# How would you regulate/limit multiple outputs from a power supply at 10A maximum?

Discussion in 'The Projects Forum' started by yoyomama, Mar 7, 2016.

1. ### yoyomama Thread Starter New Member

Mar 7, 2016
3
0
I have a single 28V power supply feeding 10 separate loads.
However, each of these outputs must be limited at 10 A maximum. That is: they cannot draw any currents higher than 10A.
When driving a single load from a single power supply, one can simply limit the power supply. However, for 10 loads, limiting the power supply at 100A won't guarantee a maximum of 10A per load.
Current limiting diodes are targeted at much smaller currents. What's out there for higher currents?

2. ### yoyomama Thread Starter New Member

Mar 7, 2016
3
0
I thought about using a LDO that can be current limited: with Vin at 29/30 V and regulating Vout to 28V, inefficiencies are "minimized". Current limit it at 10A and should get the job done.

3. ### Sensacell Well-Known Member

Jun 19, 2012
1,098
255
If you can tolerate about 100 milliohms in series, you could build 10 individual current limiters with sensing shunt resistors.

... or use a Hall current sensor.

How accurate do they need to be?
Can they just "trip out" like a circuit breaker, or do they have to limit at 10 A?
How fast do they need to respond?

4. ### TheButtonThief Active Member

Feb 26, 2011
218
38
I'm thinking a C10 MCB on the supply of each load. simple.

5. ### ErnieM AAC Fanatic!

Apr 24, 2011
7,316
1,571
Each individual output needs some sort of limiter. What kind depends on how you need the limit set, how accurate, how fast, does it need to simply limit current or to trip out and may it recover once the overload goes away.

6. ### AnalogKid Distinguished Member

Aug 1, 2013
4,284
1,172
What kind of response time do you need?

What kind of response time do you need?
Linear current limiting (out voltage sag)?
foldback current limiting (reduce to a low level and hold)?
physical circuit breaker (galvanic isolation)?
thermal or magnetic?
electronic circuit breaker (latching or auto restart)?
etc?

What are your time and money budgets? Also, what is your skill set? Are you looking for a component to be wired in, or plans for a pc board circuit?

Side comment, industrial mechanical breakers rated for DC are relatively rare. They're big in the aircraft industry, but normal industrial magnetic breakers can have surprisingly low DC ratings.

ak

7. ### TheButtonThief Active Member

Feb 26, 2011
218
38
Not true. Breakers aren't bias to either AC or DC, they work for both. They'll just be tolerant to less voltage at DC than they would at AC, for instance a C10 breaker that's good for 415VAC can cope for voltages up to 48VDC

8. ### tcmtech Well-Known Member

Nov 4, 2013
1,891
1,445
More importantly what are the loads and why do you need to limit them to 10 amps and can they handle having their voltage reduced once they reach the 10 amp current limit??

9. ### AnalogKid Distinguished Member

Aug 1, 2013
4,284
1,172
Actually, I was referring to the DC current rating of a breaker with AC-rated contacts. But to your example, I think an 88% decrease in a specification value qualifies as surprisingly low to someone who is not looking out for it.

ak

Evanguy likes this.
10. ### jlnance New Member

Dec 31, 2015
6
1
Would fuses work for what you're trying to do? They are certainly the simplest option, with the downside that they have to be physically changed after they blow.

11. ### yoyomama Thread Starter New Member

Mar 7, 2016
3
0
Thank you for all the answers.
I need to limit the charging current of some battery packs (at 10A). These are connected to a single DC power supply or to a DC grid. Currents over 10A must be limited and regulated at 10A; currents under 10A should pass without any reduction.

Fuses or circuit breakers aren't really the option, but I'll probably include a 12A or 15A after the current limiter/regulator.

Shunt current regulators (using a pass BJT or MOSFET transistor) are not a bad option. However, the fact that I have no voltage feedback to any previous regulator means that my output voltage will sag as current demand from the batteries change.

I was thinking of using a linear voltage regulator, but limit it in current. That way I'll have a feedback path to the regulator, thus regulating current and keeping voltage more or less steady. Honestly, if I could easily program the output current, the better, but I can't see that happening unless I use a more complicated design.

I would prefer a commercial/off-the-shelf option, but don't have a probably making my own pcb design.

12. ### crutschow Expert

Mar 14, 2008
12,531
3,069
What is the voltage of the batteries you are charging?

13. ### #12 Expert

Nov 30, 2010
16,026
6,540
The unlabeled transistor is any cheap, small, to-92 like a 2N3904
If the flat Darlington can't stand the heat, there is a TO-3 package.

File size:
5.3 KB
Views:
8
File size:
128.6 KB
Views:
3
14. ### Sensacell Well-Known Member

Jun 19, 2012
1,098
255
Oh yeah, it could get really hot, going to need a rather MASSIVE heat-sink.

15. ### #12 Expert

Nov 30, 2010
16,026
6,540
The next most likely option is a low drop out, 10 amp, current switcher. Got one in your inventory?

16. ### Sensacell Well-Known Member

Jun 19, 2012
1,098
255
We all do the best we can within the laws of physics.
This is a tough one.

17. ### #12 Expert

Nov 30, 2010
16,026
6,540
I guess I'm so accustomed to heat sinks that I don't feel this circuit is unusual or odd. When you think about 10 current limiters connected to a single DC supply, (I think) your mind simply has to go toward heat control. (Maybe because that's my day job.) Really. With (10) 24 volt batteries in one room, charging with 10 amps each, we're into kilowatts just on the batteries. Even if you get great efficiency with a switching design, the whole room has to be ventilated or cooled with about 10,000 BTU's of air conditioner. (Give or take, depending on whether the time is, "winter".)

18. ### AnalogKid Distinguished Member

Aug 1, 2013
4,284
1,172
2400 W is 8192 BTU/hr. That is surprisingly binary.

ak

19. ### #12 Expert

Nov 30, 2010
16,026
6,540
Kind of expected. We're talking about 1010 amps times 1010 batteries.

20. ### GopherT AAC Fanatic!

Nov 23, 2012
5,823
3,423
Interesting conversation between #12 (aka: "I'd rather build it with vacuum tubes than a Microcontroller") and a guy named "AnalogKid"