Wayne, I believe your reasoning is not practical.
Who the hell is going to sit and listen for hours on end to a sine wave at full rated power into a 2 Ohm load with a phase shift of -60 degrees?
Nobody in their right mind, I think.
Next, where's the reserve for any transient that comes along? If you want to have any, you need to decrease your AVERAGE power dissipation or the first transient coming along will fry your speakers and possibly your house too, God forbid.
You keep looking at these calculations from a theoretical, TOTALLY impractical view. You don't seem to understand that even a 25W average power level will have your walls shaking in most cases, that is, unless you have some exceptionally low effciency speakers.
The simple fact is that in most rooms, and with most speakers, 1-2 Watts of average power is already uncomfortably loud, never mind if the speakers and the amp can do much more.
Therefore, you are all right if you can ride out the transients on your capacitors. For that, you can get away with a lot less than you'd need for a full power, steady state sine wave.
For a nominally 100/200/400 W amp into 8/4/2 Ohms, the industry (specifically Harman/Kardon) uses 2x15,000 uF per channel and actually get vast numbers of watts into 1 Ohm, let alone 2. For example, my own HK 680 integrated amps, which has a 550 VA transformer for both channels, but separate windings for each channels, separate bridge rectifiers and just 2x8,200 uF caps on each channel will knock just over 500 Watts into 1 Ohm in short term peaks (well, it used to when it was new, now it needs new caps). Admittedly, those were lab resistors, but the figures are still very impressive.
I would be very surprised if John's work did any less in its respective class.
You seem to think that it's only the power supply which will deliver power, but that simply isn't so. Of course you need a capable PSU if you want a lot of power, but you also need a design which is overall capable of it as well, and I don't just mean the output stage.
I will be perfectly happy if my own design can run into 2 Ohms on a continuous basis - at least theoretically, because at that kind of power, I will surely need at least twice the already generous heat sink real estate. But it will be capable of handling loads even below 2 Ohms, even if only in impulses.
Trust me on this, a pair of 22,000 uF cans per channel is way more than enough; my entire Marantz 170 DC power amp (nominally 2x86 WPC/8 Ohms) runs on one pair of these caps and it has NEVER run out of steam yet. Nor is it likely to, unless I push it really hard and well into troubled waters.
Sorry Jacco, no division by 2.
I'll try to remember where I saw that and dig up the document, I'm sure I have it somewhere in my !"#$%&/() log.
DVV ....
Another :
http://www.profusionplc.com/images/d...ecf10-demo.pdf
When we ran Bryston amplifiers in the 80's (4B) I had requested them to increase PSU capacitance by double , they refused citing science , Julius at Audire responded and guess which ones drove and controlled the woofers better @2ohms , worst the brystons would develope PSU noises after 14-16 months and we would send them back to Bryston for recapping ( at their expense) and still they refused to increase the capacitance , citing more than enuff ...
We eventually moved them off the bass units running all Audire stuff on the bass, where they had no issues and easily outperformed the 4B's. So IMO the science is LAB junk , crowns are lab junk ( thrown out enuff to say so) successful audio products are the ones applying science with real world application , Big PSU rules in my world with my cash and I totally abhor dinky toy amps ....
Errr just saying ....
Another :
http://www.profusionplc.com/images/d...ecf10-demo.pdf
Believe me those numbers will not work in the real world at 2 ohm , and marginal for 8/4, you will hear the difference in the bass and in the grunge produced on the signal. There is a solid refinement of sound when the PSU is sized right ....
PS: Bee Geejus aren't you watching the Germany/ Dutchie match ..
Germany -2
Holland - 0
Hi Nigel;
Don't use UL, it was bad fashion. Local feedback to 1'st grids is what doctor prescribed (like was done in RH-88), but the driver should be adequate.
Search for Pyramid-VII schematic in the forum, I posted it several times. It was the best I could do woth Gu-50 tubes (EF184 were used as deivers in Pyramid-VII-M). Or I can open a thread.
Yes, right schematic, but one bridge can be used for both channels.
Don't use UL, it was bad fashion. Local feedback to 1'st grids is what doctor prescribed (like was done in RH-88), but the driver should be adequate.
Search for Pyramid-VII schematic in the forum, I posted it several times. It was the best I could do woth Gu-50 tubes (EF184 were used as deivers in Pyramid-VII-M). Or I can open a thread.
Yes, right schematic, but one bridge can be used for both channels.
Wayne, some day you will discover fully regulated power supplies. Very expensive, in effect they are just another power amp, admittedly a specific one.
That is, without any doubt, by far the best solution overall, assuming it has been as well as it should have been done. Unfortunately, it's also easily the most expensive approach, and by the time you get to 200 WPC, it starts to hurt just how expensive it is.
Krells uses that, but look at their prices; Levinson uses that, but look at their prices.
Naim used to use that, and it made them a legend. Even Marantz tried that with some of their early 80ies models, but never made any impact because they wanted to do it on a shoestring budget.
When a power amp encounters a low impedance load, which may additionally also have very significant phase shifts (since a -60 degree phase shift, not uncommon, requires twice the current, you could consider it as an equivalent 1 Ohm load), its power supplies are severely stressed and typically droop by so many volts. This changes the operation of any and all amps.
One way to make this much better is to use larger and better quality transformers. They are all too often underrated anyway (not counting the present company here) and are of just so-so quality. The next step uses separate power supply lines for the voltage gain and current gain stages, which is what I routinely do.
That way, when the current demand on the output stage is extremely high, the front end of the amp knows nothing about it because it is fully electronically regulated. This is very cost effective, because the currents we are dealing with are small, typically 20...50 mA. To be sure, this is no universal panacea, but it does go a long way in making sure the sound does not deteriorate when the amp is stressed. Or not much, at least.
Why is it not used widely then? The usual answer - price. Quite simply, it costs more and very often cannot be accommodated by the amp budget. Remember, every 1 cent built into an amp becomes (offhand) about 3.5 cents in retail.
I picked up that nasty habit from Matti Otala's IEEE text on TIM and his model amp printed in the same article. But then, I don't make amps commercially, so I can afford to be as nutty as I feel like at the moment. Same for say John on his off time, but when he's designing for a commercial product, a very different logic applies. He then has a budget he must fit the product into, while getting the most out it at the same time. Not easy, not easy at all.
As for your example, I somehow have a feeling you are using a very special woofer combination, and obviously a troublesome one. I suspect it's even a rarely encountered problem combo, which makes you the odd man out. You cannot seriously expect the industry at large to rule itself on such rare users as yourself. If they did, their already high prices would rocket into a new orbit.
For such occasions, I would always recommend you commission a specific design tailor made for you. The Good Lord knows you have your pick of the cream right here - talk to them. Building a specific amp or two is nothing new to John, Wavebourn and others, that's what they do. And before you freak out thinking about the prices, while it cannot be cheap because you seem to have a tall order, it does not necessarily have to have an out of this world price either, especially if it does the job well.
Please do not misunderstand me, I am not trying to brush you off or anything, but being involved in production myself, I know when the standard fare is no longer satisfactory and when it's time for special, even one-off models. Some of my own products were born just that way, somebody needed something way out of the ordinary, and was willing to pay for it out of the ordinary, so we (my associates and I) accommodated him. And ended up with an extra powerful line filter right there, a crying shame not to be turned into a model line.
I only wish I was always that lucky, not just that one time.
That is, without any doubt, by far the best solution overall, assuming it has been as well as it should have been done. Unfortunately, it's also easily the most expensive approach, and by the time you get to 200 WPC, it starts to hurt just how expensive it is.
Krells uses that, but look at their prices; Levinson uses that, but look at their prices.
Naim used to use that, and it made them a legend. Even Marantz tried that with some of their early 80ies models, but never made any impact because they wanted to do it on a shoestring budget.
When a power amp encounters a low impedance load, which may additionally also have very significant phase shifts (since a -60 degree phase shift, not uncommon, requires twice the current, you could consider it as an equivalent 1 Ohm load), its power supplies are severely stressed and typically droop by so many volts. This changes the operation of any and all amps.
One way to make this much better is to use larger and better quality transformers. They are all too often underrated anyway (not counting the present company here) and are of just so-so quality. The next step uses separate power supply lines for the voltage gain and current gain stages, which is what I routinely do.
That way, when the current demand on the output stage is extremely high, the front end of the amp knows nothing about it because it is fully electronically regulated. This is very cost effective, because the currents we are dealing with are small, typically 20...50 mA. To be sure, this is no universal panacea, but it does go a long way in making sure the sound does not deteriorate when the amp is stressed. Or not much, at least.
Why is it not used widely then? The usual answer - price. Quite simply, it costs more and very often cannot be accommodated by the amp budget. Remember, every 1 cent built into an amp becomes (offhand) about 3.5 cents in retail.
I picked up that nasty habit from Matti Otala's IEEE text on TIM and his model amp printed in the same article. But then, I don't make amps commercially, so I can afford to be as nutty as I feel like at the moment. Same for say John on his off time, but when he's designing for a commercial product, a very different logic applies. He then has a budget he must fit the product into, while getting the most out it at the same time. Not easy, not easy at all.
As for your example, I somehow have a feeling you are using a very special woofer combination, and obviously a troublesome one. I suspect it's even a rarely encountered problem combo, which makes you the odd man out. You cannot seriously expect the industry at large to rule itself on such rare users as yourself. If they did, their already high prices would rocket into a new orbit.
For such occasions, I would always recommend you commission a specific design tailor made for you. The Good Lord knows you have your pick of the cream right here - talk to them. Building a specific amp or two is nothing new to John, Wavebourn and others, that's what they do. And before you freak out thinking about the prices, while it cannot be cheap because you seem to have a tall order, it does not necessarily have to have an out of this world price either, especially if it does the job well.
Please do not misunderstand me, I am not trying to brush you off or anything, but being involved in production myself, I know when the standard fare is no longer satisfactory and when it's time for special, even one-off models. Some of my own products were born just that way, somebody needed something way out of the ordinary, and was willing to pay for it out of the ordinary, so we (my associates and I) accommodated him. And ended up with an extra powerful line filter right there, a crying shame not to be turned into a model line.
I only wish I was always that lucky, not just that one time.
Replaced my MC3500's with a fully regulated custom built SS amp in 1978 , the PSU had as much power transistors as the main amp, it was of 2 pc construction and was stable to 2 ohms ...
Not aware of any Krell with fully regulated supplies ... The ML2 was fully regulated, i never had one , ( have heard them a few times over the years) but did have an ML9 , different beast , I know, none in the same league as the custom built amp i had done in 78.
The woofer is not the issue, I have 3 different speaker systems, my main is very current demanding and is very much a resistor from 300-22K , below 300 I'm using dynamic woofers. I can drive this system with a Adcom 555, it just gets better with of course the Krell or my Threshold, the only time i have issues is with pretenders, which account for 80 % of the market in my opinion.
Ironically amplifiers that sound good on this speaker sounds good on the others, very rarely are they sent off by the pretenders.
No offense taken ..
Well as previously noted , i have had custom stuff done since the 70's ...
Damn Dutchies scored one , we may still have a game yet ..!!!
Germany 2-
Dutchie - 1
Ben Duncan on P234 of High Performance Audio Power Amplifiers states 10 000 uF for 100 Watts 8R . Remember that is 5000 uF in series and about 2.56 A at 4.5V ripple .
Duncan goes on to say some amps use 250 000 uF ( 125 000 in series ) . If assuming 40 A then .
40 / ( 110 x 125 000 uF ) = 2.9 V ripple
Ben says for bass it is considered required . Self says not .
Duncan goes on to say some amps use 250 000 uF ( 125 000 in series ) . If assuming 40 A then .
40 / ( 110 x 125 000 uF ) = 2.9 V ripple
Ben says for bass it is considered required . Self says not .
The link is broke. Where did you copy the text from? The link has the same board-shortened "..." thing as does the text.
Always do right-click and "copy link location" and then (if the ORIGINAL link is good) the pasted link will be good.
Yours for Sound Measurement of Quality...
These discussions of regulated power supplies are amusing to me, since with a few exceptions I've used them in very-low-cost computer-centric powered speaker applications beginning in 1994 (a two-channel system that still operates on my bench today, which needed some way of keeping the rectified a.c. adapter from soaring at low signal levels and shutting down the automotive-oriented chip amp) and including more elaborate three-piece and larger systems.
In most cases the topology was a low-dropout discrete regulator using a N channel DMOS part as the pass element. I did have some of the techs scratching their heads, as their tendency was to assume the negative terminal of the bulk cap had to be common, or close to common.
The motivation was mostly to save on iron and copper in the mains transformer. In some cases the compressor-limiter was linked to the unregulated voltage so as to prevent running out of room in the pass element. Of course the loads on the power amp channels were well-defined; the woofer was usually ported and the drive voltage highpassed. For what the customer paid, and for what was a near-field system of modest ambitions, I think they were pretty good deals.
These designs are now mostly obsolete, not because they don't work well but because the government mandates more complex and noisy switchers. This is called progress.
In most cases the topology was a low-dropout discrete regulator using a N channel DMOS part as the pass element. I did have some of the techs scratching their heads, as their tendency was to assume the negative terminal of the bulk cap had to be common, or close to common.
The motivation was mostly to save on iron and copper in the mains transformer. In some cases the compressor-limiter was linked to the unregulated voltage so as to prevent running out of room in the pass element. Of course the loads on the power amp channels were well-defined; the woofer was usually ported and the drive voltage highpassed. For what the customer paid, and for what was a near-field system of modest ambitions, I think they were pretty good deals.
These designs are now mostly obsolete, not because they don't work well but because the government mandates more complex and noisy switchers. This is called progress.
- Status
- Not open for further replies.