[Collins] Collins 62S-1 Transverter Downconversion Gain??

[Dr. Gerald N. Johnson]

I've not measured my converters that way. Way back in the 60s when I 
moved to Texas for a couple years with my new 75S-3B and a home made 
nuvistor converter and put together a 2m transmitter, I was working 
locals around Dallas and bragging about the line noise they were 
complaining about. One night one of them drove up to Allen from Garland 
and dropped off his 5722 noise generator. I soon found that I had a NF 
round about 15 dB. I worked over the converter getting it down to 3 dB 
or better (1 ma at 50 ohms), then I could hear power line noise from two 
or three different power systems some in phase with my local line and 
some 30 degrees away (no doubt the feed to my substation which probably 
had a Y primary, delta secondary transformer to give that 30 degree 
shift). And with the scope I could see noise pulses for each peak of 
both 3 phase systems.

One of those hams, W5WXV ran considerable power and good antennas (held 
a nightly sked with a station in Qunicy Illinois about 900 miles 
successful most nights) and his signal was all over the bottom 200 kHz 
of the band (no CW subband those days, he was crystal controlled on 
144.085). So with my noise generator (I found an AIL noise head, 
probably at Collins Surplus) I started adding attenuation between the 
converter and the IF but backed off when I started to affect the noise 
figure of the system. I also spent much time in the receiver and found 
its signal handling limit was the second mixer so I connected the second 
mixer grid to a tap on the IF coil and changed the mixer tube from a 
6EA8 to a 6688 or 7788 (I forget which, need a new one probably now), 
from a tube with a gm of 5500 to a tube with a gm of 50,000 which 
increased the mixer gain and reduced the mixer noise significantly.

Before the two mods, Al's signal was 20 or 30 places in 200 kHz, after 
just three, one pegging the S-meter on .085 and the other two at the 
noise level.

Some one in the area had a 62S-1 and was playing with PLL for the LO for 
better precision but it wasn't working well, he drifted around more than 
my crystals did. I don't remember him being on for band openings. I 
heard decently, better than I was heard with my 10 watts but Al heard 
much better partly because he could copy way down in the noise by ear.

According to a standard chart of receiver sensitivty vs NF and IF 
bandwidth, 4 dB and 2.1 kHz corresponds to a signal of about -138 dBm. 
Half a microvolt at 50 ohms is -113 dBm. 1.2 uv is -105 dBm. So if those 
give 10 dB S/N roughly 10 dB less would be threshold, or sensitivity, so 
-115 dBm and 6 kHz (AM selectivity) is 20 dB NF, -123 dBm is about 13 dB 
NF. Seems bad, but calaculations with low S/N get messy. I don't think a 
75S-3 for HF should have that good a NF, its not needed and not of 
benefit. I know mine does poorer after I cut the signal to the second 
mixer so much.

On 9/20/2010 6:25 PM, Pete wrote:
> Jerry,
> The pertent 62S-1 manual info is:
> 	1. RX Signal path uses a 6ER5 Triode RF Amp followed by another 6ER5
> triode as a Mixer.
> 	2. Noise Figure approximately 4 dB
> 	3. System performance is such that: 1.2uV at 2M or 6M yields 10 dB
> S+N/N on AM Signals through a 75S Receiver
> (NOTE: the 75S3 receiver manual spec is 0.5uV yields 10 dB S+N/N)

Since SSB is the mode the S-line was made for shouldn't that be in a 2.1 
kHz passband for SSB, not AM.

If the NF of the 62S-1 was made with an automated HP NF meter that meter 
takes out the factor of the IF NF and the RF/mixer gain and shows only 
the NF of the RF stage along. Its not the system NF. And I'm of the 
opinion that the 75S will have a NF more like 8 or 10 dB (still quieter 
than most HF antennas) and the front end gain in the 62S-1 is not enough 
to overcome that. Though my 6CW4 nuvistor convertor with one RF stage 
(grid driven shunt neutralized) and another 6CW4 as mixer had enough 
gain to let it set the system NF nicely with room for 10 or 12 dB 
attenuation at the IF (IIRC) with my unmodified 75S-3B.
>
> This information tells me the NF of the cascaded 62S-1/75S3 system is worse
> than the noise figure of either the 62S-1 and the 75S.
> There has to be a net translation loss from VHF to HF in order to match this
> specification.  I'm calculating there should be -8dB conversion loss based
> on these specifications in the manuals.  Interestingly I'm measuring -12 dB
> conversion loss.  But, I'd like some confirmation on what it should be.
> What do you think?
> Pete
> K5PZ
>
Not so much conversion loss as not enough conversion gain for the 2m RF 
section NF to make up for the generally poorer NF of the 75S-3.

NF of a cascade amplifier is NF of the RF stage PLUS the NF of the 
second stage divided by the gain of the RF stage. That is the Noise 
Factor (not figure) of the system is the Noise Factor of the RF stage 
plus precisely the Noise factor of the second stage - 1 divided by the 
gain of the RF stage. So the system NF is never as good as the RF stage 
unless the second stage NF is the same or better than the RF stage, or 
the RF stage has infinite gain. But if the second stage NF is as good as 
that of the RF stage, then the third stage NF is significant unless that 
third stage NF is as good or the gain of the second stage is infinite.
>
>
>
73, Jerry, K0CQ, Technical Adviser for the Collins Radio Association.