ETI 5000 MOSFET Power amp

[PMA]

Originally posted by Terry Demol


I, and plenty of others, can design an amp that gets around 0.001%
or less THD at close to full power right up to 20kHz.

But there's not much point in buying or building an amp that you
don't want to listen to.


Cheers,

Terry

Don't mix this together, Terry. The very low THD does not automatically mean it is an amp you do not want to listen to. As you know there are another factors of influence.

My question would be different: make 2 amps, both very pleasant to listen to. 1st with THD of 0.00X%, flat with power, flat with frequency. 2nd with THD of 0.X% or X%, rising with power and rising with frequency. Listen to well recorded (Bishop, Renner) Beethoven, Mahler, and tell me, which of the amps sounded better.

Cheers,
Pavel

[The Saint]

Hi Suzy

I have had heaps of experience with Laterals Fets and while they
do have problems with parisitic oscillation, they can be fixed so
they are unconditional stable, even into 2 Ohm loads.
I have designed and built amplifiers that develop as much as 1400 watts into 2 Ohm loads and they are 100% stable.

i am also very familiar with the AEM 6000 amp.
So if you require any help with it let me know.

You may find that you will have to increase the gate resistors on the N-channel fets with respect to the P-channel.
I built one of the AEM 6000 amps when they first can out.
A very nice design and sounded OK as well.

Anyway if you need help with just let me know.

[AKSA]

PMA wrote:

My question would be different: make 2 amps, both very pleasant to listen to. 1st with THD of 0.00X%, flat with power, flat with frequency. 2nd with THD of 0.X% or X%, rising with power and rising with frequency. Listen to well recorded (Bishop, Renner) Beethoven, Mahler, and tell me, which of the amps sounded better.

Aside from the narrow range of audition music, why would anyone want to build a couple of amps and then report the results to you? You have made it amply clear what you think, surely the aim is to prove something to the majority?

And is the ear the final arbiter? Or the measurements? Which is it to be?


Hugh

[darkfenriz]

Terry
Design and show me, please.
Because there are different ways to achieve such a goal, none these ways is easy, some better and some worse to other factors.

[PMA]

Keep cool, Hugh (and decent as usually).

Terry has suggested a kind of comparison of 2 amps, I do suggest another one. What I wrote works for any kind of music, the classical is the fastest to judge by ears.

[AKSA]

I'm cool as a cucumber....... age and treachery wins out against youth and enthusiasm every time......

This eternal, seemingly insoluble debate goes to the crux of how the marketers shape and exploit the market.

Now, which has it? The ears, or the meters?

Hugh

[TheFettler]

Now, which has it? The ears, or the meters?

hmm... the ears all disagree 'cos , well they're all basically different , and ( as we all know ) numbers can be made to say anything

For me at least it comes down to what I would choose to listen to.

I reckon that anyone old enough and wise enough to actually know what they are listening for, is too old to actually have the auditory range to actually *hear* it.... especially true for men, whose hearing drops off quicker than women.

When I need a second opinion about audio I always ask my wife what she hears , partly because she has better hearing , and partly because she used to work as a professional sound engineer....

Ray

[Samuel Jayaraj]

You may find that you will have to increase the gate resistors on the N-channel fets with respect to the P-channel.

Try a value of 680E for the N-channel and 470E for the P-channel FETs. This combination also sounds the best sonically in my experience.

Hai Mr.Holton, great to see you active here again. Been looking at your new offerings of all-mosfet amps. Fantastic job as usual.

[The Saint]

Thank you Sam
Nice to hear from you also

[suzyj]

What a hoot!

I've powered my monoblocks up, and am incredibly happy with them. A few lessons though:

10 Ohm 1/2 Watt resistors are no good charging the on-board electrolytics

The Vgs threshold on my MOSFETS is a little higher than the data would indicate - I had to remove the 220 Ohm resistor across the bias trimpot to get a full 100mA quiescent current.

It's a tad tempramental, and prone to taking off at 1.1MHz. I have no idea how people build these things without a good CRO. Increasing the compensation caps on the drivers from 12p to 22p helped, but hasn't cured it completely - with a speaker connected, it occasionally oscillates when I power-cycle it. For now I've just decreased the global feedback (increased the gain from 17dB to 23dB), which has cured it, but there's clearly still some work to be done optimising the compensation networks. I'd ideally like to be able to run it with more global feedback.

I'm leery of just increasing the gate resistors. That'll stop oscillations, sure, but it does nothing for slew rate.

Having hooked one up to a speaker, and played some of my favourite music (Bernard Fanning, Mike Oldfield, Suzanne Vega, James) through it, I've gotta say, I'm well impressed. It's amazingly transparent and clear. I find the limits of the speaker (not to mention the limits of my husband's patience) a long time before it starts to clip.

I've got a bit of work to do now to work out how to measure its THD and IMD performance. All the gear I have access to at work is for HF/microwave use, and has pretty poor performance at audio. I've had a fairly feeble attempt at measuring the distortion with my notebooks on-board sound card, but have just found that my notebook's sound is shocking. A loopback gives me 0.02%.

If anyone has any recommendations for a good (or at least hackable) USB sound card I'd be very grateful.

I'm _so_ happy.

Cheers,

Suzy