Aleph-X: High-Power Version

[Ian Macmillan]

Something else to consider when deciding on how much bias is required for a given application. Simulation shows this circuit to have poor performance when current limited, i.e. insufficient current for the load impedance. The waveform becomes severely distored, unlike the typical clipping displayed due to voltage limiting.

I would be interested to know whether anybody has made measurments of this condition with a real circuit since it may just be an artifact of the simulator. Unfortunately I do not have the equipment to make the measurements on my own circuit.

If this is a real effect, then more bias may be required than at first envisaged (for low impedance loads).

Finally, I seem to recall someone giving a rule of thumb for how much current from the driver is required per output MOSFET in order to drive the capacitance of the latter. Unfortunately I cannot recall the value. This may be another significant factor when cosidering a JFET front end. Can anyone help me out here?

Ian.

[loren girsley]

I'm in agreement that a jfet front end for the Aleph is a tricky endeavore. At least two pairs would need to a paralelleled and a high transconductance type from the violet bias group would be reguired like the 2SJ109 or matched 2SJ74s. The V (violet bais group are very hard to find. A mosfet follower stage between the front end with IRF610s might allow a Jfet front end and would give another 3 or 4 volts and less capacitance across the required larger value drain load resistor for the front end.

An N jfet front end would require P channel mosfets for the out put gain stage. You don't even want to go there!

The 2SJ77 might make a very good current source transistor for the front end and allow easier heat sinking than the Zetex type that has been dissicussed.

http://ezphysics.nchu.edu.tw/prophys...eet/sj77e2.pdf


It is nice to the black and white of detailed disscuss of specific approaches to this design rather the generalization in the grey areas of generality that has become much to prevalent by some posters.


Loren

[Nelson Pass]

We use JFET inputs in all the X products except the
X1000, X600 and the XA's.

I've built a number of Alephs and other 2 stage circuits,
complementary and not, with JFET diff pair inputs and they
work very well, but as the transconductance is quite low
they tend not to have enough feedback for very low
distortion. But then we don't necessarily judge an amp
on that basis, do we?

In any case, you can pretty much pop them in, adjusting the
values of the current source and load resistance to fit the
part.

[cocolino]

originally posted by Nelson Pass
We use JFET inputs in all the X products except the X1000, X600 and the XA's.

Hmmm....If I recall okay I`ve seen a picture of a X350 in a German Hifi magazine review which showed MOS-Fet`s for the front-end.
As all the front-end transistors were in TO-220 cases they must have been MOS-Fet`s

[Bricolo]

One thing I want to add:

when you calculate the output power of your aleph-x, either using formulas or an excell file, for 2,4,6,8 ... ohms, you should better make a graph with the power function to the load impedance


I've seen many of you who have made calculations for their speakers, in example going from 6 to 8 ohms, telling:
so, with X volts and Y amps, i've got:
98W in 8 ohms
97.3W in 6 ohms


but what is between 8 and 6 ohms? certainly not a straigh line, but a peak!

I've made a graph, using berekening1.xls, and the power/impedance is shown at the bottom of my post.


You see that at first you have a current limited mode, then a peak, and after that, a voltage limited mode

[Ian Macmillan]

Interesting graph.

I would be interested in whether anyone has conducted any listening tests of this circuit under different limiting conditions. For example, does the amp sound different when current limted as opposed to voltage limited. I'm curious as simulations show some pretty bizzare behaviour when current limited.

BTW, I realise that no sane person would intentionally operate their amplifier in a severe limiting condition (not good for sound quality nor speaker life!) but even so, under some music conditions an amplifier may well approach or reach such limits. I guess the real question here is that if we have to run near one or other limit, which is to be preferred?

Ian.

[Bricolo]

I think that running it far from any kind of limitation is the best


Tell me if I'm wrong, but if you run it in non limited mode, you'll have a flat top for the curve I've shown

Am I right?

If I am, that's the way we should run our amps

[Nelson Pass]

Originally posted by cocolino
Hmmm....If I recall okay I`ve seen a picture of a X350 in a German Hifi magazine review which showed MOS-Fet`s for the front-end.
As all the front-end transistors were in TO-220 cases they must have been MOS-Fet`s

Current X350's use JFETs.

[Nelson Pass]

Originally posted by Bricolo
I think that running it far from any kind of limitation is the best


Tell me if I'm wrong, but if you run it in non limited mode, you'll have a flat top for the curve I've shown

Am I right?

If I am, that's the way we should run our amps

You will not get a flat top to speak of. Pick your voltage and
bias to get the peak figure you want, and welcome to the
world of single-ended Class A.

[Bricolo]

Originally posted by Nelson Pass

You will not get a flat top to speak of. Pick your voltage and
bias to get the peak figure you want, and welcome to the
world of single-ended Class A.

do you mean that with a smaller input voltage, I'll have the same curve, but with lower power?