About VFK OK But about RKF ... BIG Trouble ?

[kagliostro]

Talking about PS a friend say me that he has an 8V loss on the B+ due to the Heater elevation Volatage Divider

so I told him to revise the values of the partition resistors and he replayed me that he can't because of a statement present on the datasheet of the 12au7 tube

in many datasheet we can find the max value for VFK (volt between filament and cathode)

BUT ...

but only in few datasheet we can find the max value for RKF (resistance between filament and cathode)

The more, he showed me the attached files

So when my friend planned his voltage divider for the elevated heaters he used a 20K resistor between the node that feeds the volage to the heaters ground reference and the ground

I confess I'm a little astonished, I have never read about this necessity 

Is this a wrong interpretation of the datasheet or it is correct and a max 20K resistor can be used 


Franco

[acheld]

Is this a wrong interpretation of the datasheet or it is correct and a max 20K resistor can be used

I'm out of my wheelhouse on this one, but my guess is that the answer is found at the bottom of voltage divider reference, which states that you should not disconnect the cathode without having a resistor between heater and cathode with a max resistance of 20k.   Thus, establishing a guaranteed leak.   

Seems to me that unless you plan on disconnecting the cathode while operating the unit, this would not be needed.   After all, we normally assume a much higher resistance between heater filaments and the cathode.

I'll be interested in other opinions.   

[DummyLoad]

So when my friend planned his voltage divider for the elevated heaters he used a 20K resistor between the node that feeds the volage to the heaters ground reference and the ground. only in few datasheet we can find the max value for RKF (resistance between filament and cathode)

Really only relevant for RF use. Heater-cathode leakage and C_h-k have an effect with RF use, hence the limits with value of R in RF applications.

With AF, you can take much more liberties so use values that are less taxing on the PS.

Talking about PS a friend say me that he has an 8V loss on the B+ due to the Heater elevation Voltage Divider

BTW, I assume that the B+ is a couple hundred volts or more, why is a 8V loss an issue?

--Pete

[PRR]

...I assume that the B+ is a couple hundred volts or more, why is a 8V loss an issue?

The PDF gives 400V.

We normally expect the full load to sag a supply about 10%. Less is expensive and more is too saggy.

8V drop from 3mA is 2,667 ohms power supply impedance, 20% sag. If we figure a "normal" 400V supply for tube amp will drop 40V at 100mA, that's 400 ohms. So why is the supply impedance so high?

I've never seen a 20k limit for tubes in audio service, though 50k may be reasonable.

[sluckey]

only in few datasheet we can find the max value for RKF (resistance between filament and cathode)

Is this a wrong interpretation of the datasheet or it is correct and a max 20K resistor can be used 

Consider this... One of the main reasons to elevate the filaments is to prevent the voltage from exceeding the maximum voltage between the cathode and filaments. The cathodyne, long tail phase inverter, and cathode followers are the usual tube circuits that have a high cathode voltage. Why do they have a high cathode voltage? Because they have a big cathode resistor. For example, the cathode followers in the 5F6A Bassman and many popular Marshall amps have a 100K cathode resistor. That kinda shoots that "20K max resistor" statement to hell, doesn't it? Ask your friend what he thinks about that.

[PRR]

In those circuits, the impedance looking into the cathode(s) is closer to 1K than 100k due to follower action, so the cathode will not get in a lot of trouble.

This may also be a "cover our butt" spec. 99% of tubes may have low H-K leakage, but there's always one in the crate with a cathode scratch or sneeze. In mass production is is wise to design to use "any" tube your distributor supplies. In boutique/DIY work you can usually shuffle V2 V3 etc to put any leaker where it does not matter.

[sluckey]

In those circuits, the impedance looking into the cathode(s) is closer to 1K than 100k due to follower action, so the cathode will not get in a lot of trouble.

I understood that R_kf parameter to be DC resistance.

[kagliostro]

MANY THANKS TO ALL YOU FRIENDS

Now my brain can accept that in our ambitus we have no such strictly restriction

I'll inform my friend

About the 8v of drop I don't remember in wich context he told me it but I don't think it was a big problem

Thanks Again

Franco

[PRR]

I understood that R_kf parameter to be DC resistance.

You can "influence" the cathode with DC if you like. If you have a good grip on the grid, the cathode can't go far away.

[HotBluePlates]

... in few datasheet we can find the max value for RKF (resistance between filament and cathode)
...
So when my friend planned his voltage divider for the elevated heaters he used a 20K resistor between the node that feeds the volage to the heaters ground reference and the ground ...

Really only relevant for RF use. Heater-cathode leakage and C_h-k have an effect with RF use, hence the limits with value of R in RF applications.

With AF, you can take much more liberties so use values that are less taxing on the PS.

I've absolutely encountered preamp tubes that hum due to heater-to-cathode leakage, and I've convinced myself this is why Fender used the crazy-large 250µF cathode bypass in the tweed Bassman amps.  What's more, David Allen contacted me maybe 15 years ago about having a case-load of new 12AX7s that hummed in his 5F6-A-based build unless he used the large bypass cap.  We discovered that it was because most of the case had high heater-to-cathode leakage, and they hummed if they weren't stoutly-bypassed.


I agree the R_fk limit cited is almost certainly about heater-to-cathode leakage:  higher-resistance means small leakage-current develops larger hum-voltage.  But if the user makes a voltage-divider from the B+ they should also tack a bypass cap of some kind across the lower-leg of the divider to kill hum/noise.  The sharp user then thinks about the AC Impedance of the resistor/capacitor combination at the frequency-of-interest, which should now be quite low.


         Paraphrasing a Sergeant-Major I used to work for:
              "You gotta be 10% smarter than the data sheets you work with."
              (said after we had trouble wrangling some piece of mechanical equipment)