Mini Aleph

[croccodillo]

Hi all,

I just acquired for a really cheap price two nice heatsink, so why not to use them for a wonderful, "little" Mini-Aleph?

THIS is the schematic I would use.
It is the standard mini-aleph schematic, as published before here in the relevant thread; I only changed the CCS in the first stage.

Just few questions:

1. Is there anybody that sells PCB already done for the mini-aleph here in Europe? I don't like to place orders in the USA, due to loooong transit time, high custom tazes here in Italy and the too high rate of lost parcels coming from there (I know the Italian Post do not like shipment from the USA).

2. I supposed three different CCS for the differential input stage, the first on the left is the "traditional" Pass's stage, slightly modified (an LM431 instead of a zener), the other two are CCS using depletion mosfets, single and cascoded.
Which do you prefer? In your opinion, which one would give the better performance? I do prefer the cascoded depletion one...

3. What about a well regulated, stabilized power supply?
My idea was to use two separate 120W toroid, each one followed by a LM317/LM7815 based voltage follower.
Imagine something like the circuits you can see HERE.

Let me know.
Ciao,
Giovanni

[Zen Mod]

well , I'm probably biased , but - why don't you make ALeph J , even as mini iteration ?

in Babelfish J , I use two bjt CCS , finding it even better than mentioned solutions ;

anyway - look for Babelfish thread ; there you can find pdfs for pcb , and make your own with toner transfer routine

[croccodillo]

Well, the idea of a cascode BJT CCS is really interesting... I also prefer BJT for CCS.
I'm developing my PCB for my future Aleph; I decided to go with separate boards, one for the driver and one for the poewr (Mosfet).
In this way I can later ad some more output mosfet on the output board without having to change any PCB.
The output board is ready (it is really trivial and no alteration are possible on the output stage); I'm now drawind the final scheamtic of the driver stage, and I have some questions for you.

Referring to THIS schematic, which is the purpose of the C1 capacitor (in the active current regulator)?
As far as I can understand, it is there to provide better power regulation to the Q3 transistor.
I also think it is used as a "bootstrap" capacitor when you have a very large positive peak at the output: in this way the output mosfet gate can be lifted well above the positive power supply voltage (in this case, 15V), and thus the output can go up to (almost) the V+ voltage before clipping, granted more output power.

Am I right?

If I'm right, and there's no other purpose for it, the schematic could be modified as per
THIS way; gone is C1, and a new power supply voltage has been added.
Of course I know lot of components values shown there are not correct (I just made a cut and paste from the previous schematic, just to show you the idea).
This has some advantages (in my opinion): having more room you can easily use a bipolar cascode CCS, or a Mosfet cascode CCS, or any other CCS you can think of.
Again, the output mosfet gate can be driven in a more effective way, having an higher available voltage for it.

What do you think, would it work? Can I proceeed with this project or am I missing something?
Any other suggestion?

The last think: I have an idea on a CCS; it should work, but I'd like to listen to your opinion.
THIS is the idea (it is not mine, by the way).

Ciao,
Giovanni

[lineup]

Yes, C1 is one bootstrap cap.
It will keep voltage constant across R8 and R12, when the upper MOSFET follows the Output voltage.

And so keep such current variations from disturb Q3.
Q3 would disturb MOSFET and I am sure it would increase distortion without C1.

[lineup]

Your idea of using one Power BJT for a 'normal' PNP currrent source is good.
MJ15025 or MJL21193 for example
Haven't this been tested before?
I would be surprised if not ....

One logical modification would be one/two P-MOSFET CCS, IRF9240 and similar.
BD140 (or MJE350 ) could be used to control the source current of MOS.
If Vbe of BD140 is 0.7 Volt, then Constant current would be like:
I = 0.7 / R-source (of the IRF9240)

[croccodillo]

Originally posted by lineup
Your idea of using one Power BJT for a 'normal' PNP currrent source is good.
MJ15025 or MJL21193 for example
Haven't this been tested before?
I would be surprised if not ....

One logical modification would be one/two P-MOSFET CCS, IRF9240 and similar.
BD140 (or MJE350 ) could be used to control the source current of MOS.
If Vbe of BD140 is 0.7 Volt, then Constant current would be like:
I = 0.7 / R-source (of the IRF9240)

I'm not sure I understood your thought... the problem is on my side, of course.
Could you sketch it?

If, as supposed, C1 is a bootstrap capacitor, would it be useful to use a floating power supply for Q3, connected between R8 and Q3 emitter?

My idea is shown HERE.

Provide a floating, well regulated 15V/50mA power supply is very trivial.

Ciao,
Giovanni

[lineup]

In output the upper mosfet works as a constant current source
I thought you wanted to try to replace this current source

Replacing the input stage CCS with a bipolar is alright.
It will work at least as good as the original.
For a supply of 15-25 Volt and 'only' like 20-25 mA, even one BD140 or MJE350 (TO126) would be enough.
( 20 Volt x 25 mA = 0.5 Watt power ) in this CCS transistor
------

As the upper mosfet via SOURCE follows the output voltage, up and down,
the voltage at GATE would be effected if we had only resistors to bias from V+

Yes, a floating regulated voltage, would do the same job as C1.
A bootstrap is equal to a constant voltage holder.
The task is to keep the GATE-SOURCE voltage of upper mosfet constant.

Using one bootstrapping capacitor is a much more simple method.
And I think this cap of a suitable value does the job almost as good as a floated constant voltage.

[croccodillo]

Thank you for the explanation.
You actually confirmed me what I supposed.
I will develop the PCB in manner it can accept both a floating supply and C1, and I will try which one works best (if there will be any difference).
As you said C1 is a much more simple way to go.

Ciao,
Giovanni