Featured Articles
Announcing Pi
Technical Specifications
Raspberry Pi FAQs
Win a Pi
Raspberry Pi Wishlist
Actions
Forum Thread Details
  • Replies 37 replies
  • Subscribers 680 subscribers
  • Views 5801 views
  • Users 0 members are here
Related

Just an interesting share about voltage input

screamingtiger

I am powering my RPI 2 with an external battery.  Its a 7.2V "LIFE" pack, which is a 2 cell battery.  The 7.2V is the max voltage, and the nominal voltage is 6.6V.  The current draw from a RPI 2 puts the voltage somewhere between the two.

 

I use a 5V linear regulator between the pi and the battery, along with a cut off micro USB cable attachment I made.  The connection between the battery and the regulator is the same as the connection between the regulator and the USB cable I modified. What this means is I can, and I did, plug the battery directly into the PI, inadvertently!  This will be an issue I fix so I don't do it again.

 

I had the pi 2 going for about 2 hours, pretty hard.   Here is what it had attached:

Chipkit Pi

Wi Pi

Wireless keyboard Dongle

OLED screen powered by the chipkit pi and connected vi I2C

A microstack GPS powered by the chipkit pi and connected via serial to the raspberry pi.

 

On top of that I was compling code and had the thread running that parses out the GPS data.

 

I DO NOT advocate trying to over power the Pi on purpose, I just wanted to share what may occur if you accidently do it like I did.  It appears the built in protection gives some leeway for errors and the Pi will just not go "poof" with a minor error like this.  Please only power the Pi with 5V from  a clean source such as a regulator.

 

Here are a list of regulators you can use:

 

http://www.hobbyking.com/hobbyking/store/__38718__Turnigy_HV_SBEC_5A_Switch_Regulator_8_42V_input_US_Warehouse_.html

 

But like these, the one I am using is 8 years old:

http://www.hobbyking.com/hobbyking/store/__4319__TURNIGY_3A_UBEC_w_Noise_Reduction.html

 

The best regulator I found is this one but it cost more and  can be configured via USB to select output voltages:

http://www.amazon.com/Castle-Creations-Bec-Switching-Regulator/dp/B000MXAR12

 

If you need serious amps:

http://www.amazon.com/Castle-Creations-Bec-Switching-Regulator/dp/B002ITEL42/ref=pd_sim_21_5?ie=UTF8&refRID=01S0CR3ZEYDVPP8TCRPC

 

The Castle Creations line is very good over all.

 

I was logged in via SSH, so I am not sure if the little square was showing up that indicates power issues.   As you can see, in the end everything is really being powered through the PI.  When I get to production I hope to connect everything directly to the external regulator except the chipkit pi.

 

So....I smelled something HOT, not burning, just like hot electronics.  My RPI B+ smells similar when I have it mounted in the Pi Tin case but this was stronger.  I then realized my mistake.

 

So I just wanted to share, plugging in a 7.2V battery will not kill the Pi if you catch it soon enough.  I venture to guess it would of kept running for days if I didn't have all the extra stuff attached to it.    I've had going for a couple days now plugged into the right power adapter, and it does not seem to be damaged.

 

Its nice to know this but not happy I took such a risk to find out!  here is a pic of what I looked like on the table.  I have the regulator installed here but you get the point.

 

PS the white tape is not being used to insulate anything, I had used to hold the wires in place is all when I had it on the table.

 

image

 

Message was edited by: Joey Thompson Added more clarification.

  • It's impressive that RasPi 2 survived 7.2V.  RasPi hasn't published the schematics for RasPi2 so we don't know what sort of protection circuits they have.  The have published the power schematics for the Model B+ and RasPi 2 is probably similar:  https://www.raspberrypi.org/documentation/hardware/raspberrypi/schematics/README.md

     

    The Model B+ has a 5V zener diode so if your power input voltage is higher than 5V the zener switches on and drops the voltage.  If your batteries can deliver a lot of current that zener will get hot.  There's also a PFET that the input current goes through, so it gets hot too.  If things get hot enough, a self-resetting fuse goes into the high resistance state.

  • The parts that were hot:

    CPU

    The HDMI port

    The sound jack.

     

    CPU for good reason, the others probably since they have metal the heat was sunk to?

  • in reply to screamingtiger

    Joey Thompson wrote:

     

    The parts that were hot:

    CPU

    The HDMI port

    The sound jack.

     

    CPU for good reason, the others probably since they have metal the heat was sunk to?

    Yes, the metal shields act as heat sinks in a well-designed board with good ground and power planes.  The USB/LAN chip used to get quite hot on some of the original boards, but that was because of a design error.  The rev 2.0 PCB fixed the error and added extra vias to conduct heat away from the USB/LAN chip.

  • This is somewhat to be expected....

     

    The 5V on the 'pi is used more or less as "unregulated power source". The CPU for instance actually runs on 1.2V. So there is a step down converter... in the CPU image that makes that 1.2V from the 5V power input. This works without trouble to below 3.0V. On the other side the specs say "max 5.5V". This is a "common spec" that the chip makers consider a "workable limit". Even if they test their chips and find they work fine up to 7V they still specifiy 5.5V which allows them to make changes that won't impact anyone who "follows the rules".

     

    That said, besides the CPU there are a couple of DCDC converters that make other voltages like 3.3V and 1.5V. The situation is exactly the same: an upper limit probably somewhere in the 6-10V range, but a specification of 5.5V.

     

    And lastly, the 5V is passed on to some interfaces that specify 5V. USB and HDMI. External hardware there might not like the high voltage you applied. This external hardware probably includes most of your external hardware as well. (i.e. powered by 5V).

     

    I've made exactly the same mistake: Connect a 2 cell LIPO to the 5V line of a project. In that case, the FT232 survived without a hitch. And that's a fussy chip: connect its reset line to the 5V and it gets really really hot...

     

    Oh! One more thing.... If John is right about the 5V zener then... First of all it must be at least 5.5V to allow for a normal 5.5V powersupply. Second, it will not divert power, but "consume as much current as necessary to force the powersupply to less than ...." (whatever the nominal voltage of the zener is). If your pack is capable of delivering 20A at 6V, that's exactly what will happen. (I didn't know they had 3.3V/cell nominal LIFE packs in that format).

  • in reply to rew

    Its a 3.3V per cell life pack and the one I had used was a 10C with capacity 2100mAh, so it can safely deliver 21A.  I have a new one that is 30C.

    Here is a big one with 30C rating that can dole out 126A safely.

    http://www.hobbyking.com/hobbyking/store/__18641__ZIPPY_Flightmax_4200mAh_2S2P_30C_LiFePo4_Pack_USA_Warehouse_.html

     

    There is no reason for me to use a LIFE pack, and I have switched over to a 2s LIPO as well.  Its just a bit less heat on a 5V regulator.

     

    BTW the external hardware all survived.  The chipkit pi has a 5V regulator on it, so it will take in much higher voltages.  Everything else was being fed off the regulator of the chipkit pi so it never saw the higher voltage.

     

    I think the Pi will run fine from a 6.6V lipo all day long.  I am pulling too much current through it even if I used 5V exactly. 

  • in reply to johnbeetem

    The poster says he was using a 5V linear regulator between the battery and the PI so the PI is not running directly of the LIPO. This is a safe way of doing this.

     

    Connecting the LIPO directly to the PI would surely void any warranty and probably kill a few chips in the process, especially if the 7.2V lipo is freshly charged

     

    the 5V into the PI is as stated many times already, directly connected to the USB power lines on the board and therefore any devices connected to this would also face the full wrath of the LIPO voltage if the 5V Linear regulator was not used. and would also be massively exceed the specification of the USB design. Any USB devices connected might survive but that would be good luck, not good design

  • in reply to Robert Peter Oakes

    Peter, if you read a bit more careful: Joey has accidentally connected the LIPO to the 5V of the 'pi. This resulted (to his surprise, and apparently yours!) in no damaged chips!

  • in reply to Robert Peter Oakes

    I had a the Wi Pi and a wireless keyboard dongle, both survived and are currently working.

     

    Im not going to pretend I know why it all worked out, John's explanation may explain it.  I just simply wanted to share my experience.

     

    What I take out of it is that if you mess up like I did, the Raspberry Pi is not so fragile and has some built in protection for people who are taking cold medicine and working on electronics, like me image

     

    I'm very glad I didn't burn it up, I probably would of given up on my project at that point if I lost everything.  The stuff in my hand in that picture is about $100 USD.

     

    Edit:  Naw, I would replaced it all  and kept going  image

  • in reply to Robert Peter Oakes

    Peter, I assure you, I am NOT lucky  image  I think it is just a good design.