Cell phone interfearance in Matchless amp?

[Tdale]

A question:

The normal Plate/cathode values are usually 100K/1,5 K. Here, the plate is 220K, but the cathode is still 1,5K. Why is that? Why isn't the cathode 3K, since the plate is 220K (about the double of the "standard" 100/1,5 pair.

Is the cathode resistor simply just smaller than normal, or does it have to do with the two plates sharing a resistor? But the cathodes shares one resistor too, so I would have thought that that would even out the required values for plate/cathode.

Did this make sense?

As for rewireing the input.. would that make them work the oposite of how they work now? As I read it, input 1 parallells the resistors and input 2 goes through only one, while the other is grounded?

[FYL]

The normal Plate/cathode values are usually 100K/1,5 K.

Yes, these are common values for a single 12AX7 section. High gain amps use 220K or more, etc.

Here, the plate is 220K, but the cathode is still 1,5K. Why is that?

Maybe Mark Sampson could answer your question. I can't.

Is the cathode resistor simply just smaller than normal, or does it have to do with the two plates sharing a resistor?

Plate and cahode resistors should be halved for parallel sections in order to keep the same working points, ie. Rp = 100K => 50 K, Rk = 1K5 => 750R.

As for rewireing the input.. would that make them work the oposite of how they work now? As I read it, input 1 parallells the resistors and input 2 goes through only one, while the other is grounded?

Stock wiring is à la Fender, with high and low inputs, both inputs use the two sections of V1. Rewiring would allow to use one or two sections with different gains and touch sensitivity.

[Tdale]

How would you wire this alternative input?

[FYL]

Something like this works very nicely:

[Rick]

I'll have him try a new power cord..

Check grounding while you're at it.

Could bad tubes cause this effect?

Hmmm, not impossible but improbable.

Also: I didn't quite get the early post, about 68K resistort, highpass etc. I know the components, but are there other values on the mentioned components, that could be changed, with a better sounding amp as a result?

The Lightning uses a very high gain first stage, you can tame it and get IMO much better sound by changing a few passive components as per the attached schemo.

Gain is reduced with the 100K common plate R - seen as 200K per section, bias is adapted, input grid stoppers are lowered.

The cathode bypass cap as well as the coupling caps can now be modified : 4µ7 will give a more dynamic sound w/o blocking disto with the new gain structure, 2n2 (or 4n7 with single coils) will extend bass response - the amp will still be chimey but in a much smoother style.

You may also rewire the input jax in order to have J1 = single section, lower gain, more laidback sound; J2 = standard parallel sections for V1, Lightning std sound.

I am coincidentally building a custom 30W version of the lightning for a repeat customer, actually it is built and I am tweaking it right now. I decided to go with similar mods of a slightly more middle ground nature with a 1K cathode resistor, 10 uF bypass cap and 2-220K plate load resistors giving 110K in parallel and effectively 220K to each plate, I think that is correct math. I will test it more thoroughly this afternoon. I'm running 2-68K grid stoppers with 34K appearing at each grid with an open jack inserted in the HI input. The low input having the typical 1Meg grid leak. 2-33Ks would, if the input wiring is equivalent, give around 16.5K grid stopping. I need to do the math to see what this does to the low pass and RF attenuation, but the 33K figure doesn't seem too out of line. The little I have tested so far seems to show a fairly noticeable reduction in gain. I went the moderate approach as I don't want to deviate drastically from the original tone/design, but some gain reduction in the 1st stage of a lightning seems to be a good thing and reduces the fizzy, somewhat harsh or electrical sounding preamp distortion to good benefit.

One thing I noticed when doing the 4-EL84 30W version is the small 100 ohm screen grid dropping resistors and only a choke to drop the voltage from the 1st node to the branching node in the PS is not enough to control screen voltage when playing at high volume. When the plate voltage swings low enough, the high sceen voltage seems to draw too much screen grid current and the amp noticeably loses power. Adding 500 ohm resistance to the screen supply corrects this. Probably why the liverpool has 1K ohm screen resistors for all 4 screens.

[FYL]

I need to do the math to see what this does to the low pass and RF attenuation

F3 is around 38 KHz or so unloaded and doesn't fall below 10 KHz with the volume down on a typical guitar.

some gain reduction in the 1st stage of a lightning seems to be a good thing and reduces the fizzy, somewhat harsh or electrical sounding preamp distortion to good benefit.

Taming the beast is a must. Early Matchless amps were quite pooorly designed.

One thing I noticed when doing the 4-EL84 30W version is the small 100 ohm screen grid dropping resistors and only a choke to drop the voltage from the 1st node to the branching node in the PS is not enough to control screen voltage when playing at high volume.

Everybody should use 1K or 1K5 5W screen resistors in CB EL84-based amps, your toobs live much longer and the amp sounds much better.

[Rick]

I need to do the math to see what this does to the low pass and RF attenuation

F3 is around 38 KHz or so unloaded and doesn't fall below 10 KHz with the volume down on a typical guitar.

some gain reduction in the 1st stage of a lightning seems to be a good thing and reduces the fizzy, somewhat harsh or electrical sounding preamp distortion to good benefit.

Taming the beast is a must. Early Matchless amps were quite pooorly designed.

One thing I noticed when doing the 4-EL84 30W version is the small 100 ohm screen grid dropping resistors and only a choke to drop the voltage from the 1st node to the branching node in the PS is not enough to control screen voltage when playing at high volume.

Everybody should use 1K or 1K5 5W screen resistors in CB EL84-based amps, your toobs live much longer and the amp sounds much better.

That is good advice, I usually do have 1K 5w screen resistors and I have like 500 of them I bought a whole big box full cheap and they work fine. I will redo the screen grids and supply and see how it sounds knowing that my toobs will be happier

[Tdale]

What's the difference between using one 220K for both plates or using one 220K for each plate? Each plate sees a 220K resistor in both cases....?

Tommy

[Tdale]

I'll have him try a new power cord..

Check grounding while you're at it.

Could bad tubes cause this effect?

Hmmm, not impossible but improbable.

Probably why the liverpool has 1K ohm screen resistors for all 4 screens.

One 1K for each screen?

[Tdale]

I'll have him try a new power cord..

Check grounding while you're at it.

Could bad tubes cause this effect?

Hmmm, not impossible but improbable.

Probably why the liverpool has 1K ohm screen resistors for all 4 screens.

One 1K for each screen?

Looked at the schem myself...much easier than to ask

[mlp-mx6]

What's the difference between using one 220K for both plates or using one 220K for each plate? Each plate sees a 220K resistor in both cases....?

Tommy

The differences include:
a) the amount of current drawn through that resistor
b) the subsequent voltage drop - higher current means lower plate voltages
c) one resistor means the grids are "mixed" at the plate, and thus the grids affect each other
d) other stuff I'm not thinking of right now

[FYL]

What's the difference between using one 220K for both plates or using one 220K for each plate? Each plate sees a 220K resistor in both cases....?

220K shared = each section sees 220K/2 = 110K. A shared plate R is a plate mixer : signals from both sections develop across it.

[Rick]

What's the difference between using one 220K for both plates or using one 220K for each plate? Each plate sees a 220K resistor in both cases....?

220K shared = each section sees 220K/2 = 110K. A shared plate R is a plate mixer : signals from both sections develop across it.

Preview
Posted: Mon Mar 22, 2010 5:21 pm Post subject:
FYL wrote:
Quote:
What's the difference between using one 220K for both plates or using one 220K for each plate? Each plate sees a 220K resistor in both cases....?



220K shared = each section sees 220K/2 = 110K. A shared plate R is a plate mixer : signals from both sections develop across it.


I see it a bit differently, at quiescence, each triode will be drawing roughly the same current through one resistor and give a bigger voltage drop across that resistor. When each plate has it's own resistor, that same 220K resistor will have roughly half the current since it is only feeding one plate and hence it will have less voltage drop, that's why a 220K for 2 plates is like a 440K for each plate, assuming equal plate currents and suppy voltages. In real life it will vary since no two triodes are identical, or not for long any way. And if they have separate and different valued cathode bias resistors and bypass caps, the intereactions and mixing will be even more complex. Let me know how I screwed this all up ...

Furthermore, to get back to the typical 100K plate load resistor for each plate, you would run something like a 56K or 50K single load resistor for a paralleled pair of plates. That will clean it up quite a bit improving headroom. The actual gain of that stage may go up with higher plate votages, I THINK, but it should run in a more linear portion of the plate curve making the stage itself "cleaner". The higher gain could drive the next stage harder losing linearity there, but that can be "fixed" with plate attenuators or grid stopper and grid leak. This is my supposition, I expect the pros to straighten me out on all this, that way I will learn even more!

[Tdale]

What's the difference between using one 220K for both plates or using one 220K for each plate? Each plate sees a 220K resistor in both cases....?

220K shared = each section sees 220K/2 = 110K. A shared plate R is a plate mixer : signals from both sections develop across it.

For some reason I don't get that... Each plate looks into a 220K resistor...both when there is only one and when there is one for each plate...

Does it matter if the triode is paralleled or not?

[Rick]

What's the difference between using one 220K for both plates or using one 220K for each plate? Each plate sees a 220K resistor in both cases....?

220K shared = each section sees 220K/2 = 110K. A shared plate R is a plate mixer : signals from both sections develop across it.

For some reason I don't get that... Each plate looks into a 220K resistor...both when there is only one and when there is one for each plate...

Does it matter if the triode is paralleled or not?

It absolutely does, in parallel the current from two plates is going through the single resistor, since the current is double, the voltage drop is dropped proportionately per Ohms law, that is why the effective resistance of a single resistor feeding 2 plates is doubled. It's creating more voltage drop than with one plate, so it acts like the resistance is higher than for a single plate. i.e., if both plates are fed each with it's own 220K resistor, in order to replace them with a single resistor, 220k/2 = 110k. So a single 110k resistor gives the same drop (roughly) to two plates as does each having their own 220k plate load resistor. You have to remember, the formula gives the value of the resistance and not the value of some arbitrary resistor already in the circuit. So saying 220k/2 is what the anode sees is not correct, 220k/2 tells you what the single replacement resistor value SHOULD be, namely 110k ohms.