Function of extra 10k Resistor on the PI

[Gary_S]

On this Marshall 4010 can anyone tell me what the second 10k resistor is for feeding the PI tube?

On watching Gerald Weber's DVD where he's talking about the long tail pair PI the 10k is there feeding the cathode before another resistor but that circuit does not have the second 10k resistor.

So i'm trying to find out what function it performs. I don't know if it's there maybe to decouple this stage from another stage's power supply or not so that there's no interference?

[tubeswell]

You talking about "R23"? That is the lower leg of the voltage divider in the NFB loop. (The nearby 100k (R22) forms the 'upper leg'.  VR6 and C17 form the filter for the presence control.)

[sluckey]

It's part of the voltage divider for the NFB. Plus it provides a dc current path for the PI cathodes.

This is the new and improved presence control. The original presence control had the PI cathode current flowing directly thru the presence pot. This put a dc voltage on that pot and caused a scratchy sound when the pot was turned. This newer presence control provides an alternate dc path for the PI cathodes and the presence tone cap blocks dc from the pot. This eliminates the scratchy sound.

[Gary_S]

Thanks guys that explains it 

As i said, i watched Gerald Weber's DVD that i have and he examines a Bassman schematic. It's the same long tail pair PI on that i think as on the Marshall but there's only the one 10k. So looking at it and trying to square that with the Marshall schematic had me wondering why the extra resistor on the Marshall. 

[Gary_S]

Ah! strange going's on here that requires some thought!!

Today i was thinking about amps as usual and resistors and their values, so i dug out my copy of Michael Doyle's latest 'History of Marshall' book. It has a lot of schematics for different models. Anyway the 2204 schematic is basically exactly like my power amp schematic posted above.

So i look at the schematic as i've done hundreds of times and i notice on the PI the two resistors that this thread asks about are not the same value!!!!

One is the 10k as usual but the other is a 4.7k. IDK why that is, but i assume it's something to do with one of the triode's having more gain than the other? I didn't know this until i watched Gerald Weber's DVD and he explained that the two plate load resistors on the PI are different values because the gain of the first is higher than the second. The RL for the first triode is an 82k the second 100k.

I always thought the 4010 schematic was exactly the same as the 2204 and was surprised, to say the least, when i saw that 4.7k instead of the 10k.

Any idea why the different value?

[sluckey]

I always thought the 4010 schematic was exactly the same as the 2204 and was surprised, to say the least, when i saw that 4.7k instead of the 10k.

What makes you think that R23 is 10K on the 4010 schematic? I can't tell what it is. And the scribbled value does not resemble the 10K that written just above for R20. The value for R23 looks like it was hastily written.

[Gary_S]

I always thought the 4010 schematic was exactly the same as the 2204 and was surprised, to say the least, when i saw that 4.7k instead of the 10k.

What makes you think that R23 is 10K on the 4010 schematic? I can't tell what it is. And the scribbled value does not resemble the 10K that written just above for R20. The value for R23 looks like it was hastily written.

Eh! thanks, appreciate the answer.

What makes me think the value is 10k is because to me that's what it looks like and i don't see what else it vaguely resembles, plus 10k is a standard value for that position in a lot of amps in the PI according to what i've read.

[sluckey]

Wasn't trying to be cute. Just saying I can't call it a 10K from looking at that schematic. Since it's supposed to be identical to the JCM-800 2204 (except in a combo cab???) I certainly would just use a 4.7K. Regardless of the actual value, the function is the same.

Any idea why the different value?

Different engineer or draftsman???

...plus 10k is a standard value for that position in a lot of amps in the PI according to what i've read.

Can you point to some schematics?

[Gary_S]

I know through that Weber video that the 5F6 Bassman has the 10k feeding the PI. I'm sure that there's others as well

[sluckey]

I know through that Weber video that the 5F6 Bassman has the 10k feeding the PI. I'm sure that there's others as well

That ain't the resistor we've been talking about.

[Gary_S]

The 10k in the Bassman is feeding current to the cathode of the PI, and here we go again back to the original question: are the two 10k resistors not performing the same function on the Marshall? That's why i asked why there were 2 10k's instead of the one on the Fender.

[sluckey]

Look at a schematic for the 5F6A and bear in mind that Jim Marshall stole this schematic and built an amp that he called the JTM-45. I want you to look at the PI cathode circuit. Note that the cathodes are connected to a 470Ω, then a 10KΩ, then a 5KΩ Presence pot. Hold that thought, especially the 5K Presence pot...

Fender abandoned the 5F6A circuit and headed in the direction of the blackface amps. But Marshall still liked the basic 5F6A circuit and continued to tweak until he had a very successful amp, the JMP Series 1987 (PLEXI). The 1987 still had the same identical PI circuit as the 5F6A. And the PI cathode current still flowed thru the 5K Presence pot. The Plexi was a very popular amp but it had one shortcoming... The Presence pot had a scratchy sound when you turned it!

The reason for the scratchy pot was because of the dc cathode current flowing thru it. So the Marshall guys scratched their heads looking for a solution. It wasn't really that tough. They just had to come up with a Presence circuit that blocked the cathode current flowing thru the pot. Their solution was to replace the original 5K pot with a 4.7KΩ resistor and put a bigger Presence pot in parallel with that 4.7KΩ resistor. The pot also had the capacitor in series (just like a simple tone control) so that no dc current could flow thru it. The Presence control worked the same and the scratchy sound was eliminated. AND... the PI circuit was unchanged. Look at the schematic for the JCM-800 2204 and you'll see that the cathode of the PI still connects thru a 470Ω, 10KΩ, and a 4.7KΩ (rather than a 5KΩ pot) to ground. No change in PI operation.

So, looking at the 2204 schematic, the 4.7KΩ resistor in the PI cathode circuit replaced the 5KΩ pot, thus keeping the PI the same as the original. Now, the 10K (I still don't believe it's a 10K) in the 4010 circuit does exactly the same as the 4.7KΩ in all the other Marshalls.

I still can't explain the difference between 4.7K and 10K, other than to say I believe it's a handwritten typo. Hopefully you now understand my point of view on this. I'm really out of words for this.

[Gary_S]

Cheers sluckey, that's some informative stuff there.  

You know IDK about that 10k and it has me wondering! it definitely looks like 10k but i notice on other Marshall schematics that it's usually 4.7k in that position on them.

Forgot to mention; GW says in that video that there's a ton of Bassman's used a 6.8k there, even though it's listed as 10k in the schematics all the time for the amp.

I emailed Marshall's tech department tonight and told them i had a 4010 and asked them what size that resistor is supposed to be because i had a schematic and it looked like 10k but wasn't 100% clear.

Hope they reply, they did once before when i asked something!

[darryl]

If you are planning to build a Marshall 4010 replica, you could install a 10k resistor in the R23 position, then temporarily attach a second 10k in parallel with the first. Once you have decided which value you prefer, make that the permanent setting.

[PRR]

Back to grampa 5F6A.

Or even a bit earlier.

In this period Fender was using NFB from speaker-tap to power amp driver.

When picking NFB resistors, you want values higher than the source Z and lower than the NFB-input Z.

Since the source is a 4 or 8 ohm tap, you can use quite low values. Actually limited by how big a resistor you want to pay needlessly for. An 800r resistor will take 1/100th of the speaker power, so for 50W amps you can use a half-Watt resistor (used to be the cheapest kind). You want a ratio of about 10, so the lower resistor is about 80. At the bench they grabbed 820 ohms and 100 ohms for some amps, 820:47 for others, to-taste.

Somewhere in there someone mis-connected the NFB from the off-side grid to the bottom of the long-tail. It was different. They liked it.

Around the same time they tried a capacitor across the bottom NFB resistor to brighten the sound. It works, but with 100r at bottom it needs an awkwardly large value, the full-up tone is too-much for most use, and 100r pots can be rare.

Everything can be scaled. They scaled the 100r up to about 5,000r, and scaled the 820r to 27K, 47K, or so. This 5K in bottom is about as high a value as can be used without hurting the NFB action. But it is high enough to allow use of an economical 0.1uFd 200V cap and fairly-standard pot.

As Steve says, that scratches. One option is to go to 10K fixed, shunted with 10K-25K pot and 0.1u cap. The large NFB resistor may be 47K or 100K.

The NFB resistor ratio is a *to-taste* choice. You can eliminate NFB and let the bass flap, increase NFB until the amp is dry and sterile... depends on your speakers, your music, and your taste. You find all sorts out there, from 2ndG Ampeg VT40 (openback 4*10) with no NFB to early Sunns with heavy NFB (large cabs).

If you eliminate NFB with the "presence pot" plan, it is not the same. This connection has to be a mistake. It does some very complicated things to the amplifier distortion-vs-frequency. Any rational designer would just say "let's wire it right so we know what is happening". Fender ran with it, and 50% of others use it, many of them correctly. If you just have pot and cap, without the resistor back from the speaker-tap, it probably does nothing at all.

Agree: with the speaker you plan to use, try both 5K and 10K. The hand-written value suggests that Jim himself (or his elves) hadn't yet decided when the schematic went to the printers. Or maybe varied depending if they got model X or model Y speakers this month, or as musical tastes changed from Bananarama to Metallica to Wham!