Viewing low power SET schematics (45, 300B, 2A3, others) thinking I'll dive in and build one, I wondered, hey, why not regulate the power supply (input/driver + output)? If regulating line-level power supplies increases the stage's musicality (unquestionably does), regulating a SET output power supply very likely will benefit the circuit by equal, perhaps greater amount. The SET topology, moreover, seems ideally suited to output power supply regulation given the modest current demands of the output tube.
Has anyone any experience regulating SET output power supplies? My personal preference leans to tube-based regulation, but I welcome any comments.
Has anyone any experience regulating SET output power supplies? My personal preference leans to tube-based regulation, but I welcome any comments.
Well, one thing to remember about SET is that it is class-A operation, and so the current draw of the output stage is fairly close to constant, thereby reducing the need for regulation. A class-AB push-pull stage, by contrast, has a constantly varying current draw, and actually would seem to benefit more from a regulated supply.
That said, you could go ahead and regulate. The problem is finding a regulator that can supply something like 400-600 volts and 120 to 150 mils of current, and not sound worse in the signal path (it is in the signal path) than the rest of the SET amp.
Good luck with the project, and keep us posted - Pat
That said, you could go ahead and regulate. The problem is finding a regulator that can supply something like 400-600 volts and 120 to 150 mils of current, and not sound worse in the signal path (it is in the signal path) than the rest of the SET amp.
Good luck with the project, and keep us posted - Pat
tubesguy said:
Hi Pat,
I've recently thought about that notion that the power supply is in the signal path. I agree, an amplifier merely reforms and modulates electricity coming from the wall, so everything---the entire power supply, all the cables, etc---is in series with the signal path. All other things being equal, a simpler signal path is IME preferred. But it occurred to me perhaps the better amplifier might one having the simplest, shortest signal path per se, combined with a regulated (and thus a somewhat complicated) power supply. My experiences experimenting with mains regulation seem to bear this out. In my system, I have a balanced transformer plugged into a PS Audio P600 plugged into a balanced transformer plugged into an Exactpower unit. I noticed better sound adding each additional component, and have many times tested this perception by simply plugging my gear directly into the wall. Here's the interesting implication: all these components (P600 etc), which combine to form a rather complex power supply, are in series with the signal path. It seems, then, that power supply regulation be a good thing, assuming it's done right.
Tom
tubesguy said:
Pat, are you speaking from experience here? My intuition tells me amplification class is probably not relevant to whether regulation might improve a circuit. What we want to know is whether non-amplification-related variations in power supply current (and voltage) are reduced by power supply regulation. I realise that in a class A circuit those variations will be a small percent of overall current drawn. Those variations, on the other hand, might constitute a quite audible percentage of overall (amplification-related + non-amplification-related) current variations. Voltage stability is a somewhat separate but related concern.
Tom
Hi,
If you must use alot of PS filtering_never hurts_put those big caps either in front or inbetween the chokes.
Use smaller value polypropylenes after the last choke for best sound.
Same goes for active regulation, bigger caps infront, smaller but fast caps behind the reg.
Beware not to put a large cap right behind a tube rectifier or you'll have yourself some fireworks.
Most people don't seem to distinguish between chokes for PS duty and anode loading anymore, using one for the other.
To my mind this is a major mistake.
With SE designs everything you use in the amp is audible to some extend.
When you decide to regulate one end you'd better regulate the lot or nothing at all IMHO.
Cheers,
If you must use alot of PS filtering_never hurts_put those big caps either in front or inbetween the chokes.
Use smaller value polypropylenes after the last choke for best sound.
Same goes for active regulation, bigger caps infront, smaller but fast caps behind the reg.
Beware not to put a large cap right behind a tube rectifier or you'll have yourself some fireworks.
Most people don't seem to distinguish between chokes for PS duty and anode loading anymore, using one for the other.
To my mind this is a major mistake.
With SE designs everything you use in the amp is audible to some extend.
When you decide to regulate one end you'd better regulate the lot or nothing at all IMHO.
Cheers,
How much resistance is there in the output transformer primary? 100-300 ohms? This resistance appears in series with the DC supply to the anode and even if you had a perfectly regulated supply the actual regulation would be limited by the transformer resistance.
A truly adventurous person might experiment with a bit of *negative resistance* in the power supply to dial out the effect of the transformer winding...
A truly adventurous person might experiment with a bit of *negative resistance* in the power supply to dial out the effect of the transformer winding...
Try the same power supply that produces the ringing at turn on and then use the stepped function of the constant current load. For example: if the normal load is 70 ma at idle, set the load current to 50 ma. After the output voltage has settled down step the load current to 90 ma. This is done by delaying the current step in the setup part of the constand current load. I think you will find that a change in load current will cause ringing also.
As you know current is not constant in a class A amp. The AVERAGE current is constant but the actual instantaneous current is always changing with program material.
The ringing seen shows us that at some frequency the power supply is resonant and has a high impeadance at that frequency. What we want is a low impedance at all frequencies so ringing is BAD.
My own preference is to use active regulation. I have built several tube amps with solid state regulators or modified tube amps to use solid state regulated +B supplies and as far as I can tell the amps become much quieter as far as noise floor and hum, the bandwidth is greater ( this you can measure) and the amp just sounds cleaner and tighter not loose and unfocused.
OK now you have my vote.
Later BZ
As you know current is not constant in a class A amp. The AVERAGE current is constant but the actual instantaneous current is always changing with program material.
The ringing seen shows us that at some frequency the power supply is resonant and has a high impeadance at that frequency. What we want is a low impedance at all frequencies so ringing is BAD.
My own preference is to use active regulation. I have built several tube amps with solid state regulators or modified tube amps to use solid state regulated +B supplies and as far as I can tell the amps become much quieter as far as noise floor and hum, the bandwidth is greater ( this you can measure) and the amp just sounds cleaner and tighter not loose and unfocused.
OK now you have my vote.
Later BZ
HDTVman said:
Hi BZ,
Not having heard a SET amp with B+ regulation, I would yet have inferred what you say about actively regulating the output supply. It seems to me the low-bandwidth, loose bass SET "sound" is not traceable to the SE of SET, or the T of SET, or to an output transformer or an interstage transformer, but to variable and highish power supply impedance of an unregulated supply. The inference seems almost a no-brainer: plus or minus so many exceptions, every other power supply is regulated, why not the B+? For SET amps, B+ regulation makes all the more sense because of the low current demands placed on the regulator.
Have you had any experience comparing transistor to tube regulators sound-wise?
Tom
Hi Tom
I can't say that I have compaired tube regulation to SS. I've always just built a SS regulator because it's cheeper and easyer. It's all about the broadband output impedance of the power supply anyway.
Later BZ
Another reason for using a regulated supply. SET amps in general don't have great Power Supply Rejection Ratios. With a regulated supply it doesn't mater.
I can't say that I have compaired tube regulation to SS. I've always just built a SS regulator because it's cheeper and easyer. It's all about the broadband output impedance of the power supply anyway.
Later BZ
Another reason for using a regulated supply. SET amps in general don't have great Power Supply Rejection Ratios. With a regulated supply it doesn't mater.
Hi,
True...One major advantage a tube regualtor has over SS is the slow turn on caused by the glowdischarge tube's inertia.
This is of course of merit when SS rectifiers are used....yes, they can be made to sound good.
It sure looks great too...
Yes, but the total quality of all the components used in the PS are still as important.
Cheers,
True...One major advantage a tube regualtor has over SS is the slow turn on caused by the glowdischarge tube's inertia.
This is of course of merit when SS rectifiers are used....yes, they can be made to sound good.
It sure looks great too...
Yes, but the total quality of all the components used in the PS are still as important.
Cheers,
I quite agree. The only way to do things is the best way you can. The quality and correctness of design of the power supply is just as important as the amplifer stages. After all it's the two working together that cause amps to "sound" a certain way. The way an amp reacts to a load/stability, it's frequency responce, the damping factor are all affected by the power supply.
By the way I always use a slow turn on with my regulators. Most of the time it's just a resistor in series with the primary and a relay that latches when the power supply has charged to 80% or so, shorting out the resistor.
Later BZ
By the way I always use a slow turn on with my regulators. Most of the time it's just a resistor in series with the primary and a relay that latches when the power supply has charged to 80% or so, shorting out the resistor.
Later BZ
serengetiplains said:
Hi Tom -
Sorry to be so late getting back. I was gone for the weekend.
My experience has been with both SETs and class AB push-pull amps. In this hypothetical, assume that both amps have a power supply with some resistive elemtents, either resistors or chokes with some finite resistance.
In the SET (class A) amp, the total current draw should be near constant. If that is the case, then signal-induced current fluctuations (what I think you're referring to as amplification-related current variations) should be near zero, always assuming a well-designed amplifier.
In the AB push-pull amp, the idling current can be fairly low, but the signal will push the amp into an operating region where the current draw is substantial. When you attempt to draw this larger current from the power supply, you get voltage sag, as a result of the resistive elements in the power supply. The higher the resistance in the PS, the larger the signal-induced B+ sag, and the more you deviate from your presumably well thought-out load lines, etc. Nasty stuff. That's why I'd suggest that the SET is in relatively less need of regulation.
On the other hand, I've done a tube-regulated SET, a spud amp using the WE417 and a big-*** surplus 6L6-based regulator. If the tube's current draw was 20 mils, I used a parallel resistor to assure a total draw from the supply of at least 80 mils per channel. It sounded pretty good.
- Pat- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.