Hey all, first post here.
I'm working on designing my first guitar tube amp. I can't seem to find much information out there to answer this question.
I know that guitar amplifiers typically connect the plate supply right to the first reservoir capacitor in the power supply, so that the choke handles less current. The Hammond 159S choke has a low DCR, 4H of inductance, and can handle a maximum of 220mA of current. Its low DCR and high current carrying ability make me wonder - are there any drawbacks to putting my plate supply after this choke and then running a dropping RC filter to the screen? My Imax is right below the max rating, and my quiescent current is much lower. In my PSUDII simulations, this works out well. In reality, that might be different. Any thoughts? For example, the high current flowing through the choke might couple with the output transformer and induce more noise than simply dealing with a bit of ripple. However, I'm using a 5U4GB with a 40u reservoir and the ripple is quite substantial before the choke (about 20Vpp).
I'm working on designing my first guitar tube amp. I can't seem to find much information out there to answer this question.
I know that guitar amplifiers typically connect the plate supply right to the first reservoir capacitor in the power supply, so that the choke handles less current. The Hammond 159S choke has a low DCR, 4H of inductance, and can handle a maximum of 220mA of current. Its low DCR and high current carrying ability make me wonder - are there any drawbacks to putting my plate supply after this choke and then running a dropping RC filter to the screen? My Imax is right below the max rating, and my quiescent current is much lower. In my PSUDII simulations, this works out well. In reality, that might be different. Any thoughts? For example, the high current flowing through the choke might couple with the output transformer and induce more noise than simply dealing with a bit of ripple. However, I'm using a 5U4GB with a 40u reservoir and the ripple is quite substantial before the choke (about 20Vpp).
If you are talking about a choke input PSU then be aware that this (if sufficiently loaded) puts out a DC voltage of around 0.9xRMS rather than the 1.4xRMS of a cap input PSU. When insufficiently loaded (that is, too little current is being drawn) the choke input supply will have a highish output impedance - DC voltage will vary strongly with DC current. At low currents it will give 1.4xRMS.
My main thought is 'why'?
He's building a valve guitar amp, so reduces the power and increases the 'quality' - neither of which are wanted in a guitar amp.
He's building a valve guitar amp, so reduces the power and increases the 'quality' - neither of which are wanted in a guitar amp.
This is 100% correct. 🙂
I'm still in the stage of conceptualizing and simulating my designs. I haven't bought any transformers or chokes yet, or filter caps for that matter. The amp isn't high power - only 30W from two 6L6GCs. I play mostly clean but I'd like a bit of breakup.
Having a signal that's free of power supply ripple noise improves the quality, yes, but I can't see any reason why eliminating this type of signal interference would make it sound worse as a guitar amp.
I asked this question because I'm using a NOS GE 5U4GB that I inherited as my rectifier (actually, I'm trying to design a switchable solid state/valve rectifier), but the 5U4GB datasheet quotes a maximum 40uF reservoir cap. My PSUDII simulations are showing there's a significant amount of ripple across this (relatively) low capacitor, on the order of 10-20Vpp, so I'm trying out ideas to smooth this without increasing my reservoir cap size.
Lets just keep any personal feelings out of this discussion. Offending posts have been removed.
Keep it on topic folks.
Because you're altering the sound by 'improving' it - thereby getting rid of some of the 'valve sound' you're presumably looking for?.
Just look how most commercial amps do it - there are good solid reasons for it (as well as cost 😀) - mainly that 'improving' means it doesn't sound as 'good' as existing amps.
Experimenting to find what works best for you is a good course of action. You can even experiment with a couple of different chokes to see the effects of differing DCR on the supply regulation (sag) in order to tailor the sound to your liking.
You don't have to live with hum if that is not part of the aesthetic for you.
I had a friend some years back when I worked in the MI industry who loved designing tube guitar amps, and designed quite a number of good sounding ones, and none of them hummed or were otherwise noisy, but sounded great. (He had a definite local following)
You don't have to live with hum if that is not part of the aesthetic for you.
I had a friend some years back when I worked in the MI industry who loved designing tube guitar amps, and designed quite a number of good sounding ones, and none of them hummed or were otherwise noisy, but sounded great. (He had a definite local following)
You're absolutely right, I'll be experimenting with the position of the choke to see where it sounds best. Leaving it in front of the plate will make the amp hum less and sag more, while putting it between the plate and screen will do the opposite. There's nothing more important than tweaking it once it's built, even with good theoretical design.
I'm still a bit surprised how much ripple there was across the reservoir cap in my simulations. Since the datasheet quotes a max of 40uF, surely it's been noticed for 5U4's using capacitor-input filters before.
Ian's Rave - I don't expect everyone to agree with me (how boring would that be)!!
Guitar players will put up with inferior power supply section and so most amp manufacturers supply just that by designing down to a price rather than up to a standard.
When building for yourself there is no reason to copy this cost driven philosophy and good reasons not to do so.
Power Supply ripple will partly cancel in the push pull output transformer, dependent on how well matched the output tubes are and on balancing the idle currents in the push pull halves.
Intermodulation products between the signal and the residual power supply ripple will NOT cancel and so a better power supply will nearly always reward you with a sound that is subjectively like a whole veil has been lifted out of the way. This is because the signal tones intermodulate with the power supply ripple to produce signal tone frequency +/- 120Hz , signal tone +/- 240 Hz, signal tone +/- 360Hz etc. because the residual power supply ripple is not just 2 x the mains frequency (that is 120Hz) but contains a heap of harmonics of that 120Hz as well.
The choke will act to reduce the 120Hz ripple but will act even better to get rid of the ripple harmonics (the 240, 360, 480 Hz etc.).
Output Tubes current is dependent mostly upon the screen voltage, the anode supply has a much lesser affect. That is why on budget driven design they will not have a choke at all or will put the choke after the main B+ feed to the Centre Tap of the output transformer. That way they can get away with a much smaller and hence much cheaper choke.
If you don't mind spending the extra cash for a bigger choke (more current capability) then you WILL be rewarded with a clearer more articulate sound by putting the choke ahead of the B+ feed to the Output Tranny.
The choke will allow you to keep the filter capacitor sizes within sensible limits too and if you really want to "go the whole hog" allow you to use a polyproylene filter cap. on the Output Tranny B+ feed without breaking the piggy bank.
These are the final 5% things and so many do not bother with them, particularly in a commercial cost driven design, but to say that they make no difference at all is in my opinion wrong.
On my own design/builds I use a choke before the B+ feed to the output tranny, I use Polypropylene final filter capacitor for this node
AND (more controversially)
I add a common unbypassed resistor in the output tube cathodes of approx 25% of the value you would use for cathode biasing. This generates some common mode feedback in the output tube pair which helps supress some 3rd and other odd harmonic distortion but is particularly effective at supressing intermodulation between the signal and the power supply ripple. I learned this from HiFi Amp design but it works well in Git. Amps too.
Cheers,
Ian
Guitar players will put up with inferior power supply section and so most amp manufacturers supply just that by designing down to a price rather than up to a standard.
When building for yourself there is no reason to copy this cost driven philosophy and good reasons not to do so.
Power Supply ripple will partly cancel in the push pull output transformer, dependent on how well matched the output tubes are and on balancing the idle currents in the push pull halves.
Intermodulation products between the signal and the residual power supply ripple will NOT cancel and so a better power supply will nearly always reward you with a sound that is subjectively like a whole veil has been lifted out of the way. This is because the signal tones intermodulate with the power supply ripple to produce signal tone frequency +/- 120Hz , signal tone +/- 240 Hz, signal tone +/- 360Hz etc. because the residual power supply ripple is not just 2 x the mains frequency (that is 120Hz) but contains a heap of harmonics of that 120Hz as well.
The choke will act to reduce the 120Hz ripple but will act even better to get rid of the ripple harmonics (the 240, 360, 480 Hz etc.).
Output Tubes current is dependent mostly upon the screen voltage, the anode supply has a much lesser affect. That is why on budget driven design they will not have a choke at all or will put the choke after the main B+ feed to the Centre Tap of the output transformer. That way they can get away with a much smaller and hence much cheaper choke.
If you don't mind spending the extra cash for a bigger choke (more current capability) then you WILL be rewarded with a clearer more articulate sound by putting the choke ahead of the B+ feed to the Output Tranny.
The choke will allow you to keep the filter capacitor sizes within sensible limits too and if you really want to "go the whole hog" allow you to use a polyproylene filter cap. on the Output Tranny B+ feed without breaking the piggy bank.
These are the final 5% things and so many do not bother with them, particularly in a commercial cost driven design, but to say that they make no difference at all is in my opinion wrong.
On my own design/builds I use a choke before the B+ feed to the output tranny, I use Polypropylene final filter capacitor for this node
AND (more controversially)
I add a common unbypassed resistor in the output tube cathodes of approx 25% of the value you would use for cathode biasing. This generates some common mode feedback in the output tube pair which helps supress some 3rd and other odd harmonic distortion but is particularly effective at supressing intermodulation between the signal and the power supply ripple. I learned this from HiFi Amp design but it works well in Git. Amps too.
Cheers,
Ian
Last edited:
I haven't really done any choke input supplies, but: wouldn't it introduce additional sag? This would probably be a wanted thing for many.
"I add a common unbypassed resistor in the output tube cathodes of approx 25% of the value you would use for cathode biasing. This generates some common mode feedback in the output tube pair which helps supress some 3rd and other odd harmonic distortion but is particularly effective at supressing intermodulation between the signal and the power supply ripple. I learned this from HiFi Amp design but it works well in Git. Amps too."
I don't fully understand. Could you supply a diagram and explanation of the signal flow? It sounds more like a 'sag' resistor to add compression, but with a totally different explanation, so I'm kind of lost.
I don't fully understand. Could you supply a diagram and explanation of the signal flow? It sounds more like a 'sag' resistor to add compression, but with a totally different explanation, so I'm kind of lost.
Local feedback reduces distortion. Unbypassed cathode resistors create local feedback. He uses this effect to deal with intermodulation distortion from the PSU.
- Status
- Not open for further replies.