The simplistic Salas low voltage shunt regulator

[RCruz]

Originally Posted by AndrewT

Yes, That's three times in as many days. I going to book myself in!

Do not even think about that Andrew.... We would all loose a crucial source of knowledge.

[AndrewT]

but as usual with internet knowledge, one must be capable of recognising if it is worth reading let alone trying to follow.

Will I know if my writings are going down the plug hole, or actually worth reading?

[RCruz]

You can be sure your writings are read and apreciated.
Not only by me but by most readers that follow your posts for a long time.
We lear a lot from you.

[ichiban]

Originally Posted by RCruz

You can be sure your writings are read and apreciated.
Not only by me but by most readers that follow your posts for a long time.
We lear a lot from you.

Ditto!!!

[j45yip]

Originally Posted by RCruz

You can be sure your writings are read and apreciated.
Not only by me but by most readers that follow your posts for a long time.
We lear a lot from you.

+3

[merlin el mago]

-1

[judby]

Hi, would you perchance have a reasonable spice model for the 9610 that you're willing to share?

Edit: Sorry, this question was for Salas, who'd told me he'd simulated "my" circuit

[judby]

Originally Posted by Salas

Picture a BJT instead of a MOSFET and its nothing more than the classic ''ring of two'' CCS concept.

Ring of two CCS for differential triode amp - stability?

Sorry for the delay, I was hit by work as my vacation ended. Now recovering from a bike accident I can get some fun done

Regarding the circuit, I was under the impression that Id is dependent on Vgs AND Vds. Looking at the 9610 datasheet I can see that, for -Vds > some 100mV, Id is practically constant given Vgs.
Then it's easy
If Vin increases, initially nothing happens in the FET part. The BJT will have to see an increased Ic - if Vce is constant (=Vgs) then Ic = (Vin - Vce) / (R2 + R3).
In order to increase Ic, Vbe must increase. But here we have exponential dependency so even the slightest increase in Vbe will make Ic increase. Because of this, the circuit is almost independent on Vin.

And one simple obvious way to improve this is to replace R2+R3 with a CCS, then Ic and also Vbe will be constant; the increased voltage drop given increase in Vin will happen over this CCS.

Thanks

And there can only be one explanation to the strange behaviour I see: instability - even if it does not show up anywhere.

Also, while looking at the 9610 datasheet I must admit, this FET has possibly only one advantage: it's cheap...

Next prototype will probably be 9240 based (got some in my stash now) and with the BJT/FET CCS with R2+R3 (R4 + R10 in the original) replaced with a K170.

[AndrewT]

Even cheaper are the N channel. They perform a bit better. Less capacitance, more transconductance.

[judby]

Originally Posted by AndrewT

Even cheaper are the N channel. They perform a bit better. Less capacitance, more transconductance.

Noted, thanks