[time-nuts] Looking for info on Scientific Atlanta 4647 IF noise test rig

[ed breya]

All this 1/f noise talk got me to thinking about some of my noise 
measurement and generation projects over the years. One project that I 
want to revisit is one for making "nearly-true" random white noise all 
the way down to DC, for comparing and testing things in the very low 
frequency 1/f realm. I have several different noise generator 
build/repair projects in the works that I haven't even looked at in 
years. One in particular is something I had pictured a while back, using 
this Scientific Atlanta 4647 IF tester thing that I got long ago, but 
could not find any info at all. I dug it out from under a pile and 
looked it over. Today I did find a little info on it (a brochure), that 
at least gives some idea of how it's supposed to work. I checked it out 
and it seems to do about what it says.

The brochure is here:

https://www.atecorp.com/atecorp/media/pdfs/data-sheets/scientific-atlanta-4647_brochure.pdf

It's just a few pages, and every other so-called "manual" that shows on 
the downloading sites seems to be that same thing or less.

I'm wondering if anyone is familiar with this or knows of any 
documentation - the ideal of course, being a pdf service manual with 
schematics. A paper manual (supposedly) popped up on ebay, but the 
picture is of an HP binder with unknown contents. I'd prefer to find a 
pdf somewhere rather than spend fifty bucks for something dubious, or 
the hassle of figuring it out.

What's nice about this unit is that it has "calibrated" and settable 
noise and carrier levels, and better yet, it seems to be in working 
order as it is. It makes pretty good (15 dB crest factor) white noise in 
a flat band about 50-90 MHz, that normally gets added to a 70 MHz IF 
carrier, in various proportions and modes for testing. It also makes its 
own 70 MHz carrier, if needed.

What I've been thinking is to simply mix the 70 MHz out with the 
noise-only (mode) out, then LPF to say 20 MHz or less, and there you go 
- a nice DC to whatever white noise source. I'll start with some 
experiments with external stuff, then probably build it into the box as 
another feature. The added items may include an amplifier to get the 70 
MHz up to 7-10 dBm for mixer drive, and the LPF(s), maybe a DC-coupled 
post amp, and some miscellaneous for routing and control. The main thing 
is that the basic box already takes care of most of what's needed, 
including space and power.

The available 50-90 MHz noise is quite stout, specified up to -73 
dBm/Hz. I looked at it on the 8566B, and the noise envelope with 3 MHz 
RBW can almost bury the 0 dBm carrier. Likewise on a scope, the nice 70 
MHz sinewave disappears into a cloud of noise as the power goes up.

The flatness, crest factor, and leveling won't be as good as the 
original signals, but I think still pretty decent. One complication is 
that this thing is all 75 ohm, so I'll probably need to do some mods 
here and there to make it 50.

Ed