1

I am trying to power my pi through the 5V pin on the header, through an lm317, for a clean 5V output. However, when I connect the 5V to the board, the pi constantly reboots itself. I am assuming this is due to low voltage, and I do have other loads in the circuit, but I believe I am well within the 1.5A max that the lm317 provides. Can someone explain to me why this is happening?

goldilocks
  • 58,859
  • 17
  • 112
  • 227
Zotto
  • 45
  • 1
  • 6
  • Sounds busy. Without more information on how your power supply circuit is constructed it's unlikely that anyone will be able to offer useful advice on what might be wrong. Can you edit your question to include a connection diagram and/or photograph of your setup? – goobering Aug 22 '16 at 16:25
  • The Pi has no 5V GPIO, they are all 3V3. Hopefully you mean one of the 5V pins on the expansion header, otherwise you will have destroyed your Pi. – joan Aug 22 '16 at 16:26
  • Im confused the pi does have 2 5V gpio pins, pins 2 and 4. – Zotto Aug 22 '16 at 16:47
  • 1
    What joan means is not all the pins on the breakout are GPIOs, although most of them are so it is often colloquially called the "GPIO header" or "GPIO breakout". But the 5V pins are not GPIOs (hopefully the wikipedia article makes it clear why); neither are the GND or 3.3V power pins. The 5V power pins are just +5V sources. They are not in any sense "general purpose input/output". – goldilocks Aug 22 '16 at 17:06
  • WRT to your question, have you verified the voltage at the pi with a multimeter? – goldilocks Aug 22 '16 at 17:08
  • Can you tell us specifically what pin you've connected to? Are you just using the LM317 or do you have caps on there? Its possible that when you start up it has a high draw which pulls the voltage down, without the caps you can pass through voltage spikes. – Ron Beyer Aug 22 '16 at 18:51
  • Yeah its just series resistors across the adj and vout pins. I read that I can connect 5V to the 5V power pin,(pin 2) and can start the pi in that manner. – Zotto Aug 22 '16 at 21:06
  • That would make interesting readng - do you have a link? – dbmitch Aug 22 '16 at 23:08
  • In several past postings on 5V supplies it has been recommended and shown that 5.25V should be the adjusted value for supply voltage. – spicetraders Aug 23 '16 at 17:13

2 Answers2

1

Rather than trying to guess why your Pi (which Pi model) doesn't work can I suggest you just give up.

The LM317 is a 40 year old chip (I didn't know you can still get them) and I haven't used one for over 30 years. They were never a particularly good regulator, and you would only use one if you really needed a variable voltage. They have a quite high Vin-Vout differential, making them even more inefficient than fixed regulators.

You MIGHT be able to get 1.5A out of the regulator, but this would require a substantial heatsink.

You would certainly need adequate decoupling to ensure stability.

Milliways
  • 59,890
  • 31
  • 101
  • 209
  • +1 for abandoning LM317. LM7805 in TO220 case with a small radiator should provide 2A and require minimum setup, while it probably costs the same amount as LM317. – Mark Aug 23 '16 at 13:33
0

A voltmeter or better yet a scope on the 5V (as previously stated) will answer if the 5V is stable enough to handle the load. The LM317 may show unloaded stable 5V but if input to the LM317 is not high enough voltage or current the output will not be stable. Input and output caps shown are standard suggested values. Not shown in circuit but a good add on for ripple rejection is a 1N4002 across the 220ohm resistor suggested PS

spicetraders
  • 141
  • 4