mhz etc, nothing noticable.

During the above test the sig gen output was kept constant.

Unscientific I know but it gave me a good idea of what's what so far as listening tests go, also remember that nothing in the receiver is screened as yet (just the metal valves) overall quite encouraging, however the acid test will be from 18 Mhz up to 30 Mhz, this is unchartered territory as I have no RF or harmonic tank coils for those frequencies yet, I will attempt to wind some next time.
I did not have time to perform the above test down at 7 Mhz but I would think the results would be even better as the inductance Q's rise.
Since I started this project I have ascertained that it performs very well (in my opinion) from MW through to 18 Mhz, the audio quality on the SW broadcast transmissions is very good when the band conditions are good.
More to come I dare say.

Lawrence.