[Premium-Rx] Noise figures on old CEI receivers

[Terry O']

Jeff is correct about the old CEI receivers.  Two of the engineers who 
designed these radios told me they optimized the designs for noise 
figure because there was no other measurable indicator of performance in 
those days.  The customers demanded measurable performance.  One told me 
their optimization extended all the way to the audio stages.  I pointed 
out how the math stacks up to favor optimization starting at the front 
end.  He responded by telling me about a design where in the field, the 
noise in the audio stage interfered with a particular application of a 
radio.  I was still skeptical, but he wouldn't reveal the details 
(Finding out how these radios were used is the hardest part of the 
research I am doing).

A retired WJ test tech told me about some of the factory alignment and 
optimization procedures used on the old UHF tuner designs they called 
the "steel box tuner."  They involve a number of things that were never 
in the manuals.  Some were straightforward, like testing all the 7077 
planar triodes and selecting for lowest noise.  Others were obscure, 
such as their use of two narrow brass straps where a thin copper wire 
would have handled the current with ease.  Part of the factory tuner 
alignment procedure involves prying these strips apart slightly or 
squeezing them together.

ACL was founded because the lead tuner designer for WJ/CEI kept 
insisting transistors were up to snuff for use in tuners but was 
rebuffed by CEI President Ralph Grimm.   So he recruited a handful of 
other engineers and left to found ACL.  ACL receivers were very good but 
the company foundered when one of the partners insisted on an expensive 
expansion into the UK.  A more restrained partner who was a superb 
engineer and the best manager resigned and returned to WJ (he retired as 
a vice-president).  This was not the only incidence of a revolving door 
in this industry.  This is why there is a significant resemblance 
between radios from different companies.

This brings me around to one thing I dislike about the concept of 
premium receivers.  The early companies Jeff listed, CEI, ACL, Nems, 
DEI... all made premium receivers.  The performance of their radios were 
head and shoulders above the consumer market at the time and they 
certainly commanded a premium price.  I don't agree with the definition 
that a premium receiver must have a microprocessor or some digital 
circuitry.  Until DSP left its infancy, the digital stuff was for 
operational convenience and was sometimes included at a cost in performance.

Terry O'


On 5/31/2018 10:33 AM, Jeff Kruth via Premium-Rx wrote:
> Hi: I have a bunch of CEI Tempest receivers that I combined into a rack, BUT . . . they are pretty much deaf compared to more modern equipment.  Because of kTBR noise the ultimate sensitivity depends on how narrow the real bandwidth of the receiver is, so something like the HP 4395A in Spectrum analyzer mode with a true RBW of 1 Hz  .......
>
>   
> _________To answer this: N=kTB depends not only on bandwidth (which traditionally is driven by modulation or data rate) but T which is T-system which is the noise figure (and some other factors) of the system. A spectrum analyzer is a great tool, but in the past was a poor receiver up to 1 Ghz or so, due to lack of preselection below 1 Ghz (which, in a dense signal envirorment would lead to a lot of spurious responses due to intermod) and the usually quite high noise figure (20 dB).__________
>   
> The old CEI receivers had noise figure well under 10 dB, more like 3-5, so they were quite sensitive. Very nice multistage preselection. Only the WJ 9080 was not preselected AFAIK.  Perhaps yours have tired front ends?____________
>   
> No question a modern FFT based analyzer can see very small amounts of energy in a 1 Hz bandwidth. But how much better is it in a 20 KHz bin size? ___________
>   
> YMMV_Jeff__________