Re: message repost

At 04:10 PM 9/4/97 -0800, Clay Cowgill wrote:
>>With a very slight modification to this design you can use the original WG
>>PNP pass transistor to boost the current on the positive rail, and the NPN
>>transistor for the negative rail (the opposite of the way WG currently uses
>>them). Assuming only a current gain of 10, for the power transistors, this
>>allows the regulators to only dissapate 500ma for 5 amps of current.
>>
>>I don't have the schematic here, (it's hand drawn) but I can scan it in this
>>weekend and send to to Al to post.
>
>Mmmmm. Good point, I'd like to see your modifications there.
>
>One other approach would be to get the regulators in a TO-3 package and
>mount them on the chassis instead of the NPN/PNP pair for the old LV
>supply. I think the TO-3 versions are rated for something like 7A
>continuous?

I couldn't find them, if I remember the TO-3 were the same ratings as the
Tab parts, the TO-3 have better coupling onto the heatsink allowing slightly
higher power dissipation. It wasn't all that different though, and the
current limiting was the same.

I have seen some TO-3, 5 volt fixed parts that can regulate up to 10 amp (in
a surplus store). They could be used as an adjustable part. I've only seen
positive regulators that do this, and the pins won't fit in a standard
socket -- too fat, and you'd still have to do something for the negative side.

The design of an external pass transistor was taken right out of a Motorola
data book (though I had to grab a little here and a little there to get the
whole design).

It's hooked up like this (Pardon the ASCII art):

    Vin -------*--------\ /---------*-------- Vout
               | v / |
               | ----- |
               | | |
               | | |
               | | |
               | | |
               *--/\/\/----*---[LM317]--*
                   R1 |
                                  |

                            Adjust circuitry

Use the NPN for the LM337 side.

As the LM317 draws current through the PASS transistor's BE junction the
transistor it turns on.

R1 is used to substain regulation at low current levels and is calculated to
have a .6v drop across it when you want the transistor to turn on. I picked
somewhere between 50ma and 100ma. Anything below this threshold will use
only the regulator. Above that the current is shared between the regulator
and the pass transistor based on the gain of the transistor. It's probably
not needed here since the HV supply makes sure you're never uses that little
current.

The output of the pass transistor shares the output of the regulator so full
regulation is maintained. A scope on the output of this shows a flat line
regardless of what is being displayed. The regulator is faster than the
noise generated by the HV, so all the HV noise goes away.

There is one problem with this circuit, it is no longer short circuit
protected (the WG design was never short circuit protected!). The regulator
protects itself but has no way of turning off the pass transistor. It takes
a second transistor and a current sense resistor to do this, it all depends
on the amount of bullet proofing you want to do.

-Zonn
Received on Thu Sep 4 17:22:42 1997