Pretected Power Supply

Using Arduino General Electronics
#1

I am making an external power supply for a Arduino and xbee. Attached is the schematic. Can someone input on the goods, bads, or changes that need to be made. The plan is to minimize the Arduino by using just the Atmel chip and crystal after prototyping. For prototyping it will have 2 opto-isolators PS2501-4A-ND, Arduino UNO, Level Logic Converter BOB-08745, Xbee series 2. Among other resistors, capacitors, and diodes. The optos are probably not necessary but are cheaper/easier/neater than using transistors for 8 inputs from +12. The power (+12) is from 12v automotive so can swing 11v-16v maybe more. I am looking for a clean protected +5v and +3.3v with minimal heat sink. I’m hoping the 12v regulator will share some “heat” that the 5v and 3.3v would have had to do with out it. Overkill is not a bad thing in my mind, considering the input voltage.

Thanks

powersupply.jpg

#2

The 7812 is rather pointless and may actually prevent any power from reaching anything. The minimum Vin for the 7812 is 14.9V. So if the car is putting out less then that, then the rest of the power supply will not receive any voltage. Now the 7805 is rated for as much a 35VDC so you don't need to worry about keeping the voltage input steady. Thats what a voltage regulator is for. To take the higher and unstable voltage and output a fairly stable lower voltage. The 7805 will get pretty hot even at 12 V so you will want a heat spreader on it.

Depending on the VR you use for the 3.3V side you might be able to use the output of the 7805 for it. If you use an LM2937-3.3 then it can handle up to 26V so going straight from the car can be done. But you will want a heat spreader for both VRs. Its not like you will need a full on heat-sink with active cooling or anything.

Just look at most car chargers for cell phones. They take the full 12V in and output 5V without heating up too much. Open 1 up and see what they have. Could just use a hacked charger for the 5V side and a simple voltage divider to give you the 3.3 V needed.

#3

Well U1, the 12vdc regulator, is not going to buy you much of anything, it's voltage drop will be too minimum so it won't share any of the heat dissipation that the 3.3 and 5.0 regulators will. Without knowing worst case current demands for the 3.3 and 5 volt regulators I can't say how hot they are going to run, but I suspect very hot. Trying to get around using adequate sized heat sinks with linear regulators is just not really possible. Look into more modern switching regulators avalible today, they have come down a lot in price from the Asian E-bay sellers and run very efficient just using their small PCB as heatsinks up to around 2-3 amps.

Lefty

#4

Thanks for the replies. The LM7812 maybe pointless, now that I think about it. Do you have any switching regulators to recommend? I have made a boost converter circuit before, to boost a piezo buzzer from 12 to 36 volts. But probably wouldn't feel comfortable making a switching circuit for this project.

#5

tnovak: Thanks for the replies. The LM7812 maybe pointless, now that I think about it. Do you have any switching regulators to recommend? I have made a boost converter circuit before, to boost a piezo buzzer from 12 to 36 volts. But probably wouldn't feel comfortable making a switching circuit for this project.

This is a good example of a typical adjustable switching voltage regulator good for 2-3 amps, buy one for the 5 and one for the 3 volt needs. There are many many other sellers of simular switching regulators. They are smaller then they look in that picture and easy to work with.

http://www.ebay.com/itm/LM2596-DC-Converter-Power-Supply-Buck-Step-Down-Regulator-In-4-40V-Out-1-5-35V-/180807722731?pt=LH_DefaultDomain_0&hash=item2a18fae6eb

Lefty

#6

Thanks that looks great! I'm going to try and find some in the USA. Should I still include the diode or will it be protected on-board? I'm sure the on-board caps smooth it out. Also what do you think about this one digi-key has on hand? PN 811-2196-5-ND

http://search.digikey.com/scripts/DkSearch/dksus.dll?WT.z_header=search_go&lang=en&site=us&keywords=OKI-78SR-5&x=0&y=0

#7

tnovak: Thanks that looks great! I'm going to try and find some in the USA. Should I still include the diode or will it be protected on-board? I'm sure the on-board caps smooth it out. Also what do you think about this one digi-key has on hand? PN 811-2196-5-ND

http://search.digikey.com/scripts/DkSearch/dksus.dll?WT.z_header=search_go&lang=en&site=us&keywords=OKI-78SR-5&x=0&y=0

That digikey regulator looks fine. Lower max current then the one I showed you and not adjustable so you still need a 3.3v version or run a 3.3 linear regulator from the output of this one, but again keep track of total current needs for your loads. An external diode is still a good idea as it protects from reverse polarity mistakes which will damage most any regulator chip or module.

Good luck

#8

Decided to go with the digi-key ones as i can get them over night OKI-78SR-5 and OKI-78SR-3.3. Also ordered the ones you suggested because they look useful just take along time to receive. I have a question about the new circuit. I think that D1 and D2 are ok, should i use the placement of D3 or D4 or not use a “clamp” at all?? Any other changes needed? Thanks again.

power.jpg

#9

Hey tnovak

Saw this post of yours and your 5V from 12V is dangerously similar to ours.

I have posted a question on my design, link below. Maybe that will help, check it out.

http://arduino.cc/forum/index.php/topic,91136.0.html

Later, BTFdev

#10

BTFdev,

Thanks for the heads up, I have already been following your post :-). Your circuit looks like what might be in the OKI-78SR-5 and OKI-78SR-3.3 that I decided to use. From the data sheet they need external protection and maybe even external filtering though...

tnovak

#11

tnovak: Decided to go with the digi-key ones as i can get them over night OKI-78SR-5 and OKI-78SR-3.3. Also ordered the ones you suggested because they look useful just take along time to receive. I have a question about the new circuit. I think that D1 and D2 are ok, should i use the placement of D3 or D4 or not use a "clamp" at all?? Any other changes needed? Thanks again.

Those diodes look good if you want to play it safe. In cars with starter motors, that's often a good idea :-) You might even want to make the bracketing diode be a TVS or Zener with about a 16 V trip point. Also, consider making the fuse a PTC resettable fuse. Those are basically little resistors that suddenly go from "very little resistance" to "very much resistance" if they overheat. Let them cool down, and they go back to normal. Makes gear much more robust and less replacement parts is a good thing :-)

#12

jwatte,

Thanks for the info, Which do you think is the better placement D3 or D4 (before or after the fuse). I looked into the PTC's as I've never used them, I'm not sure what one to use, would I go with a 2 amp trip current rating? What does the "hold" current mean?

Thanks

#13

tnovak: jwatte,

Thanks for the info, Which do you think is the better placement D3 or D4 (before or after the fuse). I looked into the PTC's as I've never used them, I'm not sure what one to use, would I go with a 2 amp trip current rating? What does the "hold" current mean?

Thanks

If used the 'safer' place for the negative clamping diodes is after the fuse, might blow the fuse on real bad negative spikes or accidental applying reverse voltage, but that's better then blowing the diode and possibly the board if added before the fuse.

Lefty

#14

tnovak: jwatte,

Thanks for the info, Which do you think is the better placement D3 or D4 (before or after the fuse). I looked into the PTC's as I've never used them, I'm not sure what one to use, would I go with a 2 amp trip current rating? What does the "hold" current mean?

Thanks

You want the clamp Zener after the fuse, because clamping will dissipate a lot of power, and you want the fuse to trip! If it's a PTC, just fixing the problem will make it recover after a few minutes.

PTCs have a "must trip" current, which is the 2A you see, and a "must not trip" current, which is the "hold current." The spec guarantees that, at the hold current, and the specified temperature, the fuse will not trip. It also guarantees that, at the specified temperature, at the specified trip current, it will trip within the specified trip time. There is some current above the "hold" current that still won't trip under most operating conditions; there is also some current below the "must trip" current where the fuse will still trip, although it may take longer than the rated time. Exactly where these limits go is variable based on variations in manufacturing. Designing a narrow gap between "hold" and "trip" and a short "time to trip" will make for a much more expensive fuse, and there are limitations to the technology they use. Also, many of them will actually change shape (expand) when tripping, and then go back when recovering, so consider that when mounting!

#15