NFB Question: What's The Tech Behind What I'm Hearing?

[rp]

I pulled the old Traynor OT I had in my 5F6A since I wanted 4/8/16 taps and it only had 8 & 16. I had Heyboer do a custom tweed Bassman OT with 2/4/8/16 taps. I 'm just using the 4/8/16 to the outs but I put a toggle on the 2 & 16 taps so I choose btwn 5F6A / JTM45 style NFB. BIG difference from one to the other.

Set on 2 Ohms it's cleaner and much louder, at modest volumes anyway. At 16 Ohms it's much less loud, warmer and a little dirtier. I can't test this one full out in my apt. for now but I'm betting the difference diminishes when cranked.

Just curious what the tech is behind what I'm hearing. Is there more NFB when on the 2 ohm tap and that's why I hear it louder & cleaner?

Anyone have any favorites (for the tap) for this circuit seeing how various years/models of 45W Fenders and Marshalls will tap the NFB on the 2/4/8 or 16?

[jjman]

2 ohm tap pushes less signal (NFB) than 16 ohm, all else equal. Less NFB means less volume reduction = more volume.

[tubeswell]

The 16R tap has a higher voltage swing and a lower current swing than the 2R tap because of the different impedance ratios. The 16R tap is 4,000:16 (i.e.: 250:1), whereas the 2R tap is 4,000:2 (i.e.: 2,000:1). The OT Pri:Sec voltage ratio is the square root of the impedance ratio, so a 800V PP swing at the primary will translate into a 50.6V PP swing on the 16R tap, but only a 17.9V PP swing on the 2R tap. (i.e.: an order of magnitude of about 2.8 x the 2R signal)

Since it is this secondary voltage swing that is being fed-back (out of phase) to the LTP cathode swing, then the effect of the 16R NFB is to reduce the signal more.

You can overcome differences by changing the NFB resistor in use with either tap in order to ensure that same strength of voltage swing ends up being fed back to the LTP tail.

In this regard, the stock 27k for a 5F6A (which acts - as part of a voltage divider with the 5k presence pot 'underneath' the LTP tail resistor - in order to feed back up to 5/(5+27)th's or 15.6% of the NFB signal swing from the OT secondary - with the pot dimed) is designed to be 'optimal' with the 2R OT secondary.

Therefore, if using a the 16R tap as your NFB voltage source, then you should aim for a voltage divider that gives 2.8 x the cut that the 27k/5k divider does. It will be easier to change out the 27k resistor to something else than to change the 5k presence pot in the leg of the voltage divider. So this means you need something like an 82k resistor in place of the 27k resistor. (because 5/(5+82) =~ 5.7%, which is close-enough to 2.8 x less than the 15.6% knock-down you get with a 5/(5+27)th's voltage divider.

Is that any use?

[teemuk]

Less NFB means less volume reduction = more volume.

It also means
- higher output impedance resulting into smaller damping factor= less linear frequency response to a reactive load
- narrower bandwidth because NFB isn't extending it as extensively
- earlier but smoother breakup characteristics, because there's less NFB to remove distortion and to linearize the signal
- less effective tone controls if in feedback path (e.g. generic resonance & presence controls)

[rp]

Is that any use?

Hells yeh!

And you obviously know your theory better than Ken Bran. Marshall didn't up the NFB resistor to 47k or 100K and move the tap to 4 or 8 ohms till the JTM50. Took 'em quite the while to figure out. No Ampgarage back then.

I'm just gonna leave it as is for now, but now I understand all the permutations.