Dual PI capacitor input questions

[RockinRocket]

This is a two input long tail PI circuit with .01uf capacitors (Lets use marshalls 18 watt as an example) question. Playing through the normal channel what happens through the other PI capacitor for the tremolo channel. Is signal being grounded from the .01uf cap to the Vol control to ground?

If so shouldn't the lite version of these amps uses a .01uf cap instead of the .1uf? Like how Matchless does there lite versions of the DC30 (lightning, Spitfire) . Seems all the 18 watt lite versions use .1 uf for the PI ground cap. Why is this? Wouldn't a .01uf to ground be a hair closer to the original in tone in this spot?

[tubeswell]

Discussions like this aways go better with a schematic.

A LTP is a differential amplifier. It amplifies the difference(s) between each grid. If there is no signal on one grid, then the 'difference' between the two grids, is the signal on the other grid.

So playing through the normal channel, the input signal is going from the normal channel's vol pot, through a .01uF cap, to one of the LTP grids. The other LTP grid is going to ground through a 500k resistance that happens to be the path between the tremolo channel's vol pot input lug and that pot's ground lug.

[pdf64]

The 0.1uF decoupling cap on the other LTP grid may be seen to be overkill, and potentially result in a slow recovery from heavy overdrive.
Merlin advises equal values on both grids http://www.valvewizard.co.uk/acltp.html whether they're coupling or de-coupling.

[RockinRocket]

Thanks guys and the link was great. Just what I thought to say the least. I don't understand why the 18 watt lite crowd is using a .1uf for the decoupling capacitor for the second grid Cg2. As well as a 8.2k dropping resistor but that's another story.

[martin manning]

The smaller cap doesn't make any real difference unless the other (unused) channel's volume is turned up, which puts some resistance between the cap and ground. In that case it acts like a HF-cut tone control. The -3dB point is up at around 10kHz, though.

[RockinRocket]

The smaller cap doesn't make any real difference unless the other (unused) channel's volume is turned up, which puts some resistance between the cap and ground. In that case it acts like a HF-cut tone control. The -3dB point is up at around 10kHz, though.

Well in this case the 18 watt light is only one channel. So the decoupling cap goes straight to ground. Whether or not there is any real noticeable difference between a .01 uf and a .1uf I may never know. It seems odd to me that all the 18 watt light schematics out there they are all using the .1 uf decoupling when a .01 uf would be a more accurate fit it seems

[pdf64]

The smaller cap doesn't make any real difference unless the other (unused) channel's volume is turned up, which puts some resistance between the cap and ground. In that case it acts like a HF-cut tone control. The -3dB point is up at around 10kHz, though.

I can't work out how that analysis has been made; would you mind expanding on it?

[martin manning]

Brute force... I compared two simulations. The low end is affected by the value of the decoupling cap just as Merlin suggests, but as resistance is added between the decoupling capacitor and ground on the unused input there is a HF shelving effect. If you have any idea where that is coming from, I'm all ears.

[pdf64]

Ha, that's put me on the back foot - can't dispute a sim!
Maybe a manifestation of the Miller effect on the common grid operation of the second triode, causing a phase shift / roll-off at high frequencies and messing up the normal LTP action?

[martin manning]

That's my suspicion too. I sent a query to Merlin about it, so maybe we can learn something here. No simulation is better than the models used in it (which are often idealized and imperfect), or the skill of the user in setting up the boundary conditions, etc. Always question a sim, IMO ;^)