I am looking at using the TL783CKCSE3 regulator

https://www.ti.com/lit/ds/symlink/tl783 ... r.com%252F

For two different applications.

1. Input 120Vdc output 105Vdc.
2. Input 61Vdc output 12Vdc.

What resistor values will I need to produce those exact voltages?

1. Needs to be exact.
2. Can be anywhere from 10-25Vdc.

Current in both applications will be well below maximum and above the 15mA minimum.

1. Shouldn't need protection if regulator shorts from input to output.
2. Protection will be needed. I'm thinking a zener between output and ground and a fuse between the regulator output and zener would work.

See section 9.2 of the data sheet.

V out = Vref ( 1 + R2/R1)

Vref is 1.25 V

More instruction there on designing whatever voltage you want.

Rich

Don't forget the heatsinks. With higher voltage drops, the power in the regulator is higher, too.

It may be helpful to add series resistors between the raw DC sources and the regulator input pins. Enough resistance to drop most of the voltage at the maximum load current.

This will reduce the power dissipated in the regulator, and limit the current in case of a short..

The voltage output is never "exact". The number of regulators used is exact, but not a voltage. The regulator's reference voltage has a tolerance, as do the resistors on the adj pin. A trim may be necessary. For a one-off, a selected fixed resistor across one of the voltage set resistors might work.

Ted

For those who don't like trip pots: https://pe2bz.philpem.me.uk/Power/-%20-%20Power-Design%20-%20-/PowerSupplyDesign/LB-46.pdf
John

I saw that formula, but wasn't sure if I was doing the calculation right.

I found an online formula for a LM-317 and here's what I got.

For an 82 ohm R1 resistor I get a resistance of 6.8K for a voltage of +105Vdc and the input B+ will be around 120Vdc.


For an 82 ohm R1 resistor I get a resistance of

820 ohms for +13.75 volts
1K for +16.494 volts
1.5K for +24.116 Volts

B+ will be around 60Vdc

So I gotta figure out what voltage is best. I'm thinking the lower voltage as the 7805 on the clock board has no heatsink.

Now are those calculations correct?

In both instances the regulators will be mounted to the chassis using a transistor mounting kit to keep them isolated from the metal chassis while providing adequate heatsinking.

The first example the regulator is pretty much just a drop in as I already have the filter caps and rectifier.

The second one I am using a 1/2 wave rectified supply and 5600uF cap as the main filter cap.

Your calculations are correct.
Vout dividied by Vref, subtract 1, multiply times R1, equals R2

Thanks.

I'll order the parts.

I just ordered some from Newark. They are clearing out inventory at super low prices.

https://www.newark.com/texas-instrument ... TL783CKTTR

Yes, it is in a surface mount pack, but for 12 cents, I can find some uses!

Rich

Cool.

I've already got a partial cart at Mouser.

EDIT:

Parts are on order.

I went ahead and bought the three resistors so I can decide which is the best choice.

I'm tempted to add two 12 volt zeners in series across the input to the clock board then add a 1/2 amp or smaller fuse so that if the regulator should short from input to output the zeners conduct and blow the fuse thusfore saving the clock board from damage since the 7805 is not rated for a 60 volt input.

I got the regulators installed.

The one for the clock works. With the 15F cap discharged (charges through three diodes and a 10 ohm resistor to 3.3Vdc) the B+ is at a steady 4.95Vdc (needed a small heatsink on the 7805 as it got quite warm during the cap charging) and once the cap is fully charged it is about 4.97Vdc so I do believe it works great. The TL783 regulator is mounted to the chassis and barely gets warm.

The one for the scope, however I initially had the 6.8K and 82 ohm resistors swapped and got about 7Vdc out. When I connected it right the regulator wasn't exactly regulating properly so I figured I must have damaged it somehow. For now I lust put back the original regulator.