Tweed Twin power reduction

[Chuck H]

"I would tell your customer simply that this is the way the amp was designed. You can’t make a house into a cat. "

I did, in no uncertain terms, trust me. I also told him to back off with the freaking constant questions and 30 minutes of texting, because after all, I am supposed to get paid for this. Do your own research, I'm not your on call oracle. We'll see where that goes, it may have solved this problem.

You could absolutely endorse the attenuator. The only reason I can conceive of being "not in favor of" an attenuator would be the possibility of user error causing an open load that may damage the amp. Otherwise I don't understand the opposition. After all, an attenuator becomes the users responsibility and is no tougher on the amp than the speaker if used correctly. A caveat being that a user may run an amp at very high levels regularly with an attenuator where that wasn't possible before attenuation. So that means the amp spends more time running harder. But that equates logically to more wear and tear naturally. Not at all caused by the attenuator.

[Justin Thomas]

I don't know what the voltages are in your Twin, but JJ 6V6s might work... Bias accordingly.

Justin

[glebert]

I don't know what the voltages are in your Twin, but JJ 6V6s might work... Bias accordingly.

Justin

When you do this do you also need to account for the higher impedance of the 6V6s, so it becomes very similar to pulling two tubes (?). If you go down to two 6V6s would you get two doublings of impedance, such that changing the speakers from parallel to series would be in the right ballpark? It would be really odd

[Justin Thomas]

It's a guitar amp. I don't worry about such things. As long as it sounds good & doesn't self-immolate, it's good enough for me.

Justin

[Chuck H]

It's a guitar amp. I don't worry about such things. As long as it sounds good & doesn't self-immolate, it's good enough for me.

Justin

Smoke 'em if you got 'em.

[HaroldBrooks]

It's a complex issue, because wattage reduction is not related linearly to sound perception. Include Equal loudness curves and the frequency amplitude perception that occurs at high volume, and you have another perception problem related to lowering the wattage, not just a volume issue.

https://en.wikipedia.org/wiki/Equal-loudness_contour

I would first find out precisely how loud is too loud for your customer, and then also show them how the Low Bass and upper treble drop off (from your ear) when you lower the watts.

Most people are expecting a lower volume to sound a big and powerful as a dimed amp, but unless there is a type of "Loudness" control like you found on old stereos, they will be disappointed.

Personally I try to introduce a lot of 'Late Stage' Low bass to my amps via cap, transformer, and speaker mods, and then dial back the bass up front with an EQ when everything gets cranked. Not really the same because I am changing the distortion content, by removing the bass up front, as opposed to the output speaker but it is better than nothing.

So, make sure HE knows what he wants, and then take the necessary steps to accommodate his request, and educate him in the limitation of sound do to factors like speaker breakup, transformer distortion, and the human ear's non-linear frequency dependent response to sound.

Lowering the volume may also require some tonal adjustments via the amp and speakers. That's one of the reasons I favor by 4" x 12" cabinets even at low volumes (but I don't like lugging them around !). You get more bass at a low volume and it sounds bigger, almost like it did when it was louder.

IMHO, that is as it's a qualitative assesement !

[Randall]

I ended up going with this Master Vol I found on Robinette's site. It's easy, and it works. I went with a 500KA pot, I didn't like the taper on the 1 Meg. It doesn't effect volume in any way that I can hear when it's full on, and lifting a leg of the pot makes no audible difference. It works by adding the out of phase PI signals, which cancel. I was a bit skeptical at first, but it works fine. Quck to do and easily replaceable to stock. I put it in the ground switch hole, which was just a tie point for the mains inlet wires anyway. Twisted the wires from switch to sockets, and we're done.

Oh, and this was in a 5F8A Twin clone.

[Chuck H]

I ended up going with this Master Vol I found on Robinette's site. It's easy, and it works. I went with a 500KA pot, I didn't like the taper on the 1 Meg. It doesn't effect volume in any way that I can hear when it's full on, and lifting a leg of the pot makes no audible difference. It works by adding the out of phase PI signals, which cancel. I was a bit skeptical at first, but it works fine. Quck to do and easily replaceable to stock. I put it in the ground switch hole, which was just a tie point for the mains inlet wires anyway. Twisted the wires from switch to sockets, and we're done.

Oh, and this was in a 5F8A Twin clone.

[ATTACH=CONFIG]54023[/ATTACH]

This is just the simplest post PI master circuit. But why the hell is it implemented at the grids instead of behind the grid stop resistors!?! Any benefit of power tube clipping is omitted in this scenario. But then, tone is in the ear of the beholder. If your customer is happy then you're done.

[Randall]

"This is just the simplest post PI master circuit. But why the hell is it implemented at the grids instead of behind the grid stop resistors!?! Any benefit of power tube clipping is omitted in this scenario"

I wondered this myself, but I'm not sure I understand it. Actually, since a 5F8-A only has grid stop resistors on one tube of each pair, I connected at pin 5 of the one that has no grid stop, where the grid stop of the other tube also connects. So, in essence ach pair is connected both ways, if that makes sense.

As far as power tube clipping goes, in this scenario I don't think that is going to be an issue considering the nature of taming a high power twin.

http://www.prowessamplifiers.com/sch...5f8a_schem.pdf

[nickb]

I ended up going with this Master Vol I found on Robinette's site. It's easy, and it works. I went with a 500KA pot, I didn't like the taper on the 1 Meg. It doesn't effect volume in any way that I can hear when it's full on, and lifting a leg of the pot makes no audible difference. It works by adding the out of phase PI signals, which cancel. I was a bit skeptical at first, but it works fine. Quck to do and easily replaceable to stock. I put it in the ground switch hole, which was just a tie point for the mains inlet wires anyway. Twisted the wires from switch to sockets, and we're done.

Oh, and this was in a 5F8A Twin clone.

[ATTACH=CONFIG]54023[/ATTACH]

It works because the output impedance of the PI and the MV pot form a potential divider, not due to mixing of phases.

[Leo_Gnardo]

This is just the simplest post PI master circuit. But why the hell is it implemented at the grids instead of behind the grid stop resistors!?! Any benefit of power tube clipping is omitted in this scenario. But then, tone is in the ear of the beholder. If your customer is happy then you're done.

Just ran across this type of MV as a factory install on a rare Risson amp the other day. FWIW it was on the 'other side' of the grid stop resistors, away from the grids. If it sounds good, go for it. Next step would be the more commonly seen dual volume pot PPIMV circuit.

[Chuck H]

It works because the output impedance of the PI and the MV pot form a potential divider, not due to mixing of phases.

Isn't the output impedance of the "circuit" the PI output impedance parallel with the 220k bias resistance? And where is the signal "divided" to? A more easily understood voltage divider divides to 0V, but in this case it would have to be divided to the opposing phase. That seems like local NFB and should reduce the circuit impedance IIRC. I'm seeking clarity because I always saw this circuit as phase dependent like Rob's site says.

[Helmholtz]

Isn't the output impedance of the "circuit" the PI output impedance parallel with the 220k bias resistance? And where is the signal "divided" to? A more easily understood voltage divider divides to 0V, but in this case it would have to be divided to the opposing phase. That seems like local NFB and should reduce the circuit impedance IIRC. I'm seeking clarity because I always saw this circuit as phase dependent like Rob's site says.

The MV is wired as floating variable resistor connecting between (ideally) equal amplitude, opposite phase voltages. So signal voltage across the MV resistance corresponds to twice the value at each PI output. Also plate signal currents double.
This acts as if separate resistors having half the actual MV resistance were wired between each PI output and ground, acting as voltage dividers together with the PI output impedances. So the MV acts as the lower part of a differential voltage divider.

It may help to understand the principle to visualize the center of the MV resistance as grounded. As long as PI signals are symmetrical, grounding the MV center would not change anything. There is no "phase mixing".

At zero setting of the MV both PI outputs are essentially shorted. It is a rather rude method, causing extreme variations of the AC loadlines of the PI triodes.


I used this kind of MV around 1970 with my AC30, just shorting the series cap in the "Cut" circuit.

[Justin Thomas]

I thought that was usually called a "Crossline Master Volume," and was (in)famous for use in Matchless amps? Seemed like a lot of people didn't like them because they didn't sound good at bedroom levels, though they weren't intended for that kind of attenuation anyway? More for "taking a little off the top?" I know in my Spitfire-w-Cheese (a Chuck H. Design) I liked the Master at 6 or above, though at extreme attenuation it made some pretty cool noises...

Seems to me that it works by shorting the PI outputs together, therefore works via phase cancellation? The main issue with it being, the location of the pot depended on whether your amp was fixed or cathode nias? But I forget the details.

Justin

[Dave H]

Isn't the output impedance of the "circuit" the PI output impedance parallel with the 220k bias resistance? And where is the signal "divided" to? A more easily understood voltage divider divides to 0V

Yeah, but the output impedance of the usual LTP PI is only around 40K so the PI 'wins'
A 40k resistor to ground from each output of the PI would divide the signal by two as would an 80k between the outputs. One problem I can see with low MV pot settings is bass roll off (because of the coupling caps). If the PI plates are at the same voltage the MV pot could be connected directly between the plates to avoid this.