6S6B Plate Resistance

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Murdock

Well-known member
Joined
Jan 28, 2015
Messages
844
Location
Germany
Hey,

Im trying to calculate the plate resistance of the 6S6B tube to find the suitable output transformer specs.
Can someone please take a look and tell me, if my math is correct?
I drew some lines on the grid curves



I calculated it as follows: the red line is the load line. 120V B+ with a 100k Ohm plate resistor limits the current to 1.2mA.
Idle state is the green line which is around 37V and 0.8mA.
Then I drew the blue lines which represent a 5V voltage change in each direction.
So for a 10V change in plate voltage the current change is 0.7mA.
10V/0.7mA = ~14k Ohm plate resistance.
This in parallel with the 100k Ohm plate resistor equals around 12.2K Ohm Zout.
Is my math correct? Is that the way to calculate it?
And is it normal, that the more negative you go on the grid the higher the plate resistance gets? Under the same B+ conditions.
On the -2V grid curve I got around 16.5K Ohm.
 
Hi,
you can also put a 1Mohms resistor to load the plate (plate to ground), inject and adjust a 1khz sinus voltage at the grid (via a 100pf cap) to have 1Volt across the 1Mohms , then decrease the value of the load resistor until you have 0,5v, then the value of that resistor is close to your output impedance.
you need a meter with an impedance greater than 1Mohms.
regards
Fred
 
granger.frederic said:
Hi,
you can also put a 1Mohms resistor to load the plate (plate to ground), inject and adjust a 1khz sinus voltage at the grid (via a 100pf cap) to have 1Volt across the 1Mohms , then decrease the value of the load resistor until you have 0,5v, then the value of that resistor is close to your output impedance.
you need a meter with an impedance greater than 1Mohms.
regards
Fred

Thanks alot for the tip!
You wrote "decrease the value of the load resistor". Do you mean the 1M resistor? So use a 1M poti?
So I disconnect the circuit from the transformer and put a 1M poti across plate and ground and decrease it until I have 0.5V across it?
 
Ok, do I have to remove the 100k plate resistor?
And should I inject the 1Khz sinus voltage at the grid with the capsule connected or should I replace it with capacitors?
 
leave the 100k
disconnect the capsule and replace it with a 100pf cap, then inject 1khz before that cap to the grid
the circuit must be powered on
 
I tried two methods and each gave me different results...

I built the C version with cathode bias and with a 1.6k Ohm cathode resistor my voltages are spot on.
- 43V at the plate so 0.73mA current through the 100k Ohm plate resistor with 116V B+
- 56V at the capsule.

Only cathode is at 1.2V instead of 1.6V.

I only have a 500k pot here, but I doubt that is an issue, right?
I disconnected the hot side of the primary and put one side of the pot after the output cap and the other side to ground.
Is that right? Or does the pot have to be right at the plate before the output cap and 100k Ohm resistor?

With these specs and frederics way of measuring it I get around 6k Ohm...


The other method I saw in a post from Dave.
I reconnected the primary to the circuit.
Put an 1k Ohm resistor across the output XLR at the PSU and put the 500k pot in parallel. Adjusted the 1Khz to get around 300mV at the XLR. Then turned down the poti until I had 150mV and measured the poti afterwards. Got around 90 Ohm.

As I have a 10:1 transformer this would mean the plate resistance is about 9k Ohm instead of 6k...
I don't know which method is more accurate. 

Either way, I think a 10:1 ratio is too high for this circuit. Will probably use a 7:1 transformer. 


PS: I didn't use the 5Meg feedback resistor which is said to lower plate resistance a bit. C4 is about 5pF. 
Does C4 modify plate resistance? If yes, does higher value mean higher resistance or the other way around?
 
If you want to be a bit more accurate:
-disconnect transformer
-put 1Mohm resistor load
-put your scope across it
-inject 1khz and increase V input, until you see clipping at the output
-adjust R bias until you have no clipping at your max input voltage
-then measure that Vmax out
-adjust (decrease) R load until you have half Vmax out
-measure R load at 1/2 Vmax out, this value is in the ballpark of Z out

regards
Fred
 
RuudNL said:
You don't have 73 mA through the 100 kOhm resistor...  :p
::)
Forgot the zero... Thanks for pointing that out.

Does anybody know, why the measurements give such different results than what I calculated with the grid curves?
 
Murdock said:
Does anybody know, why the measurements give such different results than what I calculated with the grid curves?

the curves are write with some 6s6b Tube somewhere in a labor in russia. tubes have big tolerances. the curves are a benchmark.
the best way is to write your own curves with a good old tube tester. my measurements with my tube tester are different to the numbers in the datasheet of each tube.

best
Andreas
 
Back
Top