[ARC5] 10 meter BC-454 a bit more

[Kenneth G. Gordon]

On 6 Mar 2013 at 21:55, Richard Knoppow wrote:

> Which two tubes? 

The two we have been talking about.

> If you mean 12SK7 and 12SG7 there is a 
> lot of difference.

Gm for the 12SG7 is almost exactly double that for the 12SK7.

Transconductance is 2000 for the 12SK7, and 4000 for the 12SG7.

It is 4900 for the 12SH7, 9K for the 6AC7, 11K for the 6AG7, and over 13K 
for the 7963 submini dual triode.

>  Pentode tube noise is calculated on the 
> basis of Gm and what is called partition noise, which is 
> dependent on the ratio of the screen and plate voltage.

Yes.

>  It 
> is the partition noise which makes pentodes noisier than 
> triodes of the same Gm.

Yes. The more grids, the more noise, according to Terman.

>  There are formulas for calculating 
> the noise of a tube as an "equivalent noise resistance". 
> This is the thermal or Johnson noise of a perfect resistor 
> at the ambient temperature. Without calculating the 
> partition noise the relative noise will be roughly the ratio 
> of the Gm, which is considerable between these two tubes.

2X. Partition noise is identical according to the tube manuals. 

>     Gain is also a matter of the Gm so a 12SG7 should have 
> more gain than a 12SK7 provided both are biased for maximum 
> gain.

Yes. Provided. In this case, they are biased identically, so I maintain that the 
gain is not substantially different.

One added factor is that the screen voltage for both tubes is 1/2 that listed in 
the tube characteristic manuals: i.e. 75 VDC instead of 150 VDC.

>  The 6SG7 or 12SG7 is a very good RF tube.

Not too shabby, but certainly not perfect either. I wish there was a tube with 
the transconductance of the 6AC7 (9K) or even the 6AG7 (11K), or better 
yet, the 7963 (13K), but which was a remote cutoff pentode type.

> The 
> miniature type 6BA6/12BA6 is similar but has lower 
> interelectrode capacitances making it more suitable for 
> higher frequencies.  RCA gives the limit for the 6SG7 as 18 
> mhz although they were often used at much higher 
> frequencies, at least to 30 mhz, while the 6BA6 is good to 
> above the FM band.

This is one reason I was surprised that the 12SK7 works as well as it does at 
30 MHz.

>     In conventional receivers the noisiest source is usually 
> the mixer. The worst are pentagrid or hexode mixers.

And the 6/12K8 is one of the noisiest mixers available. As I remember it, 
ENR is something like 360K ohms, which is WAAAAAAY up there.

>  The 
> quietest are cathode coupled triodes but they have low gain 
> and often load the RF circuits due to low plate resistance 
> and miller effect.

Yes, but even Collins used straight triode mixers in some of their better 
receviers....like the R-390.

> A cathode coupled pentode is not too bad 
> but these were not often used.

The ARRL Handbooks had several receivers which used a 6AC7 as such a 
mixer. Those receivers worked quite well. I have never seen a commercial 
receiver which had one of those and always wondered why.

> National did it in the HRO.

Another unusually sensitive and quiet receiver.

>     I will have to dig out the pertinent formulas from the 
> Radiotron handbook and calculate the noise of some of these 
> tubes.

Terman's book has those.

>     FWIW, the best criterion of noisiness is noise factor. 
> NF is a measure of the noise of a receiver (or amplifier) 
> compared to the theoretical noise input.

S+N/N ratio is a pretty fair comparison and is fairly easy to do. Ultimate NF is 
not as easily measured with amateur equipment, as I remember it.

Again, I could be wrong.

Ken W7EKB