[R-390] R-390A IF gain setting and S/N measurement

[Larry H]

As you know, I used Roger Ruszkowski's procedure to rate tubes for noise and gain and improve one of my 390As S/N (to an average of .5 uv) and good sensitivity.  See details in my previous post subj: 'Re: [R-390] R-390A s/n improvement 6HA5 mixer - CONCLUSION' on Aug 8 at 9:35 PM.  I found this procedure while on my hunt through the 'Pearls' and 'Y2KR3' manual for information on how to set the IF gain a little more precisely than commonly known (1. 'not too high and not too low' and 2. Chuck Rippel's '-4 to -7 db on RF noise' and 3. that -7 v on diode load with 150 uv IF in is too high).  I didn't find my answer, but found info that helped me understand it.

We've been told that if the IF gain is turned too high the S/N is negatively affected and that if we turn it too low the Dynamic Range is negatively affected.  I believe that we all understand S/N, but why is DR important.  It is the ability of the rx to handle very strong signals without distortion (Intermodulation and others).  This is important when you are trying to receive that very strong signal and want minimum distortion.  This is where the IF gain comes in.   RXs with a high DR receive weak and strong signals well.  In order to receive a very strong signal, the AGC must be working correctly.   I use two tests to see if it is:  


1.  Apply an unmodulated signal to the balanced antenna in starting at 1 uv and step it up by 10 while watching the c/l meter and measuring the AGC on terminal 3/4.  At 100,000 uv in, the c/l meter should be close to max (not real important) and the agc v should be around -15 or -16 (important).

2.  Turn the modulation on and scope the line out.  There should be no audio distortion at any sig gen level up to and including .1 volt.

If I could test for InterModulation Distortion (IMD), I would, but don't have the equipment.  However, if your rx is in good condition (no weak tubes in the RF path, resistors, caps and tuned circuits are good), then IMD should not be a problem as long as the AGC is good. 

I wanted to understand the affect of adjusting the IF gain, so while I was stepping up the RF sig gen level, I recorded the AGC voltage with the gain set to min, center, and max. Here's the results:

    RF in Uv: 1     10     100     1,000    10,000  100,000

        min:  0     2.3    5.4      9.2     12.4     16.1

        ctr:  0     3.0    6.0      9.7     12.9     16.8

        max:  0     5.0    7.9     11.2     14.2     17.5


I scoped the line out while doing this and found no audio distortion on any setting.  I then measured the Diode Load on min and 3/4 max for -7 v and the RF in was 1.7 uv to .11 uv, with .8 uv with the adjustment in the center.  Notice I stopped at 3/4 max. This is because when I went past 3/4 max, the noise was obviously overwhelming.  This is the point that is obviously too high.
Now for the affect on 'MEASURED' S/N:


    min: .39, center: .54, 3/4 max: .68


I call it 'measured' S/N, because this does not agree with what I see on my scope.  I could not hear any difference in S/N when I adjusted the IF gain from min to 3/4 max, so I thought I'd see if I could see a difference on my scope.  I injected a low level modulated signal into the balance input and set it so I could see about 1/3 noise on the modulated signal.  This turned out to be .7 uv.  I then slowly adjusted the IF gain from min to 3/4 max while watching the scope.  Well, surprise - the noise on the signal did not change any, and it would have been easy to see if it did.  I proved this by turning the gain up past 3/4 and did see increased noise.

So here's my conclusion: Find the noise threshold and keep it below that.  I could not tell if the DR was hampered in min or not, but I did not see any signs of it while testing or in real listening conditions.  I think you should try setting the gain for 1 uv RF in = -7 v on diode load and keep it below the noise threshold in any case.

Regards, Larry
________________________________
From: Larry H <dinlarh at att.net>
To: R-390 Forum <r-390 at mailman.qth.net> 
Sent: Thursday, July 14, 2016 8:40 PM
Subject: Re: [R-390] R-390A IF gain setting and S/N measurement


Thanks for all your great input.  It has been very helpful.  This has prompted me to read through the Pearls and Y2KR3 again this past few weeks looking for an alternate and/or a more precise way to set the IF gain.  I always thought that the 150uv/-7v setting was a little high, but not sure how much.  And, I knew that setting it too low was not good either.  And, thanks to Charles Steinmetz, I now know that too low negatively affects its Dynamic Range. 

So, here's what I found:

1. A procedure by Chuck Rippel says to set it between -4 to -7 db on the line meter on RF noise (no antenna).


2. A lot folks agree that the 150uv/-7v setting was a little high.

3. A lot folks agree that keeping the gain as low as possible (bot not too low) improves the S/N.

4. A few folks agree that setting it too low negatively affects its Dynamic Range.

I found a lot of good info written by Roger Ruskowski, but it was mostly on how to measure RF and IF/audio for 30:1 and how to get there, but nothing on how to set the IF gain more accurately.

I found some very helpful info by Dallas Lankford, but there again nothing on how to set the IF gain.

Did I miss finding on how to set it? 


My tech manual says that after setting it for 150uv/-7v, and if 1 to 4 uv input to the balanced input yields -7v on the diode load, that is good, leave it.

I like Chuck's setting in 1 above, as that seems to be fairly close.  IE, a little less gain than the above 150uv/-7v book setting.  But that's a wide range and this is affected by the quality of the RF deck. 

Because the above are too granular, what I'm looking for is something like this: 

'If the S/N is OK, then set the IF gain for -7v with @ 2uv average input to RF balanced connection (with the sig gen impedance matched, of course)'.

Does this sound fairly close?

What I'm 'feeling' is that 2uv might be the sweat spot - this would yield a good balance between DR and S/N.

I do not know how to tell when the DR is being negatively affected.  How can I tell?  I assume this is about keeping distortion to a minimum on strong signals.  Can this be easily measured?

Comments please.

Regards, Larry