I undertsand all , even my position with AP.
the "speed" is all, I insist (and is also demonstrated with other measures) that listening is different. If the voltage does not drop during the transient, then the current rises. (rises very rapidly near the corner of the peak)
This overcomes the mechanical compression of a woofer and reproduction is more original than the real dynamics. This causes an immediate difference to listening, even at low power.
In reality we are saying the same thing, the problem is that you do not believe that is possible with a SMPS. (In fact, you're right) but is just why I developed this SMPS. research and high costs do not believe can be justified from 100 SMPS for diy, then the goals are other, much higher
Just does not happen every day to see this measure.
On resistive load, a generator at 400Hz, start, the driver amplifier to 1.76 V using time as the x axis.
Graph show input signal level Vs. output (Vp) of amplifier (mospower AB Class)
the other graph is zooming.
signals with variable frequency (music), the ratio varies continuously significantly (also in the segment low power)
Maybe a diy, do not think that behind an SMPS, we wanted to find a real remedy to the problem of Unregulated psu.
The green trace, although aumentere power decreases as the (little), has maintained a good ratio close to linear REF
I do not follow the Voodo or magicians, but only science and technology.
Regards
On resistive load, a generator at 400Hz, start, the driver amplifier to 1.76 V using time as the x axis.
Graph show input signal level Vs. output (Vp) of amplifier (mospower AB Class)
the other graph is zooming.
signals with variable frequency (music), the ratio varies continuously significantly (also in the segment low power)
Maybe a diy, do not think that behind an SMPS, we wanted to find a real remedy to the problem of Unregulated psu.
The green trace, although aumentere power decreases as the (little), has maintained a good ratio close to linear REF
I do not follow the Voodo or magicians, but only science and technology.
Regards
Attachments
Last edited:
I never said that it is impossible with a SMPS; IMO what can be done with a SMPS can also be done with a linear power supply. The technology is different but will have the same results if designed properly.
Dynamic loudspeaker compression is caused by heating of the voice coil. A hot voice coil will have a higher impedance; this higher impedance will in fact cause lower power demand from the amplifier so that the compression is in fact a mutual relationship. Loudspeaker compression is more an electromechanical problem, not a mere mechanical problem.
Uhm..i agre only if we have linear with big capacitors bank, as all professional audiophile i know. but as you see on graph, this solve 50% of problem. (and have big dimension).
Today demand of market is, how develop new amplifier without big 800VA trafo in side? (with same or superior performances) put a smps for ...lamps? ahahah!
This amplifiers design is a prefect example of technology allowing us to build a better amp with less. New chips are offering us performance in the past was only able to offer with desecrate components. That is what I like about the SMPS power supply it is able to offer the performance that in the resent past was only able to achieve with a fair size slug of iron and a health bank of caps.
Bill
Bill
Last edited:
Roberto, thanks for the graph, I am slowing getting there. In a word, the SMPS has better transients than 600VA+100KuF. If we take a capacitor bank made of fairly good 10x10k Vishay BC 63v caps + the trafo, even the cost of only these two parts will be more than the SMPS. It's a no brainer, really... at least for me! I can see guys arguing and being defiant of light devices vs heavy jewelry that looks "oh so macho" but I am sold on your concept, great stuff!
Last edited:
I already confirmed that the SMPS has the big advantage of weight and volume, and unquestionably the quality of SMPS's has reached the quality for audio.
But what are we discussing here: quality or economics?
Be a big man and admit that with a quality regulated supply (yes: heavier, more expensive, larger) the same quality can be obtained (I'd say at least the same quality); that at least is my opinion, and that is what I (try to) bring up here.
The choice of power supply is up to the amp builder of course; I have no interest in that regard.
Last edited:
Pieter, we discussing "quality" but in a new way.
There is only one way to ensure that transients are not changed (from PSU).
currently this is provided by a transformer with a fast linear regulator. (on a fast amplifier). but this space and high costs, so it's easier to convince that the psu is not important.
Any other statement taken on non-scientific publications, (as linear Unregulated is better) as the guru of sound, built in the past 30 years, balls over balls. (the mystery, ignorance infiltrates good)
Do not be surprised to find even distorted concepts in scientific publications.
especially on the relationship between perceived sound and electric signal.
If you want to be ahead of others, no one will help you.
Regards
Roberto,
First of all: your English is improving dramatically
So you agree that the SMPS has no inherent advantage over a quality linear supply, am I right?
By the way, when I look at the graphs above I see that the unregulated 100kuF supply is equally fast as the SMPS up to 100 mS (x10?), but I don't know how to interprete this correctly, and how fast "things go wrong" up to 280 or so mS. Could you explain?
Last edited:
Uh oh - better start sending out mass emails to all those engineers selecting SMPS then...
No, I mean technical advantages, not economical advantages (price; weight; volume).
smps GB
Hi everyone,
Sorry to say but I'm going to have to withdraw from the GB for the smps... nothing to do with the resent discussion, just I'm about to sell my apartment and am packing things away etc.
I'm hoping to get a place with enough room for a dedicated home cinema / listening room
Hi everyone,
Sorry to say but I'm going to have to withdraw from the GB for the smps... nothing to do with the resent discussion, just I'm about to sell my apartment and am packing things away etc.
I'm hoping to get a place with enough room for a dedicated home cinema / listening room
Guys lets not get personal here. Perhaps if a stalemate has been reached take it off line or agree to disagree.Early results from the bench
After correcting the orientation of Z1 so that the LME is not on continuous mute, I got the following preliminary test results.
Test 1:
PSUs: LME +/-65V regulated; OPS +/- 50V monitored & controlled
Output: 30Vrms @ 8ohm @ 1KHz (1Vrms below clip)
THDN: 0.003%
Test 2:
PSUs: LME +/-50V regulated; OPS +/- 50V monitored & controlled
Output: 28.8Vrms @ 8ohm @ 1KHz (1Vrms below clip)
THDN: 0.003%
Test 3:
PSUs: LME +/-65V regulated; OPS +/- 60V monitored & controlled
Output: 33.5Vrms @ 8ohm @ 1KHz (1Vrms below clip)
THDN: 0.004%
Test set: HP339A; two HP6294A (for LME PSU); 72K uF caps per rail monitored and adjusted by variac (for OPS).
Observation: test 1 Vds = 7.6V peak vs. test 2 Vds = 9.3V peak => 1.7V difference
Comment: a) I am very please with the performance of the amp. Hats off to Owen. b) I am also convinced that the separate, higher LME supply is of value for several reasons, but it looks like reducing the voltage drop accross the output device is not of significant value.
What are the rest of you finding?
After correcting the orientation of Z1 so that the LME is not on continuous mute, I got the following preliminary test results.Test 1:
PSUs: LME +/-65V regulated; OPS +/- 50V monitored & controlled
Output: 30Vrms @ 8ohm @ 1KHz (1Vrms below clip)
THDN: 0.003%
Test 2:
PSUs: LME +/-50V regulated; OPS +/- 50V monitored & controlled
Output: 28.8Vrms @ 8ohm @ 1KHz (1Vrms below clip)
THDN: 0.003%
Test 3:
PSUs: LME +/-65V regulated; OPS +/- 60V monitored & controlled
Output: 33.5Vrms @ 8ohm @ 1KHz (1Vrms below clip)
THDN: 0.004%
Test set: HP339A; two HP6294A (for LME PSU); 72K uF caps per rail monitored and adjusted by variac (for OPS).
Observation: test 1 Vds = 7.6V peak vs. test 2 Vds = 9.3V peak => 1.7V difference
Comment: a) I am very please with the performance of the amp. Hats off to Owen. b) I am also convinced that the separate, higher LME supply is of value for several reasons, but it looks like reducing the voltage drop accross the output device is not of significant value.
What are the rest of you finding?
The supply voltages limit the peak output voltages from the amplifiers.
An audio replay system, rarely reaches these peak output voltages. We generally design our systems to avoid clipping in normal use.
The result is that the only time that our amplifiers get near to clipping and thus use up that PSU voltage capability is for short term transients. By definition, these transients exist for short periods of time and thus place very little demand on the long term capability of the PSU.
It's the stored charge in the medium and high frequency decoupling that supplies the current for the short term transients.
For these reasons small variations in the choice of PSU voltage at the design stage will have little effect on the audio performance heard from the speakers.
An audio replay system, rarely reaches these peak output voltages. We generally design our systems to avoid clipping in normal use.
The result is that the only time that our amplifiers get near to clipping and thus use up that PSU voltage capability is for short term transients. By definition, these transients exist for short periods of time and thus place very little demand on the long term capability of the PSU.
It's the stored charge in the medium and high frequency decoupling that supplies the current for the short term transients.
For these reasons small variations in the choice of PSU voltage at the design stage will have little effect on the audio performance heard from the speakers.
Andrew,
I basically agree, but at the same time I am afraid that your remarks will restart the power supply discussion, and I don't think that's a good idea....