Voltage regulator

Hello

I'm currently building a project for wireless controlled leds, for fun and as a learning experience. I'm trying to build my own pcb later on so i'm not using an arduino but just the atmega 328p and needed components. The leds are at 12v and i'using the atmega and an xbee module at 3.3 volts.

So far this is working. But the voltage regulator gets very very hot, 85C last time i checked. I'm currently using an ld1117. The load would be 200mA max. I'm guessing it gets hot because of the voltage difference (12v --> 3.3v).

I want to build my own pcb later on based on smd components, so i'm guessing that an ld1117 in smd (dpak package) would get even hotter ? is this correct ?

My problems are now: a) is this normal ? (i'm guessing yes) b) is this safe ? c) how could i prevent this ? or what are my alternatives ? on this point i'm a little lost currently.

Thanks in advance.

tulpix: Hello

I'm currently building a project for wireless controlled leds, for fun and as a learning experience. I'm trying to build my own pcb later on so i'm not using an arduino but just the atmega 328p and needed components. The leds are at 12v and i'using the atmega and an xbee module at 3.3 volts.

So far this is working. But the voltage regulator gets very very hot, 85C last time i checked. I'm currently using an ld1117. The load would be 200mA max. I'm guessing it gets hot because of the voltage difference (12v --> 3.3v).

I want to build my own pcb later on based on smd components, so i'm guessing that an ld1117 in smd (dpak package) would get even hotter ? is this correct ?

My problems are now: a) is this normal ? (i'm guessing yes) b) is this safe ? c) how could i prevent this ? or what are my alternatives ? on this point i'm a little lost currently.

Thanks in advance.

It is perfectly normal, altough at 85C I would defenitivelly not say its safe.

SMD components for such power dissipation are a poor choice. It wont get any hotter, the power dissipation would be the same, but the PCB might be damaged over the time or the regulator might start throttling the output to avoid a further temperature increase. At this point strange things might start happening with your arduino.

You need to make sure the PCB can safelly handle the power requirement without too much temperature increase. if not attach the regulator to an heatsink as per below example http://www.ebay.com/itm/AMS1117-DC-DC-Power-Module-3-3V-with-Heat-sink-/130920716277?pt=LH_DefaultDomain_0&hash=item1e7b7b97f5

If it was my project I would ditch the linear regulator all togueter and would instead use a switch mode equivalent. that would solve your heating issues altoguether.

Here are some examples:

3Pin replacement switch mode regulator http://www.ebay.com/itm/3-Terminal-3-3V-1A-Switching-Voltage-Regulator-Power-Supply-/261213515448?pt=LH_DefaultDomain_0&hash=item3cd189eeb8

Or if you have a PCB you could use one of these: http://www.ebay.com/itm/5pcs-MP2307-3A-DC-DC-step-down-power-module-KIS-3R33S-for-LED-GPS-MP3-/171212078291?pt=LH_DefaultDomain_0&hash=item27dd08fcd3

casemod:
It wont get any hotter, the power dissipation would be the same.

If he’s moving from a TO-220 to an SMD package, it certainly will get hotter! Power dissipation may be the same, but the SMD package likely has worse thermal characteristics than a TO-220.

What are you powering off this regulator that takes 200 mA?

Jiggy-Ninja: If he's moving from a TO-220 to an SMD package, it certainly will get hotter!

We dont know that much do we? 200mA @ (12V-3.3) = 1.75W, well within the thermal ratings ratings of a SMD component, properly soldered to a PCB with a ground plane for heat dissipation. If this canot be achieved OR the load is over 3W a TO-220 with an heatsink would be required (altough at such power levels a SM regulator would likelly be a better option)

Jiggy-Ninja: What are you powering off this regulator that takes 200 mA?

Good Point. Is this regulator, at 85C drawing 200mA or are you designing your project so that it can draw up to 200mA? If so how much is the circuit (on average) drawing now?

You also need to take into account that the Xbee is likelly to use short bursts of power during transmission that your ampmeter might not be able to measure. This may cause either damage to the regulator or undervoltage

12V to 3.3V is a lot. For a board mount, this is good: http://www.digikey.com/product-search/en?pv1525=94&FV=fff40042%2Cfff800df&k=murata+oki&mnonly=0&newproducts=0&ColumnSort=0&page=1&stock=1&quantity=0&ptm=0&fid=0&pageSize=25

www.pololu.com has little regulators too that mount with 3 pins.

Hello

thanks a lot for your quick replies.

to answer a few questions.

Jiggy-Ninja: What are you powering off this regulator that takes 200 mA?

casemod: Good Point. Is this regulator, at 85C drawing 200mA or are you designing your project so that it can draw up to 200mA? If so how much is the circuit (on average) drawing now?

Currently i'm powering an atmega328p and an xbee module (more or less 60 mA) an 6 transistors (15mA base current per transistor) so that gives me 130 mA (more or less). I will hook up some status leds (did not do that yet) so i'm guessing 150 mA tops and to be safe i used 200 mA max from this circuit. Average values are more difficult to give. But i can say it should draw between 60 mA (transistors off) to 130 mA (transistors fully on) in my current configuration.

--

But after looking at your suggestions i think i'm moving to a different regulator to be on the safe side. I searched mouser a little bit and found this: http://www.mouser.lu/ProductDetail/Murata-Power-Solutions/MEF1S1203SPC/?qs=sGAEpiMZZMvGsmoEFRKS8FOVoxBG8r/rvhRYYQXry8AfGPl2qdFwBg==

Would that be a suitable component for this ?

That regulator is suitable, but bare in mind the following points:

It has isolation, so you need to attach both GND's toguether. It does not have short circuit protection - you will burn it right away in case of a short or overload. If this is predictable chose the 3W model and use a resetable fuse rated for 400mA Use a good capacitor to filter the input and remember that type of regulators has a minimun load, so ALWAYS load the output with 5% rated current as minimun, otherwise the voltage will go higher (possibly about 5-6V).

Using a buck type regulator, as the ones recomended earlier will mostly elliminate such issues.

It seems that you are using the ATMEGA328 straight from 3.3V. If so are you de-rating the clock to 12 or 8MHz?

Regards

casemod: That regulator is suitable, but bare in mind the following points:

It has isolation, so you need to attach both GND's toguether. It does not have short circuit protection - you will burn it right away in case of a short or overload. If this is predictable chose the 3W model and use a resetable fuse rated for 400mA Use a good capacitor to filter the input and remember that type of regulators has a minimun load, so ALWAYS load the output with 5% rated current as minimun, otherwise the voltage will go higher (possibly about 5-6V).

Using a buck type regulator, as the ones recomended earlier will mostly elliminate such issues.

It seems that you are using the ATMEGA328 straight from 3.3V. If so are you de-rating the clock to 12 or 8MHz?

Regards

Thanks for the detailed explanation. Since i never worked with this type of regulator i was unaware of almost all of it :)

after consideration of your points i decided for a regulator mentionned in a previous post. I think i found it on mouser: http://www.mouser.lu/ProductDetail/Murata-Power-Solutions/OKI-78SR-33-15-W36-C/?qs=sGAEpiMZZMt6Q9lZSPl3RZClZUSM9XSruU79hDuX5cI%3d Would this type eliminate all your considerations ?

and i have a few other questions: what do they mean with voltage nominal ? in this case: "Input Voltage-nominal: 24 V". Are there any considerations when powering such a device from 12v then ? Which capacitors should be used with such a device ? just to make sure: this eliminates the need for a linear voltage regulator altogether ?

--

For your question, yes i power the atmega directly from 3.3 volts and i use it with a 8mhz crystal.

i took the time to do a diagram in eagle (my first) of the circuit as it currently works on my breadboard. Hope that helps. Note: its a bd137 transistor not a bd139.

thanks

for some reason attaching a picture seems not to work. sorry here is a link to the picture https://dl.dropboxusercontent.com/u/25825280/initial%20breadboard.PNG

tulpix: Thanks for the detailed explanation. Since i never worked with this type of regulator i was unaware of almost all of it :)

after consideration of your points i decided for a regulator mentionned in a previous post. I think i found it on mouser: http://www.mouser.lu/ProductDetail/Murata-Power-Solutions/OKI-78SR-33-15-W36-C/?qs=sGAEpiMZZMt6Q9lZSPl3RZClZUSM9XSruU79hDuX5cI%3d Would this type eliminate all your considerations ?

and i have a few other questions: what do they mean with voltage nominal ? in this case: "Input Voltage-nominal: 24 V". Are there any considerations when powering such a device from 12v then ? Which capacitors should be used with such a device ? just to make sure: this eliminates the need for a linear voltage regulator altogether ?

--

For your question, yes i power the atmega directly from 3.3 volts and i use it with a 8mhz crystal.

i took the time to do a diagram in eagle (my first) of the circuit as it currently works on my breadboard. Hope that helps. Note: its a bd137 transistor not a bd139.

thanks

They mean that the operational voltage of the regulator suits a 24V system. Such equipments, when operating in DC need to comply with an allowable working voltage spec ranging from 18 to 32V, which the regulator specs meet, BUT not with a 36V nominal system that can typically reach 40+Volts

The graps on Pag 4 show that the max efficiency is archieved at about 300mA+ and show efficiency for different operating voltages., with higher being at 7V and just a bit less at 12V. For peace of mind you can verify that it will deliver the rated 1.5A current, however for protection of your equipment I would use a 500mA resetable fuse just in case something malfunctions (the arduino has one for the USB input, for example).

The device has a built in capacitor and also on pag 4 you can see the ripple as having a max 23mV swing. If this is acceptable (I dont know the requirements of your project/components) just use a general purpose 10V electrolytic storage capacitor and the usual decoupling near the chip. No other special requirements. If you are using the analog inputs, at 10bit resolution, the usual disclaimer applies for the AVCC supply filtering as with any other power source

Yes, that device eliminates the need for a linear regulator alltoguether, assuming you need a single 3.3V supply rail.

Thanks for your detailed explanations and help.

I just ordered some parts and will experiment with them.