I'm in the design stages of my first push-pull guitar amp. I want it to be very simple, something along the lines of a Fender 5E3, or one of the early Gibson or Vox designs.
I'm going to be using 6GK6s for the output pair, feeding an 8K plate-to-plate transformer salvaged from a Bogen Challenger 6v6pp amplifier. I'll use a 12AV7 or 12AY7 for the input, with a 12AX7 as the phase splitter, pretty much following the 5E3 schematic.
I've modeled two different power supplies with parts I already have on hand. One uses just CRC filters, like the original amps. The other adds a small PIO motor-run cap and a choke after the rectifier tube. Both produce just about the same voltage, with similar loads on the power transformer.
The power supply with the choke has less than half the ripple of the chokeless version. Will this be an improvement, or will I lose some of the desirable properties of the supply fluctuating under load by adding the choke?
Here are my initial simulations. Opinions appreciated.
I'm going to be using 6GK6s for the output pair, feeding an 8K plate-to-plate transformer salvaged from a Bogen Challenger 6v6pp amplifier. I'll use a 12AV7 or 12AY7 for the input, with a 12AX7 as the phase splitter, pretty much following the 5E3 schematic.
I've modeled two different power supplies with parts I already have on hand. One uses just CRC filters, like the original amps. The other adds a small PIO motor-run cap and a choke after the rectifier tube. Both produce just about the same voltage, with similar loads on the power transformer.
The power supply with the choke has less than half the ripple of the chokeless version. Will this be an improvement, or will I lose some of the desirable properties of the supply fluctuating under load by adding the choke?
Here are my initial simulations. Opinions appreciated.
Attachments
If the ripple is reduced, surely it is good. But, be careful with CLC design: it is a passive filter and it always has a frequency of resonance near 1 / 2 pi sqrt (L * C) where C is the two caps in series. If this frequency is high (low L and or low C), the amplifier entirely may oscillate given a certain Q of such a filter, or have large output (of the supply) dips and or peaks of voltage when the musical program involves the frequency above mentioned. Use as much L and second cap cell as you can to reduce this happens. And the first L cannot go so high as the value recommended by manufacturer as a maximum for the tube being used.
Good luck.
Good luck.
Changes to filtering affect the feel and bass response of a guitar amp. These effects are about the same magnitude as you might expect when changing a speaker. The supply with the choke will be a little cleaner sounding, without the choke, a little more aggressive and screaming quality to the overdrive tone. The RC filtering before the main DC output will reduce this somewhat. The vintage 5E3 used 16uF filter caps.
BST. Welcome back to the fun stuff.
The power supply has a lot to do with the tone quality of a guitar amp. Many of the old amps that we know and love either used a CRC filter, or if a choke was used, the B+ tap on the OPT is connected before the choke. These choices were for cost reasons but the created part of the sound that we have grown to like, and try to recreate.
The CRC is cheap, electrolytics were not the greatest back then, and the usual size was a 16 to 30 uF with a 40 uF being huge.
The B+ tap before the choke allows a much cheaper choke to be used since most of the amps B+ current flows to the plates of the output tubes and does not go through the choke.
Both of these techniques unintendedly increase the B+ supplies tendency to sag under load. This increases the amps ability to sustain a note. Why? The designers of the old stuff never intended them to be driven far into saturation, but when you overdrive an amp the B+ current goes up reducing the B+ voltage. The drop in B+ reduces the gain and headroom through the entire amp including the preamp, increasing the distortion. As the note dies out the overload diminishes, reducing the B+ current, which increases the B+ voltage, and boosts the gain. This dynamic compression effect has a lot to do with the power supply.
Some ripple can be tolerated in a P-P amp. Again it will affect the sound. At "normal" playing volumes the ripple voltage is less than the B+ headroom in the output stages and the ripple doesn't have much effect. When the amp is cranked to 11 and driven with 97 effects boxes chained together the output stage is attempting to act like a pair of switches chopping the B+ up and feeding it to the speaker through the OPT. There is no more ripple cancellation and much of the ripple gets passed right on through to the speaker. You don't hear it because the speaker is eating the full cranked to 11 power of the amp, but its effects are there in the form of IMD. The guitar's tone has been modulated with 60, 120, 180 Hz (and so on) tones, but only on the signal peaks when the amp is clipping (or near clipping). This can be seen on an FFT analyzer, and it imparts a grainy sound to the distorted peaks.
A "soft" power supply will increase the dynamic compression and IMD effects and is often favored by blues and some rock players. A "hard" power supply will offer more overall power since it does not sag under load. This makes the amp seem "faster" since the note's attack doesn't get rounded off. Metal players like this. Some amps even had a switch to select a tube or solid state rectifier allowing the player to choose his sound.
Installing the choke and PIO will tend to make the rectifier circuit harder with less sag. I have found that you can simulate all day long, but you just got to play with some parts to find out what works for you. If you tend toward a cleaner sound and don't dial up 11 daily, add the choke. If you want some scream, leave it out. If you don't know.....add another switch, after all this one doesn't have a $100 budget.
The power supply has a lot to do with the tone quality of a guitar amp. Many of the old amps that we know and love either used a CRC filter, or if a choke was used, the B+ tap on the OPT is connected before the choke. These choices were for cost reasons but the created part of the sound that we have grown to like, and try to recreate.
The CRC is cheap, electrolytics were not the greatest back then, and the usual size was a 16 to 30 uF with a 40 uF being huge.
The B+ tap before the choke allows a much cheaper choke to be used since most of the amps B+ current flows to the plates of the output tubes and does not go through the choke.
Both of these techniques unintendedly increase the B+ supplies tendency to sag under load. This increases the amps ability to sustain a note. Why? The designers of the old stuff never intended them to be driven far into saturation, but when you overdrive an amp the B+ current goes up reducing the B+ voltage. The drop in B+ reduces the gain and headroom through the entire amp including the preamp, increasing the distortion. As the note dies out the overload diminishes, reducing the B+ current, which increases the B+ voltage, and boosts the gain. This dynamic compression effect has a lot to do with the power supply.
Some ripple can be tolerated in a P-P amp. Again it will affect the sound. At "normal" playing volumes the ripple voltage is less than the B+ headroom in the output stages and the ripple doesn't have much effect. When the amp is cranked to 11 and driven with 97 effects boxes chained together the output stage is attempting to act like a pair of switches chopping the B+ up and feeding it to the speaker through the OPT. There is no more ripple cancellation and much of the ripple gets passed right on through to the speaker. You don't hear it because the speaker is eating the full cranked to 11 power of the amp, but its effects are there in the form of IMD. The guitar's tone has been modulated with 60, 120, 180 Hz (and so on) tones, but only on the signal peaks when the amp is clipping (or near clipping). This can be seen on an FFT analyzer, and it imparts a grainy sound to the distorted peaks.
A "soft" power supply will increase the dynamic compression and IMD effects and is often favored by blues and some rock players. A "hard" power supply will offer more overall power since it does not sag under load. This makes the amp seem "faster" since the note's attack doesn't get rounded off. Metal players like this. Some amps even had a switch to select a tube or solid state rectifier allowing the player to choose his sound.
Installing the choke and PIO will tend to make the rectifier circuit harder with less sag. I have found that you can simulate all day long, but you just got to play with some parts to find out what works for you. If you tend toward a cleaner sound and don't dial up 11 daily, add the choke. If you want some scream, leave it out. If you don't know.....add another switch, after all this one doesn't have a $100 budget.
Tubelab.com makes a very good point - it isn't just about sag here, it's quite posible that ripple really does make a contribution to the sound of a great guitar amp - have a look at this:
AC Ripple in a Class AB Power Amp
These fine fellows got to play about with the power supply to a Fender Bassman 5F6-A, with facilities provided by a german university, supposedly to try to "reduce noise".
I have to say that I have some suspicions that maybe they might have known that the way to reduce hum at least, in a fender amp is to re-work it so that the chassis isn't used as signal conductor!
Getting back to the OP's ripple reduction, the conclusions of the Berlin Technical University Project WILDCAT group were that getting rid of all ripple, even if sag is provided, wrecks the sound of the amp!
The article doesn't state what method was used for this necessarily subjective assesment, which is a shame - but it's noteworthy that their results were apparently a surprise to them (often a good sign).
Having studied a little of the acoustics of musical instruments I find it quite feasible that power line harmonic intermodulation could be an integral part of the sound of good guitar amps.
I have even heard it suggested that different power line frequencies might contribute to differences between "US" and "British" recorded rock guitar sounds.
Furthermore, I suggest that there might be musical value in providing power with not only programmable sag, but also ripple, adjustable for depth, waveform and frequency.
This is clearly something that needs further research, and any university willing to offer a suitably attractive package, first rate facilities and free beer to a technically literate long-haired oddball to boost their reputation and help them seem hip should contact me immediately.
AC Ripple in a Class AB Power Amp
These fine fellows got to play about with the power supply to a Fender Bassman 5F6-A, with facilities provided by a german university, supposedly to try to "reduce noise".
I have to say that I have some suspicions that maybe they might have known that the way to reduce hum at least, in a fender amp is to re-work it so that the chassis isn't used as signal conductor!
Getting back to the OP's ripple reduction, the conclusions of the Berlin Technical University Project WILDCAT group were that getting rid of all ripple, even if sag is provided, wrecks the sound of the amp!
The article doesn't state what method was used for this necessarily subjective assesment, which is a shame - but it's noteworthy that their results were apparently a surprise to them (often a good sign).
Having studied a little of the acoustics of musical instruments I find it quite feasible that power line harmonic intermodulation could be an integral part of the sound of good guitar amps.
I have even heard it suggested that different power line frequencies might contribute to differences between "US" and "British" recorded rock guitar sounds.
Furthermore, I suggest that there might be musical value in providing power with not only programmable sag, but also ripple, adjustable for depth, waveform and frequency.
This is clearly something that needs further research, and any university willing to offer a suitably attractive package, first rate facilities and free beer to a technically literate long-haired oddball to boost their reputation and help them seem hip should contact me immediately.
I found a link to the German university in a mainstream engineering publication several years ago. I read it and set out to play and yes you can hear, and see (with an FFT analyzer) the effects of ripple in a guitar amp.
The report you linked mentioned that a class A P-P amp is relatively immune from ripple. This is true, but most of the "class A" P-P guitar amps sold are NOT class A amps (AC 30 comes to mind) and NO amp, not even an SE amp remains in class A when it is driven into hard clipping.
I have had the same thought.
That one too. I have attempted to build a mosfet based "Sagulator" to emulate these effects but I believe it is a bit more complex than I first thought. Merely imposing a sine wave on the B+ line does not get the sound right. It has to look like power supply ripple. I think a programmable power supply "vibrato" circuit might be worth an experiment.
While chasing my masters degree and possible PHD in engneering, I mentioned teaching a course in vacuum tube audio. Despite having several interested students, the best I could get out of the engineering department was a laugh.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.