Overcurrent sensing using a comparitor

[fr0st]

I had an idea for shutting down a SMPS if there was too much current draw.
The idea is to put a comparitor measuring the voltage drop of the fets so that when a certain amount of Voltage drop is obtained from overcurrent the comparitor would read high and a microcontroller or similar would shut down the amp.
I figured the connections would be have the + on the non switched side of the fets and a multiturn trimpot setting the voltage and the - on the switched side of the fets.
Would this work?
(I hope I explained clearly )

 

[JOE DIRT®]

frost...good concept and I would do it unabtrusive with audio and keep anything out of the "chain"....a simple signal diode or transisitor can sense temp....also a temperature ic (dallas)...can sense an overcurrent...I have more data if you need it. I just try and keep it simple for audio

DIRT®

 

[jackinnj]

isn't this already available

the current mode control on the primary side should do this, shouldn't it? i mean, it was originally set up with current control to prevent BJT devices from getting fried if the transformer went into saturation -- you just have to set the "sense" resistor for the error amp to the correct value and it will reduce the PWM.

are you trying to do current mode control with an LM3524 or SG3524? you can buy a unitrode (TI) controller for about the same amount of money and implement a design from TI's application notes.

 

[fr0st]

The main idea was to alert a microcontroller that it reached a certain current point so it can shut down and do a self check or something (havn't figured that far ahead).

 

[JOE DIRT®]

I use these temp controllers all the time you can find the application notes on maxims site...they are the Dallas semi ds1821

 

[peranders]

The main idea was to alert a microcontroller that it reached a certain current point so it can shut down and do a self check or something (havn't figured that far ahead).

A switched PS have allways a max current limit and also limiting. A PS without this is dangerous to use. Fuses can be too slow for this.

It's nothing wrong to have an extra current limitation but the PS itself must have one.

 

[jackinnj]

The main idea was to alert a microcontroller that it reached a certain current point so it can shut down and do a self check or something (havn't figured that far ahead).

A microcontroller is too slow -- and involves too many parts -- the current mode controllers I refer to measure on each pulse of the pulse width modulation and adjust the pulse width -- just go to On-Semi or Texas Instruments websites for several good tutorials -- current sense PWM chips aren't expensive.

 

[peranders]

Quite right, speed is important. We talk micro seconds, one or few!

Check also Linear Tech for current sensor chips.

 

[JOE DIRT®]

ummm Jack the controllers I use have response time of logic states in nanoseconds not hours

 

[jackinnj]

I haven't really gone over 4MHz with a microchip controller, although I have some 20MHz chips to fool around with at some point.

I wouldn't use a PIC for a battery charger or a PWM supply controller since I would have to write the code, and TI or Maxim has already done it for me. Sloth and laziness are the mother and father of all invention.

Then again, people on the PICList seem to do it all the time. Horses for courses.