ATXRaspi is not supplying enough power [solved]

[nate456789]

Hello
I am using it on a Pi2 running rasplex.
Also have Flirc USB remote receiver and a cana kit Wifi dongle.
I have tried a couple good quality supplies 2+ amps and various high quality power cords.
When I power the system thru the atx I get the rainbow box upper right corner.
When I power it directly into the original pi port everything works fine.
Is there something I can tweak on the ATX to get the correct power needed.
Thanks

[Felix]

The rainbow means the voltage is supposedly dipping below some threshold, something like 4.75 if I'm not mistaken.
From my research the Pi will actually work down to 3.3v although the rainbow box will show.
Are there stability issues?
Can you actually check what the voltage is?
Does the rainbow box stay on top of any GUI you open?

[nate456789]

I measured the voltage across the Pi out pins with everything connected and the USB dongles connected and the voltage said 5.17 volts.
Doesn't make sense to me.
When I run from the normal power in usb plug I don't see the Rainbow box.
The box is visible from boot and flashes on and off every 30 sec or so on top of everything until the pi is shut down.

Power into the Pi is on Pins 1 and 3. Is there another pin that I need to supply 5v to in addition?

[Felix]

Indeed weird, I would expect the box to show only if the voltage is low.
The install page shows how to wire the 5vin to the Raspi header, or you can always use the USB connector if you have that. Only 2 pins need to be powered, GND and 5V.

[tonyp7]

I am facing the exact same issue.

When powering through the PI directly, never any issue (I have the 5.1V 2.5A official PSU). If I go through ATXRaspi, the yellow rainbow box tends to appear. The worst part? Nothing is plugged on this PI. Just the HDMI, RJ45 and a PS4 controller through bluetooth. Absolutely nothing else is plugged.

How much power can the ATXRaspi officially supply?

[Felix]

Please remember it's not a power supply, just a switch.
The switch can let through a few amps of 5V, plenty for a Pi especially the newer ones.
Can you attach a photo of your ATXRaspi + wiring + Pi?
Also a closeup photo of the back of the board so we can observe the soldering points.
A good solder job is essential to avoid any issues.

[tonyp7]

Hi Felix,

First of all, thank you very much for answering.

This is the setup:


And soldering job


You mean ATXRaspi is really just a switch and all the power of the PSU should get supplied to the PI?

[Felix]

You mean ATXRaspi is really just a switch and all the power of the PSU should get supplied to the PI?

Exactly right, the switcher is a power P-MOSFET.
Solder job looks great, cabling and USB looks OK. I would try another USB cable if you have one, see if it makes a difference. Another option is to use jumper cables from the +/- output to your Pi GND/5V on the GPIO, see if that is different.
Are you doing any overcloking or anything special? I've done that myself on the Pi3 and no issues at all.
Can you also please measure your 5V rail on the GPIO and see what the voltage is?

[tonyp7]

Hey Felix, I read 5.05V on the GPIO.

It could be the USB cable indeed. I'll get some more jumper cables and report back.

Thanks!

[perky]

You're not getting power from PC USB port, or possibly an unpowered hub? Is it possible the source of your power is limited buy that source to 100mA?
Mark.

[tonyp7]

No, it's coming straight from the PSU.

Anyway, I have finally got a chance to go to a nearby electronic shop and got some jumper cables so I could power the PI directly through GPIO: now it's working fine (no voltage drop warning so far).



Seems like micro USB was at fault here.

Thanks for the support guys, I'll make sure to post some pics once it's properly enclosed in a PSX case!

[perky]

ATX supplies don't regulate very well unless they have a decent load, they're designed for high currents. It's a good idea if you're only driving a few 10s of mA to put a high power ceramic resistor (maybe 10W, 4R7) across the +5V rail, and of course connect the PS_ON# signal to ground to turn it on.
Mark.

[Felix]

@perky: this is about ATXRaspi, the little switch board I make. Confusingly the name includes "atx"

@tonyp: Great, I would suspect the usb cable itself has some issue... but if you have this in a case then it doesn't matter if you use straight jumpers to the GPIO to power it.

[perky]

@Felix: Ah, OK. It's worthwhile point though about ATX PC supplies, they often contain saturable core reactor elements as part of the regulation circuitry. Sorry for the confusion ;-)

[Felix]

@perky: no problem, thank you for your valuable input!