Amplifier Ratings....

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Amongst amplifier specifications are power delivered into varying load impedences, 8 and 4 ohms being the most common.
Some amplifiers double rated power into 4 ohms WRT 8 ohms and some achieve less than doubling.
I have always found amplifiers that double power into half impedence load to sound more correct.
Is this lacking of power delivery a fundamental distortion mechanism ?.
 
Hi Eric

Beneath PSU sag, there is also the SOA protection circuitry for the output devices, that may reduce power into low impedance loads.

Badly engineered SOA protection circuitry can be detrimental to sound.

BTW: There's an interesting series about such protection circuits in some of the later issues of EW+WW (written by a diyAudio member).

Regards

Charles
 
You're right up against one of the things that I get obsessed about...and that there aren't any accepted ways to measure.
If you define distortion in the broader sense of anything that isn't simply a magnified version of what went into the amp, I'd say that, yes, it is a distortion mechanism. A volume related one, at least.
My thought experiment runs like this: If the amp doesn't have sufficient bias to double into a 4 ohm load and the speaker has a dip in the impedance curve to...I dunno, let's make it 3.5 ohms...then there will be a lessening of dynamics when the amp tries to deliver more power than it is capable of delivering. The voltage may be there, but the amp current limits at moderate to high volumes.
There ought to be some way to do a broadband burst--white noise, probably--into a known load, where you know what went in and can measure what happened at the other end. It would need to be done at a fairly decent volume (pick a number...what say we standardize at 75% of full rated 8 ohm power?), so that could get pretty obnoxious to the tester. I'm getting ringing in the ears just from thinking about trying this with something like an X-1000. I guess a test chamber would be in order.
Weren't some of the old Apogee speakers 1 or 2 ohms for a fair portion of the spectrum? They might be a good start. Or perhaps it could be done with a non-trivial dummy load on the bench.
Besides, even with class A amps, more bias sounds better.
Does that come anywhere near what you were thinking?

Grey
 
It could be voltage drops through the amplifier, particularly the output stage. For instance a Triple Darlington like the LeachAmp will have a greater voltage drop, and therefore less power, into a lower load. I think I simulated 125W and 240W for 8 and 4 ohms at the onset of clipping using ideal rails of 50V. The nearly double current through the 10 ohm base resistors made that much of a difference. . . . (2V p-p) :rolleyes:

But I would assume you mean more egregious descrepancies like 100W/8, 150W/4?

Mark Broker
 
I don't think that the limiting role in the Leach Amp, at least, is the voltage drop of the triple Darlington. It is more with the voltage rails driving the voltage amplifiers.

Consider the output voltage swing of the input diff amp stage. If we assume a peak input voltage of 1 volt, and assume a diff amp stage gain of 2 (per Leach), the output of the diff amps will have a maximum 2 volt swing in both directions. I haven't calculated it, but I assume the collector resistance is such that it will drop several volts at zero input, and since it has to accomodate the voltage amplification of the input, the quiescent collector voltage is probably around 48-52 volts with a 58 volt power supply. The second VA stage will have its maximum possible peak output reduced to this level, minus the Vbe drop plus some headroom, thus producing a maximum output to the first Darlington driver to the 45-50 volt range, I estimate.

The Vbe drops of the Darlington's will lower the ultimate output voltage more, but certainly not by 10 volts as we see with the VA stages. What one could do, and I brought this up on a thread a while back, is to operate the VA stages at a higher voltage, using a low power transformer, so that the full 58 VDC is available for power delivery via the Darlingtons. In other words, if we could run the VA stages off of, say 68 volts, and run the drivers off of 58 volts, we can maximize the utilization of the power transformer. The drawback is that we need that second, lower power transformer, and we would need to recalculate the values of our components in the VA stages.

There has also been some discussion about the need for the 10 Ohm base resistors. I don't think they are really needed, and Leach only included them on Version 4 and later boards. They are there as insurance to help prevent any possibility of transistor oscillations. If you keep the base leads short, one should not have a problem.
 
Jeff, I had actually used the Leach Amp as an example of possible discrepancies for the "double-down" rule, as Eric had questioned above. I didn't mean to suggest that the main reason its power rating is rather low compared to other amps using 60V rails is voltage drops through the triple darlington. FWIW, I had simulated about a 44.5Vmax into an 8 ohm load, and about 43.5Vmax into a 4 ohm load with ideal 50V rails. I would agree with your assessment of the amplifier, though. :)

As opposed to a seperate power supply transformer, have you considered using a voltage doubler off the main power supply transformer? There's a superb example of one in the Pass A75 writeup. :cool:

Mark Broker
 
I think all amps will deliver double the power into half the load as long as your not talking about maximum power.

It will only depend on the damping faktor how much voltage is "lost" on it´s way to the speaker.

An amp that delivers 100 watts into 4 and 8 ohms will when tested at 50watts/8ohms will deliver 100watts when the load is changed to 4 ohms.

I think what Eric hears and what I also hear is that amps with big current capability are a lot more powerfull when driving low impedance speakers.

You only have to look at a speaker with a 2 ohm minimum, here a so called 100watt/8 ohms amp will only deliver 25 watts instead of the 400 that are called for.
This means that when driving these speakers the power into 8 ohms should not go any higher than 25/4=6.25 watt. Any more than this will give compression into impedances lower than 2 ohms.

william
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.