high or low grid leakage?

[conisat]

I am just pondering on my next amp design, and was wondering on the effect of grid leak resistances, I donot like copying amp designs and have so far managed to succesfully design and build 3 amp,s this year all of which have small grid leak values (ranging from 82-220Kohm). Can any one provide information on the advantages and dissadvantages of large values(1-1.5Meg).

[loudthud]

If you dig into the spec sheets for power tubes, there is usually a maximum value given for "Grid #1 circuit resistance", usually different values for fixed and cathode bias. There is probably a reason given deep within the RDH but I always assumed it had something to do with protecting the tube. Preamp tubes (triodes) are a little different. There is a technique called "grid leak bias" where the cathode is grounded and a large grid leak resistor is used, usually 5Meg or higher. Some grid current flows causing a negative voltage to appear on the grid. Because you don't want to short this negative voltage to ground, you need a cap for the input signal. Some of the very early Fender Tweed amps used this type bias on the first stage.

http://www.el34world.com/charts/Sche...DELUXE_5C3.pdf

These amps have a slightly different tone than the normal Fender input stage. I've build two 5E3 variant amps where one input used this type bias. You can hear the difference. I think some Fuchs amps use it also. The penality is noise. The higher the resistance, the more noise. Some Sunn bass amps used 150K on the first tube with a 47K "stopper" giving 197K input impedance. Have fun, experiment, try something new. Daz reported in this thread

http://music-electronics-forum.com/s...ad.php?t=12939

that he used lower than normal grid resistor to knock the gain down between stages and liked the results.

A word of caution: When using high values, use appropriate cap values. The Traynor YBA-1A (5F6A variant) used 4 Meg volume controls with .022 coupling caps. This gives a little more gain on the front end. Most people feel this makes the amp too "spongy" and change to 500K controls. Changing the caps to .002 would probably do the same thing. What happends is the power supply pumping that gets to the preamp tube's B+ gets coupled to the grid of the next stage. Remember that coupling caps and the grid resistor of the following stage form a high pass network.

[Robert M. Martinelli]

Hi,
my advice, FWIW, would be to use more efficient and consistent biasing methods, as the "grid leak" bias has been almost universally abandoned, because it can vary from tube to tube and it can also drift over time. My votes go to the traditional cathode resistor ( plus bypass cap when needed ) or a diodes/LED arrangement.

I understand and appreciate your will to design something different and to keep yourself from copying existing designs, but amp designs don't have to be necessarily complicated or "revolutionary" to sound good...often it's the other way around - The first Marshalls were more or less copies of a 1959 Fender 5F6-A.....

Very often a good ear together with some tweaking of an existing design can lead to spectacularly good results

Just thinking out loud....

Best regards

Bob

[Sonny ReVerb]

I like to fool with my own circuit ideas, as well. However, most tube technology was 'perfected' by the 1950's. Leo didn't invent the wheel

Here's a quote from 'Tube Terminology for Dummies' (no offense):

The grid ("leak") resistor, typically sets the impedance of the stage and biasing. It is interesting because it and the previous stage's plate resistor form a voltage divider on the signal. What this means to you is that the grid leak resistor can be used to control the level into the stage. Low grid leak values will attenuate the signal into the tube stage. If you look at different tube amp schematics, you can see where they control the level into the stage by using different values for the grid leak resistor. There is a maximum value that you need to adhere to. Check the datasheet for the tube you are using to see the typical value of the grid leak resistor.

Here's what Morgan Jones' 'Valve Amplifiers' has to say about grid leaks.

[tubeswell]

The grid leak resistor also affects the input impedance of the stage that it is in front of. The lower the grid leak value, the lower the input impedance, and the heavier the load that the stage presents to the stage that is in front of it. This affects the overall efficiency of the stage through loss of signal voltage. The valve wizard has a downloadable chapter on triode gain stages which has an excellent technical explanation of all this, here:

http://www.freewebs.com/valvewizard1/gainstage.html

[Chuck H]

A point on always trying to design outside the box...

The challenge to the artist is to create something new without creating something strange. It's very easy to be innovative if there are no criteria put on performance. Most players, designers, engineers and artists of any walk have to learn their craft by what has gone before. Otherwise ther is no access to a venue in the first place. Only if you have achived the skills to better your predecessors and contemporaries can you expect to always get satisfactory results by always designing around what has gone before. And maybe not even then. As was stated before, most of the best in tube tech has been exploited. And chances are better that average that even if you use some pre existing elements in your designs they will need to be tweaked away from being an identical circuit before your done. Any knowledge you have on designing tube amps comes from your familiarization with useful circuit designs. Use them. To let any kind of pride get between you and the best designs you can build is just counter productive.

Chuck

[Robert M. Martinelli]

Thanks Chuck,
this is exactly what I was trying to say in my previous post, but, as always, my rather poor English hinders my freedom of expression....

As to the term "grid leak" to which the thread title refers, to avoid confusion, in this case I'd rather talk simply about "grid resistor", which I find more appropriate, as the term "grid leak" in theory books refers to a biasing method, while the "grid resistor" only sets the stage's input impedance and the amount of load imposed to the the previous stage, while bias in preamp stages is set by a cathode resistor ( or by other means, but all cathode-related ).

Regards

Bob

[Steve Conner]

To let any kind of pride get between you and the best designs you can build is just counter productive.

This sentence is kind of self-contradictory, too. If you take pride and ego out of the equation, what's left to drive you to build the best design you can? Designing tube amps is hardly some kind of service to humanity: they're expensive luxury goods made from obsolete technology. It's not like they also purify water in Bangladesh or whatever.

If you asked me to show you a selfless amp design that really serves the needs of the guitar playing public, I'd say some little 15 watt Crate or Valvestate. Chipboard cabinet, TDA2030 power amp, that kind of thing. It couldn't have been much fun to design within the price constraints, but it's probably what the majority of players end up using.

[Robert M. Martinelli]

Hi Steve,

If my understanding of what Chuck was willing to say is correct ( if not, please bear with me ) he doesn't say we don't have to be proud of what we're doing ( as you rightfully say this can indeed be a great "propellant" ), he probably means that the excessive "will" to design/build something unusual "at all costs" can keep one away from his goal instead of helping him to reach it.

As I already said, often a good ear together with good technical knowledge/skills has brought to spectacular results by simply tweaking some existing designs ( just like J. Marshall did back in the early '60s ).

Cheers

Bob

[FUCHSAUDIO]

for the record

I do not use this in any Fuchs amps. Not sure where that came from, but it's incorrect.
af

[mooreamps]

I am just pondering on my next amp design, and was wondering on the effect of grid leak resistances, I donot like copying amp designs and have so far managed to succesfully design and build 3 amp,s this year all of which have small grid leak values (ranging from 82-220Kohm). Can any one provide information on the advantages and dissadvantages of large values(1-1.5Meg).

most amps are designed with a fairly high impedance ; as to interface with the volume pot inside the instrument. I remember when the first Valve Junior come out, with it's 128 K input impedance. One of the first mod's people stated doing, was pull the input resistors and replace with the more traditional values.

A small grid leak would give you a little bit of bias, and thus the player would be able to over-drive the first gain stage much easier ; compared to a very large grid leak. So for clean tones, a large grid leak. For more over-drive tones, a smaller grid leak.



-g

[daz]

I don't know WTF i'm doing 1/2 the time. But for what it's worth, i found that those 100k grid leaks were the long awaited magic bullet i found to get large amount of preamp OD from only 2 tubes after unsuccessfully trying everything else imaginable. Nothing else worked. Either issues with blocking distortion or horrid tone, etc. once those 100k's went in in the place of the voltage dividers that were there before, everything came together and thats when the amp first started sounding right.

[Robert M. Martinelli]

daz,
FWIW I don't know what I'm doing 99% of the time, so I think you definitely have an egde

In your case I think that those 100K grid resistors simply acted as a correct
load for the preceding stage together with setting the following stage's input impedance, thus matching impedance between the two stages; as I told already, in these cases I find more correct to refer to these resistors simply as "grid resistors", because the "grid leak" was a biasing method, now almost universally abandoned.

Cheers

Bob

[BackwardsBoB]

Then what do you call a grid stop resistor?

Imagine a sine wave of fixed amplitude. We go through one asymmetrically clipping triode, giving the peaks of one polarity a haircut, and drive a second triode with the same characteristics, giving the peaks of the other polarity (do to the inversion in the first stage) a haircut. The result looks more like a constant amplitude square wave.

If we modulate the input sine wave amplitude at a moderate frequency (10 Hz?), and send it through the first stage, we'll get an output at the plate that's varying amplitude on one polarity of the signal swing, and fixed amplitude (cllipped) on the other polarity of the signal swing (imagine a child's drawing of a city skyline). It includes the modulation frequency (different buildings are different heights). If we can get an offset version of this signal to the grid of the next tube, we'll get the tops clipped off the tops of the buildings, and wind up with a constant amplitude "square wave". We have to have a grid leak resistor and a coupling cap, and these, being a high-pass filter, can filter out the modulation signal, resulting in a much different waveform, where both polarities of the signal are modulated in amplitude. It's as if someone drove each building exactly half-way into the ground, then exposed the buried part so that you could still se it. Running through the second stage like the first, the previously unclipped polarity of the waveform (the tops of the buildings) gets a haircut to a constant amplitude, but the other polarity (now underground), passes through. Since the other polarity is now modulated (taller building were driven deeper than shorter buildings), the output is modulated, by up to half as much as the input signal, even though both sides of the signal were clipped hard.

Picking and string dynamics have the same effect as our simple modulation. A coupling capacitor/grid leak resistor combo will preserve more dynamics like attack and decay if both are large value, and yield a more compressed tone for smaller values. If both are large, very low frequencies will pass easily, but a low-cut can be performed before any stage that won't be driven into saturation or cutoff.

Note that if you are restricted to a range of grid leak resistor values (and you are), the value of the coupling capacitor can also be adjusted, and that if the associated input is driven into saturation so that there is appreciable grid current, this will charge the blocking capacitor, reducing output until the cap is discharged through the grid leak, so a smaller grid leak provides faster recovery. This is in addition to the obvious changes in interstage filtering and attenuation.

[Robert M. Martinelli]

Then what do you call a grid stop resistor?

Hi Bob,
I'm not questioning the resistor's function, I was simply saying that the term "grid leak" was not appropriate IMHO, because ( even in technical reference books ) this term has been always used to identify a biasing method. Biasing has to do with setting a tube's "center" or "quiescent" point of operation; so, to avoid confusion, I find that "grid stopper resistor" or, more simply, "grid resistor" is more appropriate.

Cheers

Bob