Spitfire filtering vs 18w

[bgkyt1]

Hi,

I notice some amps (marshall 18w,ampeg j12) use the same filter cap for the preamp and phase inverter, but the Spitfire phase inverter has its own filter cap, even though the dropping resistor values (from the previous node) and cap values are the same.
What is the benefit of giving the PI its own cap (not sharing it with the preamp -- on amps with just one gain stage) if the dropping resistor and cap values are the same?

thanks...

[pompeiisneaks]

Generally the PI is considered part of the power amp, and has much higher current and potential for sending noise into the preamp, if not decoupled. On lower gain amps it's probably not as 'necessary' but any time you can decouple the PI from the preamp, it's a good idea.

If you can decouple every single stage it's even better. Dumble did for each tube, (B+ for anodes, B+2 for screens, B+3 for PI, B+4 for the OD, and B+5 for the preamp (as well as another for the fet).

Looking at the trainwreck amps, as well, it has the same 'per node' type of filtering.

Some of the older amp designs were built around saving cost over optimal filtering. The only thing that can be said about stronger filtering, is that it does stiffen up the power supply and if you're looking for sag, you'll lose some of that with this type of filtering (and even more so with a Solid State rectifier.)


~Phil

[tubeswell]

Generally the PI is considered part of the power amp, and has much higher current and potential for sending noise into the preamp, ...

Its actually about the same current as a pre amp tube for both triodes (in the above schematic, 62V/48k2 = 1.3mA going through the cathode at idle)

(FWIW I too prefer to decouple the PI from the rest of the amp)

[bgkyt1]

thanks for the info. i appreciate the quick responses.

[Colossal]

The Spitfire is also derived from the Vox AC power supply with C-L-C and then PI and preamp nodes branching from the screen node. Marshall maintains the serial RC filtration scheme.

[wpaulvogel]

Also just a side note but the Marshall 18 watt circuit is basically a carbon copy of the Watkins Dominator 20T. I love Marshall amps but they copied in the beginning.
It’s always best to use separate nodes for filtering gain stages.

[gingertube]

To avoid "Motorboating" style of oscillation as bypass caps get older, it is essential to run no more than 2 amplifier stages from any power supply node, and those 2 stages MUST be of opposite signal phase.

If the 2 stages are the same phase, then any "modulation" of the B+ due to signal current in one stage becomes positive feedback for the other.

Since the Phase Splitter essentially has both phases of the signal it should have its own power supply node and nothing else should share that node.

This is just good design although I have seen this essential "good design" rule broken by some by some big names.

One example which comes to mind is the "Burn" Channel of a Fender SuperSonic. 3 cascaded gain stages on a single power supply node. The first and last have the same phase. It came to me to "repair" with guess what? an oscillation problem. As I wasn't prepared to redesign/rebuild, I replaced every bypass cap on that node and then added a 470nF/630V film cap to 0V at the top of the 3rd stage anode load resistor. That is, I treated the "symptom but not the disease".

Cheers,
Ian

[bgkyt1]

good info, thanks!

[JD0x0]

This is just good design although I have seen this essential "good design" rule broken by some by some big names.


Cheers,
Ian

And that's why it's best not to just copy stuff, and understand what you're building. I suspect this is often done to save a few pennies, much in the same way other 'big names' will run underrated diodes, because they saved a penny across the entire amp... And then you see what happens, like the supersonic (which has a notoriously high noise floor.. hmm..) then the customer has to pay hundreds of dollars to fix flaws caused by cost cuts of a few cents.