[ARC5] Receiver sensitivity

[Richard Knoppow]

        The older methods of checking for signal to noise are mostly 
useful for seeing if the receiver is working correctly but can not be 
used to compare receivers of different types.  A more modern measure is 
noise factor AKA noise figure. This is a measure of the ratio of the 
output noise to that of a "perfect" or noise-free receiver. The 
procedure is basically to use a noise generator with a known output 
level to set the receiver output level and then measure the output noise 
with a terminating resistor of the same value as the source impedance 
the receiver will operate from. The theoretical thermal agitation noise 
of the resistor can be calculated and the gain of the receiver is known 
from the calibrated noise source. From this the level of the noise from 
the receiver can be compared with the level that would be present if the 
receiver was noise-free.  Because a random noise source is used to 
calibrate the gain the bandwidth and, indeed, shape of the band pass is 
of no consequence so receivers or amplifiers with different bandwidths 
can be compared. A typical vacuum tube communications receiver will have 
a NF of something like 6 to 10 db.
     The NF can be made very low for such applications as radio 
telescopes by cooling the input device to nearly absolute zero. For a HF 
communication receiver the typical input noise from the antenna is far 
above the receiver noise, at least to perhaps 10Mhz. Above that the 
receiver noise becomes more important and at perhaps  50Mhz will dominate.
     One can calculate the "equivalent noise resistance" of tubes and I 
think also semiconductors.  This is a resistance with a thermal 
agitation noise equivalent to the noise from the tube if referred to the 
grid.  Triodes are quieter than pentodes and the noise goes down as the 
transconductance goes up. Multi-grid mixers are probably the noisiest of 
all tubes. The old 6K8 has nearly 300K ENR while a high Gm triode may 
have a couple of hundred ohms. Part of the noise comes from the division 
of currents at the grids. Multi-grid mixers have some virtues but noise 
is not one of them.
On 12/10/2016 3:00 PM, Kenneth G. Gordon wrote:
> On 10 Dec 2016 at 14:38, Richard Knoppow wrote:
>
>>     Another thought that has nothing to do with speakers:
>>     When you speak of sensitivity what exactly are you measuring?
>>     Manufacturers specified different techniques, some of which are
>> useful only for determining if the gain is normal. Sometimes people plug
>> similar tubes into the RF stages and find they make the sets "hot" but
>> never measure signal to noise ratio.
>
> Yes. I have wondered about this for a long time too. One standard is to measure the signal
> vs the signal-plus-noise. I.e., the ratio of the two.
>
> I have never yet investigated this aspect of sensitivity.
>
> I HAVE measured MDS or Minimum Discernible Signal for several of my receivers. But this
> value is somewhat subjective. What exactly does "discernible" mean to you? Are you
> partially deaf as I am? Or do you have more acute hearing? Is "hearing" the signal the same
> as being able to communicate with another station when signals are that weak?
>
> I have always viewed noise generated internally by a receiver as being of far greater
> importance with regard to communications than bare sensitivity alone. This is why I have
> done everything I could with my receivers to minimize internally generated noise.
>
> I hate multi-grid mixers for this reason. "Converters" are inherently superb noise generators.
> Also, the more grids in a tube, the more noise. For instance, the 6/12K8 is the noisiest mixer
> tube listed in Terman, having an equivalent noise resistance of something like over 840,000
> ohms, as I remember it. The 6AC7 has an ENR of something like 2000 ohms in comparison.
>
> Also, all else being equal (which it seldom is) the tube with the higher value of
> transconductance is generally far less noisy than a tube with the lower value of
> transconductance. And this is when the tube is used as either an amplifier or a mixer.
>
> Of course, in most cases, tubes with higher transconductance ratings are also easier to
> overload by a strong input signal, but in practice here, I have seldom had that sort of
> problem.
>
> Ken W7EKB
>
>
> ---
> This email has been checked for viruses by Avast antivirus software.
> https://www.avast.com/antivirus
>
>

-- 
Richard Knoppow
1oldlens1 at ix.netcom.com
WB6KBL