Power Supply Schematic - STEPS or TREAD - Page 2

Quote:

Originally Posted by Avro_Arrow 

Here is a LED CCS to replace the CRD:

 

 

I hope that is enough information

A suitable LED would be 630-HLMP-D101 on mouser

I used that in my "cagematch" amp. Its nice. 

You could also use a JFET in CCS mode. 

Quote:

Originally Posted by scootsit 

That would work! I should match transistors, right?

 

I didnt. 

The LED voltage will have an effect on the CCS current too so if you do, match your LEDs too. 

I buy my LEDs in bulk from MPJA. about $2 for 100 :)

http://www.mpja.com/LEDs/products/116/

Quote:

Originally Posted by scootsit 

Couldn't I eliminate everything and just make R2 a varisistor?

Do you mean a trimpot? Then yes. A trimpot would also save you the trouble of matching transistors & LEDs.

To repeat, a varistor is totally not what you want at all. 

Quote:

Originally Posted by scootsit 

I think I'm going to make it easier on myself and just go with the 1N5292, which yields slightly higher current, but Pete Millett, in his original article played with. Anyway, I'm building this sucker on protoboard, so I want to lay out my boards before I place any orders.

 

Thanks for the help, guys.

I guess none of the above matters anymore :p

Maybe try for a ~1.5ma CRD - or go back to the BJT CCS with a trimpot. Either one is a total departure from the original design, BUT it has some advantages in actual use for me anyways. 

After a bit of time with a classic Millet hybrid I kind of got in the opinion that the only "millet hybrid tubes" worth using are 12fk6 - the 12AE6 and all of the others have too much gain for me and I preferred the way the 12fk6 sounds to boot. The info I have found for the 12fk6 specifies 1.5mA idle current, so its easy to feed it well. 

http://www.nj7p.org/Tube4.php?tube=12fk6

Why can't you use a current mirror? See the PIMETA v2 class A biasing circuit for the concept.

Bonus: it eliminates two resistors and an LED over duplicating the whole CCS.

Thats a really good idea...I'm going to look at that tomorrow.

Quote:

Originally Posted by tangent 

Why can't you use a current mirror? See the PIMETA v2 class A biasing circuit for the concept.

 

Bonus: it eliminates two resistors and an LED over duplicating the whole CCS.

@Tangent

Is this what you were thinking about?

Something like that, yeah.

The LED resistor is probably too high to get illumination or good current regulation.

I'm not certain the polarities are all right, but it looks like you've simulated it, so I assume you're satisfied on that front. I guess you had to flip this around because of your different application.

Yes, I simulated it. The values are for an input voltage of 24 volts and an output of 580uA on AM2 and AM3.

AM2 and AM3 would be the connection to the anode of the tube.

This is what I was think...

Sweet-looking optimization!

By the way, never mind about the too-high resistor value. I didn't click on your first schematic above and get a closer look. I probably saw "221k".