Hi Mel
For pre-amp tubes that typically draw 10mA or less, you can get away with CCS. For "wimpy" high mu triodes a shunt can work superbly well too.
But for an end stage these strategies won't work... You're pulling 10's or even 100's of mA's so the power supply WILL need to swing.
So since it must swing, the "best" solution is a power supply that responds very quickly to current draw. So quick that you don't hear it.
In Class A amplifiers, the power supply is fully in the sound path, so designing one that does not respond quickly is just not going to work well.
Choke input is NICE since it does a lot of things for you:
1. Saves all sorts of potential energy for your amp to draw incredibly quickly as it is needed.
2. Allows you to use lower uF caps, and perhaps even Film caps, which sound WAY better. and are not nearly as parasitic as other caps during recharge. 🙂
3. Takes a huge load off your power transformer.
3. Reduces your B+ - which incidentally lets your B+ immediately "spring" up as your tubes need it.
4. Takes a ton of load off your rectifier.
5. Doesn't color the sound one little bit like some less-well considered sand solutions might. (tip of the hat to the FEW good ones out there).
6. Can even filter B+ ripple (!)
I could go on and on.... but I think you get the idea. A class A single ended power amplifier (not a pre-amplifier) LCL will sound much better than CLCL assuming it is well designed - EVERY time.
That's been my personal experience...
If you are building a phono or pre-amplifer then things are much easier with the PSU... you can use CLCRC then a shunt or constant current source, gyrator.. (good schematics exist for that) and you probably won't hear a difference.
Ian
For pre-amp tubes that typically draw 10mA or less, you can get away with CCS. For "wimpy" high mu triodes a shunt can work superbly well too.
But for an end stage these strategies won't work... You're pulling 10's or even 100's of mA's so the power supply WILL need to swing.
So since it must swing, the "best" solution is a power supply that responds very quickly to current draw. So quick that you don't hear it.
In Class A amplifiers, the power supply is fully in the sound path, so designing one that does not respond quickly is just not going to work well.
Choke input is NICE since it does a lot of things for you:
1. Saves all sorts of potential energy for your amp to draw incredibly quickly as it is needed.
2. Allows you to use lower uF caps, and perhaps even Film caps, which sound WAY better. and are not nearly as parasitic as other caps during recharge. 🙂
3. Takes a huge load off your power transformer.
3. Reduces your B+ - which incidentally lets your B+ immediately "spring" up as your tubes need it.
4. Takes a ton of load off your rectifier.
5. Doesn't color the sound one little bit like some less-well considered sand solutions might. (tip of the hat to the FEW good ones out there).
6. Can even filter B+ ripple (!)
I could go on and on.... but I think you get the idea. A class A single ended power amplifier (not a pre-amplifier) LCL will sound much better than CLCL assuming it is well designed - EVERY time.
That's been my personal experience...
If you are building a phono or pre-amplifer then things are much easier with the PSU... you can use CLCRC then a shunt or constant current source, gyrator.. (good schematics exist for that) and you probably won't hear a difference.
Ian
Last edited:
Using the stepped load in the PSU simulator is changing a DC condition.
Now monitor a class A amplifier's power supply DC current draw during normal listening conditions. It stays relatively constant, hence using a stepped load in Duncan's PSU software tells you nothing useful for a Class A amp.
Linear Power Supply Design
"Class-A Amplifiers
The other common amplifier type is Class-A. These amps draw a large current on a continuous basis, and place a completely different loading on the supply. The current pulses are gone from the supply leads, but the rectifier and filter now must handle the maximum current on a continuous basis.
The continuous load creates a new set of constraints on the design of a power supply, and the use of a Class-A amp implies that the builder already wants the very lowest noise. Although the noise of the power supply DC output (hum/ ripple) will normally be low because of extensive filtering, regulation or a capacitance multiplier, the switching noise of the diodes in the rectifier can become more than a nuisance if proper care is not taken."
You want to give the AC current a low resistance through the power supply so high quality decoupling caps are important.
With a cap input filter he just needed a bit more tweaking to get the rms current down closer to 230mA, possibly just lowering the capacitance and also adding some serial resistance with the rectifiers.
Now monitor a class A amplifier's power supply DC current draw during normal listening conditions. It stays relatively constant, hence using a stepped load in Duncan's PSU software tells you nothing useful for a Class A amp.
Linear Power Supply Design
"Class-A Amplifiers
The other common amplifier type is Class-A. These amps draw a large current on a continuous basis, and place a completely different loading on the supply. The current pulses are gone from the supply leads, but the rectifier and filter now must handle the maximum current on a continuous basis.
The continuous load creates a new set of constraints on the design of a power supply, and the use of a Class-A amp implies that the builder already wants the very lowest noise. Although the noise of the power supply DC output (hum/ ripple) will normally be low because of extensive filtering, regulation or a capacitance multiplier, the switching noise of the diodes in the rectifier can become more than a nuisance if proper care is not taken."
You want to give the AC current a low resistance through the power supply so high quality decoupling caps are important.
With a cap input filter he just needed a bit more tweaking to get the rms current down closer to 230mA, possibly just lowering the capacitance and also adding some serial resistance with the rectifiers.
Based on Soulmerchant's advice and lots of reading at VinylSavor's blog I'm pretty sold on choke input.
In searching for power transformers I failed to appreciate that my mains voltage (124-126VAC) is going to significantly alter my secondaries. I'd like to use the Lundahl 1651--it's spec'd at 500V unloaded without a CT rated from 115V on the primary. It's also generously overrated at 430mA. Given my increased mains voltage this puts me at 543.5V unloaded on the secondary. It also offers four additional 6.3V filament taps, so I'll only need to buy one additional filament transformer (two taps for the 300Bs, one for both drivers, and two for the 6AX4-GTB bridge). I'll likely have to use voltage dropping resistors on the filaments due to the increased mains, putting me at about 6.8V.
I'll use a Lundahl 1673 for the input choke (10H, 200mA recommended, saturates at 290mA), followed by maybe a cheaper Hammond choke or another Lundahl, I'm not sure yet.
Does anyone foresee a problem using this setup? My primary concern is that I'm depending on my elevated mains voltage to boost the secondary to 543.5V unloaded. Dipping any lower puts by B+ under 400 VDC and I'd rather not go that low for my 300Bs.
In searching for power transformers I failed to appreciate that my mains voltage (124-126VAC) is going to significantly alter my secondaries. I'd like to use the Lundahl 1651--it's spec'd at 500V unloaded without a CT rated from 115V on the primary. It's also generously overrated at 430mA. Given my increased mains voltage this puts me at 543.5V unloaded on the secondary. It also offers four additional 6.3V filament taps, so I'll only need to buy one additional filament transformer (two taps for the 300Bs, one for both drivers, and two for the 6AX4-GTB bridge). I'll likely have to use voltage dropping resistors on the filaments due to the increased mains, putting me at about 6.8V.
I'll use a Lundahl 1673 for the input choke (10H, 200mA recommended, saturates at 290mA), followed by maybe a cheaper Hammond choke or another Lundahl, I'm not sure yet.
Does anyone foresee a problem using this setup? My primary concern is that I'm depending on my elevated mains voltage to boost the secondary to 543.5V unloaded. Dipping any lower puts by B+ under 400 VDC and I'd rather not go that low for my 300Bs.
You can use a graetz bridge (i.e. virtual center tap using silicon diodes) and two damper diodes and two silicon diodes. I do this all the time (not because of the cost, because I am too lazy to put in so many sockets).
To simulate a graetz bridge, I use the full wave rectifier option in PSU - it measures pretty close to a "full wave" rectifier in my experience.
I like the lower input cap in your simulation. Hammond chokes can work too, but I prefer Lundahl. Thomas has his own line of mains transformers, and I find them to be excellent quality and easier to work with than Lundahl. They might even be cheaper too..
I like to use a smaller value first choke, then a not-too-big C1, then a bigger filtering choke followed by bigger C2.
To simulate a graetz bridge, I use the full wave rectifier option in PSU - it measures pretty close to a "full wave" rectifier in my experience.
I like the lower input cap in your simulation. Hammond chokes can work too, but I prefer Lundahl. Thomas has his own line of mains transformers, and I find them to be excellent quality and easier to work with than Lundahl. They might even be cheaper too..
I like to use a smaller value first choke, then a not-too-big C1, then a bigger filtering choke followed by bigger C2.
Hi, yes the Class A definition is of course textbook correct. It makes certain important inferences. I particularly like this one:
Now read this:
If it sounds incredible, do the maths yourself. Low DCR in the power supply will clearly be important in Class A (as you note).
Also, can the mains transformer alone deliver the goods to faithfully reproduce the transients that makes our Class A amplifier truly Hi-Fi? What else can be done? Besides filter ripple, a choke can also be used to store potential energy, and release it VERY quickly. Not using this strategy is in my humble opinion, a pity. Usually people want 300b because they want some Power! 🙂
Also, if the mains transformer is already close to max current rating, then it really won't cut it when actual music is being played on that Class A amplifier. It will get VERY WARM. Performance will decline.. It might even vibrate.
I know since I did stuff like that with Hammond transformers in the days before internet.
Now read this:
If it sounds incredible, do the maths yourself. Low DCR in the power supply will clearly be important in Class A (as you note).
Also, can the mains transformer alone deliver the goods to faithfully reproduce the transients that makes our Class A amplifier truly Hi-Fi? What else can be done? Besides filter ripple, a choke can also be used to store potential energy, and release it VERY quickly. Not using this strategy is in my humble opinion, a pity. Usually people want 300b because they want some Power! 🙂
Also, if the mains transformer is already close to max current rating, then it really won't cut it when actual music is being played on that Class A amplifier. It will get VERY WARM. Performance will decline.. It might even vibrate.
I know since I did stuff like that with Hammond transformers in the days before internet.
Last edited:
But you're still not taking into account the stored energy of the output transformer...
You don't seem to be taking into account that the output transformer does this as well.. You're talking about the energy stored in the choke in the B+ supply.. What about the energy stored in the output transformer? The power stage draws current from here.. Does it not?
No, we have not got that far yet, but of course the OPT is Also in the signal path. if you want to simulate the whole circuit then you need to use spice. There are people who only ever seem to simulate and never build anything though...
I personally prefer to build things.
A choke input filter will not have any regulation against line fluctuations, it is only good at regulating large current (load) changes, which his 300b amp won't have.
To say a 1 watt amp needs a reserve power of 100 watts for transients made me lol.
If the OP wants to build a choke input supply so be it, but I just don't think it has much advantage in this application.
To say a 1 watt amp needs a reserve power of 100 watts for transients made me lol.
If the OP wants to build a choke input supply so be it, but I just don't think it has much advantage in this application.
But aren't we that far?? If you're simulating changes in the amount of current being drawn by the power stage, then you should be taking into account the output transformer, and its stored energy. The anode of the power tube is connected to one side of the primary of the output transformer, B+ is connected to the other side of the primary. Current drawn by the power stage is via the output transformer. The B+ supply is only part of the equation, and so simulating what will happen to the B+ supply without taking into account the output transformer and its stored energy doesn't really tell the whole story in my opinion.
Last edited:
Hoj
I agree with you! But look at the original poster's post and explain how he can put his the OPT into PSU-II. 😉
I would suggest that if you use cheaper OPT (say 16H primary inductance or less) then choke input becomes a must. Why? because it can make your amp sound like you are using a far higher quality OPT.
The common mistake by many is to think that the power supply is not in the signal path. I tried to explain this in different ways. Some posters here don't want to believe it, but:
In single-ended Class A amplifiers the power supply is fully in the signal path.
That pretty much sums it up. Its a mistake to think that the load is anything else but inductive. You need to realize that if you plot a load line with a ruler that is not the way the amplifier is working. Its only a design starting point.
Last time I checked, expensive OPT's can be really expensive... but decent chokes are not that expensive. And anyway, if you are spending $1'000 or more on fancy OPT's then you can certainly afford to build a choke input supply while you're at it.
If you use spice you can consider these specs as pretty darn good for 300b OPT:
SA08 SINGLE ENDED OUTPUT TRANSFORMER
Primary induction 25H - VERY nice specs for 300b - unfortunately not cheap. Most 3.5k OPT's designed for 300b are around 16H.
There is a whole thread on best OPT's for 300b here. Its an interesting read. George (TubeLab) also has a section on budget OPT's that quite useful for US hobbyists.
I agree with you! But look at the original poster's post and explain how he can put his the OPT into PSU-II. 😉
I would suggest that if you use cheaper OPT (say 16H primary inductance or less) then choke input becomes a must. Why? because it can make your amp sound like you are using a far higher quality OPT.
The common mistake by many is to think that the power supply is not in the signal path. I tried to explain this in different ways. Some posters here don't want to believe it, but:
In single-ended Class A amplifiers the power supply is fully in the signal path.
That pretty much sums it up. Its a mistake to think that the load is anything else but inductive. You need to realize that if you plot a load line with a ruler that is not the way the amplifier is working. Its only a design starting point.
Last time I checked, expensive OPT's can be really expensive... but decent chokes are not that expensive. And anyway, if you are spending $1'000 or more on fancy OPT's then you can certainly afford to build a choke input supply while you're at it.
If you use spice you can consider these specs as pretty darn good for 300b OPT:
SA08 SINGLE ENDED OUTPUT TRANSFORMER
Primary induction 25H - VERY nice specs for 300b - unfortunately not cheap. Most 3.5k OPT's designed for 300b are around 16H.
There is a whole thread on best OPT's for 300b here. Its an interesting read. George (TubeLab) also has a section on budget OPT's that quite useful for US hobbyists.
Last edited:
I quoted that from the TubeLab site linked in an earlier post. Kinda outrageous sounding isn't it?
After I read it the first time, I actually sat down and did some maths. 😉
You're right. He can't put them into PSUDII.. And so all I'm saying is that simulating changes in current with PSUDII isn't really simulating what will happen in the amp..
You make it sound like the whole power supply is in the signal path which to my knowledge is completely wrong. The only thing that can be superimposed into the signal is the 120Hz current pulses but that is not what we are talking about. The signal current loop is as shown.
As I said before the decoupling caps in the PS need to be of high quality because they are in the signal path.
Yes, for single ended Class-A power amplifiers the whole power supply is in the signal path. Far more than the last filter cap in your illustration above.
I would suggest you try it some time. From your previous posts it is clear that you didn't try it yet...
If you must, please go ahead and ask diy members with more experience though. Start a new thread perhaps if you feel inclined too.
I would suggest you try it some time. From your previous posts it is clear that you didn't try it yet...
If you must, please go ahead and ask diy members with more experience though. Start a new thread perhaps if you feel inclined too.
Last edited:
- Status
- Not open for further replies.