For folks who audited the Grand Integra M510 many times, it may be somewhat uncomfortable so witness an Onkyo power amplifier with a single pair of output devices.
Even for the ones who saw budget amps as a Rotel RB850 hit the market it's somewhat odd.
Not so strange if one looks at the extension of the market, a couple of hundred dollars for a power amp that fits in a home theatre surround setup is good enterprising.
I can't imagine anyone in his right audio mind to hook up a power amp, that's not even power rated at 4 ohm, to heavy duty loudspeakers.
The "audiophile" solid state types i see at this forum (and others) in general build way less undersized power amps for the kind of loudspeakers they have (though often mismatched, e.g. a SE class A amp combined with << 90dB/2.83V loudspeakers that have a minimum impedance under 4 Ohm)
Plenty of multi-channel power amps with single pair output stages on the market, but the M-282 is indeed an eyebrow raiser for the oldfashioned power orientated (A1943/C5200).
Goes to show what a softy powersupply and early/rigid 9-pin IC protection can achieve.
(car lounge : Chevy LS9 and aftermarket crate engines, lower $/hp than the '90s 502ci mk6)
Last edited:
There is another point to multiple devices . Better Ft . Like Douglas Self did many NE5532 as a power amp , one can do multiple drivers as outputs . D44/45 perhaps ? Even BD139/140 with care . I suspect some of the super transistors are exactly that .
As you say the budget Rotels had 8 outputs total ( RA 820 , super fast and enough for 100 W amp ) . RA 820 was a conservative 25 W ( 35 W 4R and great into 1 ohm on bursts ) .
http://www.rotel.com/content/manuals_archived/ra820b.pdf
As you say the budget Rotels had 8 outputs total ( RA 820 , super fast and enough for 100 W amp ) . RA 820 was a conservative 25 W ( 35 W 4R and great into 1 ohm on bursts ) .
http://www.rotel.com/content/manuals_archived/ra820b.pdf
Last edited:
I think it's because it's black.
There will always be a way to rationalize anything. Personally, I think it's an outrage from a company I believed to be wiser than most. Period.
The key downside to multiple output devices is pairing them. When your output switches, it has to be in as much unison as possible.
The rest is all advantages, bless their little emitters.
I think that's what emitter resistors help with . As far as I know 0R33 will swamp any differences . Mostly the voltage points are very similar on the same device type . Not so MOSFET's . Do NPN and PNP ever make true pairs ? Some still argue for Quasi comp .
Talking of multiples ( forgive ) .
Kawasaki 48 Cylinder Runs! - YouTube
Talking of multiples ( forgive ) .
Kawasaki 48 Cylinder Runs! - YouTube
A few oldies are.
Example : http://www.diyaudio.com/forums/solid-state/135892-2sa968-2sc2238.html#post1709514
Well back to ripple in power supplies. If you set 2% ripple as your goal UNDER FULL LOAD, then the ripple will be much less at normal output levels. That of course will depend on you amplifier topology.
Most of those funny amplifiers with the red hot glass transistors use more filter stages for the input and gain stages than the output stage. The more normal solid state amplifiers often have an additional RC filter on the input stage.
Now for those who pay attention to critical band research around 30 db of suppression of the low frequency ripple may just be adequate for most users! This seemingly high level would be masked by the actual music content and from the effect of what is shown in the Fletcher Munson research.
Now my OPINION is that those with experience or training can detect artifacts which would be caused by 2% ripple. That would be -34 db at full power without feedback.
But the vast majority of audio gear in production really does have 2% or more power supply ripple at full load.
So a reasonable experiment would be to design a power supply with 2% ripple and a moderately high value resistor bypassed by a switch to a larger capacitor bank that would reduce the ripple to .2%. As the resistor should keep the bank charged you could switch it in and out without harm. Then one could actually listen to see what effect the larger bank really did produce.
Of course that would take all of 10 minutes and B... about it wastes much more time.
Most of those funny amplifiers with the red hot glass transistors use more filter stages for the input and gain stages than the output stage. The more normal solid state amplifiers often have an additional RC filter on the input stage.
Now for those who pay attention to critical band research around 30 db of suppression of the low frequency ripple may just be adequate for most users! This seemingly high level would be masked by the actual music content and from the effect of what is shown in the Fletcher Munson research.
Now my OPINION is that those with experience or training can detect artifacts which would be caused by 2% ripple. That would be -34 db at full power without feedback.
But the vast majority of audio gear in production really does have 2% or more power supply ripple at full load.
So a reasonable experiment would be to design a power supply with 2% ripple and a moderately high value resistor bypassed by a switch to a larger capacitor bank that would reduce the ripple to .2%. As the resistor should keep the bank charged you could switch it in and out without harm. Then one could actually listen to see what effect the larger bank really did produce.
Of course that would take all of 10 minutes and B... about it wastes much more time.
Last edited:
Thanks pretty much what I was thinking, although maybe with more steps (kinda complicated, that). Would be a nice test of what more stored energy could do.
If you can hear 10/1 then you might want to do more. But as a simple proof of concept it is easy to do.
2% ripple from the PSU does not necessarily equate to -34dB below full output. It could be significantly better or worse than this even before feedback is taken into account. Let the ripple into an input stage and things could be a lot worse. On the other hand, some stages have inherently good PSRR (e.g. pentode output) so will be better.simon7000 said:
On a practical note this might help . Hope it opens ? These people sell worldwide . I guess there will be a time limit for it working . Apologies in advance if it has expired ? It should be in ascending price . It is 22 000 uF and above > 40 V .
Your Search Results | Farnell United Kingdom | Results
Your Search Results | Farnell United Kingdom | Results
@Pano
I did what you proposed, and here’s what I came up with:
Loudspeaker – various, see below for description;
Room – app. 14 m2, app. 41 m3, no drapes, one wall ¾ glass, other wall in full wood covering, very small carpet, rather full of whatever (my own Kingdom Of Organized Chaos);
Average RMS voltage = 2 ... 2.8 Vrms, depending on mood, L/S= B&M 1041
Average RMS voltage = 3 ... 3.7 Vrms, depending on mood, L/S= JBL Ti600
Average RMS voltage = 4.5 ... 5.2 Vrms, depending on mood, L/S= AR 94
Speakers:
B&M= B&M Acoustics Monitor 1041, eff. 92 dB/2.83V/1m, nominal impedance 8 Ohms, minimum impedance 6.5 Ohms, worst case phase shift -25 degrees;
JBL=Ti600 floorstanders, nominal impedance 8 Ohms, minimum impedance app. 4.5 Ohms, worst case phase shift app. -45 degrees, efficiency app 91 dB/2.83V/1m
AR = Acoustic Research 94, nominal impedance 8 Ohms, minimum impedance 3.9 Ohms, worst case phase shift app. -54 degrees, effciency app. 89 dB/1W/1m (manuf. Spec.).
In fact, I think AR should have declared the 94 as a 6 Ohm speaker, because that’s where its impedance curve is like 75% of the time, rising above 8 Ohms only at 20 kHz.
The measured speaker data dates back to 2002, when we were cooking up the 1041 monitor, and took the time and trouble to measure the other two.
If I've missed something, or you simply want to know more, don't hesitate to ask.
I did what you proposed, and here’s what I came up with:
Loudspeaker – various, see below for description;
Room – app. 14 m2, app. 41 m3, no drapes, one wall ¾ glass, other wall in full wood covering, very small carpet, rather full of whatever (my own Kingdom Of Organized Chaos);
Average RMS voltage = 2 ... 2.8 Vrms, depending on mood, L/S= B&M 1041
Average RMS voltage = 3 ... 3.7 Vrms, depending on mood, L/S= JBL Ti600
Average RMS voltage = 4.5 ... 5.2 Vrms, depending on mood, L/S= AR 94
Speakers:
B&M= B&M Acoustics Monitor 1041, eff. 92 dB/2.83V/1m, nominal impedance 8 Ohms, minimum impedance 6.5 Ohms, worst case phase shift -25 degrees;
JBL=Ti600 floorstanders, nominal impedance 8 Ohms, minimum impedance app. 4.5 Ohms, worst case phase shift app. -45 degrees, efficiency app 91 dB/2.83V/1m
AR = Acoustic Research 94, nominal impedance 8 Ohms, minimum impedance 3.9 Ohms, worst case phase shift app. -54 degrees, effciency app. 89 dB/1W/1m (manuf. Spec.).
In fact, I think AR should have declared the 94 as a 6 Ohm speaker, because that’s where its impedance curve is like 75% of the time, rising above 8 Ohms only at 20 kHz.
The measured speaker data dates back to 2002, when we were cooking up the 1041 monitor, and took the time and trouble to measure the other two.
If I've missed something, or you simply want to know more, don't hesitate to ask.
Last edited:
How would it be worse? A class D amplifier is usually the worst case. For that it would really be -34 db.
The point was that 20 db is around the minimum acceptable level.
I see the voltages climb as the speaker sensitivity drops.
Why do i get the feeling no one is willing to go to the bulls eye on this
OK with separate stages , full regulation on the VA stage, would this mean we can isolate the outputs as far as ripple content ?
- Status
- Not open for further replies.