Help understanding single-signal reception via i.f. regeneration, please ..

[KC9KEP]

Hello all,

I have added an "optional" i.f. regeneration component (C15) to my 1941 Single-Signal receiver.

http://www.bignick.net/images/RadioPix/7_tube/schematic_7_tube.jpg
http://www.bignick.net/Morgan_Radio/Radio_6.htm

It provides a degree of regeneration in the i.f. stage, sharpening response.

The handbook tells me to re-align the i.f. stage after the installation of the feedback capacitor (C15), which I did.  The i.f. peaking adjustment was much more touchy than it had been when aligning the i.f. without the feedback component.

Now, when I advance the r.f. gain, the audio will squeal if adjusted too high, which is expected because the i.f. stage breaks into oscillation.  The optimum r.f. gain setting is just prior to the i.f. stage breaking into oscillation according to the handbook.

The receiver sounds differently than it did without the feedback component, as expected.  The sharper i.f. has a narrower band pass and hence less background noise.

The whole idea of adding the i.f. regeneration is to allow the operator to be able to tune to a CW signal on just one side of zero-beat instead of on both sides, as is the common case.

The problem is that I still cannot seem to achieve the single-signal reception that the sharper i.f. tuning is supposed to provide.
For that matter, I cannot locate any theory of circuit operation that explains why this sharpened regenerative response will create single-signal reception.

Does anyone have experience with this circuit/methodology or could point me to a reference?
The ARRL handbook talks about phasing of crystal filters in the i.f. which makes sense, but I do not see why a sharp regenerative i.f. peak could provide single-signal reception?

Thanks!
KC9KEP

[G3RZP]

Firstly, where do you set the BFO? Suppose the IF response peak is at 455, then the BFO needs to be at 456 or maybe 455.8, depending on the tone you like.

The idea of regeneration is to increase the effective Q of the IF transformer by reducing losses - this is effectively by putting a negative resistance across the coil. Increased q means a narrower bandwidth - the 3dB bandwidth of a circuit is the tuned frequency divided by twice the Q. A problem is that the response of coupled tuned circuits is a function of coupling factor and Q, known as kQ. When kQ =1, the circuits are critically coupled: anything over that, and you start getting two peaks. If the IF transformer was originally critically or overcoupled (which was done for some broadcast receivers), increasing Q by regeneration may make the situation somewhat worse. An apparent narrowing of the passband may be because the other peak has moved far enough away.

The 1948 ARRL handbook has a receiver using a regenerative IF to get a single signal response. The specified Millen IF transformers appear to have quite loose coupling, judging by the distance between the coils. That article also makes the point about tuning the BFO away from the IF peak.

Does this help?

[W8JI]

If you tune the BFO far enough away, any receiver can have single signal. Of course only a dog can hear the CW notes and that does not cure the real issue, selectivity.

:-)

Maybe Tom has the BFO wrong, or maybe he wants single signal with signals too close for the selectivity slope (shape factor). His -30 dB points might be too far away, because a Q multiplier on one stage does not do much for the ultimate rejection. It just makes the peak "peakier".

[KC9KEP]

OK gentlemen,

Thank you for your expedient and thorough responses!

I did peak the i.f. transformers for 455 kHz.  Now, I'll take some measurements
to see where I've got the BFO set.  I had simply adjusted the BFO "by-ear" to
achive a "pleasing" CW tone, but I do not know exactly where it is.

(I am using the i.f. transformers (Millen 64456 64454) as sepcified by the Handbook.)

I'll see if I can locate the 1948 ARRL handbook to review their i.f. information too.

73!

--KC9KEP

[N2EY]

As noted, the BFO needs to be offset from the IF peak to give single-signal reception.

The regenerative IF and Q multiplier schemes increase selectivity by increasing the effective Q of an IF tuned circuit. This helps, but it is effectively a "one-pole filter". The shape factor of such methods is usually very high, and the rejection of the "unwanted sideband" is far from perfect. The Ancient Ones did it that way because the alternatives were even more expensive.

73 de Jim, N2EY

[W8JI]

As noted, the BFO needs to be offset from the IF peak to give single-signal reception.

The regenerative IF and Q multiplier schemes increase selectivity by increasing the effective Q of an IF tuned circuit. This helps, but it is effectively a "one-pole filter". The shape factor of such methods is usually very high, and the rejection of the "unwanted sideband" is far from perfect. The Ancient Ones did it that way because the alternatives were even more expensive.

73 de Jim, N2EY

Absolutely. The peak is profound, but the skirts remain far too wide.

For single signal reception the attenuation 2x tone frequency removed is infinitely more important than a narrow peak from the one or two pole filter offered through regeneration.

I think perhaps multiple stage regeneration would be better. Back when I was poor, uneducated, and stupid I added a 5 tube AC/DC receiver to an old Zenith SW receiver. I changed the mixer from a 12SA7 to a 12SK7, and attached it to the 455 IF that formerly fed the diode detector tube of the Zenith. I added shielded cable from the output of the AC/DC last IF stage back to the mixer plate of the Zenith. This added feedback around the entire IF path, as well as extra cans.

This was far more effective than my Q multiplier. (I built a BFO/Q multiplier box so the Zenith could do CW).

I wish I had time for fun, but with my wife's job exported to China retirement is now postponed forever.