Separate names with a comma.
Discussion in 'Amps and Cabs [BG]' started by terbay, May 5, 2005.
You mean a conventional power supply, not a linear one.
Don't listen to power supplies, listen to amplifiers.
Hey basstrader.... Not in the least, this was exactly what I was hoping for...
A linear supply IS a conventional supply and you do listen to the power supply.
The Transistors/Tubes Get their power from the supply. You CAN hear it whan a supply is under powered, noisy, has failing caps, a shorted choke...etc. the power supply IS the amp. The rest of the circuitry gets the power from the supply, modified by the signal from your bass, and shoves it into your speakers.
A conventional supply is not linear.
If you can hear the power supply, your amp is either defective or you're driving it into clipping.
One difficulty with the switching type POWER SUPPLY, is that a typical "conventional" switching supply cannot "ride through" current peaks as a "linear" (big heavy iron transformer) supply can. Most of the time, the output capacitors are much smaller than with a linear supply.
In fact, there are certain large disadvantages to a large output capacitor on a switching supply.
But, for audio, there are big advantages to having a large capacitor on the output, since it provides a source for what may be very large transient currents. That tends to be why units with big iron and large capacitors make the best subwoofer amplifiers.
A 500W amplifier has the capability of regularly producing 1000W peaks, normally, and may have a higher transient power peak under some conditions. Obviously more power=more current, in general.
If the switching supply has small output capacitors, it must directly supply whatever current is demanded. If it is not properly designed for the purpose, it may not be capable of supplying that much current, causing a transient clipping condition.
There are several ways around that problem, but it is a very hearable limitation with some of the available amplifiers that have switching power supplies.
Just like amplifier topologies, there are many good and bad power supply examples. Noise on the power supply rails often gets to the output, and an insufficient supply will degrade output quality when the amplifier is pushed. However, there is nothing inherently inferior about a switching supply in audio applications. They main problem is that they are much more difficult to design, so the likelihood of having a substandard one is much higher. Properly designed, a switching supply can provide cleaner, more solid supply rails, while saving a lot of size and weight. By the way, the conventional power amplifier supply is not linear, but rather its a line frequency transformer, rectifier and filter. Linear implies that a regulator has also been included to stabilize the output voltage. Normally regulators are only used for the low-level circuits (preamp and intermediate driver circuits).
Jerrold, if the switching supply behaves this way someone had a misunderstanding of how to best design it for the task at hand. You are correct that the large output capacitors in a conventional supply provide a lot of peak ride-through capability, but a properly designed switching supply will have vastly superior ride-through capability. This is achieved with a combination of high peak current capability and energy storage on the primary side of the transformer. There are two things in your favor that allow things to be much smaller while achieving this advantage. First, a switching supply is almost always regulated, so a feedback loop is naturally working to hold the voltage up. Second, capacitive energy storage is much more volumetrically efficient at high voltage (on primary side). This all goes out the window if the supply goes into current limit during a transient, but then, the supply wasnt designed properly for the application.
Don't forget, Jerrold, that a typical switch-mode power supply used in a power amp has a big primary reservoir in addition to the secondary reservoirs (the output capacitors) on the supply rails.
Yes, the "output capacitors" are smaller because they're refreshed some hundreds of thousands of times per second instead of 120 times per second.
With a conventional (non-linear, BTW) supply, sub frequencies at lower than 2x the mains frequency are especially taxing on the reservoirs, which is why they need large capacitances. That's much less of a concern when the power supply's frequency is itself higher than even the spectral content of of the transients.
You're right that the switch-mode supply's transformer and switching devices must have the current capability to refresh the rail reservoirs under heavy loads, but IMHO that goes without saying (if the amp is well designed they're fed from a sizable reservoir anyway, and the power lost in the resistance of the transformer will be orders of magnitude lower than in a conventional line-frequency transformer). Conventional supplies need adequate current capability also; only the time frame is different.
When I say a conventional supply is non-linear, I'm looking at it as a peak rectification issue. This is common with most AC-powered electronic devices, whether they have a switching supply or a conventional one. The phenomenon is that the supply conducts little or no current from the AC mains except at the peaks of the AC voltage waveform. The heavier the current demand on the power supply, the more the reservoir voltage will droop in between voltage peaks, and so the supply conducts current over a wider angle of the voltage waveform. Ironically, the heavier the current demand and wider the conduction angle, the more linear the supply is, and yet it's also more likely to distort the AC voltage waveform. And another irony is that the electronic supplies that most resemble a linear load on the AC mains are switch-mode supplies--with power factor correction.
All I know about it is that I have a little toy-like Teac amp that can almost hang with my McIntoch Mc30s in my home stereo. I did say almost, but the Teac cost me $70. Has the Tripath chip in it, 3-ch (two boards).
I must say, this thread is proving to be very informative! Thank you all for sharing. I feel like I am talking to Jim Bergantino!
[BTW, as sarcasm is often lost or applied innappropriately via on-line discussions, allow me to be clear that I thoroughly enjoy talking to Jim B. and I value his opinion. Furthermore, I consider it very kind on his part to take the time to talk to me, just as it is very considerate of Bob, Jerrold, and the rest of you to take the time out of your day to "talk" to us here!]
The Tripath design by the way, is what Tripath calls class T. Its a switching amp, but its not PWM.
You guys are going to have to clarify some stuff for me...
A conventional supply is...
A Linear Supply is...
A switching supply is....
I have been under maybe the mistaken conclusion that a linear supply consists of
1, a really big transformer
2, a bridge rectifier
3, a bunch of filter caps
4, sometimes a choke
And that a linear supply is the conventional way to provide DC for the amp...???
Hey, are the bipolar transistors on a special medications? Are they looked down upon by other components?
Where do I take my amp to find out if its bipolar in anyway?
That is hilarious....
Almost as good as the teddy bear that Texas Instruments used to give it's sales people as handouts.
It was white....
A Bi-Polar Bear.....GROAN>>>>
There are some very good sounding bipolar amps, IMHO. My old Adcom GFA-555II was pretty darn good for its day, and I think that the bipolar Adcom amps (535, 545, 555, etc.) sound better than their newer, MOSFET amps (5300, 5500, etc.).
That's a conventional supply.
An example of a linear supply would be a transformer powering an incandescent light bulb.
A switch-mode supply generally contains the equivalent of a conventional supply, except made for a much higher frequency, plus a means for converting incoming mains-frequency AC (50 or 60 Hz) into DC and then into high-frequency AC for transformer and the rest of the supply.
Both Bob Lee and 44ME make good points...some of which I mentioned...
The issue of "ride through" is related to the "hold-up" time of an SMPS... i.e. is there enough energy stored somewhere for the output to remain in regulation (if it is regulated......) or at an adequate voltage (if not regulated) for at least one full half cycle of mains voltage.
Stored at the input, the CV^2/2 factor for energy storage can help. But the entire path through to output must be made to carry the required current without excessive drops, etc. A linear supply can be fairly simply made to have no trouble with this, although it will be heavy and bulky.
In order to work "as well as" a "linear" supply, the SMPS needs, one way or another, to be capable of providing the maximum "legitimate" current that the amp needs. That can be problematic, but is doable. (Fault current is another issue entirely.)
Now a linear supply typically isn't regulated. Emulation of a linear supply with an SMPS therefore does not "require" regulation. In fact, there are some disadvantages as well as advantages to regulation.....
Suffice it to say, a conventional "200W" SMPS as the typical power supply designer would design it, likely will not properly power a "200W" audio amplifier.
Bob certainly knows this, QSC does use SMPS for some products....and theirs work.
However, nearly any SMPS currently provided in an audio amp will balk at providing long-term near full power output at the lowest impedance. Not that it couldn't be designed to do so, but that they rarely are.
Linear supplies are also rarely designed to provide full power long term. But for a number of technical reasons, even underpowered linear supplies are usually quite a bit more forgiving of overloads. You credibly may notice that difference, one way or another.
For that matter, you can often hear the difference between weaker and stronger linear supplies.....it can have an impact on the audible performance of the unit.
Our friends in europe have essentially mandated the use of SMPS for higher powered equipment. And those rules are coming here to the USA next. So, you will be seeing more and more SMPS, be they good, bad, or indifferent.
Back on the issue of class-D vs mosfet (or other linear) amplifiers, there are some similar issues with Class-D and transient currents that have to be considered.
It still is quite a bit easier to design a linear amp to produce a given power with massive overload capability, than to do the same with a Class-D. But of course it is possible to do either way.
I'm learning a lot from this post. I'm wondering if any of you could put this information into some practical sonic terms for me by naming some of the major brands of bass and power amplifiers out there and noting which class--A, A/B, D, etc.--each happens to be. I have played many amps over the years, and I know what I like and don't like sound-wise, but I have not categorized any of those amps by class, because I simply have not had the information. Such knowledge could be useful next time I'm fiending for a new head or power amp!
Thanks for your help!
Ok I got it...