diyAudio (
-   Tubes / Valves (
-   -   Fixed bias design question (

Rob11966 30th April 2009 12:53 AM

Fixed bias design question
This is a simple one I think.

With regards to fixed bias (Triode SE), presuming that the bias supply voltage is appropriate for the tube and application, why do you need a resistor prior to the grid?

I have seen schematics with fully adjustable fixed bias with the appropriate bias voltage applied directly to the grid and other schematics with a 100K + resistor in series.

What is the rationale for the resistor if the exact bias voltage can be applied directly to the grid?

Thanks in advance,


Eli Duttman 30th April 2009 01:06 AM


A grid leak resistor is not mandatory. The secondary winding of an interstage trafo works too. Two things are necessary: an appropriate load must be presented to the driving circuitry and DC must be blocked. The "simple" way to satisfy those requirements is cap. coupling and a grid leak resistor.

Be aware that "fixed" bias usually entails a reduced max. allowable value for the grid leak resistor. There are issues with the tube running away should the grid leak resistance be too big.

Hearinspace 30th April 2009 01:21 AM

Re: Fixed bias design question

Originally posted by Rob11966

What is the rationale for the resistor if the exact bias voltage can be applied directly to the grid?

Remember that bias voltage is net grid to cathode voltage so the bias supply has to be referenced to the audio circuit (ie. cathode) ground. Without sufficient impedance the audio signal being fed to the grid is simply shunted to audio circuit ground via the bias supply. Look again at the schematic and you'll see it.

Rob11966 30th April 2009 03:39 AM

2 Attachment(s)

That makes sense.

So taking the attached schematic by way of example, I have modified for fixed bias. Have I done this correctly? With respect to R1 in the modified schematic, how is the value calculated?

Once again, thanks in advance - these simple problems are slowly advancing my understanding...



Rob11966 30th April 2009 03:41 AM

2 Attachment(s)
and the modified version -

Sorry, R1 as referred to above is my grid stopper resistor.

Funker 30th April 2009 03:53 PM

Hi Rob,
you should insert a grid resistor between you bias supply and the gridstopper. The bias supply is ground for the audio ac, so the output from the driver will shorted to ground without a grid resistor. Keep the grid resistor asl low as possíble, refer to the tube data sheet. As Eli Duttman wrote , a too high grid resistor is a risk of a tube run away .

Here a simplified schematic of a bias supply. The bias voltage should be adjustable by a potmeter for proper settings.

I would leave the circuit as it is, because kathode bias preventing the tube under all circumstances from run away and comensate the aging until the end of its lifetime. Its worth to do so even that you need a higher supply voltage of the amount of the bias supply .

regards from Hamburg


Hearinspace 30th April 2009 05:00 PM

Rob, I wrote this late last night and waking up in front of the computer this morning with it still on the screen unposted. Got to get out the door to work so will post as is. Hope it's useful.
Well, I'm late back from a night out and am not entirely uhhh, stable, so if there are any glaring errors here I hope others will issue corrections. There are also different takes on this process. But in general-

First of all don't confuse a grid stopper with the resistor we're talking about here. They're not there for the same purpose and in the case of a 300B it's not likely you'll need a grid stopper so the schematic should show the capacitor connected directly to the grid and the resistor in series with your bias supply output.

The highest Rg admissible is often cited in the data sheet. If you exceed that value there is the risk that with grid current flow the bias becomes unstable, possibly resulting in what is known as runaway or thermal runaway. Basically the tube eats itself. I tend not to go above half the max. The lower in resistance value you go the less that can be a problem but you increase the demand (loading) on the driving stage, eventually getting to a point (dependent on the stage in question) where that also becomes a problem. Somewhere in between the two extremes are points that work well for the two issues above as well as working well with the value of the coupling capacitor and the circuit's own inter-electrode capacitances to provide a filter that has acceptable time constants/frequency roll-offs.

For Low Frequency characteristics you look at the coupling cap and the grid resistor as a voltage divider on the output of the driving stage. The reactive impedance of the cap forming the upper half , Rg the lower. This means you want to have the capacitor's reactance be small in proportion to Rg at frequencies you want to hear. The ratio of the two determines the amount of reduction. For example, where the coupling capacitor's reactance is equal to Rg the signal will be reduced by half.

For high frequency roll-off the calculation is a little more complicated as you can take the Miller plus grid to cathode capacitive reactance (Xc) of the driven tube added to the plate to cathode Xc of the driver as the lower half of a divider and the driving stage's rp, rL and the Rg all in parallel as the upper. It's also possible to add the Xc of the socket elements. I learned to add the reactance for 10pF.

There are a number of other factors, for example it has been noted that in amplifiers where there are three or more stages sharing a common power supply the size of coupling capacitor has bearing on overall stability and it is considered wise to keep it as small as is reasonable (ie. bearing in mind that as it gets smaller the grid resistor needs to be increased in value).

On the other side, the lower the grid resistor the less susceptible the circuit will be to rfi, electrostatic fields and some sources of hum . . . . . . and blah blah blah.

Remember, there is always a balancing factor to be considered. Another example is that for low frequency extension you want a higher value for Rg but for high values of Rg on peak signals if the coupling cap charges due to grid current and needs to discharge through Rg that it will take a longer time during which distortion is increased.

dizzy yet?

Easy rule of thumb? Here are two different ones - there are others.

Value of Rg to be 10 times the plate impedance of the driver stage then size the cap to suit.


Size Rg (provided of course that it doesn't exceed the Max .) so that in parallel with the plate load of the driver it forms an impedance roughly five times the operating rp of the driver.

Note that in some circuits the differences in sound with changes in value can be enough to change the sound of a circuit from one you like to one you don't and vice versa so once you get an amplifier circuit that you have decided is basically a keeper it really can be worth playing with and fine tuning the values.

Sheldon 30th April 2009 05:53 PM

Hearinspace describes the issues well. I'd add one more tip. When you design your grid supply, the final cap to ground (the one connected to the bottom end of your grid resistor) should be large enough so that the corner of the final supply cap/grid resistor, is well below your lowest frequency of interest. Use a good quality cap here. A good electro is fine, but you may want a film bypass too. If you do this you can assume that the AC path to common is low impedance here.


Originally posted by Funker
I would leave the circuit as it is, because kathode bias preventing the tube under all circumstances from run away and comensate the aging until the end of its lifetime. Its worth to do so even that you need a higher supply voltage of the amount of the bias supply .
I like fixed bias, but for the reason Funker cites, I would not use a regulated supply here, unless the B+ is also regulated. That way, if the line voltage varies, both B+ and your bias supply will shift in the same proportion. This provides some automatic compensation.


Rob11966 1st May 2009 10:07 AM

Thanks guys,

Hearinspace - thanks for the comprehensive reply. A very good effort considering the circumstances of the night!

I am following most now.

Funker, I hear what you are saying and I am sure that you are correct but I want to play around with fixed bias on this one, more as an experiment than anything else.

So it looks like I can connect the cap directly to the grid and only need one resistor between the bias and the grid. Is there a standard value or is there a rule for calculation?

Sheldon, I have seen some designs without the capacitor to ground. I think the simple SE is an example of this. Is it absolutely required?

Thanks again for the replies.


jmbee 1st May 2009 02:35 PM

original spec. from Western Electric for 300B are:
maximum grid circuit resistance for
fixed bias ......... 0.05 megohm
self bias............0.25 megohm
very different!!


All times are GMT. The time now is 03:33 AM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 18.75%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio