Pentode Noise Tests

[PRR]

> maximum gain is at minimum gm

Yes; but Gm only varies 7%. I think it is swings and roundabouts: as one factor changes, another changes, so the net change is small.

> K voltage as near as possible to 1.12V

Where does that magic number come from? Some assumed peak input voltage? (for 1:7? 1:10? 1:20 iron? For far-mike ribbon on harpsichord or dynamic on full orchestra?)

My thought is that Rg2 is not a primary parameter, Vg2 is. Rg2 is useful for self-bias because a hot cathode will suck more Ip and Ig2, the lower Vg2 means lower current, so tube interchange is improved (variability is less). Rp is not a primary parameter, as long as Vp lands in a flat spot of the plate curves and has some swing both ways. (Not a lot of swing because this will surely be followed by another stage; 10V swing is plenty.) And rp has not fallen so far as to reduce gain much.

To go too far afield: Vg1, Vg2, and gain are interrelated. Vg1 will tend to be near Vg2/Mu(g2), and gain will approach Vp/Vg1. For max gain, set Vg2 as low as possible for the required Ip(peak). For large power out (not needed here), set Vg2 higher. I have no idea where along this line might be lowest EIN.

 

[DaveP]

> K voltage as near as possible to 1.12V

Where does that magic number come from?

Sorry, I should have explained that better.  It has no special significance , it was a random voltage that was near 1V on the first test with 1M so I tried to keep it the same for the rest of the test to be a constant and not another variable.

If I have understood you correctly PRR, you are saying that the performance of the tube is voltage  dependent rather than resistance or current.  I thought the gain of a pentode was RL/ Cathode resistance or gm times RL.

I think the current ratios (from the resistor ratios) have a bearing too as these charts show.

6.8 times seems the optimum for 220k

Whereas the best current ratio looks like 5.8

DaveP

 

[abbey road d enfer]

If I have understood you correctly PRR, you are saying that the performance of the tube is voltage  dependent rather than resistance or current.

PRR mentioned the role of Vg2 rather than Ig2. Vg2 acts as a Gm multiplier. The screen grid also has a transconductance attached to it; RF designers designers have frequently used G2 as an input (RF mixers).

I think the current ratios (from the resistor ratios) have a bearing too as these charts show.

I'd rather think it's the Vg2 variations that are put in evidence.

 

[DaveP]

Vg2 acts as a Gm multiplier.

So reducing its voltage, reduces its contribution to gm and noise by making it nearer to a triode perhaps?

DaveP

 

[DaveP]

I modeled this schematic on Duncan's Tonestack programme

Peak at around 2.5k, drops off either side around 40dB, similar to  A weighting curve.

This would be better than nothing I guess

DaveP

 

[CJ]

"in a 'usable circuit' the source impedance (presumed low) will shunt the grid resistor. You really do need to short the grid if you want useful noise measurements of tubes.."

would be interesting to see noise specs for pentodes in other circuit applications, seems like we are limiting our discussion to a first stage with 1:10 transformer, 

 

[emrr]

How many instrument input pentode amps?  Preamps with direct to grid DI inputs and 1M-2M2 grid resistors? 

 

[abbey road d enfer]

I modeled this schematic on Duncan's Tonestack programme

Peak at around 2.5k, drops off either side around 40dB, similar to  A weighting curve.

This would be better than nothing I guess

DaveP

At its peak, this circuit has 14dB attenuation; unless you precede it with some gain, the signal to the meter may be under its intrinsic noise.

 

[abbey road d enfer]

How many instrument input pentode amps?  Preamps with direct to grid DI inputs and 1M-2M2 grid resistors?

The impedance presented by an electric guitar is typically about 100k; that's what the grid sees.

 

[DaveP]

would be interesting to see noise specs for pentodes in other circuit applications, seems like we are limiting our discussion to a first stage with 1:10 transformer,

CJ, I worked these backwards from from the published E.I.N. by subtracting the IPT gain, not sure if it tells us much.

DaveP

 

[emrr]

...so 100K has an observational value...

 

[CJ]

wonder what pentode noise specs would look like in triode mode?

 

[CJ]

favorite guitar input stage of all time?  VOX AC 15, pentode input, 1M grid resistor,

would love to see noise vs screen voltage specs as these stages have a tendency to go microphonic, finding that lowering screen voltage helps to prevent this,

don't really care about actual noise figures, more interested in the  relationship between different values of circuit components/voltages vs noise.

 

[CJ]

another example of a pentode with a 1M grid resistor,  V76, 2nd stage,

 

[CJ]

need noise specs for power pentode phono preamp 

 

[ruffrecords]

wonder what pentode noise specs would look like in triode mode?

Many of them perform really well. The 6AU6 used to be a favourite back in the day. The BBC designed a whole mixer around it (featured somewhere in RDH I think).

Cheers

ian

 

[abbey road d enfer]

...so 100K has an observational value...

For DI boxes and guitar amps, yes. But noise is not as much a concern in these applications, since the signal is quite larger.

 

[abbey road d enfer]

another example of a pentode with a 1M grid resistor,  V76, 2nd stage,

2nd stage is a totally different issue, since the source impedance (i.e. the output impedance of the 1st stage) takes over the grid resistor, and the level is already elevated.

 

[CJ]

has link to DIY filter and amp

https://tavishdesign.com/pages/audio-tube-noise-measurments

 

[abbey road d enfer]

has link to DIY filter and amp

https://tavishdesign.com/pages/audio-tube-noise-measurments

Interesting but "C" weighted, which is intended for low-speed tape recorders. Cannot be used for comparison with theoretical formulae, or takes tedious calculations.
No schemo for the 100x amp.