I've been searching for new amps for some time and had not even considered Krell (actually, I wondered if maybe they had gone out of business). Then I saw a YouTube interview with their COO and he was talking about their Class A iBias technology (lots of forum members have posted about the magic of pairing Pass Class A amps with ML's but I don't think I can deal with the heat output). Anyway, Krell's claim is their new amps can run in Class A to full rated power by adjusting the current supplying the output transistors to only the current they actually need, thus reducing both the power draw of the amp as well as the associated heat output. They go on to explain that they calculate the bias at the output stage which then makes the real-time demands of the speaker a part of the calculation. I doubt I am explaining this as well as they could so here's a link to their one page explanation of iBias Technology.
I'm not an engineer, and no kind of amplifier expert so I'm trying to figure out if this is just marketing spin, or if this actually works (and specifically, works well with ML electrostats). Also, I'm struggling to understand how if they measure this at the output stage so they can include what the speaker is calling for, it seems that they would be measuring and then sending back instructions to change the bias on the output transistors after the actual signal had already passed through them. IOW, it sounds like they are measuring and changing the bias 'after the fact' if that makes sense. Can someone help me understand this, and whether this sounds like a valid approach?
Also, does anyone have direct experience using one of the Krell amps with the iBias technology to drive their ML electrostats?
I'm not an engineer, and no kind of amplifier expert so I'm trying to figure out if this is just marketing spin, or if this actually works (and specifically, works well with ML electrostats). Also, I'm struggling to understand how if they measure this at the output stage so they can include what the speaker is calling for, it seems that they would be measuring and then sending back instructions to change the bias on the output transistors after the actual signal had already passed through them. IOW, it sounds like they are measuring and changing the bias 'after the fact' if that makes sense. Can someone help me understand this, and whether this sounds like a valid approach?
Also, does anyone have direct experience using one of the Krell amps with the iBias technology to drive their ML electrostats?