Arduino Forum

Using Arduino => General Electronics => Topic started by: eloso on Jan 04, 2013, 05:25 am

Title: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 04, 2013, 05:25 am
I need a 12VDC voltage regulator that can produce at least 3A. My DC Source is 18.5volts 3.5A. I'm powering gsm shield, arduino, solenoids and etc. I only have here two LM317 regulator and their max current is only 1.5a based on the datasheet. The electronic shop near me doesn't have regulator that can produce more than or equal to 3A. My circuit has a master switch and backup battery that will turn on when the main supply is off. This is the schematic if i will use two regulators with master switch and back up, it is very complicated, and i want to reduce it as shown in the left of the picture by using only one regulator.

(http://i539.photobucket.com/albums/ff351/tophi45/testing_zpsaf67dcdb.jpg)
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 04, 2013, 05:44 am
my bad, i forgot the voltage drop of zener diode. so the left schematic in the picture is invalid. -_-
Title: Re: Combining the Ampere of two voltage regulator?
Post by: be80be on Jan 04, 2013, 05:45 am
You don't want to do that there not going to do what you think. One will try to put out more the it should and basically cutoff the second one. See there not going to supply the same voltage.

I would do this use a PNP to get you more current
Title: Re: Combining the Ampere of two voltage regulator?
Post by: retrolefty on Jan 04, 2013, 05:55 am
Well the following are 12vdc 3 amp linear regulators in TO-3 package:
78H12
NTE933

and lambda sometime ago use to have a 12vdc 5 amp regulator in a TO-3 package that you can sometimes find sold as new old stock:

http://www.ebay.com/itm/LAMBDA-LAS1912-12-VOLT-5-AMP-TO-3-REGULATOR-NOS-78H12-NTE933-EQUIVALENT-/380547135339?pt=LH_DefaultDomain_0&hash=item589a60776b

Lefty
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 04, 2013, 06:23 am

You don't want to do that there not going to do what you think. One will try to put out more the it should and basically cutoff the second one. See there not going to supply the same voltage.

I would do this use a PNP to get you more current


can i use npn transistor only? i have a couple  of TIP120 here.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 04, 2013, 06:33 am

Well the following are 12vdc 3 amp linear regulators in TO-3 package:
78H12
NTE933

and lambda sometime ago use to have a 12vdc 5 amp regulator in a TO-3 package that you can sometimes find sold as new old stock:

http://www.ebay.com/itm/LAMBDA-LAS1912-12-VOLT-5-AMP-TO-3-REGULATOR-NOS-78H12-NTE933-EQUIVALENT-/380547135339?pt=LH_DefaultDomain_0&hash=item589a60776b

Lefty


do you have alternative solutions than buying a new regulator? :D
Title: Re: Combining the Ampere of two voltage regulator?
Post by: be80be on Jan 04, 2013, 06:41 am
Get 2 tip42
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 04, 2013, 08:39 am

Get 2 tip42


so i cannot use npn transistor?
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 04, 2013, 11:10 am

You don't want to do that there not going to do what you think. One will try to put out more the it should and basically cutoff the second one. See there not going to supply the same voltage.

I would do this use a PNP to get you more current


i simulate that circuit using circuit wizard and it is not working.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: be80be on Jan 04, 2013, 12:42 pm
I guess they put that in there data sheet for looks It works.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: jackrae on Jan 04, 2013, 12:44 pm
Let's go back a few steps

Question - Why do you think you need a 3A regulated supply ?  What is your actual load ?  Just because the power supply can provide 3A+ doesn't mean your regulator has to be rated at that figure.  If 6 volts meets you back-up needs than "losing" 12.5 across a linear regulator is a shear waste of power  (12.5 x 3 = 37.5 watts).  If you really need 6 or 12 volts  (you seem to be a bit mixed on voltages) then I'd suggest a switched mode regulator that offers over 90% efficiency with very little heat.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: dhenry on Jan 04, 2013, 01:00 pm
Quote
my bad, i forgot the voltage drop of zener diode.


Not to mention that finding zeners that can flow 1.5a/each is quite hard.

If you really need 3amp, you can look into the datasheet of your voltage regulator to see  how to expand their current capabilities.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: be80be on Jan 04, 2013, 01:15 pm
Quote
If you really need 3amp, you can look into the datasheet of your voltage regulator to see  how to expand their current capabilities.


I posted that for the OP he said in some kind of simulator it didn't work
Title: Re: Combining the Ampere of two voltage regulator?
Post by: dhenry on Jan 04, 2013, 01:29 pm
Quote
it didn't work


thank god we don't live in a simulator.

Quote
so i cannot use npn transistor?


Yes, you can.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 04, 2013, 02:18 pm

Let's go back a few steps

Question - Why do you think you need a 3A regulated supply ?  What is your actual load ?  Just because the power supply can provide 3A+ doesn't mean your regulator has to be rated at that figure.  If 6 volts meets you back-up needs than "losing" 12.5 across a linear regulator is a shear waste of power  (12.5 x 3 = 37.5 watts).  If you really need 6 or 12 volts  (you seem to be a bit mixed on voltages) then I'd suggest a switched mode regulator that offers over 90% efficiency with very little heat.


I have a solenoid lock,alarm,relay rated at 12v. A couple of 5 inputs to arduino. The gsm shield required is 1.5A. Assuming the arduino is 500mA, gsm is 1.5A, the solenoid lock is 200mA, alarm is 300mA, relay is 50mA then the total is 2550mA or 2.5A. The remaining 500ma is just for extra. The purpose of the back up battery is not to power them all but to only power the arduino and gsm, so that the gsm can still send message and call.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: dhenry on Jan 04, 2013, 02:31 pm
Quote
The gsm shield required is 1.5A.


A nice heater element.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: be80be on Jan 04, 2013, 02:43 pm
Here a NPN setup (http://www.reuk.co.uk/OtherImages/78xx-Regulator-high-current.gif)
Title: Re: Combining the Ampere of two voltage regulator?
Post by: dhenry on Jan 04, 2013, 03:15 pm
Quote
Here a NPN setup


The output would fluctuate with the load.

You need 1 small signal pnp, + 2 resistors.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: SirNickity on Jan 04, 2013, 09:31 pm
The GSM shield requires 5v @ 1.5A?  Really?

I would use an unregulated PSU for the solenoids, because there's not likely to be any good reason to worry about the exact voltage going through a coil.  Same for the relays.  As long as it's rectified and filtered to remove any serious ripple, they should be perfectly happy.  A simple transistor (and protection diodes) to fire them would suffice.

That leaves the control circuitry and the GSM module.  At a load of 1.5A, you really need to move to switching regulators.  That's way, way too much current to stick with linear.  You'll be generating a TON of heat.  That means, you're wasting quite a bit of power, and you need sufficient means to remove that heat.  Those LM78xx chips say they'll drive 1.5A, but don't even think about asking them to do it IRL.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 05, 2013, 12:47 am
Yes, because i will use calls too.
http://tronixstuff.wordpress.com/2011/01/19/tutorial-arduino-and-gsm-cellular-part-one/ (http://tronixstuff.wordpress.com/2011/01/19/tutorial-arduino-and-gsm-cellular-part-one/)

Quote
The GSM shield can often require up to 2A of current in short bursts - especially when turned on, reset, or initiating a call. However your Arduino board can only supply up to just under 1A. It is highly recommended that you use an external 5V power supply capable of delivering 2A of current - from an AC adaptor, large battery with power regulator, etc. Otherwise there is a very strong probability of damaging your shield and Arduino. Ignore this at your own risk. When connecting this supply DO NOT use the DC socket on the Arduino. Instead, connect the 5V (positive) from the supply to the 5V pin on the GSM shield, and the negative to the GND pin.


I'm using linear regulator because it is easy to make and it is only a school project. So i will only use it once. I already have a big heatsink and thermal paste so generating heat is not a problem.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: Docedison on Jan 05, 2013, 03:05 am
Follow the circuit posted that uses a PNP 'wrapped' around the regulator. It has short circuit protection, is a reprint from the National Semi Linear app notes book (1 of 3) and I would LOVE to get copies of those in digital format $$$... It works very well as I've built several dozen copies of it in the past 20 or so years... and I can tell you from personal experience that although later an that same app note AN??? it (the app note)
shows a version that used an NPN transistor as the pass transistor. The NPN requires either a variable regulator or a diode in the ground lead of the regulator to correct for the .7 V loss in the resultant emitter follower that results from adding the NPN transistor. The NPN version doesn't work as well and requires more parts. The single one I made was real problematical.. the current limiter was an issue... Use the PNP.

Bob
Title: Re: Combining the Ampere of two voltage regulator?
Post by: SirNickity on Jan 05, 2013, 04:59 am
I'm using linear regulator because it is easy to make and it is only a school project. So i will only use it once. I already have a big heatsink and thermal paste so generating heat is not a problem.


I think you under-estimate how hot a regulator will get passing 1.5A of current.  I'm too lazy to do the math, but if it's dropping any voltage whatsoever, it may be impractical to keep the die temperature cool enough to survive.  I built a 28v-to-15v regulator to supply 100mA or so of current, and the little 1"x2"x0.5" heatsinks got hot enough that I couldn't touch them comfortably.  You're asking for a lot more current than that.  Trust me -- that "1.5A" spec on the regulator is nothing short of wishful thinking.  There's no way you'll keep the case cool enough that the die won't damage itself at that kind of load.

That is unless the 1.5A spec is for a very brief, worst-case, transient load... and average current is really *significantly* less.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: be80be on Jan 05, 2013, 05:37 am
Do you really think you'll use that much power. The link the OP posted said bust up to 2 amps The Lm78 can handle 2.2 Amps in short burst.

These burst of the  SM5100B cellular module only last 4.6 mS
Title: Re: Combining the Ampere of two voltage regulator?
Post by: krupski on Jan 05, 2013, 08:23 pm

You don't want to do that there not going to do what you think. One will try to put out more the it should and basically cutoff the second one. See there not going to supply the same voltage.

I would do this use a PNP to get you more current


Here's a circuit that I use for more regulated power... the small resistor values cause the 78XX part to provide current limit (around 10 amps with these values) and since the regulator and pass transistor are on the same heatsink, the thermal protection of the 78XX is also provided.

Note the way the output is connected... this provides remote current sensing which compensates for drop along long wires if they are used.

Only disadvantage to this circuit is that a minimum load of a few milliamps is required because of the design of the 78XX regulator. If you use an LM-317 instead, then this is not a problem.

(edit to add): If you use a 79XX regulator and an NPN 2N3055 and reverse the plus and minus terminals, you get the same circuit.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: pekkaa on Jan 05, 2013, 09:07 pm

thank god we don't live in a simulator.


You don't really know that. (Sorry, off topic, but couldn't resist.  ;))
Title: Re: Combining the Ampere of two voltage regulator?
Post by: dhenry on Jan 05, 2013, 10:08 pm
Quote
Here's a circuit


The 0.15ohm ressitor needs to be 100x of its current value.

You can buffer that pnp with a power npn / n-ch mosfet so you can use a smaller pnp device instead.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: dhenry on Jan 05, 2013, 10:09 pm


thank god we don't live in a simulator.


You don't really know that. (Sorry, off topic, but couldn't resist.  ;))


The latest on that is that they have proposed tests to see if we are real or not.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: larryd on Jan 05, 2013, 11:50 pm
As long as there is a ground connection between the power supplies you don't have to connect the Positive outputs together. Therefore you can have 1,2,3,4,5... regulators connected together (as long as there is a common ground)!
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 06, 2013, 02:35 pm

I'm using linear regulator because it is easy to make and it is only a school project. So i will only use it once. I already have a big heatsink and thermal paste so generating heat is not a problem.


I think you under-estimate how hot a regulator will get passing 1.5A of current.  I'm too lazy to do the math, but if it's dropping any voltage whatsoever, it may be impractical to keep the die temperature cool enough to survive.  I built a 28v-to-15v regulator to supply 100mA or so of current, and the little 1"x2"x0.5" heatsinks got hot enough that I couldn't touch them comfortably.  You're asking for a lot more current than that.  Trust me -- that "1.5A" spec on the regulator is nothing short of wishful thinking.  There's no way you'll keep the case cool enough that the die won't damage itself at that kind of load.

That is unless the 1.5A spec is for a very brief, worst-case, transient load... and average current is really *significantly* less.


I see, but based on the replies in my topic i can draw a lot of current from pnp transistor not the regulator. My heatsink is 80mmx41mmx15mm. Thank you for the warning. Now i will focus on adding a pnp transistor.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 06, 2013, 02:39 pm
I tried this circuit a while ago in a breadboard and it is not a working. I tried to draw 1 ampere by using a load of 10 ohm 5 watt resistor and the regulator gets very hot the pnp transistor is very cold and it seems it is not the one supplying it. I'm sure my connections are correct.

http://www.zen22142.zen.co.uk/Circuits/Power/boosti.htm (http://www.zen22142.zen.co.uk/Circuits/Power/boosti.htm)
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 06, 2013, 02:42 pm

You don't want to do that there not going to do what you think. One will try to put out more the it should and basically cutoff the second one. See there not going to supply the same voltage.

I would do this use a PNP to get you more current


what is the watts of the resistor?
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 06, 2013, 02:50 pm


You don't want to do that there not going to do what you think. One will try to put out more the it should and basically cutoff the second one. See there not going to supply the same voltage.

I would do this use a PNP to get you more current


Here's a circuit that I use for more regulated power... the small resistor values cause the 78XX part to provide current limit (around 10 amps with these values) and since the regulator and pass transistor are on the same heatsink, the thermal protection of the 78XX is also provided.

Note the way the output is connected... this provides remote current sensing which compensates for drop along long wires if they are used.

Only disadvantage to this circuit is that a minimum load of a few milliamps is required because of the design of the 78XX regulator. If you use an LM-317 instead, then this is not a problem.

(edit to add): If you use a 79XX regulator and an NPN 2N3055 and reverse the plus and minus terminals, you get the same circuit.


I have LM317 here.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: dhenry on Jan 06, 2013, 02:52 pm
Quote
I tried this circuit a while ago in a breadboard and it is not a working.


That circuit is designed for the pnp to kick in at load current > 700mv/1ohm = 700ma. Anything below that, the voltage regulator is doing the heavy lifting.

If you want the pnp to kick in at a lower current, increase that 1ohm resistor.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 06, 2013, 03:24 pm

Quote
I tried this circuit a while ago in a breadboard and it is not a working.


That circuit is designed for the pnp to kick in at load current > 700mv/1ohm = 700ma. Anything below that, the voltage regulator is doing the heavy lifting.

If you want the pnp to kick in at a lower current, increase that 1ohm resistor.


I see. What do you suggest?
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 06, 2013, 03:31 pm
so the regulator can be very hot at 700mA? i almost got skin burned when i touched it. Im using tip42c pnp.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: be80be on Jan 06, 2013, 03:44 pm
(http://pad1.whstatic.com/images/7/73/ATX-to-Lab-Supply-1.JPG)
http://www.wikihow.com/Convert-a-Computer-ATX-Power-Supply-to-a-Lab-Power-Supply
As much problems your having find you a old PSU from a computer kind with a switch to turn on and use that
most have up to 30 amp output on 5 volt and about half of that on the 12 volt.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: dhenry on Jan 06, 2013, 05:48 pm
Quote
What do you suggest?


You want to the regulator to run 50 / 100ma at least. So I would use a resistor of 8 - 10ohm, give or take. Higher where the voltage dropout is big.

Quote
so the regulator can be very hot at 700mA?


Yes, particuarly if the Vin-Vout is large. Your heatsink is puny for this type of applications.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 06, 2013, 06:27 pm


You don't want to do that there not going to do what you think. One will try to put out more the it should and basically cutoff the second one. See there not going to supply the same voltage.

I would do this use a PNP to get you more current


Here's a circuit that I use for more regulated power... the small resistor values cause the 78XX part to provide current limit (around 10 amps with these values) and since the regulator and pass transistor are on the same heatsink, the thermal protection of the 78XX is also provided.

Note the way the output is connected... this provides remote current sensing which compensates for drop along long wires if they are used.

Only disadvantage to this circuit is that a minimum load of a few milliamps is required because of the design of the 78XX regulator. If you use an LM-317 instead, then this is not a problem.

(edit to add): If you use a 79XX regulator and an NPN 2N3055 and reverse the plus and minus terminals, you get the same circuit.


what is the use of 0.1 ohms resistor?
Title: Re: Combining the Ampere of two voltage regulator?
Post by: Docedison on Jan 06, 2013, 10:03 pm
It is used for stability and t provides some short circuit protection for the pass transistor with a little negative feedback.

Bob
Title: Re: Combining the Ampere of two voltage regulator?
Post by: krupski on Jan 07, 2013, 01:22 am

Quote
Here's a circuit


The 0.15ohm ressitor needs to be 100x of its current value.

You can buffer that pnp with a power npn / n-ch mosfet so you can use a smaller pnp device instead.



No it doesn't. It's chosen so that when the load is around 10 amps, the load that the 78xx sees is 1 amp. The idea is to carry over the current limit and thermal protection built into the 78xx without needing extra components.

I've been using this circuit for literally decades... believe me, it works.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: krupski on Jan 07, 2013, 01:25 am

so the regulator can be very hot at 700mA? i almost got skin burned when i touched it. Im using tip42c pnp.


Note that my drawing shows both the regulator and pass transistor on the same heatsink. The idea is for both devices to share the same temperature... that way if the pass transistor gets too hot, it heats the 78xx regulator and trips the thermal protection. The circuit is designed to pass along the current limiting and thermal protection of the 78xx regulator to the pass transistor.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: dhenry on Jan 07, 2013, 01:34 am
Quote
It's chosen so that when the load is around 10 amps, the load that the 78xx sees is 1 amp.


With the 78xx flowing just 1amp, the 0.15ohm resistor is providing a voltage of 150mv: how is it going to open up the pnp?

Quote
I've been using this circuit for literally decades... believe me, it works.


I am sure it works. You may want to write the manufactures a note so they can rewrite the datasheet.
Title: Re: Combining the Ampere of two voltage regulator?
Post by: be80be on Jan 07, 2013, 01:57 am
Here one for the OP LM317
Title: Re: Combining the Ampere of two voltage regulator?
Post by: eloso on Jan 07, 2013, 02:52 am
any idea the required watts of the resistor?
Title: Re: Combining the Ampere of two voltage regulator?
Post by: dhenry on Jan 07, 2013, 02:56 am
Tiny: the voltage across it is 700mv at most.


Quote
Here one for the OP LM317


Sounds like high time for them to update their datasheet.

BTW, you can use a n-ch mosfet in place of the npn.