Cold clipper switch on Jet City Custom5

[Sansteeth]

Hey folks,
first post here, really exciting! 

I'v been working all afternoon on a Jet City Custom5 that I have lying around, it's 5W amp switchable to 2W in triode mode with two 12ax7 and one 6L6 or El34 in the power section.
Sound is far from great, real fuzzy and muddy stock so I bought the JCM800 mod kit from Epic Tone from before I wanted to actually understand what was going on in an amp.
Anyway, still not satisfied and I had the parts lying around for a cold clipper switch. I thought I'd give it a try before gutting it or selling it as is.
I followed Rob Robinette's wiring for a cold clipper switch on the cathode bias of the second gain stage
(https://robrobinette.com/How_the_Marshall_JCM800_Works.htm#2203_Master_Volume)
I located the cathode pin on my second triode, sure enough I had soldered that 10k cathode bias resistor for that JCM800 mod. I de-soldered it and soldered on a DPTD (on-off-on) switch the 39k resistor in the middle, 12K on one side and 4.7K on the other, giving me 4.1K on one side, 9.03K on the other and 38.8K in the middle, measured across the switch.
However, when I fire it up and play I get more or less the same level of distortion on both sides of the switch (while it should give me a clean-is and Marshall tone) while the cleanest setting is in the middle with the 39K (the Soldano cold clipper).
I'm not sure where to look at this point as the results don't make any sense to me. It's the first time I'm messing around with cathode bias so I'm not even sure what kind of different I should hear between 4k and 9k cathode bias resistor but I'm pretty sure the 39k should make the amp roar while it does the exact opposite now...
Would you folks have any leads? I made a mistake somewhere, but I can't find where!
Thanks a lot!

Disclaimer: This is my first time tracing the signal on a double sided PCB, I cannot find a schematic for this amp.

[sluckey]

I'm not even sure what kind of different I should hear between 4k and 9k cathode bias resistor but I'm pretty sure the 39k should make the amp roar while it does the exact opposite now...

It's behaving exactly as it should. The bigger you make the cathode resistor, the less it's gonna roar. In fact, make it large enough and it won't even purr.

[dude]

I was just reading about that clipper switch, according to Robinette you should hear differences between the clipping, (Marshall 10K vs, 39K of the Soldano), go over the wiring, I would think on a PCB board this mod wound be a little challenge gettin gwired correct. Rob is usually on target, check your wiring again.

[ac427v]

Mine works the same way. I was surprised to find that both cold clipper settings were fairly clean at low to moderate settings of the volume pot.
The second stage with the huge cathode resistor adds very little gain to the signal. The famous Marshall tone comes when I turn the Master Volume down to 2 or 3 and crank the first volume to 9. I replaced the 4.7k with a 2.7k resistor for the 1970 Bosstone fuzz sound. Now that's gain!

[PRR]

4k or 39K cathode resistor is MASSIVE negative feedback, low-low distortion, up to a point.

Past that point it tends to hard-clip. (Like an opamp. So why not use an opamp??) But you may need more front-end gain than a cheap 3-watt amp has to get the many-K cathode stage out of its clean zone.

[Sansteeth]

Hey there,
thanks a lot for the responses,

It's behaving exactly as it should. The bigger you make the cathode resistor, the less it's gonna roar. In fact, make it large enough and it won't even purr.

I'm a bit confused because Rob refers to the 39k cathode resistor as the "Soldano" cold clipper, which means it should have a lot of gain on tap. My understanding was that the value of the cathode resistor becomes so high that it operates near cut-off and clips on "the cold side".
I also read somewhere that very low AND very high cathode resistor value will make the signal clip, either in the troughs or the hills of the signal (hence the symmetrical clipping people are looking for)

I was just reading about that clipper switch, according to Robinette you should hear differences between the clipping, (Marshall 10K vs, 39K of the Soldano), go over the wiring, I would think on a PCB board this mod wound be a little challenge gettin gwired correct. Rob is usually on target, check your wiring again.

Yes I will go over the wiring again tomorrow, now I'm just a bit more confused about the results I'm supposed to get.
I guess my notions a little shaky on the cathode bias resistor business

Mine works the same way. I was surprised to find that both cold clipper settings were fairly clean at low to moderate settings of the volume pot.
The second stage with the huge cathode resistor adds very little gain to the signal. The famous Marshall tone comes when I turn the Master Volume down to 2 or 3 and crank the first volume to 9. I replaced the 4.7k with a 2.7k resistor for the 1970 Bosstone fuzz sound. Now that's gain!

Mmmmh interesting... I would have thought the cathode bias would have an effect on gain no matter where the knobs
are...

4k or 39K cathode resistor is MASSIVE negative feedback, low-low distortion, up to a point.

Past that point it tends to hard-clip. (Like an opamp. So why not use an opamp??) But you may need more front-end gain than a cheap 3-watt amp has to get the many-K cathode stage out of its clean zone.

ahah here's the logic behind this mod: I think this amp sounds bad, I think that short of gutting it and making something entirely different, it's sort of hopeless.
I heard about the cold clipper as being one of the things that make the JMP sound. Having the parts I thought I'd give it a try purely for the learning experience, getting to understand cathode bias better in the gain stages. A JMP has enough gain for me and I have little interest in Soldano's sound, so I'm not looking for a sound, I was just trying to mod an amp, understand the mod, fail to mod it and learn from the failure 
Your mention of negative feedback in this context caught my attention though, how does a high value cathode resistor provoke negative feedback?


I'm going to look some more into it tomorrow hopefully!
Thanks a lot for your help

[ac427v]

Your mod was a great success at showing how uber size cathode resistors affect an amp. You were expecting the opposite result but that's what experimenting is all about. The cold clipper stage does not increase the voltage of the guitar signal (much). It changes the sound of the signal. Marshall and Soldano use that stage along with several other stages that make higher gain (due to smaller cathode resistors) to create their distinctive sound. Your amp is not hopeless. It may have the potential to have much of that sound if you do some other mods too. Hoffman has one Jet City schematic in his library. Is this model similar to yours? If so, it has the basic structure you can use to make a low watt version of the Marshall cold clipper sound. Sleep on it and see if another day shows some fun possibilities.

[66Strat]

The un-bypassed cathode resistor introduces degenerative feedback. It takes more of a voltage swing on the input grid of a triode with an un-bypassed cathode to get the same gain as a like biased triode with a bypassed cathode resistor. The attached might be useful to you.

[Sansteeth]

Hey,
okay so I think I really misunderstood what I was going to get from those mods, and with all the replies I got, it sort of starts to make sense.

Hoffman has one Jet City schematic in his library. Is this model similar to yours?

It does look like it, at least the two first gain stage (I didn't trace the whole thing, I'm getting dizzy looking under and above the dual-sided PCB board at quick intervals  ). Resistor and cap numbers also compute to a certain extent. This confirms the location of the cathode bias resistor. I talked briefly with the person providing those kits and it seems like my wiring is good (I was not so worried about the wiring on the switch itself as it is easy to measure with DMM whether I messed up or not).

@AC427v
Your comment yesterday made me want to try different gain knob/master knob position and indeed while the 39k is just cleaner than the two other settings, I do get different tones for the position 1 and 3 (4k and 9k) if I cook the amp a little (master volume seems to play a bigger role in accentuating those differences than gain knob for some reason).
If I dime the gain knob though, the differences between position 2 and 3 more or less disappear.

I'm really confused though about cathode bias resistor value. Rob Robinette, and the person who sold me the kit, referred to the 39k cathode resistor as the "Soldano" cold clipper. My assumption here being: people looking for what Mike Soldano did with his designs are usually not the kind of demographic looking for a cathode resistor switch that gives them a cleaner tone than stock. However, you folks are telling me that 39k is severely restricting the tube's ability to distort. I mean, it makes sense:
- According to Ohm's law, the higher the resistance of the cathode bias resistor, the higher the voltage drop across that resistor
- The cathode becomes extremely positive in relation to the grid, and the electrons are not to excited about traveling all the way to the plate if the cathode is overly positive
- plate current is low

Please, tell me if my reasoning is flawed, but then I'm not sure why most Fender preamps have 820-1.5k cathode resistors, Marshall have 2.7k, 10k and 820 cathode resistors and a Soldano SLO has a bunch of 1.8k and a 39k.
According to that reasoning, the Fenders should have the most gain, right? What am I missing here? I know all three designs have different amounts of gain stages, but if all else being equal, why would you throw a 39k cathode resistor in a random amp and call it the Soldano cold clipper when, more often than not, it will do the exact opposite of what people might expect?

I still believe I got to go back to the books, because if I had a strong notion of what the cathode bias resistor does, it would make sense straight away. 

Thanks a lot!

[PRR]

Fenders use 100k and 1.5k resistors because that was in the Sylvania tube manual as a "good" operating point.

AFAIK, classic Fender used 820 in only two places: when being cheap and biasing two triodes from one resistor (820 == 1.5k/2), or in the stage before the cathode-follower where a 1.5k led to a too-high interstage voltage.

[ac427v]

Wow, I really admire the effort you are putting into understanding this. And cool that the pico valve schematic is helpful. I don't envy you trying to follow and mod  a circuit on a two sided board.
Page 2 of the pico schematic shows someone's effort to increase the gain of the first stage by adding a 1uf cap parallel to the 1.5k cathode resistor. That cap will greatly increase the amplification of the first stage. When they turn up the gain pot, the signal hitting the second (and following) stage will have greater voltage and result in more distortion. Adding the bypass cap to the cathode increased gain even though the resistor value did not change. The Soldano Slo also increases the value of the plate resistor which increases gain. The number of preamp gainstages in a circuit will affect gain. The attenuation between stages (like the blackface tonestack) affects gain. So comparing the gain of a Fender v Marshall v Soldano requires knowing all of those factors.To get a sense of what the 39k cathode sounds like in a Soldano you could max your gain pot. And use a boost pedal with your guitar. That would kindof copy the effect of Soldano's two stages that come before the cold clipper stage.

[Sansteeth]

Hey,
thank you for bearing with me!
I looked into my notes and somehow, even though somewhere inside my head, I know that the higher the value of the cathode bias resistor, the lower the gain in a power tube bias context, my brain refused to bridge this knowledge of power tube bias with preamp tube bias whereas it's exactly the same operation 
I guess the whole: "You want want Soldano type distortion? Put a 39k bias resistor on the second stage" thing got me doubting everything I knew. But you're right @ac427v, a circuit has to be considered in its entirety, with its extra gain stages, attenuation, plate voltages.
With that being said, I think you folks helped me solve my problem, wiring was good, concept was good, my expectations and my understanding was way off!

While we are still on the subject of cathode bias, if I may ask one more general question about it:

I understand that, with all else being equal, decreasing the value of the cathode bias resistor will make the cathode less positive compared to the grid and that way the plate becomes more attractive for the electrons, increasing current, which increases plate dissipation (since it is the product of plate voltage x plate current) which means more (clean?) headroom.
By the same principle, a high value cathode bias resistor will reduce headroom. But that would then mean that increasing the value of the cathode bias resistor also increases distortion?
For all the explanations I have found, I find the terminology of hot and cold bias really confusing as there seem to be a confusion between what techs mean and what guitar players mean by those terms (same goes with the word GAIN, which means distortion for most guitar players). As a guitar player, I always understood "hot bias" as a bias that runs the tube at their max dissipation, making the power tubes distort at high volumes and a "cold bias" as a high clean headroom kind of set up, but poor in harmonic content (because the tube is not "cooking").
But when I start reading people who know a thing or two about tube amps, a hot bias seem to mean the highest (clean?) headroom and cold bias is a bias that makes the tube distort early. Which in my mind kind of confirms the terminology of a "cold" clipper in a preamp context since I thought it was making the tube operate near cutoff, clipping the bottom part of the signal.
So yeah, I'm confused as heck... (for those who weren't paying attention  )

EDIT: @ac247v adding a bypass capacitor increases the amplification because offers an alternative "easier way" for the electrons to ground, making the cathode bias resistor less efficient at keeping a high voltage drop across it, am I getting that right?

[shooter]

have you read this from Merlin?

http://www.valvewizard.co.uk/Common_Gain_Stage.pdf

[sluckey]

Let it go. Buy a couple books from the Valve Wizard. Read for 6 months.

[ac427v]

You're getting there. Here are my reactions based on my seat of the pants learning:
You're right. That amp will never sound like a Soldano. Not enough tubes. But don't throw out the baby with...The components of the Pico circuit are very similar to a Marshall JCM800. If yours is like the Pico you are 90% rockin. Solder a 1uf cap in parallel to the 1.5k cathode resistor on the first stage. Turn the gain to 10 and you have a raucous music machine. If you are really ambitious set up the first stage exactly like a Marshall with a 2.7k cathode resistor and .68 cathode bypass cap.You're right. Your struggle with understanding the theory is that many amp terms have been so misused that they don't accurately communicate. What does "hot bias" really mean? "cold clipper"? "headroom"?Headroom (maximum clean volume) comes from middle bias. It is the musical equivalent of Goldilocks' porridge. Bias too hot and you get loud distortion. Bias too cold and you get quiet distortion.Then there is "gain". Gain is an increase in voltage in a stage. If the stages that come later in the circuit can handle more voltage that added gain creates more clean volume. If the stages that come later can't handle that much signal voltage they will distort. Confusing?
The experts here have an incredible combination of theoretical training and practical experience. I scratch my head but usually learn something every time they respond to a problem. Thanks for listening.

[2deaf]

Hey.  Have you read that Common Gain Stage thing yet?

[Sansteeth]

Hello there,

Hey.  Have you read that Common Gain Stage thing yet?

Let it go. Buy a couple books from the Valve Wizard. Read for 6 months.

have you read this from Merlin?
http://www.valvewizard.co.uk/Common_Gain_Stage.pdf

Funny you should mention M. Blencowe, I actually have the second edition of "Designing Tube Preamps for Guitar and Bass" which has the Common Gain Stage chapter. I read it up to the biasing sub-section when I was far from home, far from any tube amp, DMM or any tools allowing me to understand the theory with practice.
I had to put down the book at some point as I figured I was punching above my weight reading about things like cutoff and saturation points of a tube without any sense of what such points sounded like and that I should start experimenting and see for myself what does a resistor here, a capacitor there does.
I certainly want to understand the theory behind it (I hope I made that clear with my annoying questions) but at the time, I wasn't sure where I was going with a book on designing preamps when I still hadn't put my soldering iron inside a tube amp.
That being said, this is a great reminder that now that the real world slapped me in the face with my shaky notions of bias, I can go back to the Merlin's book, along with Richard Kuehnel's and Robert Megantz'!

You're getting there. Here are my reactions based on my seat of the pants learning:
You're right. That amp will never sound like a Soldano. Not enough tubes. But don't throw out the baby with...The components of the Pico circuit are very similar to a Marshall JCM800. If yours is like the Pico you are 90% rockin. Solder a 1uf cap in parallel to the 1.5k cathode resistor on the first stage. Turn the gain to 10 and you have a raucous music machine. If you are really ambitious set up the first stage exactly like a Marshall with a 2.7k cathode resistor and .68 cathode bypass cap.

The kit I was sold did make me convert the preamp into a jcm800 preamp (with slight variations, a 2.2k cathode resistor that was stock instead of a 2.7k) but I'm mostly there. I also have negative feedback on the 8ohm output. It still is a mud beast. Works okay on a fender jaguar (single coil) with standard tuning, but my C standard Les Paul does not sound as tight and percussive as it does on my Marshall JMP, surprising uh? 
I will leave it there as I think it should still be about fun, and dual sided PCBs ain't fun. the lack of a PI, a design with an EL34 giving 5W in pentode mode is probably not helping the sound either...

Thanks a lot for your combined help, I have this sweet Guyatone Jazz reverb that has a tremolo that is acting up, and a tremolo channel that is less gainy than the normal channel, I'll come back at a later time to bother you some more with that.

Sansteeth

[alathIN]

For all the explanations I have found, I find the terminology of hot and cold bias really confusing as there seem to be a confusion between what techs mean and what guitar players mean by those terms (same goes with the word GAIN, which means distortion for most guitar players). As a guitar player, I always understood "hot bias" as a bias that runs the tube at their max dissipation, making the power tubes distort at high volumes and a "cold bias" as a high clean headroom kind of set up, but poor in harmonic content (because the tube is not "cooking").
But when I start reading people who know a thing or two about tube amps, a hot bias seem to mean the highest (clean?) headroom and cold bias is a bias that makes the tube distort early. Which in my mind kind of confirms the terminology of a "cold" clipper in a preamp context since I thought it was making the tube operate near cutoff, clipping the bottom part of the signal.
So yeah, I'm confused as heck... (for those who weren't paying attention  )

The way you say it in your second paragraph is exactly how Rob explains the function of a cold clipper.
You are biasing the tube cold with the larger cathode resistor.

Minimum distortion occurs about halfway between cutoff (so cold the electron flow stops) and saturation (max hot).
You can get clipping at either extreme.
People usually think of hotter bias generating more distortion - which it can obviously when you are running the tube near saturation - and also you get more output so you're driving all the other components of the amp a little harder. So yes, you can get distortion that way.

The cold clipper is a different kind of distortion. The cool bias means that the down lobes of the signal get clipped (you're reaching cutoff) but the upward lobes are just getting in to the cleanest part of the range and come through clean. This is why it's "asymmetric" distortion - it's only clipping off the bottom lobes of the sine wave. It can be very overdriven and still sound musical because one side of the signal isn't distorted (or isn't distorted as much).

An important feature of Marshall cold clipper amps is that they use four triodes in the preamp stage and the triode after the cold clipper is biased warm. I think this is what keeps it from sounding thin and tinny. You get the robust closer-to-saturation tones, and drive everything downstream, PLUS the asymmetric "cold clipped" distortion.

I think this is what ac427v was getting at when he said "not enough tubes." The cold clipper is fed by one triode and its output is boosted by two more.

I too have noticed that the higher "Soldano" cathode resistor values just sound thin at lower volume settings. But when I turn that volume knob up, the differences become more obvious. I am not sure if this works as well in a regular preamp with 3 triodes.
Also if you wire your cold clipper switch the way Rob suggests, the middle position is not "medium" cold clipping because it connects the highest value resistor only - the other two positions parallel the 39k with a lower value resistor. So you get a three position switch that goes low - high - medium.

[2deaf]

It doesn't take as much grid-to-cathode voltage to put a tube into cut-off when it is cold biased as compared to center biased, but what it takes to get that grid-to-cathode voltage needs to be examined.

How about some diagrams.

[2deaf]

As you increase the resistance of the cathode resistor, gain decreases, internal plate resistance increases, and the amplification factor decreases.  These things affect what it takes to get the necessary grid-to-cathode voltage for cut-off. 

How about some more diagrams.

   

[2deaf]

So as you increase the cathode cold-clipper resistor, it takes more input voltage to achieve cut-off clipping.  Rock and rollers don't care about no harmonic distortion, they just want signal clipping.  Your amp doesn't amplify the signal enough with the first stage to get any significant clipping with any of the cold-clipper resistors mentioned.

[silverfox]

As you increase the resistance of the cathode resistor, . . . .internal plate resistance increases,

How about some more diagrams.

 

I don't think so. the plate resistance, in my mind, should not change bc you add a resistor to the cathode. Perhaps internal impedance, changes; But I don't believe that either. Could be wrong.


silverfox.

[2deaf]

I don't think so. the plate resistance, in my mind, should not change bc you add a resistor to the cathode. Perhaps internal impedance, changes; But I don't believe that either. Could be wrong.

Did I use the wrong terminology?  I meant r_p here in the States or r_a elsewhere.

[silverfox]

Is the plate resistance value, to the left side, the resistor that is connected to the plate? Or is it an actual change in resistance? The trans-conductance looks like that is what is changing. Likely due to the shift in voltage either towards, or away from the plate, depending on the cathode resistance.


Just asking since I don't know for sure, but it seems to me, the internal resistance of the tube is not going to change depending upon the cathode resistor. Barring frequency influence- But that would be impedance of the plate depending on cathode resistor.


silverfox.

[2deaf]

Is the plate resistance value, to the left side, the resistor that is connected to the plate? Or is it an actual change in resistance? The trans-conductance looks like that is what is changing. Likely due to the shift in voltage either towards, or away from the plate, depending on the cathode resistance.

The scale on the left is the internal plate resistance, r_p .  There are three curves for r_p for three plate voltages, E_b :  100V_dc , 200V_dc and 300V_dc .  The scale on the lower right is transconductance, g_m .  There are three curves for g_m also at plate voltages of 100, 200, and 300V_dc .  The scale at the upper right is the amplification factor, u, and there are the usual three curves for it.

As the cathode resistor increases, the positive voltage at the cathode increases.  The grid bias is the grid relative to the cathode, so the grid volts scale at the bottom becomes more negative as the cathode resistor increases.  You can see what happens to the three parameters as the grid volts become more negative.   

[silverfox]

Is the plate resistance value, to the left side, the resistor that is connected to the plate? Or is it an actual change in resistance? The trans-conductance looks like that is what is changing. Likely due to the shift in voltage either towards, or away from the plate, depending on the cathode resistance.

The scale on the left is the internal plate resistance, r_p .  There are three curves for r_p for three plate voltages, E_b :  100V_dc , 200V_dc and 300V_dc .  The scale on the lower right is transconductance, g_m .  There are three curves for g_m also at plate voltages of 100, 200, and 300V_dc .  The scale at the upper right is the amplification factor, u, and there are the usual three curves for it.

As the cathode resistor increases, the positive voltage at the cathode increases.  The grid bias is the grid relative to the cathode, so the grid volts scale at the bottom becomes more negative as the cathode resistor increases.  You can see what happens to the three parameters as the grid volts become more negative.

Sorry, totally off subject, but, stored energy in tube- fascinating, thank you.

[Sansteeth]

Hey there,
much as been happening on here while I wasn't looking!
Re- reading Designing Tube Preamps for Guitar and Bass's first chapter much so much more sense in the light of what you folks have been telling me.

I will refrain from asking a whole bunch of new questions and try to digest what has been said on here first.

As you increase the resistance of the cathode resistor, gain decreases, internal plate resistance increases, and the amplification factor decreases.   

Thanks a lot 2deaf for your contribution, those diagrams are giving me much food for thoughts, it also makes me see triodes in a slightly different way so it's a little bit of a challenge as I'm not super comfortable yet as to where Ohm's law can be applied and where it can't. I'm also not sure I get why the input voltage is changing from diagram to diagram, is the bias resistor responsible?
I'm also discovering Vgk (grid-to-cathode voltage) which I thought was the same thing as Vg (grid voltage)...

 I can see though that plate resistance increases if the value of the bias resistor increases, because as we get closer to the cutoff point, the grid curve slopes are less and less steep, right? So plate resistance is a "constant" if we take a bunch of points on a straight-ish section of the tube, but as soon as we get in the curvy parts, we witness non-linearity and plate resistance does change, am I looking at it the right way?

So as you increase the cathode cold-clipper resistor, it takes more input voltage to achieve cut-off clipping.

okay, this one is going to take me a minute 

Minimum distortion occurs about halfway between cutoff (so cold the electron flow stops) and saturation (max hot).
You can get clipping at either extreme.
People usually think of hotter bias generating more distortion - which it can obviously when you are running the tube near saturation - and also you get more output so you're driving all the other components of the amp a little harder. So yes, you can get distortion that way.

The cold clipper is a different kind of distortion. The cool bias means that the down lobes of the signal get clipped (you're reaching cutoff) but the upward lobes are just getting in to the cleanest part of the range and come through clean. This is why it's "asymmetric" distortion - it's only clipping off the bottom lobes of the sine wave. It can be very overdriven and still sound musical because one side of the signal isn't distorted (or isn't distorted as much).

Okayyy, so cold bias does make the tube clip, just the bottom part (I'm gonna have to find out why...  ) and a hot bias, when the amp is pushed, will clip both lobes of the sinewave.
Now I'd like to understand what cold-bias clipping sounds like.
The thing that is confusing as heck for me is that, in my mind, if the tube cuts off, there is no current, which means no signal, not "clipped signal, just no signal...

Also if you wire your cold clipper switch the way Rob suggests, the middle position is not "medium" cold clipping because it connects the highest value resistor only - the other two positions parallel the 39k with a lower value resistor. So you get a three position switch that goes low - high - medium.

yes! I've noticed that, it was kind of messing with my OCD as it wasn't wired low-medium-high but I thought it wasn't the time to get creative just yet

thanks again for your contributions, this is pretty much as confusing as it is exciting! 

[sluckey]

The thing that is confusing as heck for me is that, in my mind, if the tube cuts off, there is no current, which means no signal, not "clipped signal, just no signal...

Well, that's true. No signal passes when the tube is cut off. The plate just sits there at a high dc level.

But the grid signal is still changing and as soon as the negative peak rises above the cutoff point, the tube starts conducting again and passing signal. As the grid signal continues to change, going through the positive peak then sliding down the slope toward the negative peak, the tube still conducts, passing signal. UNTIL... the grid once again dips below the cutoff point. Now once again, the tube is cutoff and will remain so until the grid voltage rises above the cutoff point. The plate signal will be flattened on top. Just think, Here come ole flat top, he come groovin' up slowly.

[2deaf]

Clipping is the absence of change over time in an otherwise changing signal.  The signal changes until the tube goes into cut-off and then the output remains constant while time goes on.  Think "Here comes a flattop, he was movin' up with me".

[2deaf]

I'm also not sure I get why the input voltage is changing from diagram to diagram, is the bias resistor responsible?
I'm also discovering Vgk (grid-to-cathode voltage) which I thought was the same thing as Vg (grid voltage)...

The tube responds to the difference in voltage between the grid and the cathode.  V_g on the characteristic curves is the grid voltage relative to the cathode voltage as opposed to just the input voltage.  V_a on the characteristic chart is the voltage on the plate relative to the voltage on the cathode as opposed to the voltage from plate to ground.

As the diagrams show, a signal at the input causes signals at all resistors in series with the tube.  One of those resistors is the cathode resistor and a signal appears across it that is in phase with the input signal.  As the input signal rises, the signal across the cathode resistor also rises.  The grid sees the difference between these two signals and amplifies that.  A larger cathode resistor will develop a larger signal across it so that it will take a larger input signal to get the same grid-to-cathode signal.

The Point of Clip diagrams show the conditions where the tube first goes into cut-off.  Multiple factors come into play at the same time making it difficult to predict what input signal will be necessary just by the seat of your pants.

[shooter]

Multiple factors come into play at the same time making it difficult to predict what input signal will be necessary just by the seat of your pants.

That's why some lab rat with a flat-top invented the oscope 

[2deaf]

I can see though that plate resistance increases if the value of the bias resistor increases, because as we get closer to the cutoff point, the grid curve slopes are less and less steep, right? So plate resistance is a "constant" if we take a bunch of points on a straight-ish section of the tube, but as soon as we get in the curvy parts, we witness non-linearity and plate resistance does change, am I looking at it the right way?

The chart I posted in Reply #22 shows that the "constants" are just about anything but constant, but that's a whole nother matter. 

The internal plate resistance manifests itself as the tangent to one of those V_g curves at a specific point (talking about the chart posted in Reply #29).  Any one of those curves can have the same slope on the tangent at some point, but we aren't operating at just any point.  Once you draw your load line, you can see what point you're at on the curve by the intersection with the load line.  The more horizontal the tangent is, the greater the internal plate resistance.

The 39K cathode resistor puts you right around the -4V curve.  The tangent to the operating point is getting pretty horizontal, which means the internal plate resistance is getting pretty high.

 

[Sansteeth]

Hi,
do not mean to bump up a thread about something that has been "solved", but I hate to abandon a thread without at least saying thanks to the people who took time to reply.

But the grid signal is still changing and as soon as the negative peak rises above the cutoff point, the tube starts conducting again and passing signal. As the grid signal continues to change, going through the positive peak then sliding down the slope toward the negative peak, the tube still conducts, passing signal. UNTIL... the grid once again dips below the cutoff point. Now once again, the tube is cutoff and will remain so until the grid voltage rises above the cutoff point. The plate signal will be flattened on top. Just think, Here come ole flat top, he come groovin' up slowly.

Clipping is the absence of change over time in an otherwise changing signal.  The signal changes until the tube goes into cut-off and then the output remains constant while time goes on.  Think "Here comes a flattop, he was movin' up with me".

Rather than "cut-off means no signal", I was assuming "cut-off means no sound", but that's because of my poor understanding of what a waveform actually represents. DC means no sound, but AC, no matter how clipped or jagged in the tops and troughs, is sound, or am I missing the mark again?

Multiple factors come into play at the same time making it difficult to predict what input signal will be necessary just by the seat of your pants.

That's why some lab rat with a flat-top invented the oscope 

Yeah, figured I might look at what that cold clipper does through my oscilloscope (signal generator at the amp input, Oscope at the speaker output, with a dummy load), and sure enough the 39k does clip the bottom of the wave earlier than the others, even though it sound less distorted than the other two positions that clip the bottom later, but always in conjunction with clipping the top around the same time (even though the clipping looks different, bottom is flat, top is jagged).

The tube responds to the difference in voltage between the grid and the cathode.  V_g on the characteristic curves is the grid voltage relative to the cathode voltage as opposed to just the input voltage.  V_a on the characteristic chart is the voltage on the plate relative to the voltage on the cathode as opposed to the voltage from plate to ground.

Oh wow, so Vp on the chart is not the value of the B+ you feed the plate with (what you would call plate-to-ground voltage I guess) but it is the differential between cathode and plate? 

I can see though that plate resistance increases if the value of the bias resistor increases, because as we get closer to the cutoff point, the grid curve slopes are less and less steep, right? So plate resistance is a "constant" if we take a bunch of points on a straight-ish section of the tube, but as soon as we get in the curvy parts, we witness non-linearity and plate resistance does change, am I looking at it the right way?

The chart I posted in Reply #22 shows that the "constants" are just about anything but constant, but that's a whole nother matter. 

The internal plate resistance manifests itself as the tangent to one of those V_g curves at a specific point (talking about the chart posted in Reply #29).  Any one of those curves can have the same slope on the tangent at some point, but we aren't operating at just any point.  Once you draw your load line, you can see what point you're at on the curve by the intersection with the load line.  The more horizontal the tangent is, the greater the internal plate resistance.

The 39K cathode resistor puts you right around the -4V curve.  The tangent to the operating point is getting pretty horizontal, which means the internal plate resistance is getting pretty high. 

yessss, I got that right! 

[2deaf]

Oh wow, so Vp on the chart is not the value of the B+ you feed the plate with (what you would call plate-to-ground voltage I guess) but it is the differential between cathode and plate? 

B+ isn't the plate-to-ground voltage for the gain stage you're talking about.  B+ is usually the positive DC voltage at the first power supply node after the rectifier, but people frequently use B+ to mean any positive DC supply voltage at any power supply node.  There are three positive DC voltages involved here:  The supply voltage, the plate voltage, and the cathode voltage.  V_p or V_a on the chart is the difference between the plate voltage and the cathode voltage.  The cathode voltage is very small compared to the plate voltage in your case, so plate-to-cathode is close enough to plate-to-ground for the thickness of the lines you're going to draw.

B+ comes from the use of batteries to supply voltages for tubes.  The "A" battery was for the heaters, the "B" battery was for the positive supply, and the "C" battery was for the negative grid bias voltage. 

[alerich]

I'm a bit confused because Rob refers to the 39k cathode resistor as the "Soldano" cold clipper, which means it should have a lot of gain on tap. My understanding was that the value of the cathode resistor becomes so high that it operates near cut-off and clips on "the cold side". I also read somewhere that very low AND very high cathode resistor value will make the signal clip, either in the troughs or the hills of the signal (hence the symmetrical clipping people are looking for)

I always understood that the original 10K cold bias stage in the late 70s Marshalls was an attempt to suppress the gain of the preamp after the first two stages had been cascaded in the original master volume design. That's why people will reduce that resistor a bit to add more gain to the preamp. In an interview I read with Mike Soldano he said that the original design clipped in a very asymmetrical manner. He said he experimented with different values and settled on 39K as the value that yielded more symmetrical clipping that he was looking for.

[GlideOn]

I read through this and can't believe the most obvious thing wasn't answered.

Soldano 39k Cold Clipper works because it has an additional gain stage (4 total) driving the preamp.

Marshall does not (only 3).

It's like saying two cars both have V6 Engines, but one makes much more horsepower and torque because it has a turbocharger.

Beyond the surface similarities, they are very different amps.

IMHO, don't think the 39k is as significant a factor as the other dead-center bias 1.8k cathode biases I see in V1, V2 and V3 gain stages. That's far more Fender than Marshall. This makes for a more percussive, somewhat more HiFi, clinical "symmetrical" clipping preamp tone compared to a Marshall, but with additional gain stages it makes it very fluid and defined. Very clever, but very different approach vs Marshall.

To not totally 'speed' into a muddy mess at high volume from the additional gain stage, once of the cathodes is scaled way back. Hence the 39k V2 (V3 in Soldano) "Cold Clipper" cathode. Maybe amounts to about 5% of the overall tone. Not a huge deal.

In layman's terms: If you're gonna give a car more horsepower, you also need to give it better brakes and tires. That's pretty much what Soldano is doing.

We get fixated and mystified on a single value component when there's so much else to understand about the sum of the parts working together. I've been there too.

IMHO, the Marshall DSLs, JVMs are really its' contemporaries, not the JMPs or JCMs.

If you really want to have that fluid gain tone AND classic Marshall character, check out the HotMod EVO V2 made by Legendary tones. Simply install into the V2 preamp tubes slot and you gain an additional gain stage, just like (mostly like) the Soldano! No internal modding or soldering required.

Ironically, Soldano was also the first to release an add-on tube gain stage but was far more expensive and utilized a hard to find tube.