The PPIMV (Post-PI-Master-Volume) thread.

[Sazabi]

Well, there 2 questions in my post.
About Lar/Mar wiring and about my problem with Rich mod.
And, actually, it's still not clear for me. If wiring is same (except for parts nominals), than why Rich Mod has caps lugs direct connection to valves and Lar/Mar doesn't (I've circled questioning point on picture).

Looking at Ceriatone wiring diagram, there's no direct connection from 0.022uF caps (in my case those are 0.1uF) to valve pin. And this is Lar/Mar mod.
But Rich mod does have such connection.
My wiring is just the same as on attached picture. Checked 3 times. Can wrong nominals affect bias? I haven't checked dual pot readings.

[CoffeeTones]

The .022uF - 0.1uF cap connections do not go to the power tubes. The power tube connections are not drawn in the Rich / Metro layout at all, it has been implied and assumed that the builder knows, the power tube connections have been left in the standard location, which is connected to the 220k resistors. It's drawn in a confusing way, with disregard to novice use, very likely to separate it from the LarMar and facilitate leaving the 220k resistors on the board. Think outside the box and draw on paper, both the LarMar and Rich / Metro, in a way you understand and it will become clear to you.

[Sazabi]

I guess my English isn't good enough to understand me correctly:). Because you sad just what I did already. In stock 0.1uF caps connects to one side of 220kOm resistors and that side has connection to valve pin. In Rich mod ppimv connection from resistor to valve remains and connection from cap goes through 500kOm pot.
In Lar/Mar ppimv caps remains on stock position, but resistors are removed. More of that, connection from caps to valves removed also (red circle on my picture). I've already found that from wiring photos and other wiring diagrams.
So thanks for at least trying to help me:) But one question is solved.
Bias problem is still here. I will rewire ppimv for LarMar version and try again.

[CoffeeTones]

Your English is fine. I understand but I still think you are confused. Both circuits are the same, only the LarMar has the safety resistors moved to the dual pot.

why Rich Mod has caps lugs direct connection to valves and LarMar doesn't? Looking at Ceriatone wiring diagram, there's no direct connection from 0.022uF caps (in my case those are 0.1uF) to valve pin. And this is Lar/Mar mod.

The Rich / Metro doesn't have the output caps wired directly to the power tubes, the wires to the power tubes go from the 220k resistors on the board (the original wires are left in place, from the 220k resistors to the power tubes) then NEW jumper wires lead from the 220k resistors to the center lugs of the Rich / Metro dual pot. New connections are made between the PI output caps and the dual pot's outer lugs, with no other connections. (the same as the LarMar)

On the LarMar dual pot, the wires to the power tubes go directly from the center lugs, where the 220k resistors are also mounted / connected, and do not need NEW jumper wires to go to the board, because those connections and safety resistors are on the LarMar pot already.

This is why your bias is not working correctly. You have no direct connection from the power tube grids to the 220k bias resistors, as you wired it.

The LarMar does not need wires to go to the board because those connections and safety resistors are mounted on the LarMar dual pot already, however the two circuit connections are the same.

But Rich mod does have such connection.

The Rich / Metro doesn't have the PI output caps wired directly to the power tubes, each PI output cap leads through a wire, directly to the outer lugs on the dual pot only. This is the same as the LarMar. The opposite, outer lugs are connected through a different wire, to the bias output voltage and the other end of the 220k resistors. The NEW jumper wires, to the safety resistors are not needed in the LarMar because the safety resistors are mounted on the pot, not the board. Both circuits have the same connections, only different values and the LarMar does not use jumper wires to the safety resistors and the wires going to the power tubes lead directly from the center pot lugs, not the board as in the Rich / Metro master.

[Sazabi]

Let's try it another way:) Here're pictures of my amp and wiring. Is that correct?
http://i.imgur.com/gxMWCGZ.jpg
http://i.imgur.com/779KqSU.jpg
Sorry, don't know why, but pics not scaling so direct link.

[CoffeeTones]

This is correct http://s1140.photobucket.com/user/Fotoe ... u.jpg.html

[Sazabi]

This is correct http://s1140.photobucket.com/user/Fotoe ... u.jpg.html

Yep, just as it's in my amp. Already sorted things out a bit:)
So few other questions:)
Lar/Mar uses 250K pot and 2.2M resistors, that gives ~220K max for bias.
Rich Mod uses 500K pot and 220K resistors, and that gives ~150K max for bias. I think this is why my bias floated away.
But how people use Rich Mod than?!
Even Maestro Metropoulos does - https://www.facebook.com/MetropoulosAmp ... 09/?type=1

[CoffeeTones]

Bias current should not be affected.

[Sazabi]

Bias current should not be affected.

In case total max resistance remains near 220K, but if it goes, i.e lower?

[CoffeeTones]

Should not make a difference. By the way, each wire pair(to the capacitors, to the resistors, to the power tubes) needs to be shielded or twisted. The amp is usually quieter and sounds much better that way. Especially at low PPIMV settings.

[Sazabi]

Ok, progress.
Re-soldered with another pot and resistor (as in Lar/Mar version, 250K pot and 2.2M resistor). Plus shielded cables and 5.6K grid resistors (to eliminate cosmos sounds:)). Bias not back to normal.
So I still don't understand how pair 500K pot and 220K resistors are working in other amps.
And looks like bias IS dependent on this pot/resistor pair.
http://i.imgur.com/xI34Cfh.jpg

[CoffeeTones]

Question for all readers - Has any one else seen an effect on bias current with the LarMar or Metro PPIMV? I've not experienced any dramatic effect.

[CoffeeTones]

Ok, progress.
Re-soldered with another pot and resistor (as in Lar/Mar version, 250K pot and 2.2M resistor). Plus shielded cables and 5.6K grid resistors (to eliminate cosmos sounds:)). Bias not back to normal.
So I still don't understand how pair 500K pot and 220K resistors are working in other amps.
And looks like bias IS dependent on this pot/resistor pair.
http://i.imgur.com/xI34Cfh.jpg

Do you have an outer shield wire on each pair, going to ground? Shielded / twisted pairs are better than those that are not paired, but it's preferred to ground the outer shield if it exists. It is also preferred not to reference the outer shield to the bias output voltage, as seen in some layouts on the internet.

[Ondrej]

Hi guys,

as was pointed out earlier in this thread, George mentions that he prefers the 500k Type 2 PPIMV over any other type. From what I read, both LarMar and Rich/Metro are Type 2 PPIMV, i.e. the bias resistor (typically 220k) is replaced by a pot (ideally 220k, but hard to get, so non 220k is used) with a parallel resistor to it (to get close to 220k log taper and to protect the tubes in case the pot wiper looses connection).

The only difference to my knowledge (as was also pointed out in this thread) is in the values of the pot and the parallel resistor:

LarMar: 250k pot + 2200k parallel resistor (soldered directly to the pot lugs)
Rich: 500k pot + 220k resistor (original bias resistor in the Plexis)

I have made a simple excel calculation of how both variants behave in terms of the taper, the input resistance and the output resistance. The together with the PI coupling caps, the bias resistors form an RC high-pass filter, whose cut-off frequency you probably want to keep fixed. The images below show my excel sheets in terms of all the observed variables.

RichMetro.png

Rich/Metro analysis

(53.68 KiB) Not downloaded yet

LarMar.png

LarMar analysis

(51.88 KiB) Not downloaded yet

So what's interesting is that LarMar is close to an ideal 220k pot in all aspects, i.e. constant input resistance (constant cut-off frequency of the RC filter), log taper output resistance. On the other hand, the Rich variant has rather inverse-log taper and the output resistance 152k at full power. On the other hand, the input resistance varies significantly going from 500k to 152k. This changes the cut-off frequency of the RC filter, although these frequencies are probably way too low to get noticed.

One thing I forgot to include in the analysis is the voltage division curve (in dB). In LarMar, it's logarithmic, while in Rich variant, it's close to linear (S-shaped).

So the question is, what makes George choose the Rich variant over the LarMar variant? Could it be the cut-off frequency compensating the NFB effect when on lower volumes?

Thanks

[Doug H]

I'm not sure what rich version is, but a 500k pot would put a larger load on the PI as it's turned down. I can see how some of that might be desirable through some of the sweep.