Time for the tech-heads...

We all know to tune the antenna and IF circuits to get the best sensitivity, but does the strength of the local oscillator have any effect?

Obviously if the level is zero, you don't receive anything, but is there an ideal level?

Should the level depend on the station's signal strength?

Does a stronger oscillator make for a better (or worse) receiver?

I've wondered about this for a time, but of course radios don't have adjustments of the level without messing up the tuning range.

Any thoughts on this?

In many Pentagrids the modulation falls with frequency increase. In sets with AGC / AVC it controls the signal strength, so it is difficult to have over modulation in a set like that, if it is designed properly. but the ideal as far as I can see is to avoid over modulation and many ABC transmitters actually broadcasted (may still) with under modulation.

Over modulation should, I would think, result in distortion.

Marc

Hi Robbert,
Don't know for sure .... but I would suspect that the Local oscillator signal level should be comparable to the modulated carrier signal level that is being tuned into. Obviously, this will vary.
The oscillator 'mixes' with the carrier to produce the IF ... this is a frequency related mixing, not amplitude, therefore I am guessing that the amplitude of the resulting IF will be equal to the smallest of the two mixed signals.
Only my guess ...

A too powerful LO will result in nuisance levels of RFI. Other problems I expect would be unwanted harmonics and distortion from non-linear operation.

It would be interesting to experiment with this.

I don't think that there is much need to experiment to see what happens; Over modulation is pretty much a known. Once the audio escapes the envelope you start to get phase & bandwidth issues along with spurious signal.

Full AM modulation gives the loudest signal at the second detector, which is why the commercials push it & then there is signal compression to add in. Distortion will result from signal clipping in the amplifiers if the bandwidth is exceeded.

You would need an Oscilloscope to see it all & there are books on Oscilloscopes by "Rider" & others that show over modulation patterns.

I have a Home made generator here that deliberately uses a dual gate FET as an autodyne mixer. With it I can control the modulation level from none to over done.

Marc

This thread is not about modulation. I want to know if the level of the local oscillator signal produced in the mixer valve has an effect on the radio's overall sensitivity.

I guess the real sensitivity is achieved in the front end.
And would not be directly effected by the level of oscillator injection.
Apart from other rubbish(noise), on the actual osc signal and the noise produced by the mixer.
Interesting topic though.
JJ

Hi Guys, a very interesting question and I understand exactly what robbbert is asking.
I have some measurements to do with this topic from one of my superhet builds as back then I just did not quite understand how a convertor works. I still don't really get the maths fully but now can sort of see how it all happens.
I'll do a little paper on these tests when I dig them out and decode my gibberish and I'll attach them to this comment tomorrow as a PDF.
Lets define what the question means, " does the STRENGTH of the local oscillator have any effect on SENSITIVITY?"
I understand this this to mean: That if you vary the local oscillator amplitude will that vary the sensitivity?
The local oscillator strength could be measured by the AC voltage at the oscillator tank or at the convertor plate.
The sensitivity of the set could be expressed as the millivolt audio output for a given microvolt RF input. (how loud it sounds)
As we are not looking for absolute values if we could by some means vary the oscillator strength and ASSUMING ALL OTHER THINGS REMAIN CONSTANT like the AM signal amplitude and the oscillator frequency and 10 other things, and then read the AUDIO OUTPUT as we vary the oscillator frequency we will get an answer.
THE SHORT ANSWER is the local oscillator amplitude in the sets I have built has a "goldilocks" value. That is you need enough but not too much or too little. The amplitude can vary over a wide range and so long as this is somewhere near equal to, but can give and take a lot, of the value of the incoming AM signal then beats will be produced to give the resultant IF frequency value.. As the amplitude of the AM modulation still remains pretty constant , then the audio output will not vary much and so the "sensitivity" of the set will not vary much.
Obvioulsy the incoming RF AM signal will vary in amplitude thus varying the proportionate size or percentage of the modulation but from an overall view this is what the AGC control system as a negative feedback loop is there for! As we are all playing about with radios or components that are already sized and proportioned to work and not building our set and coils from scratch then convertors and IF sections just work as they should. it is only when you go out of your way to vary or have a faulty part that varys things that you find out if you have too much or too little!
But coming back to the SHORT ANSWER if you vary the oscillator amplitude WITHOUT AGC you can 'drop out', or,' overload' some part of the loop such as the IF stages if you go out of the goldilocks band. If you engage the AGC then you virtually have to SHUT DOWN the local oscillator or somehow screw it so high it blasts over the RF AM signal before it becomes a problem.
I have some photos somewhere from that superhet build that illustrate the audio variation with oscillator variation and i'll drop them in in my PDF..
Cheers Fred.

Ok I have now decoded my scribbling and found the jpgs and have put together an essay on the subject.

Superhet Oscillator Output

Cheers Fred.

I have had a snowy picture on a television receiver caused by a noisy oscillator transistor a couple of times.
This symptom is usually caused by the R.F. transistor BF200 etc.
While not the mixer, it is still one part of the two signals to be mixed.

Wayne.

Thanks Fred, something I've wondered about, and you've answered it very clearly. But some of us ARE making radios from scratch, hand-wound coils and all.

Obviously if you do not think that the sets own oscillator is about modulation, then it is quite clear that you do not fully grasp how a superheterodyne mixer works and you will not be alone.

The IF signal passed through the entire IF stage is a modulated signal and it can be messed up in the Pentagrid. If it is the RF stages will distort. If it under modulates, or is receiving an under modulated signal that reflects on the volume. AGC can only compensate so much. The rest of the sensitivity relates to the Q of the tuned circuits and their bandwidth. Related to Fred's comment I have had to adjust the bias in an autodyne (screen grid) set due to distortion from signal overload (no AGC).

Hand wound coils are interesting. I recently made one for a transmitter antenna & fishing line played a role. The SWR was not quite as good as a commercial one the guy had but the field strength was actually greater.

Marc

Document uploaded to Post 8.

Thanks Fred and Brad.

Going by that essay it appears as long as the local oscillator is running normally, everything will be roses.

But it's possible that things might work better if the local and the RF inputs were at the same level. AGC should keep the RF somewhat constant, so in the end it should all just work out.

Hello Robbbert.

Yes, the oscillator activity does affect the sensitivity. Oscillator activity was traditionally measured by inserting a current meter between the oscillator grid resistor and the earthy connection. There is an optimum value. For a 6AN7A for example, using a 22k grid resistor it is about 360 micro amperes. At this level, the conversion transconductance, gc, is at maximum, at about 0.75 mA/V. In practise though, the grid current does vary quite a bit over the tuning range and this will affect the sensitivity to an extent. At 700 uA, the gc drops to about 0.69 mA/V, and at 200 uA, it drops to 0.6 mA/V. These figures are from a graph in the 1965 edition of the Miniwatt Designers Handbook. If the grid current stays within these limits over the tuning range, I wouldn't worry. A radio will work with a lower grid current but the oscillator might drop out over some of the tuning range under conditions of low line voltage and an ageing valve.

Cheers, Stuart

The IF is produced by mixing the incoming modulated RF with he LO which is tracked to have the IF offset relative to the RF frequency at all times.
When those two frequencies are mixed, the IF is produced, and conveniently, the heterodyne frequency is modulated with the intelligence on the incoming RF.
So ask yourself if you adjusted the amplitude of the incoming RF, would that have an effect on the amplitude of the IF?
If you theoretically sent a pure sine into RF, and modulated the LO (don’t do that) would that make any difference to the way a superhet works?

The answer to those, and many more questions will never come from me but you can send an already IF modulated signal through the antenna or LO.