scaling or voltate divider step down using LM2596?

[Peter Stein]

I have been trying to use an adjustable LM2596 ebay style module to scale the input voltage with out using extra opamps and cctry. Has anyone tried to do this and been successful. My application is to feed say 30V into the module and have a 1/3rd of that voltage at the output. On the surface it seemed easy - just use a 3 to 1 voltage divider resistor network accross the input voltage - then connect the FB pin (i.e. remove the pot) to the voltage divider in some way - however that won't have any regulation function, as the real output is not connected to the FB. So simplest extra parts solution is the best I can hope for.

Cheers

Peter
(previous post flew out of my PC before I finished)

 

[ronsimpson]

The LM2596 does not allow you to play with the reference voltage. ( 1.235V)

The output voltage is sent into the error amplifier and compared against 1/3 the input voltage.
Another way to say it: Don't use 1.235 volts as a reference but use a fraction of the input voltage.
You need a part with a unconnected error amplifier. Both inputs need to be brought out of the IC.

How accurate do you need to have the 3:1 ratio?

 

[ronsimpson]

OK here is a circuit that mostly works. If Vout is low then FB is low and the PWM must work harder. If Vout is high then FB is too high and the PWM will reduce the duty cycle. If you PWM's FB pin is something low like 1.25 volts or 2.5V this thing should work.

 

[Peter Stein]

The LM2596 does not allow you to play with the reference voltage. ( 1.235V)

The output voltage is sent into the error amplifier and compared against 1/3 the input voltage.
Another way to say it: Don't use 1.235 volts as a reference but use a fraction of the input voltage.
You need a part with a unconnected error amplifier. Both inputs need to be brought out of the IC.

How accurate do you need to have the 3:1 ratio?

Not that accurate - 5-10% would be fine. Yes you hit the nail on the head - not having access to the ref voltage is the issue - just thought I could find a cheats way around the limitation.

Cheers

Peter

 

[Peter Stein]

OK here is a circuit that mostly works. If Vout is low then FB is low and the PWM must work harder. If Vout is high then FB is too high and the PWM will reduce the duty cycle. If you PWM's FB pin is something low like 1.25 volts or 2.5V this thing should work.
View attachment 89543

That may be a better way of going than using an Opamp comparator - but was still hoping for less parts to make it go

 

[ronsimpson]

but was still hoping for less parts to make it go

Well: R3, R4 you need any way to get the 1/3.
D1 is there to temperature compensate. You could drop it out.
R1 could be removed.

Don't use the 3.3V , 5V or 12V versions of the IC. Do use the -ADJ version.

 

[Peter Stein]

Well: R3, R4 you need any way to get the 1/3.
D1 is there to temperature compensate. You could drop it out.
R1 could be removed.

Don't use the 3.3V , 5V or 12V versions of the IC. Do use the -ADJ version.

I will give that a go - as I doubt if I can get away with less and still get a reasonable result. I knew I had to use the ADJ version - and from the limited Ebay specs, I thought I could just remove the normally included pot and redirect the wiper connection - I just did not think it through properly before I ordered a few of these to play with.

Thanks for the input and suggestions.

I always considered the idea of scaling a voltage would be a common problem/need - to efficiently drive a lower voltage motor/lamp etc from a higher variable one - - and on one of my searches of this forum I found another use - and that is of a tracking pre regulator for low noise efficient variable DC supplies.

Cheers

and thanks
Peter

 

[Peter Stein]

I have been testing the idea for the last few days - and all seems good. I toasted one regulator board when ramping the input volts up to 55V so I guess the ebay ones using what is labelled as LM2596 HVS are just fakes or rejects!

But good solution thanks Ron