Step down converter using LT3840

Thread Starter

vicdlr

Joined Jan 23, 2012
2
[Moderator's note: You were posting in an old thread and your conversation quickly turned into a diversion of the regular thread, so this thread has been created especially for you.]

Hi Scott,

I'm thinking of using the LT3840 to make a current limited regulator with input at 48V from a battery and output of 10-25A @ 36-40V, and plan to use a Hall Effect current sensor instead of a shunt resistor. Do you think it's workable?

TIA,
Vic
 
Last edited by a moderator:
Hi Vic,

I am really not an expert to this (the mentioned prototype was for my bachelor thesis). I also used 48V input (from a switched mode power supply) and achieved output of 0.05A-5A @ 2V-45V by adding some external circuitry.
For your approach you need to make sure that your MOSFETs are able to switch such high currents while staying in specified temperature range (maybe additional cooling could be necessary).
Also take into account that a shunt resistor has an (approximately) constant system response over all frequencies while a Hall effect current sensor has limited bandwidth. Thus, its own system response could have an unwanted impact on the whole system's behaviour. So I am not sure if the resulting system would be stable.

I think that Linear Technology's support can give you more useful answers than me ;-)
 

Thread Starter

vicdlr

Joined Jan 23, 2012
2
Ok, yeah, Linear support from Singapore said it can't be done without saying why. <BG> I guess will have to find out for myself. Appreciate your comment on current sensing.

Thanks!
Vic
Hi Vic,

I am really not an expert to this (the mentioned prototype was for my bachelor thesis). I also used 48V input (from a switched mode power supply) and achieved output of 0.05A-5A @ 2V-45V by adding some external circuitry.
For your approach you need to make sure that your MOSFETs are able to switch such high currents while staying in specified temperature range (maybe additional cooling could be necessary).
Also take into account that a shunt resistor has an (approximately) constant system response over all frequencies while a Hall effect current sensor has limited bandwidth. Thus, its own system response could have an unwanted impact on the whole system's behaviour. So I am not sure if the resulting system would be stable.

I think that Linear Technology's support can give you more useful answers than me ;-)[/QU
 
Top