Constant Current Supply with Switching Regulator

Hi,

I am currently attempting to design a power supply for a battery charger project. I require a constant current in the range of a few amps. I originally intended to use an LM317 but linear regulators get really hot. What I'd like to do is use a switching voltage regulator to create a constant current source.

I found the LT1074 switching regulator (here) that can supply up to 5 amps, and has an input voltage of up to 64v. I expect that if I modify the feedback circuit to represent the current instead of the voltage then I should be able to regulate the current.

I propose to insert a 1 ohm current sense resistor in the output path and measure the voltage across it using an instrumentation amplifier. The gain of the inst. amp will be proportional to the current set point, Gain=2.21v / (Iout * Rsense) The 2.21v is from the error amplifier inside the package. See the attached image


I'm also hoping to power this circuit off of mains voltages using a rectifier bridge. I am wondering if this modification should work or if there is a better option, or a package that is specifically designed for this purpose.

Any and all comments and pointers would be appreciated

The concept with a current sensing resistor and an error amplifier controlling the FB pin seems to be the right way to go ..
Attached is a working example of similar project, where the current–to-voltage conversion is sensed by one Opamp and the signal is then amplified by the second Opamp ..
Whatever you do keep in mind that the control signal is somewhere around 2.2V, so it is a good idea to employ an Opamp that can easily handle signals close to the GND rail, especially if you consider supplying this Opamp of just single voltage ..

Rgds,
IanP
:wink:

Ah, yes this is pretty well what I have done. I'm just using an in-amp package instead of making one. I have another issue though, since my goal is to make this plug into the mains voltages, the in-amp needs to be able to handle 60v. The AD620 that I was considering can only handle ~36 volts and I can't seem to find any thing that *will* handle these voltages.

I'm thinking that I need to make a fake ground like I've done here

but I'm unsure how this will effect the operation of the switcher.

The ground pin is the ground reference for the entire package. Does this mean that the switcher will not be able to output a voltage less than 30 volts?

Any thoughts?

---------- Post added at 16:36 ---------- Previous post was at 16:34 ----------

Also, I have no idea what the Vc pin is. There was no mention of it in the data sheet that I could find.

If you find a current-mode controller and can tolerate some
slop, you may be able to get a fixed current by simply rigging
the error amp as a follower and driving it with a trim-pot. Some
use external sense resistors and some use "lossless" internal
current sensing.

For very high currents and low loop bandwidths there are also
Hall sensors which can give you isolated feedback and no
ohmic losses.

If 317 got hot,you can use a transistor(2SB817) as a pass transistor,Regarding constant current switching supply,that can be made with Tl494 or Ka7500 ic.but this curcuit need high frequency transformer(ferrite core chopper transformer,almost use in cpu smps).