These amps have a PWM switching power supply but a more conventional MOSFET amplifier. I'm considering adding more big caps to the power supply output caps, across the rails between the power supply and the amplifier stages. I'm hoping it improves the amp's ability to handle big short-duration (20-millisecond) demands.
I accumulated 5 of the Soundcraftsmen PR1800 amps also rebadged as the MTX 900X2. These are the bigger brother (twice the output, twice the power) commercial versions (balanced inputs) of the more popular half-rack sized PCR800 and PCR860 (1/4 inch inputs only). There was also a PCR1600 (no balanced inputs) of which there were even fewer made.
The original review of the baby-brother half-sized PCR800 and PCR860 amps by Hircsh-Houck Labs in Stereo Review in 1984 said that "thermal sensors on critcal parts of the amplifier will increase the fan speed with rising temperature...continued high-level operation causes the supply gradually to reduce the voltage to the output stages, thereby limiting the maximium deliverable power and the corresponding heat generated to safe values without shuting the amp down."
This might be a different viewpoint on the same reality...possibly rephrased as that there are sufficient MOSFETs and fan-cooled heatsink to accomodate the full output capability of the switching power supply. Beyond that, with extended demand the power supply output drops (whether on purpose or not I really don't know).
The review also said: "When we drove the PCR800 at one-third rated power for the standard FTC pre-conditioning process, its fan switched to high speed within a minute or two. Before the required hour had passed, the PCR circuits had reduced the amps supply voltage sufficiently to limit the output to perhaps 30 or 40 watts"... This I don't quite understand, as I really expected the reviewer's output figures to be for continuous output. I guess just under an hour is nearly continuous for most real material crescendos. I've owned several varieties of these Soundcraftsmen amps with switching-mode supplies and I've never seen one in this reduced-output mode.
The review's figures show that these amps have very little dynamic headroom with 20-millisecond tone bursts.
As background, I like class "D" amps, but have to adjust my required specs for them because they do not accomodate the short-term volume demands of music. A switching-mode amp that can output 1000 watts often cannot make a decent-sounding 1002 watts output for even a very short sudden note attack. A more conventional 1000 watt amp might handle a very short demand and reproduce a decent if slightly compressed output far exceeding its continuous rating. So when I buy a switching-mode amp I usually get about twice the continuous power rating I would for a more conventinal amp. Those phony-baloney short-term burst power figures that unscrupulous advertisers post are sometimes actually a valuable real-world number for comparing amps if the conditions are carefully stated.
Despite the conventional MOSFET amplifier stages, these Soundcraftsmen switching-mode power supplies behave a little like those class-D amps. The power supply is fully regulated, so within its operating range it avoids compression due to dips in the rail voltage. Nice and pnchy as long as you don't approach the power limits. But though it has two large electrolytics, they are not all that big for an 1800 watt amp. Like Hirsch indicated, they don't handle a big 20-millisecond demand.
The early amps have been upgraded to the late-model 'soft-start' circuitry to limit the inrush current when you turn the switch on.
I'm considering adding more large electrolytics to the power supply. Would there be reliability implications? Would it likely improve that dynamic headroom figure? There's a LOT of MOSFETs in these things (and I'm not certain as my memory fails me but I think they're in some decent old-school transistor cases that transfer heat better than the convenient modern tab case formats); would the additional electrolytics be likely to stress the MOSFETs and cause them to deteriorate? Is it likely to over-stress the PWM power supply? When the additional caps really come into play, the rail voltage will drop less but take longer to recover; is this likely to be audible as a delayed "breathing" compression? Would that be more objectionable than the more immediate compression that probably occurs now?
I accumulated 5 of the Soundcraftsmen PR1800 amps also rebadged as the MTX 900X2. These are the bigger brother (twice the output, twice the power) commercial versions (balanced inputs) of the more popular half-rack sized PCR800 and PCR860 (1/4 inch inputs only). There was also a PCR1600 (no balanced inputs) of which there were even fewer made.
The original review of the baby-brother half-sized PCR800 and PCR860 amps by Hircsh-Houck Labs in Stereo Review in 1984 said that "thermal sensors on critcal parts of the amplifier will increase the fan speed with rising temperature...continued high-level operation causes the supply gradually to reduce the voltage to the output stages, thereby limiting the maximium deliverable power and the corresponding heat generated to safe values without shuting the amp down."
This might be a different viewpoint on the same reality...possibly rephrased as that there are sufficient MOSFETs and fan-cooled heatsink to accomodate the full output capability of the switching power supply. Beyond that, with extended demand the power supply output drops (whether on purpose or not I really don't know).
The review also said: "When we drove the PCR800 at one-third rated power for the standard FTC pre-conditioning process, its fan switched to high speed within a minute or two. Before the required hour had passed, the PCR circuits had reduced the amps supply voltage sufficiently to limit the output to perhaps 30 or 40 watts"... This I don't quite understand, as I really expected the reviewer's output figures to be for continuous output. I guess just under an hour is nearly continuous for most real material crescendos. I've owned several varieties of these Soundcraftsmen amps with switching-mode supplies and I've never seen one in this reduced-output mode.
The review's figures show that these amps have very little dynamic headroom with 20-millisecond tone bursts.
As background, I like class "D" amps, but have to adjust my required specs for them because they do not accomodate the short-term volume demands of music. A switching-mode amp that can output 1000 watts often cannot make a decent-sounding 1002 watts output for even a very short sudden note attack. A more conventional 1000 watt amp might handle a very short demand and reproduce a decent if slightly compressed output far exceeding its continuous rating. So when I buy a switching-mode amp I usually get about twice the continuous power rating I would for a more conventinal amp. Those phony-baloney short-term burst power figures that unscrupulous advertisers post are sometimes actually a valuable real-world number for comparing amps if the conditions are carefully stated.
Despite the conventional MOSFET amplifier stages, these Soundcraftsmen switching-mode power supplies behave a little like those class-D amps. The power supply is fully regulated, so within its operating range it avoids compression due to dips in the rail voltage. Nice and pnchy as long as you don't approach the power limits. But though it has two large electrolytics, they are not all that big for an 1800 watt amp. Like Hirsch indicated, they don't handle a big 20-millisecond demand.
The early amps have been upgraded to the late-model 'soft-start' circuitry to limit the inrush current when you turn the switch on.
I'm considering adding more large electrolytics to the power supply. Would there be reliability implications? Would it likely improve that dynamic headroom figure? There's a LOT of MOSFETs in these things (and I'm not certain as my memory fails me but I think they're in some decent old-school transistor cases that transfer heat better than the convenient modern tab case formats); would the additional electrolytics be likely to stress the MOSFETs and cause them to deteriorate? Is it likely to over-stress the PWM power supply? When the additional caps really come into play, the rail voltage will drop less but take longer to recover; is this likely to be audible as a delayed "breathing" compression? Would that be more objectionable than the more immediate compression that probably occurs now?
You have to remember that an SMPS runs at high frequency so the ripple is high frequency when compared to a linear mains PSU and so the time taken to recharge the reservoir caps is much shorter.
You really need to put a scope on the rails, with the amp delivering full power at say 20hz and see what the supplies look like.
I can understand what you are asking... but is this a real or imaginary issue... thats why you need to scope the supply.
Hard to say whether more capacitance would upset the PSU... its during start up where there may be problems where the capacitance may be seen as a "short" or low impedance load on the rails. What happens then depends on the PSU... probably nothing more than a few fractions of a second to bring the rails up... or does it see the extra capacitance as a short or low impedance load and go into a limit mode... or worse... impossible to say.
You really need to put a scope on the rails, with the amp delivering full power at say 20hz and see what the supplies look like.
I can understand what you are asking... but is this a real or imaginary issue... thats why you need to scope the supply.
Hard to say whether more capacitance would upset the PSU... its during start up where there may be problems where the capacitance may be seen as a "short" or low impedance load on the rails. What happens then depends on the PSU... probably nothing more than a few fractions of a second to bring the rails up... or does it see the extra capacitance as a short or low impedance load and go into a limit mode... or worse... impossible to say.
With the age of these amps, renewing the capacitors would bring an improvement. Adding further capacitance on an SMPS is likely to make it unstable or cause it to shut down.
Unfortunately I sold my oscilloscoope. It needed work too, and I didn't need another project. This is just a thought for a possible future project sometime in the next year.
Of course it has plenty of capacitance to take care of the high-frequency ripple and plenty to handle some demand well beyond the dip between power supply cycles. And if I'm satisfied with the output there's no real problem. I'm just wondering whether I can get these amps to sound much "punchier" by improving their dismal short-duration dynamic power rating. They have a wonderful continuous power rating, 900 watts a channel at the optimal impedance; but there are 100-watt amps with a better short-burst dynamic peak power rating.
I look at it like a hose into a lake with a dam. If you 'regulate' the hose valve to put in more water when the lake level drops, the lake doesn't have to be very big. But if you want to occasionally dump a few million gallons over the dam, it really helps to make the lake really big. But for the lake level not to drop and be a problem, the lake has to be really big. A medium-size lake might be worse than a small one. The small lake would fail to supply my peak demand dump over the dam and recover quickly with the hose valve turned on all the way. It limited the peak demand. The medium-size lake would dump more over the dam but leave the lake nearly empty for a while as the hose takes a longer time to refill the lake. The really huge lake satisifes the demand for water to dump over the dam without "noticeably" lowering the lake level, and if there's enough recovery time before the next peak demand the hose will recover the lake level.
So there ar three issues:
1) Will much larger caps really improve the sound?
2) Will that overwork the power supply?
3) Will that overwork the amplifier?
...assuming I don't constantly over-work these things ridiculously, I suspect:
1) It will occasionally make it sound much punchier.
2) I suspect that if I don't abuse it, the power supplies will be OK. They have soft-start so I think they'll be OK at turn-on. After that, the caps will usually smooth out the demand, only occasionally increasing the total demand.
3) I know the MOSFETs do degrade when pushed too hard continuously, but there are a LOT of them and I suspect they'll be OK.
Heck, this is a forum of other people who just can't leave well enough alone. But I'm not going to do this immediately anyway, I've got other projects first. And if I get another one of these cheap on eBay I can just add another amp instead of modding these.
Of course it has plenty of capacitance to take care of the high-frequency ripple and plenty to handle some demand well beyond the dip between power supply cycles. And if I'm satisfied with the output there's no real problem. I'm just wondering whether I can get these amps to sound much "punchier" by improving their dismal short-duration dynamic power rating. They have a wonderful continuous power rating, 900 watts a channel at the optimal impedance; but there are 100-watt amps with a better short-burst dynamic peak power rating.
I look at it like a hose into a lake with a dam. If you 'regulate' the hose valve to put in more water when the lake level drops, the lake doesn't have to be very big. But if you want to occasionally dump a few million gallons over the dam, it really helps to make the lake really big. But for the lake level not to drop and be a problem, the lake has to be really big. A medium-size lake might be worse than a small one. The small lake would fail to supply my peak demand dump over the dam and recover quickly with the hose valve turned on all the way. It limited the peak demand. The medium-size lake would dump more over the dam but leave the lake nearly empty for a while as the hose takes a longer time to refill the lake. The really huge lake satisifes the demand for water to dump over the dam without "noticeably" lowering the lake level, and if there's enough recovery time before the next peak demand the hose will recover the lake level.
So there ar three issues:
1) Will much larger caps really improve the sound?
2) Will that overwork the power supply?
3) Will that overwork the amplifier?
...assuming I don't constantly over-work these things ridiculously, I suspect:
1) It will occasionally make it sound much punchier.
2) I suspect that if I don't abuse it, the power supplies will be OK. They have soft-start so I think they'll be OK at turn-on. After that, the caps will usually smooth out the demand, only occasionally increasing the total demand.
3) I know the MOSFETs do degrade when pushed too hard continuously, but there are a LOT of them and I suspect they'll be OK.
Heck, this is a forum of other people who just can't leave well enough alone. But I'm not going to do this immediately anyway, I've got other projects first. And if I get another one of these cheap on eBay I can just add another amp instead of modding these.
Hmm...I'd hope the bigger caps would not make the SMPS unstable, that would be a definite bummer but I don't really believe it. But yes it could cause problems at start-up or if they ever really drained in use; lots more demand on the power supply. The soft-start circuit should help, or I could switch in the additional caps after some delay for initial turn-on.
Instead of speculating based on reviews of related-family amps I should just try to get the real schematics to start with.
I should ask this over on the power supply forum. There's such an optimal-design philosophy of putting in caps just big enough to handle the ripple between cycles. But in audio we're used to unregulated supplies and larger caps, not as RF filters or 60-sysle filters but almost as batteries to better meet the power demands of real music, where there's loud note and percussion attacks with rests and quieter in-between.
Instead of speculating based on reviews of related-family amps I should just try to get the real schematics to start with.
I should ask this over on the power supply forum. There's such an optimal-design philosophy of putting in caps just big enough to handle the ripple between cycles. But in audio we're used to unregulated supplies and larger caps, not as RF filters or 60-sysle filters but almost as batteries to better meet the power demands of real music, where there's loud note and percussion attacks with rests and quieter in-between.
With 900 watts output on offer have you actually measured to see how close you come to that ?
Maybe you could adapt the technique here to get an idea,
http://www.diyaudio.com/forums/soli...er-do-you-really-need-domestic-listening.html
Maybe you could adapt the technique here to get an idea,
http://www.diyaudio.com/forums/soli...er-do-you-really-need-domestic-listening.html
- Status
- Not open for further replies.