Power Select Circuit - PChannel FET

Status
Not open for further replies.

wuchy143

Member
Hi all,

I'm trying to find a good p-channel mosfet candidate for a +5v power select circuit. Essentially the design will have two independent +5V power inputs. The circuit I attached will select the power source which has power on it and will draw from that source only.

The problem I'm facing is having very little body diode voltage drop. I want my circuit to see as close to +5v as possible. You will see in my schematic that due to the normal .6v drop across the body diode I get ~4.4v out for power to my design.

Are there any p-channels out there that can handle at least 2 Amps but rated for 5 for a decent de-rating?

Thanks.
 

Attachments

  • Untitled.jpg
    82.7 KB · Views: 147
V2=5V, C1 is high because M1 is on, current flows backwards through M2 to V1. Most MOSFETs have a diode inside that allow current to flow backwards.

I think your comparators should run from 15 volts supply. Use N-MOSFETS and reverse the inputs of U1, U2 to make the N-fets to work.
 
The body diode isn't the culprit.
Two problems with your sim; 1) the drain-source connections are reversed from what they should be, 2) if you don't specify the PMOS type you get weird results.
 
Last edited:
Two alternative approaches might to have the two supplies isolated from each other via schottky diodes so you only lose .1 to .3 volts and if 5.0 volts is critical, add a simple switcher boost circuit after the diodes that can be set to 5.00 volts.
 

Attachments

  • Draft4.asc
    3.1 KB · Views: 118
Last edited:
alec_t said:
The body diode isn't the culprit.
Two problems with your sim; 1) the drain-source connections are reversed, 2) if you don't specify the PMOS type you get weird results.
Click to expand...

Oops. I picked a "real" pchannel fet and now my output is exactly +5v and not the 4.4v I was seeing earlier. The circuit is working how I intended.

The only issue I can see with my circuit is when power is applied to both inputs M2's body diode will potentially conduct if for whatever reason the drain is less in voltage than the source(even by a little). Any suggestions?

Any other weaknesses in the design?
 
When I first looked at your sim, I didn't pay any attention to you voltage dividers and thought that you might get your sensor switching from one supply to the other due to loading changing the voltage, but I couldn't make it happen, instead,(with the FETs in correctly) what I see is it locks onto V1, even when V1 voltage drops a full volt below V2.
 

Attachments

  • Power_Swithc_Sense j.asc
    3.4 KB · Views: 103
This might be what you want to do in a IC.
 

Attachments

  • IR5001s.pdf
    145.9 KB · Views: 160
ronsimpson said:
This might be what you want to do in a IC.
Click to expand...
I agree.
I have simulated a simpler circuit which I think tries to do what you want.
Trying to use two p-channels for active ORing, as in the attachment below, doesn't work. When a supply is off, the body diode for that supply's MOSFET will conduct, shorting out the other supply, assuming its MOSFET is on. Note that, in the attached sim, v(out) goes to approximately half the supply voltage when the other one is off. This is due to almost 100A flowing through one or the other of the body diodes. In reality, of course, either the MOSFET would die, or the power supply(s) would die or go into current limiting.
 

Attachments

  • power switch.asc
    2 KB · Views: 184
  • power switch.png
    19.2 KB · Views: 120
Last edited:
I use sync. rectification and NMOSFETs to get low voltage drop diodes. There are a number of this type of ICs. Here is the data sheet on one. I usually am switching at 200khz or faster and get very low voltage loss at high current.
 

Attachments

  • AN-1087.pdf
    373.6 KB · Views: 258
Status
Not open for further replies.

Similar threads

M
  • Question
Power factor
Replies
2
Views
832
rjenkinsgb
  • Question
IR circuit issues
Replies
16
Views
1K
eTech
E
C
  • Locked
  • Question
Battery Charger Circuit
Replies
10
Views
2K
rjenkinsgb
J
  • Question
Variable frequency 555 circuit
2
Replies
29
Views
2K
Papabravo
C
  • Question
One thing I don't understand about jfets
Replies
7
Views
1K
Papabravo
Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…