Psu with rotary switches

[Simon123]

Hello!
I wana build power supply on which i can set the voltage with rotary switches.
i will use lm317 as cinstant 100mA supply and resistors to make voltage drops over it.
91ohm and trimmer for 10,20 volts, 10ohm for 1,2,3,4,5,6,7,8,9 volts and 1ohm for100mv,...
When switch is shorted out there is 0volts on switch.
How can i make output stage,so it wont have any voltage drop, or maybe couple of milivolts.
I tried emitter follower circuit, but it isnt stable, depending on load and it has voltage drop.
And i will use 24volt transformer, which will give me about 35volts dc and 8 amps.

Thanks!

[magnus0re]

Your description falls a bit short.

sounds like you want to make something like this: https://en.wikipedia.org/wiki/File:LM317_1A_ConstCurrent.svg

the lm317 has a dropout voltage of 1.5-2.5 volts, depending on current.

24 volt trafo * 1.41 = 33,9 V ~ 34V
rectifier diode drop: 2V  (source 1n4001-4007 datasheet)
34-2 V = 32 V  trafo->rectifier
32 - 2 Volt = 30 V rectifier -> lm317
So, you'll get about a max of 30 V, with low current out. 

BUT this configuration is not a current regulator, it is a current limiter. so current depends on the load UNTIL the load is so heavy that I*R > 1.25 volts , then the current regulation kicks in.

nothing of that design has anything to do with volts.

if you need a power supply, you need a voltage power supply with some protection, not a current limiter.

my 0.02$

[Simon123]

I want to use lm317 as 100mA current limiter connected to the switches.
With ohms law i calculated resistor values for difrent voltages.
At 10volts and 100ma i need 100ohm resistor and there are 2 to be able to heve 10 and 20 volts(i can dial up to 29.9 with other two switches.)
At 1volt i need 10ohm resistor and 0.1volt i need 1ohm resistor.
So, i i add up for example 100ohms for 10 volts and 50ohms for five volts i get  150ohms, -ohms law 150*0.1 =15 volts drop.
This is just the regulation part, but i dnot know how to make output part with which is stable and amplifies my voltage from 100ma to max 8amperes.

[Simon123]

Ignore op amp and transistors-output stage, beacuse this is what i dont want.

[magnus0re]

have you read the datasheet for the lm317 ? if not, here it is:http://www.ti.com/lit/ds/symlink/lm117.pdf.

as you have stated it, My understanding of your design is:

100ma through resistor gives voltage.

then amplify voltage AND current to a power output of max 8Amps.

instead of doing this, you can simply use the circuit from Figure 39 in the datasheet. The voltage is controlled from R5, and you can swap that out with switches that make before break.

[Simon123]

But then it wont be very acurate and resistors wont be in linear values.
Now i have found this circuit http://terpconnect.umd.edu/~toh/ElectroSim/Booster.html
Hope itl work.


Yea, it works, thanks anyway.

[Simon123]

Ok, now i have schematic and there are some problems.
Which op amp would work on 35volts and  would not die on 30volts input?
I want something affordable.

Theres also problem when rotating switches, when non of the contacts are togeather and the voltage goes to maximum? Ill proably use zener, transistor and connect it to voltage reference output.
It should short out, when it excedes about 30v(when i rotate the switch).Would it work?

Thanks!

[Dave]

LM317 is not a temperature stable current source. Using resistor in the order of 1 and 10 ohms is ridiculous with mechanical switches, the contact resistances will ruin everything for you.

Here is a quick sketch of a better circuit:


Now this is just a circuit that will give you an adjustable set point for your output. You still need the part of the circuit that will take this voltage and use it to regulate the output. Think of it as a replacement for a DAC, if you were to make a digitally programmable power supply.
You want to use SHORTING (a.k.a. MBB - make before break) switches for the voltage adjustment decades, otherwise the output voltage could go wild while you are changing values.
There is no need for trimmers in a 3-decade system. You can just buy a bunch of 1% resistors and pick the best ones with a multimeter. Shouldn't be a difficult task.

Let me know if you have any trouble selecting the components and their values, I'll walk you through it.

[Simon123]

Ok, i still have problem finding op amp, that could work on these voltages.

[dr_p]

LM12 


but why use the op-amp straight from the rails and directly driving the output ?


It's my understanding you need an adjustable regulated 8A voltage supply.


So why not build/buy/get any fixed power supply  and interfere with the feedback to achieve your desired voltages.


Something like this is a basic regulated supply:


YMMV

[Dave]

If you are going to make an 8 amp power supply, you are going to need voltage sense lines, otherwise the bottom voltage decade will be completely irrelevant at full load.
You also need to add current limiting, because things could be torn apart if you happen to overload it (and it will happen).

[Simon123]

If you looked at the schematic you could see, that i have overload protection with flipflop.
what do you mean about voltage sense, beacuse opamp at transistors should make output at transistor the same as input?

[Dave]

I was thinking more like adjustable current limiting / cutoff, so you can protect your DUT as well as your power supply.

As for the sense lines, this is what the output of an Agilent E3634A (goes to 7A) looks like:


Four output jacks. Interesting, isn't it? I'll let you figure this one out yourself.

[Simon123]

So should i connect all inverting(-) inputs on opamps togeather and have it as sense line on positive side and disconect ground connection on voltage selection part and also have it as a sense line?
i would be proably using thick wires and i dont see very big advantage with sense lines, specialy beacuse its resolution is 100mV.