Grid Resistor Value Determinations

General discussion area for tube amps.

Moderators: pompeiisneaks, Colossal

Post Reply
SilverFox
Posts: 222
Joined: Fri Feb 10, 2012 6:03 pm

Grid Resistor Value Determinations

Post by SilverFox »

How does one determine what value grid resistor should be used on the output tubes? For example 6L6 vs. KT88? Is it the plate voltage or the tube characteristic?

silverfox.
pdf64
Posts: 2688
Joined: Sat Mar 12, 2011 9:23 pm
Location: Staffordshire, UK

Re: Grid Resistor Value Determinations

Post by pdf64 »

Control or screen grid?
If control grid, the grid stopper or grid leak / referencing resistor?
If the latter, it's generally specified as a limiting value on the manufacturer's info, eg http://www.mif.pg.gda.pl/homepages/fran ... k/KT88.pdf
So Rg-k limit is 220k in cathode bias, 100k in fixed bias.

Regarding the screen grid stopper resistor, Merlin outlines the process http://www.valvewizard.co.uk/se.html
Last edited by pdf64 on Tue May 10, 2016 3:13 pm, edited 1 time in total.
User avatar
martin manning
Posts: 13081
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: Grid Resistor Value Determinations

Post by martin manning »

The value depends on the tube type, and you can find the recommended not-to-exceed value in the data sheets. Many guitar amp designs have used higher than recommended values without issues, however.
SilverFox
Posts: 222
Joined: Fri Feb 10, 2012 6:03 pm

I'll look at the information you posted

Post by SilverFox »

What these questions relate to is the changing of output tubes from KT88 to 6L6. I'm trying to determine if I need to change the grid stoppers to somehow match the 6L6 output tubes.

Another question- If there is no substantial current flowing in the bias circuit, how do different value bias feed resistors affect the bias on the output tubes? For example, 220k vs 100k on the same tubes in question. The KT88 have 220k and 6L6' bias circuits I remember use about 100k.

I'll take a look at the recommended readings, and thanks.

silverfox.
tubeswell
Posts: 2337
Joined: Thu Mar 27, 2008 6:42 am
Location: Wellington. NZ

Re: I'll look at the information you posted

Post by tubeswell »

SilverFox wrote:...
Another question- If there is no substantial current flowing in the bias circuit, how do different value bias feed resistors affect the bias on the output tubes? For example, 220k vs 100k on the same tubes in question. The KT88 have 220k and 6L6' bias circuits I remember use about 100k...
Grids are (practically) open circuit if they have enough grid resistance. But a path is needed to replenish electrons that get displaced (as a result of the heat experienced at the grid, which occurs from being in close proximity to the cathode). The more tube current there is, the more likely it is that there will be grid current, therefore faster replenishment of electrons to the grid is needed in order to ensure the tube doesn't loose (too much) bias voltage. Bigger tubes typically have more tube current, and therefore require less grid resistance to maintain stable operation.
He who dies with the most tubes... wins
SilverFox
Posts: 222
Joined: Fri Feb 10, 2012 6:03 pm

Most informative

Post by SilverFox »

So then it sounds like the wrong grid resistor could cause distortion and/or red plating depending upon which way the error goes. Too high a value, overheating, too low, distortion?

silverfox.
User avatar
jjman
Posts: 753
Joined: Sat Oct 06, 2007 2:33 pm
Location: Central NJ USA

Re: I'll look at the information you posted

Post by jjman »

SilverFox wrote:............Another question- If there is no substantial current flowing in the bias circuit, how do different value bias feed resistors affect the bias on the output tubes? ..............silverfox.
The flow of current in the bias circuit does not flow thru the grid / bias feed resistors. The bias circuit's current flows from the tranny to a diode to resistors to ground. Voltage is on (both sides) of the bias feed resistors but no current flows thru them. Like testing a battery for voltage, while there is no current.

All of the current of the signal flows from the input cap(s) thru the grid (leak) resistors (aka bias feed) to the ground in the bias circuit. During distortion from overdriving the grid, some signal current flows thru the stoppers and into the grid.
If it says "Vintage" on it, -it isn't.
tubeswell
Posts: 2337
Joined: Thu Mar 27, 2008 6:42 am
Location: Wellington. NZ

Re: Most informative

Post by tubeswell »

SilverFox wrote:So then it sounds like the wrong grid resistor could cause distortion and/or red plating depending upon which way the error goes. Too high a value, overheating, too low, distortion?

silverfox.
Yes too-high-a-resistance causes build up of positive grid charge, resulting in loss of bias and eventual thermal runaway - depending on how much charge builds up c/f gets neutralised. Assuming an appropriate bias voltage is maintained, then too-low-a-resistance attenuates the signal, resulting in lowering of signal output.
He who dies with the most tubes... wins
pdf64
Posts: 2688
Joined: Sat Mar 12, 2011 9:23 pm
Location: Staffordshire, UK

Re: Grid Resistor Value Determinations

Post by pdf64 »

Yes, if the grid leak value is lowered too far then due to this heavy loading, the phase splitter may hit the rails and overdrive before full power is achieved.
That's the principal on which the crossline type 3 master volume is based.
User avatar
trobbins
Posts: 194
Joined: Tue Jun 01, 2010 8:51 am
Location: Australia

Re: Grid Resistor Value Determinations

Post by trobbins »

The grid stopper normally provides a minimum resistance at higher frequencies, that deters the chance of feedback causing oscillation. That function usually warrants at least a few hundred ohm, and 1k is usually a practical minimum used by commercial amps.

Grid current starts to noticeably flow as Vgk approaches 0V - there are a few contributions for grid current leading up to about 0V, but then 'diode' like current flow starts to dominate and grid current has to increase substantially to push Vgk higher than 0V.

For a capacitor coupled driver, any such grid current flow during the peaks of signal waveforms causes a shift in the average voltage across the driver stage coupling capacitor, and that is the mechanism that leads to blocking distortion. A higher value of grid stopper softens the level of grid current during signal peaks, introducing a form of peak compression, and a lessening of blocking distortion level.

The grid stopper plus the grid leak resistor form the total resistance seen by the static level of grid current wanting to flow due to internal grid behaviour. This behaviour normally causes only a minute grid current level, but a valve suffering from higher internal gas density (eg. from outgassing when the tube gets overheated or from seal leaking, and the getter hasn't soaked excess gas up; or from bits of cathode oxide that accumulate on the grid wire) can cause increased current flow. The max rated grid resistance is a value that normally restricts the rise above 0V in grid voltage to just a small value as a valve normally ages. Similar to a coupling capacitor that starts to leak more, the grid voltage can creep up in value, which makes the bias voltage a bit hotter and internal dissipation level rises, which then causes even more grid current, and that all causes a walk-away/run-away failure of the valve.

The 6L6 or KT's require 100k grid resistance max, although the designers allow that to go to 120k due to resistor tolerance, so nowadays with better resistors that means 120k is reasonable.

The grid leak value is a major signal loading on the driver stage, so lowering the leak value in order to raise the stopper value causes a drop in amplification. As most amps don't go above 10k grid stoppers. then loss of signal gain is not such a big concern (but for 6550 it is a bigger issue).

There are other issues if you were trying to maximise high frequency performance for hi-fi, but that is probably not a worry for you.
User avatar
martin manning
Posts: 13081
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: Grid Resistor Value Determinations

Post by martin manning »

trobbins wrote:The 6L6 or KT's require 100k grid resistance max, although the designers allow that to go to 120k due to resistor tolerance, so nowadays with better resistors that means 120k is reasonable.
I haven't seen this mentioned before. Do you have a reference?
User avatar
trobbins
Posts: 194
Joined: Tue Jun 01, 2010 8:51 am
Location: Australia

Re: Grid Resistor Value Determinations

Post by trobbins »

Interestingly I hadn't seen or appreciated that before either - but I saw it explicitly stated in the Genalex KT88 March 1959 Issue 2 datasheet. It would have to be implicitly at least 10%, and given the era then it's prettty likely that all datasheets would have been basing recommendations on the nominal use of 20% resistors.

Links I just located are:
http://www.newsensor.com/pdf/genalex/kt88-genalex.pdf
https://frank.pocnet.net/sheets/084/k/KT88.pdf
Post Reply