Why some amps use grid leak as the input stage?

[Alan0354]

I have been studying cascaded frontend schematics lately, Badcat uses a grid leak as the first stage. What is the advantage? From my understanding, the input impedance is quite low on top of input drawing current when signal goes positive at the grid. That will make the guitar sound very muddy. It is known that the output impedance of the pickup rises with frequency, a low impedance load will cause voltage divider effect on the highs.

It is proven if you use 100K pot for volume and tone, you'll reduce a lot of highs of the guitar signal, this is even more obvious for humbuckers that optimized for loading of 250K(500K volume and 500K tone in parallel).

Is the designer purposely use this to lower the highs for the cascade amp?

[Chuck H]

Very few modern amps use grid leak bias for the input stage. The reason is just that they like the way it sounds with their amp. And yes, it's muddy (or "smooth" depending on the connotation you're after). It was mentioned to you in another thread that most modern amps trim lows early on and highs later. Well... Not to take that back, but... An oft used design ideology for "classic" gain amps is to increase mids. A lot of boutique type amps do this to increase sustain and gain without making the amp flabby or harsh. I wouldn't call the Bad Cat a high gain amp. I'd put it right there in that vein of boutique types. Even early Boogie amps qualify. Often times gain is increased and top end bleeders are used to get the mid accentuation for these types of designs. Honestly, I think these amps are starting to go out of favor. An amp that's known for tone should do more than just a fluty LP neck pickup tone. Some players call that rolled back guitar tone into an overdriven amp "woman tone". Well, there was a long era when that was all the rage and many a Carlos wanna be tried anything to make their guitar sound "fat" and "rich". Which is just another way of saying 'lots of mids and less definition.' It's not really a contradictory philosophy for amp design, just another flavor.

JM2C

[uneumann]

I have been studying cascaded frontend schematics lately, Badcat uses a grid leak as the first stage. What is the advantage? From my understanding, the input impedance is quite low on top of input drawing current when signal goes positive at the grid. That will make the guitar sound very muddy. It is known that the output impedance of the pickup rises with frequency, a low impedance load will cause voltage divider effect on the highs.

It is proven if you use 100K pot for volume and tone, you'll reduce a lot of highs of the guitar signal, this is even more obvious for humbuckers that optimized for loading of 250K(500K volume and 500K tone in parallel).

Is the designer purposely use this to lower the highs for the cascade amp?

Well a few points about this. Your conclusions are not quite on target, I believe.
First - the input impedance of a grid leak stage is very high - it has to be >1 M to create a grid leak bias to begin with.
Second, true an overdriven stage (or any sort) will create grid current, and at that point in the waveform, the input impedance will drop. BUT this is true for any stage grid-leak or otherwise.
Third - the lower impedance due to grid current causes the wave to be clipped. As in all stages, a clipped wave is distortion, and creates MORE high frequencies (harmonics), not less.

So - be careful, you're mixing up different ideas and concepts. Clipping due to grid current does not have the same impact as a static lower impedance of a resistor produces.

[Alan0354]

Mesa sound is exactly what I dislike. To me, it's not mid rangy, it does not have any mid and highs!!! You ever put a Mesa in higher gain setting and roll back the volume of the guitar. It is as muffled as if you roll the tone knob all the way down. There is no tone, just mud. Not that I do volume roll off on my guitar to get clean. I don't like that at all.

That's good to know, I was going to try the grid leak and I just have enough question not to. Now I got the answer.

And yes, after you and JMF's insightful posts, I change the design philosophy, it's day and night difference just like that. Now my next quest is how to increase the bass. I think the tone is there, I just need a little more bass.

[uneumann]

Mesa sound is exactly what I dislike. To me, it's not mid rangy, it does not have any mid and highs!!! You ever put a Mesa in higher gain setting and roll back the volume of the guitar. It is as muffled as if you roll the tone knob all the way down. There is no tone, just mud. Not that I do volume roll off on my guitar to get clean. I don't like that at all.

That's good to know, I was going to try the grid leak and I just have enough question not to. Now I got the answer.

You may not want to try a grid-leak stage, but decide for the right reasons. Grid-leak bias has NOTHING to do with tone. It's simply a way to create a bias voltage. The tube doesn't care whether it's biased by grid-leak, or a cathode resistor, or a static DC voltage.

Don't jump to false conclusions -- The fact that Mesa uses a push-pull power stage doesn't mean you don't like or want to use a PP power stage. All amps use similar circuits and stages and their tone differences are rarely due to any one simple factor.

[Alan0354]

I understand the impedance is lower only when the grid conducting current. But the grid is conducting current to develop the voltage. Also as Chuck said and we all know, the normal high output hb easily get 500mV peak. I looked at the vintage strat single coil, I can easily get over 100mV peak when I hit the strings.

I was referring to Mesa because Chuck mention about the high gain amp that has mid heavy characteristics and Mesa is one of them.

Well, what is the obvious characteristics of using grid leak biasing on the first stage? What is the advantage of using it?

[uneumann]

I understand the impedance is lower only when the grid conducting current. But the grid is conducting current to develop the voltage. Also as Chuck said and we all know, the normal high output hb easily get 500mV peak. I looked at the vintage strat single coil, I can easily get over 100mV peak when I hit the strings.

I was referring to Mesa because Chuck mention about the high gain amp that has mid heavy characteristics and Mesa is one of them.

Well, what is the obvious characteristics of using grid leak biasing on the first stage? What is the advantage of using it?

Right - so look at the 1st stage of the Flexi (below)





THe input resistor to ground is 6.8M. That's a high inpedance. The microamps that flow from the grid are just that... tiny microamps of current. If the stage developed 1v bias (about what it does), that means about 0.15 microamps are flowing from the grid. That's just no enough to create the low impedance you're considering. Remember, this current is a different grid current phenomenon than you're thinking about when then grid is pulled positive. In this case the grid is actually moving negative to create the -1v bias Vgk. The cap on the input makes sure no DC current flows in or out of the jack. The only DC path is from the grid through R4 to ground. A grid left alone (or floating) picks up electrons that strike the grid as they fly by on their way to the plate. That's the grid current that causes grid leak bias. The size of the resistor (R4) is what determines the bias voltage. There is no problem getting 1 volt or more with this method. It's no different than using a 1.5K cathode resistor to create 1.4v bias. Note also the flexi stage uses a 100k load resistor. That's also an indication that plate current is about the same as a cathode bias stage.

Now the other kind of grid current is caused when a signal pulls the grid positive. That current will add to the micro-amps used to build bias and then the grid starts conducting more substantial current (milliamps) and its impedance drops quickly - but only for that portion of the wave that causes current to flow.

As for a 100mv or 500mv input - or several volts of input from a pedal. These inputs will pull on the grid exactly the same way, whether its grid leak bias or cathode bias.

[Alan0354]

Ok, so what is the advantage of using the grid leak configuration over the traditional cathode bias? How does it sounds different?

[Chuck H]

Well... For lager input signals the tube is basically biased hot, so the onset of clipping will be early and asymmetrical. This is good for generating 2nd harmonic when the tube is overdriven by a hot humbucker or boost pedal. The input impedance is high, so the effect of impedance rising with frequency is a little more notable when the tube is operated within it's distortion threshold. But as noted, at the onset of distortion square waves tend to sound edgy. Since it's asymmetrical I would expect this to be a sort of smooth, edgy tone. Perhaps it could make for a nice dynamic range dependent on pick attack in the right amp. Nice for some styles but not really right for a metal amp. This is all just speculation since I've never dealt with a grid leak bias input stage. Since THD and Bad Cat have both put their stamp of endorsement on it I think it could be worth experimenting with if your after a blues, hard rock, some pop and indie sort of sound. Probably loads of fun with something like a Gretsch semi hollow loaded with FilterTrons

[uneumann]

Ok, so what is the advantage of using the grid leak configuration over the traditional cathode bias? How does it sounds different?

It doesn't - that's the point. Any difference in sound is not directly due to the method of bias.

One main impact of GL bias is stability and repeatability. Those fractions of microamps that create the bias are a side-effect of the tube physics and manufacture. I suspect tube makers don't monitor that spec tightly so you'll find relatively wide variations in that current between different tube makers and different ages of a tube. The variations in current lead to varied bias. I've noticed considerable variations in tone between the 12ax7s that I put in the Flexi. I suspect I could get the same effects by varying the cathode bias resistor of cathode-bias circuits. The point is its not the GL circuit that causes the tone to change, it's the fact that the GL circuit doesn't set a consistent bias that causes the tone to change. Its a bit like throwing the dice on bias voltage each time you try a tube.

A lesser issue is the high input impedance itself. Unless some care is taken to control the impedance and wire the grid circuit carefully, you may see more noise pickup from surrounding circuits or RF sources. The Flexi uses a 68K series resistor and adds a 22pf cap from plate to grid. Both control HF roll-off and help reduce noise. These are lesser issues with a standard 1M from grid to ground on most stages. The flexi does use a 1M resistor as the pickup load, just like most amps.

An input cap is also needed for a GL stage, but you don't need a bypass cap. Neither really has a tone impact but it's a difference.

So - its not like there is some "grand" difference between GL and other bias methods. They simply set the bias voltage using different circuit methods. There are some side issues that can and do impact tone, but tone variations are due mainly to the lack of consistency and control over bias voltage. IMO - that's why few companies use it anymore. There is little to gain and lots to lose. If someone plugs in a new tube and the tone change is unexpected or unwelcome, the manufacturer has to deal with an unhappy customer. In my case, I don't mind... in fact I like the roll-the-dice aspect. I understand what's happening and why and I just plug in another tube until I'm happy again. I've had the Flexi for a while and I honestly can't find any reason not to like it. The GL stage just adds a bit of unique "character" to the amp.

[Chuck H]

It doesn't - that's the point. Any difference in sound is not directly due to the method of bias.

One main impact of GL bias is stability and repeatability. Those fractions of microamps that create the bias are a side-effect of the tube physics and manufacture. I suspect tube makers don't monitor that spec tightly so you'll find relatively wide variations in that current between different tube makers and different ages of a tube. The variations in current lead to varied bias. I've noticed considerable variations in tone between the 12ax7s that I put in the Flexi. I suspect I could get the same effects by varying the cathode bias resistor of cathode-bias circuits. The point is its not the GL circuit that causes the tone to change, it's the fact that the GL circuit doesn't set a consistent bias that causes the tone to change. Its a bit like throwing the dice on bias voltage each time you try a tube.

A lesser issue is the high input impedance itself. Unless some care is taken to control the impedance and wire the grid circuit carefully, you may see more noise pickup from surrounding circuits or RF sources. The Flexi uses a 68K series resistor and adds a 22pf cap from plate to grid. Both control HF roll-off and help reduce noise. These are lesser issues with a standard 1M from grid to ground on most stages. The flexi does use a 1M resistor as the pickup load, just like most amps.

An input cap is also needed for a GL stage, but you don't need a bypass cap. Neither really has a tone impact but it's a difference.

So - its not like there is some "grand" difference between GL and other bias methods. They simply set the bias voltage using different circuit methods. There are some side issues that can and do impact tone, but tone variations are due mainly to the lack of consistency and control over bias voltage. IMO - that's why few companies use it anymore. There is little to gain and lots to lose. If someone plugs in a new tube and the tone change is unexpected or unwelcome, the manufacturer has to deal with an unhappy customer. In my case, I don't mind... in fact I like the roll-the-dice aspect. I understand what's happening and why and I just plug in another tube until I'm happy again. I've had the Flexi for a while and I honestly can't find any reason not to like it. The GL stage just adds a bit of unique "character" to the amp.

I would contend a lot of this. I don't think owning an amp that happens to have this sort of bias on the input stage is, in itself, knowledge of the circuit or even what it's doing. It is only one part of a whole amp after all. You suggest that the tube doesn't care if it's grid is negative WRT the cathode or if the cathode is positive WRT the grid. At face value that IS accurate. But in this case it's very misleading. A tube that is biased 1.8V positive at the cathode is a very different thing than a tube that's biased .6V negative at the grid. How different? 1.2V of bias! The cathode biased arrangement typically employed generates about three times the bias voltage of grid leak circuits. So the GL tube is biased hot. That means it's going to saturate a good deal before it cuts off. That's a lot of asymmetry compared to a cathode biased triode. Then there's the fact that the cathode doesn't fluctuate, but stays at a steady 0V. With cathode bias the DC voltage jumps up and down due to the cathodes resistance in the current path. This affects the envelope of attack. and compression characteristics. Then, of course there is the huge input impedance of the GL bias. We've already agreed on that part. Does the GL bias triode sound different than a cathode biased triode? I'm certain of it. I've just never experienced it for myself. I've never swam in the Indian ocean either but I know it's wet.

[Alan0354]

So nobody did a direct comparison of the sound yet? So how it is justify using GL bias if it is against all odds for consistency? AND Badcat use it on their production amp!!! I don't imply that it is justified when they use it. There are something kind of funny with other part of their circuit also.

[Chuck H]

I don't think consistency comes into play here! The amps are designed with whatever tubes are chosen by the designer. These are likely to be the tubes that are sanctioned by the manufacturer. The fact that there is inconsistency between tube brands tells more about the nature of contemporary tube production than the proficiency or intent of the engineering of the designs. I think it might be a good idea to have a variable grid leak resistor for these designs so that a consistent and intended bias can be satisfied. Maybe a 2.2M resistor in series with a 2M pot. Simple and effective and now any tube can be made to perform as intended. But players DO like to experiment with different tubes and tout the superiority of one over another based on what happened in their amp. It's all erroneous of course because most modern production tubes don't technically qualify as their designated part number as per the specs. Many even reprint vintage specs for a given design that aren't accurate for the product they make. As designers and players our only options are to experiment and settle or adapt our environment and resources to our needs. Two things humans are known to do.

[Alan0354]

I don't think consistency comes into play here! The amps are designed with whatever tubes are chosen by the designer. These are likely to be the tubes that are sanctioned by the manufacturer. The fact that there is inconsistency between tube brands tells more about the nature of contemporary tube production than the proficiency or intent of the engineering of the designs. I think it might be a good idea to have a variable grid leak resistor for these designs so that a consistent and intended bias can be satisfied. Maybe a 2.2M resistor in series with a 2M pot. Simple and effective and now any tube can be made to perform as intended. But players DO like to experiment with different tubes and tout the superiority of one over another based on what happened in their amp. It's all erroneous of course because most modern production tubes don't technically qualify as their designated part number as per the specs. Many even reprint vintage specs for a given design that aren't accurate for the product they make. As designers and players our only options are to experiment and settle or adapt our environment and resources to our needs. Two things humans are known to do.

This is very interesting. I kept questioning whether different brand of the same tubes really sound better or worst. I heard a lot of people swear by a brand, to be honest, I hear a difference between 6V6 by Ruby vs JJ 6V6S, just slightly different, but I think you can easily adjusted it out.

[soundmasterg]

I've never experimented with grid leak bias in my own projects....not yet anyway. Its on the list of things to try out some day. One of the best blues tones I ever got was with a 1949 Fender Pro....yes it used grid leak bias....I've also heard that a lot of ham radio transmitters used grid leak bias....I've also heard the bit about how it is often noisier and more susceptible to hum than the typical cathode biased stage....knowing the tricks Andy used at THD to make it work are helpful so thanks for sharing!

Greg