That's almost what I did on my PP amp but I used one CCS per tube and coupled the cathodes with a two 1000UF in series, midpoint connected to ground with 1M. So maybe I already have a PP amp with SE signature. It could be true, this amp sound so different from the standard PP circuit it replaced... (resistor bypassed with cap per tube)
I'm using this at the moment without the zeners. I can see the transistion from class A to AB1 in the speaker. It wobbles...
I'm using this at the moment without the zeners. I can see the transistion from class A to AB1 in the speaker. It wobbles...
Last edited:
Are you driving both inputs ? For the most SE signature the undriven half wants to have a much higher gm than the driven half. This makes it follow the driven half more perfectly. Hence a pentode, or even a MOSFET, on the undriven side makes for a more pronounced SE signature. In my RH slayer design I apply schade feedback to the pentode been driven (turning it into a pseudo triode), and leave the other half as a pentode.
I used the illustrated arrangement on almost all of my builds up until the last few. It works very well, but most of the time I drove both inputs.
Shoog
I used the illustrated arrangement on almost all of my builds up until the last few. It works very well, but most of the time I drove both inputs.
Shoog
Last edited:
When you use a power resistor or light bulb as the dc current compensation load on a SE driven PP OPT, the problem is that you will loose a lot of output power on that low impedance current sink.
The AC which is on the power triode side will be present on the inverted side as well. Square that voltage and divide it by the resistance and you will know how much power is not going into your speakers.
A high impedance current sink is the way to go. Can be sand or pentode (triode will be too low impedance).
I have seen amps with a 300B on one side and an undriven EL34 as a current sink on the other.
The AC which is on the power triode side will be present on the inverted side as well. Square that voltage and divide it by the resistance and you will know how much power is not going into your speakers.
A high impedance current sink is the way to go. Can be sand or pentode (triode will be too low impedance).
I have seen amps with a 300B on one side and an undriven EL34 as a current sink on the other.
Here is a link to one of my light-bulb amps.
Audio ratbag: Lumiere
Yes, it wastes output power, but efficiency is usually not a goal with SE amps.
Curiously there is a difference in sound between using light-bulbs and using resistors. This amp has a switching arrangement where I can try both and I happen to prefer the light-bulb sound which has a more bell like quality.
The big concern about using bulbs is the temperature coefficient of the tungsten in the bulb. The resistance varies with temperature and is non-linear. I suspect this is not really a huge issue since the tungsten will have some thermal mass and the resistance swing while playing music will not be as great as would occur if you were taking resistance measurement readings with a few seconds between readings.
Sometimes I wonder if my preference for the bulbs comes from the minor change in resistance. When you think about it, direct heated triodes also have changes in current occurring when the music plays and the filament material will also have a temperature coefficient. Of course this is just speculation on my part, but I wonder if this is why some of us like the sound of direct heating.
ray
Audio ratbag: Lumiere
Yes, it wastes output power, but efficiency is usually not a goal with SE amps.
Curiously there is a difference in sound between using light-bulbs and using resistors. This amp has a switching arrangement where I can try both and I happen to prefer the light-bulb sound which has a more bell like quality.
The big concern about using bulbs is the temperature coefficient of the tungsten in the bulb. The resistance varies with temperature and is non-linear. I suspect this is not really a huge issue since the tungsten will have some thermal mass and the resistance swing while playing music will not be as great as would occur if you were taking resistance measurement readings with a few seconds between readings.
Sometimes I wonder if my preference for the bulbs comes from the minor change in resistance. When you think about it, direct heated triodes also have changes in current occurring when the music plays and the filament material will also have a temperature coefficient. Of course this is just speculation on my part, but I wonder if this is why some of us like the sound of direct heating.
ray
I am driving both inputs but i could ground one grid and decouple the cathodes with a switch.
I came across this nice article...
Switchable Single-Ended, Push-Pull Guitar Amp
So, for a fixed bias amp where both cathodes are connected together and each grid get its neg. voltage through a separate winding of the IT: Would it be sufficient simply to ground one grid by connecting a cap netween grid and ground ? That is it ?
The cathodes are not isolated through sep. resistors as shown in the patent as there are no cathode resistors at all in my fixed bias amp....
I wanted to see how the sound changes and how badly my speaker needs the power/damping factor of the PP....
Switchable Single-Ended, Push-Pull Guitar Amp
So, for a fixed bias amp where both cathodes are connected together and each grid get its neg. voltage through a separate winding of the IT: Would it be sufficient simply to ground one grid by connecting a cap netween grid and ground ? That is it ?
The cathodes are not isolated through sep. resistors as shown in the patent as there are no cathode resistors at all in my fixed bias amp....
I wanted to see how the sound changes and how badly my speaker needs the power/damping factor of the PP....
The CCS on the undriven side of the OPT does work. Remember that using half of a P-P OPT gives you 1/4 of it's impedance. That 6600 ohm OPT becomes a 1650 ohm OPT.
Any kind of resistive load in place of the CCS will suck up a good chunk of your output power. In the case of a pure resistor it will eat about half of it.
I have seen the "use a second tube and don't drive it" trick used, and it appears to work but I couldn't make it sound good in a little 2 watt guitar amp I built, so.....
I made it into a self split output stage. It made a bit more power but still wouldn't scream......'
I wound up using a simple mosfet split load PI for a push pull amp and I could overdrive it for all sorts of good sounds. That's the way the amp has been for 3 years now.
Mesa has been "patenting" all sorts of stuff that has been commonplace in the guitar amp world for years. That whole switching of modes stuff is "obvious to anyone who is skilled in the art" thus making the patent invalid. They are betting on intimidation and the "we are bigger than you are" legal stance.
You could do that, but it may (and probably will) upset the balance of your PI since one of it's outputs is now tied to an AC ground. This may or may not matter. If the PI doesn't distort, it will work OK. Otherwise use a switch to disconnect the coupling cap from the PI and connect it to ground on the undriven side.
Any kind of resistive load in place of the CCS will suck up a good chunk of your output power. In the case of a pure resistor it will eat about half of it.
I have seen the "use a second tube and don't drive it" trick used, and it appears to work but I couldn't make it sound good in a little 2 watt guitar amp I built, so.....
I made it into a self split output stage. It made a bit more power but still wouldn't scream......'
I wound up using a simple mosfet split load PI for a push pull amp and I could overdrive it for all sorts of good sounds. That's the way the amp has been for 3 years now.
Mesa has been "patenting" all sorts of stuff that has been commonplace in the guitar amp world for years. That whole switching of modes stuff is "obvious to anyone who is skilled in the art" thus making the patent invalid. They are betting on intimidation and the "we are bigger than you are" legal stance.
You could do that, but it may (and probably will) upset the balance of your PI since one of it's outputs is now tied to an AC ground. This may or may not matter. If the PI doesn't distort, it will work OK. Otherwise use a switch to disconnect the coupling cap from the PI and connect it to ground on the undriven side.
Well, the idea in my case is more like a case study to try and not to use this permanently...I want to experiment with all kind of setups to see what is sound how keeping as many parameters fixed and one variable at a time.
So, currently I use a PP-Setup. But I want to see what will happen when driver stays the same etc., but last stage is SE, so that you do not judge other effects.
But the impedance thing I guess destroys the idea of using a second tube (in my case a 300B) as a dummy to keep the transformer happy. Currently I run the 300B with a 6k transformer in PP. So, if I would geound one grid, we have only a 1k5 OPT the remaining is playing into. This will not sound good.
So, currently I use a PP-Setup. But I want to see what will happen when driver stays the same etc., but last stage is SE, so that you do not judge other effects.
But the impedance thing I guess destroys the idea of using a second tube (in my case a 300B) as a dummy to keep the transformer happy. Currently I run the 300B with a 6k transformer in PP. So, if I would geound one grid, we have only a 1k5 OPT the remaining is playing into. This will not sound good.
If your OPT has multiple taps on the OPT secondary you can try connecting the 8 ohm speakers to the 4 ohm output tap. This will double the load that the tube sees making that 1K5 into 3K.
If your 300B push pull amp is biased such that it stays in class A for all (or most) of it's usable power range, the tubes see 3K each effectively at normal loading.
If your 300B push pull amp is biased such that it stays in class A for all (or most) of it's usable power range, the tubes see 3K each effectively at normal loading.
I experimented with such designs. The last breadboard of such configuration used triode on one side, and modulated MOSFET CCS on another. Then I loaded a stage on CCS, increasing B+ twice. The effect was the same, except no current through OPT. Then I turned all around, making it as a cathode follower loaded on modulated CCS. Then I switched to source follower loaded on a modulated CCS. The final version was an augmented source follower loaded on a modulated CCS. B+ went down to 150V, for 100W output (400+ W dissipation), and an output transformer had been eliminated.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.