Half-power operation
Moderators: pompeiisneaks, Colossal
Re: Half-power operation
Yes, you guys are right I should have left the "SPL" part off. I told ya this was confusing.
Last edited by UR12 on Tue Mar 04, 2008 3:58 pm, edited 1 time in total.
Re: Half-power operation
Good thread. My $.02 is that 50/100 watt configurations are a waste of time although they may make for good marketing (like the "Triode" switch). If the math arguments presented earlier don't convince you then here's another argument which is a bit counter to some I read earlier, but it is my experience: I put a cathode lifter on my Twin. In 2 tube mode it made the amp lifeless. There was not much drop in volume, but the transients and presence were affected. I got some compression, but the amp had no authority in half power mode. I could not change the 4 ohm inpedance of the OT and did not go to one speaker, so maybe it would sound better doing one of those things. The other very real possibility is I needed new tubes and running 4 was masking the real problem. Regardless, though it was a failure for me, others say the tone does not change (but as we see, the volume does not change to any useful degree either) so why bother.
Re: Half-power operation
While all the math stuff is good to know, this is, in my opinion, the best argument against cathode lift or triode/pentode switching. I have never seen it improve the sound/feel of an amp, and the difference in volume is next to nothing.Tonegeek wrote:Good thread. My $.02 is that 50/100 watt configurations are a waste of time although they may make for good marketing (like the "Triode" switch). If the math arguments presented earlier don't convince you then here's another argument which is a bit counter to some I read earlier, but it is my experience: I put a cathode lifter on my Twin. In 2 tube mode it made the amp lifeless. There was not much drop in volume, but the transients and presence were affected. I got some compression, but the amp had no authority in half power mode. I could not change the 4 ohm inpedance of the OT and did not go to one speaker, so maybe it would sound better doing one of those things. The other very real possibility is I needed new tubes and running 4 was masking the real problem. Regardless, though it was a failure for me, others say the tone does not change (but as we see, the volume does not change to any useful degree either) so why bother.
I think my favorite lower volume solution is the z airbreak. Let me qualify this by saying that I never heard any other attenuators or heard the z model with any amps other than the Route 66 and the Galaxie, but on all but the lowest settings it did a great job without sucking tone.
Re: Half-power operation
Gerald Weber (him again) says he once did a blind taste test with a musician where they compared amps playing into the "correct" impedance, into half the impedance and into twice the impedance. The guy universally preferred the sound of double the impedance (my experience too, at least so far as Super Reverbs at 4 ohms go (way nicer than 2 ohms, IMO).)
Final dB math check: +6dB corresponds to a doubling of voltage or current (amplitudes) but a quadrupling of power (proportional to the square of the amplitude). For Sound Power Level (which should be written SWL), +3dB is double. For Sound Pressure Level (SPL), +6dB is double. The relationship between SWL and SPL is frequency -- if 100 watts of power produces a certain SPL at 440Hz, the same power will produce double the SPL at 220Hz. I think this is why we like amps tweaked for more highs.
Final dB math check: +6dB corresponds to a doubling of voltage or current (amplitudes) but a quadrupling of power (proportional to the square of the amplitude). For Sound Power Level (which should be written SWL), +3dB is double. For Sound Pressure Level (SPL), +6dB is double. The relationship between SWL and SPL is frequency -- if 100 watts of power produces a certain SPL at 440Hz, the same power will produce double the SPL at 220Hz. I think this is why we like amps tweaked for more highs.
Re: Half-power operation
IMHO i agree with you guys that a 1/2 power switch and pentode triode switches aren't worth the effort. I think it is more marketing hype. The other side of the coin is that if you spend the money on a matched set (quad) of tubes and you run the amp with the same two disconnected for a while you are not letting the tubes get the same amount of use and you will have one pair wear faster than the other. If you are going to go 1/2 power then at least swap the tubes around like you rotate your tires so they all wear the same.
Re: Half-power operation
OK, well I happen to have an amp on the bench right now in which the customer insisted on a half-power switch.
So would just disconnecting the grids (or grounding them) on 1/2 of the tubes have the same effect as lifting the cathodes in terms of perceived output level? I think one reason why cathode-lift 1/2 power schemes don't work too well is because they reduce the load on the PS by some large percentage close to 1/2, and therefore affect dynamics and transient response, not to mention letting B+ rise in most circumstances. But grid-lift of the outside power tubes would seem to allow the amp to still run the same ps load, same OT load, but two tubes are idling and the other two are working.
just a thought. I can't be [among] the first [million people] to think of this so there must be some major flaw in it that I don't see immediately.
So would just disconnecting the grids (or grounding them) on 1/2 of the tubes have the same effect as lifting the cathodes in terms of perceived output level? I think one reason why cathode-lift 1/2 power schemes don't work too well is because they reduce the load on the PS by some large percentage close to 1/2, and therefore affect dynamics and transient response, not to mention letting B+ rise in most circumstances. But grid-lift of the outside power tubes would seem to allow the amp to still run the same ps load, same OT load, but two tubes are idling and the other two are working.
just a thought. I can't be [among] the first [million people] to think of this so there must be some major flaw in it that I don't see immediately.
Re: Half-power operation
You can't just ground out the signal on two tubes, because that will also ground out the signal on the remaining two. And if you just disconnect the grids but leave them open, I think there may be a noise issue (anybody else know for sure?) So you'd have to disconnect two grid wires AND ground the grids themselves. What is that -- a 4P4T switch?
You're right about the difference in loading that pulling two tubes or disconnecting their cathodes produces. There's a pretty detailed discussion of this in a Plexi in TUT3 (which I don't have handy at the moment).
You're right about the difference in loading that pulling two tubes or disconnecting their cathodes produces. There's a pretty detailed discussion of this in a Plexi in TUT3 (which I don't have handy at the moment).
Re: Half-power operation
no it's just one DPDT switch, on-on variety with no center detent.
I have tubes 1, 2, 3, and 4. 1&4 are the outside pair, 2&3 the inside pair.
There is a 10K resistor from the PI coupling cap to each of the grids (four total resistors).
So my switch's common pin would go to the 1&4 grids, and one side of the switch would be ground while the other would be the 10K grid resistor coming from the PI. So switch in the first position, grids are connected to the grid resistor as normal. Switch in the second position, 10K grid resistor is floating and the grids of the outside pair of tubes are grounded.
In this case I'd only be grounding the outside tube grids, and leaving the 10K grid resistors floating. My guess is that the biggest problem would be leaving those R's floating, as they may pick up some noise that gets coupled (through the grid R) directly into the active tube pair's grids. I am inclined to try it.
I haven't read TUT3 or any of the O'Connor books... guess I should.
I have tubes 1, 2, 3, and 4. 1&4 are the outside pair, 2&3 the inside pair.
There is a 10K resistor from the PI coupling cap to each of the grids (four total resistors).
So my switch's common pin would go to the 1&4 grids, and one side of the switch would be ground while the other would be the 10K grid resistor coming from the PI. So switch in the first position, grids are connected to the grid resistor as normal. Switch in the second position, 10K grid resistor is floating and the grids of the outside pair of tubes are grounded.
In this case I'd only be grounding the outside tube grids, and leaving the 10K grid resistors floating. My guess is that the biggest problem would be leaving those R's floating, as they may pick up some noise that gets coupled (through the grid R) directly into the active tube pair's grids. I am inclined to try it.
I haven't read TUT3 or any of the O'Connor books... guess I should.
Re: Half-power operation
Okay, I got it. Worth a try, anyway. If you can wait till tonight, I'll see what O'Connor says in TUT3 and send you PM.
Re: Half-power operation
well we should also realize that for an amp with NFB, this also changes the relationship. In order for that amp to sound good in 2 tube mode I suspect it the switching also needs to adjust the NFB resistor, and possibly the precense control circuit.I put a cathode lifter on my Twin. In 2 tube mode it made the amp lifeless. There was not much drop in volume, but the transients and presence were affected. I got some compression, but the amp had no authority in half power mode.
This says nothing as to the pre-amp being designed for a certain load, which I can say from my builds (7 dumble clones so far, with 6v6, 6l6, el34, 6550) it does make a difference, and the voicing does need some minor tweaks to be its best with a given tube compliment.[/quote][/code]
it really is a journey, and you just cant farm out the battle wounds
Re: Half-power operation
yeah I'm in no hurry. It'll be about two weeks before I have to make this decision. I'll probably just try it with flying leads first and see what works best.
Re: Half-power operation
[/code][/quote]briane wrote:well we should also realize that for an amp with NFB, this also changes the relationship. In order for that amp to sound good in 2 tube mode I suspect it the switching also needs to adjust the NFB resistor, and possibly the precense control circuit.I put a cathode lifter on my Twin. In 2 tube mode it made the amp lifeless. There was not much drop in volume, but the transients and presence were affected. I got some compression, but the amp had no authority in half power mode.
This says nothing as to the pre-amp being designed for a certain load, which I can say from my builds (7 dumble clones so far, with 6v6, 6l6, el34, 6550) it does make a difference, and the voicing does need some minor tweaks to be its best with a given tube compliment.
Only vaguely related, but hopefully a quick question: if you know that a 100k/4.7k resistor pair is used with a .1uF cap in a presence circuit for a 100W 4-tube amp, how do you scale that for an amp that only uses 2 tubes? 200k/9k with a .05uF cap? Or does it matter?
-g
Re: Half-power operation
It's not the tube complement that sets the sensitivity of the loop, it's the output impedance. If you have only one set of output taps, the feedback resistor (the larger value) OR the shunt resistor (the smaller value) should vary with the load (if you change them both by proportional amounts, you haven't changed anything except the time constant with the presence cap). For example, if you have a 100K feedback resistor on an amp with an 8-ohm speaker and you change the speaker to 4-ohms, you'd drop the feedback resistor to 47K to maintain the same ratio between the speaker and the resistor (yes, one is an impedance and one is a resistance, but that's more or less the way they've always done it). It's much easier if you have multiple output taps -- just wire the feedback network to one tap and "set it and forget it."
The change in tube complement may produce a small change in the power amp's overall phase shift, but probably not enough to hear at guitar frequencies.
EDIT: I'm going to correct myself a little here. Changing from four down to two output tubes also cuts the transconductance in half, so there is also a reduction in how much gain the feedback loop trims off the output. Not enough I think to justify doubling the feedback resistor (while leaving everything else the same), but you could certainly experiment with 120K or 150K as feedback resistor values. Don't change anything else though. There is, technically, a change in the behavior of the coupling caps between the PI and the output tubes with any change in the feedback, but to compensate for it, you would need cap values that don't exist (like .082725uF).
The change in tube complement may produce a small change in the power amp's overall phase shift, but probably not enough to hear at guitar frequencies.
EDIT: I'm going to correct myself a little here. Changing from four down to two output tubes also cuts the transconductance in half, so there is also a reduction in how much gain the feedback loop trims off the output. Not enough I think to justify doubling the feedback resistor (while leaving everything else the same), but you could certainly experiment with 120K or 150K as feedback resistor values. Don't change anything else though. There is, technically, a change in the behavior of the coupling caps between the PI and the output tubes with any change in the feedback, but to compensate for it, you would need cap values that don't exist (like .082725uF).