Milica
Posts: 25
Joined: Sat Jan 05, 2019 3:00 pm

GPIO only has current at 5V pins (none at 3.3V pins)

Sat Jan 05, 2019 8:13 pm

After working fine a few days ago my Pi 3 b+ will not start up today (has the red light constantly on but no ACT light flashing). I have tested using a multimeter the various GPIO pins and found I have 5V power at the first two pins on the outer row (classed as pin 2 and 4 I think) but NO POWER at pin 1 and 17 (meant to be 3.3V I believe).

Note that I also have the mains power with dual microB USB power cable attached, so would this necessarily cause a problem booting up anyway? My GPIO pins are only really being used to power my LCD screen - when it was working the other day I had the red jumper on the 5V pin but it is not working today.

pcmanbob
Posts: 6898
Joined: Fri May 31, 2013 9:28 pm
Location: Mansfield UK

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Sat Jan 05, 2019 10:25 pm

If your pi has no voltage on the 3.3v power pins on the gpio header it's faulty, some pi3B+ have suffered sudden failures in this way, but it's also possible that if you shorted one of the 3.3v pins to ground or a 5v pin that you caused the failure.

Your only option realistic option is to replace your pi.
We want information… information… information........................no information no help
The use of crystal balls & mind reading are not supported

Milica
Posts: 25
Joined: Sat Jan 05, 2019 3:00 pm

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Sat Jan 05, 2019 10:31 pm

pcmanbob wrote:
Sat Jan 05, 2019 10:25 pm
If your pi has no voltage on the 3.3v power pins on the gpio header it's faulty, some pi3B+ have suffered sudden failures in this way, but it's also possible that if you shorted one of the 3.3v pins to ground or a 5v pin that you caused the failure.

Your only option realistic option is to replace your pi.
Thanks pcmanbob. That's the conclusion I was starting to get to as well.

pcmanbob
Posts: 6898
Joined: Fri May 31, 2013 9:28 pm
Location: Mansfield UK

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Sat Jan 05, 2019 10:33 pm

Milica wrote:
Sat Jan 05, 2019 10:31 pm
pcmanbob wrote:
Sat Jan 05, 2019 10:25 pm
If your pi has no voltage on the 3.3v power pins on the gpio header it's faulty, some pi3B+ have suffered sudden failures in this way, but it's also possible that if you shorted one of the 3.3v pins to ground or a 5v pin that you caused the failure.

Your only option realistic option is to replace your pi.
Thanks pcmanbob. That's the conclusion I was starting to get to as well.
If you are sure it's nothing you did to cause the failure and your pi is not that old talk to your supplier.
We want information… information… information........................no information no help
The use of crystal balls & mind reading are not supported

User avatar
JohnBeardmore
Posts: 211
Joined: Thu Nov 15, 2012 11:03 pm
Location: Derbyshire UK.
Contact: Website

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Sat Jan 05, 2019 11:53 pm

So why have "some pi3B+ have suffered sudden failures in this way" ?

This is interesting. We made a board of support hardware for our Pi projectcts (open hardware if anyone's interested), which has various 5V PSU options. These power the Pi, which in turn provides 3.3V power of a hardware watchdog timer, some A to D converters and a D to A.

This worked fine with the Pi 2B and the Pi3B, but was pretty unreliable with the 3B+ with the 3V3 rail freqently going down. Sometimes it would recover when the power was cycled, but we replaced a high proportion of them and still had problems.

In the end we reverted to the Pi3B and six Pis have been collecting 78 chanels of data without missing a beat. We've come to the provisional conclusion that the 3B+ is a bit of a dog, and won't be buying more of them unless we get to the bottom of this problem.

Any comments from other users and the RPF welcome !

Cheers, J/.
Author of oBeMS open source Building energy Management System.
Automatic Meter Reading (AMR), Building Management System (BMS),
Building Energy Management System (BEMS), Infrastructure Control System (ICS).
See: http://t4sustainability.co.uk/oBeMS/

User avatar
Burngate
Posts: 6000
Joined: Thu Sep 29, 2011 4:34 pm
Location: Berkshire UK Tralfamadore
Contact: Website

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Sun Jan 06, 2019 11:41 am

As I understand it:
The 3B+ features a new DC-DC chip, which is supposedly bullet-proof but has on occasion died for no apparent reason, or isn't as short-circuit-proof as one might expect.
Talks are ongoing between R-Pi & chip manufacturer about the situation.
Someone - Hippy? - was collecting annecdotal information from posts on the forum.

Milica
Posts: 25
Joined: Sat Jan 05, 2019 3:00 pm

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Sun Jan 06, 2019 12:42 pm

Many thanks pcmanbob, John Beardmore and Burngate for your input and advice. Reading around it does seem you are right, there appear to be many other people going through the same thing.

Will this in itself be the reason that I am unable to get the Pi to boot up? I just assumed that since I am getting power at the mini USB (where I have plugged the mains supply in) of 5V the Pi would at least boot up!?

pcmanbob
Posts: 6898
Joined: Fri May 31, 2013 9:28 pm
Location: Mansfield UK

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Sun Jan 06, 2019 12:48 pm

Yes the missing 3.3v is the reason your pi will not boot, the 3.3v is used to power most of the chips on the pi.
We want information… information… information........................no information no help
The use of crystal balls & mind reading are not supported

Milica
Posts: 25
Joined: Sat Jan 05, 2019 3:00 pm

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Sun Jan 06, 2019 12:53 pm

pcmanbob wrote:
Sun Jan 06, 2019 12:48 pm
Yes the missing 3.3v is the reason your pi will not boot, the 3.3v is used to power most of the chips on the pi.
:thumbsup:

hippy
Posts: 5928
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Sun Jan 06, 2019 3:55 pm

JohnBeardmore wrote:
Sat Jan 05, 2019 11:53 pm
So why have "some pi3B+ have suffered sudden failures in this way" ?
No one knows. RPT are investigating this and other issues relating to the PMIC with MaxLinear, the manufacturer of the chip.
Burngate wrote:
Sun Jan 06, 2019 11:41 am
Someone - Hippy? - was collecting annecdotal information from posts on the forum.
More noting them in passing. Multiple reports of 'unexplained failures', often on the same day, raised some internal mental flag, and later I set about quantifying the issue.
Milica wrote:
Sun Jan 06, 2019 12:42 pm
there appear to be many other people going through the same thing.
Not that many. I have noted about 50 posts of 'unexplained failures'. That's a very small number compared to how many 3B+ have been sold.

I only recorded those where it was certain the user had a 3B+, were adamant they had not shorted anything out, where it appeared they had not done anything wrong, and the Pi had booted successfully then failed.

I may have missed some, even discounted some I shouldn't have, but I doubt it would alter the numbers significantly. There may be some I shouldn't have recorded because I never tracked the threads.

Looking at those I recorded; reports of 'unexplained failure' increased from June and peaked in August last year, which may or may not be significant - 2 in Mar 2018, 3 Apr, 2 May, 7 Jun, 9 Jul, 14 Aug, 3 Sep, 2 Oct, 4 Nov, 1 Dec, 2 Jan 2019.

If I were to have recorded those where the PMIC failed after a short it would probably double the number. That's based on the posts I recall discounting as 'not unexplained', where it wasn't clear the user had a 3B+, or where a short had been admitted to or was suspected.

But, doubling to 100 would still be a small number - 0.01%, one in 10,000, if a million sold, and it's probably more than that.

hippy
Posts: 5928
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Sun Jan 06, 2019 4:13 pm

JohnBeardmore wrote:
Sat Jan 05, 2019 11:53 pm
We made a board of support hardware for our Pi projectcts ... which has various 5V PSU options. These power the Pi, which in turn provides 3.3V power of a hardware watchdog timer, some A to D converters and a D to A.

This worked fine with the Pi 2B and the Pi3B, but was pretty unreliable with the 3B+ with the 3V3 rail freqently going down. Sometimes it would recover when the power was cycled, but we replaced a high proportion of them and still had problems.
There have been reports of HATs and add-on boards which use the 3V3 supply of a Pi preventing some Pi 3B+ and Pi 3A+ from booting. Those boot and work fine once the HAT or board is removed. You may have been encountering this or a related issue.

This is something the RPT are investigating . The current workaround appears to be to disconnect the Pi 3V3 supply and provide 3V3 from a separate regulator powered from the Pi's 5V supply. It might be worth trying that with a Pi 3B+ which is proving problematical with your board.

If your boards were just losing 3V3 while running it might be worth contacting the RPT to let them know of your issue as they might want to include that in their investigations or it may help with them.

User avatar
JohnBeardmore
Posts: 211
Joined: Thu Nov 15, 2012 11:03 pm
Location: Derbyshire UK.
Contact: Website

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Fri Jan 11, 2019 12:54 am

We bought a bunch of 3B+ boards back in early summer. We used 6 on a farm for logging air quality in chicken sheds, one as a data logger in the chemistry lab, and had a few spares.

Of the 7 used, I'd guess four have been prone to inexplicable crashes on the farm sometimes with silly 3V3 rail voltages, which have sometimes been OK after cycling the power, but which have failed in the end, and been replaced from the pool of spare 3B+s.

A colleague and I spent many man days traveling to the farm to chase odd hardware faults (best part of a 100 mile round trip), but reliability remained poor, which in turn slowed down development of the 'distributed data logging' thing. I don't care about the cost of the Pis. I do care about the cost of that time and the damage those delays have done to the project.

Further to those failures, the 3B+ in the lab refused to boot the other day. Of the 7 3B+ units then, four were prone to unexplained crashing typically every few days then failed completely, and the one in the lab, failed completely over a 7 month period. That's 5 out of 7 !

We've developed similar systems before using the original 3B. We've had no failures, and very few unexplained crashes, so we decided to revert to using the original 3B Pis on the farm. Since we reverted, we've had no unexplained crashes and no failures, I was able to finish the software over christmas, and data is flowing reliably.

It may be that only a small number / proportion of people are reporting this in RPF forums. Many people who buy a Pi may never use these forums. A million 3B+s may have been sold (?) butI don't see a million users on threse forums.

A big percentage of the 3B+ units we've bought have failed. I wonder if newbies who buy a Pi and find it breaks after a while assume that it's something they've done, or a duff SD card or whatever ? It must be very dispiriting for them unless they have the resources and tenacity to get to the bottom of what's going on.

Cheers, J/.
Author of oBeMS open source Building energy Management System.
Automatic Meter Reading (AMR), Building Management System (BMS),
Building Energy Management System (BEMS), Infrastructure Control System (ICS).
See: http://t4sustainability.co.uk/oBeMS/

hippy
Posts: 5928
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Fri Jan 11, 2019 7:06 pm

JohnBeardmore wrote:
Fri Jan 11, 2019 12:54 am
A big percentage of the 3B+ units we've bought have failed.
It may be worth contacting your retailer, emailing the Foundation to let them know, or reporting your post and asking if a mod can advise what to do.
JohnBeardmore wrote:
Fri Jan 11, 2019 12:54 am
I wonder if newbies who buy a Pi and find it breaks after a while assume that it's something they've done, or a duff SD card or whatever ?
Some may but I would have expected most to have come and 'vented' here or on social media when it does happen and we would see more posts if it were a common problem. But no one can know for sure.

PeterBell
Posts: 9
Joined: Mon Nov 17, 2014 4:46 am
Location: Philippines (since 2007)

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Fri Aug 16, 2019 7:43 am

I have a Pi3B+ which failed suddenly after several months use. It was running LibreElec from micro SD card. No expansion boards installed on the GPIO pins - simply powered from a genuine Pi power plug, with HDMI and network connected. It was working perfectly well then, one day several months ago, after a powercut (common occurrence here in Philippines) it failed to start up. A little investigation revealed that there was no power on the 3.3V pins on the GPIO connector. Is there any way to repair this, or is it simply fit for throwing in the bin?
One other Pi3B+ running LibreElec and one running in a PiTop are still fine, as are five Pi1Bs and five Pi2Bs.

Has this issue been overcome in the Pi4?

pcmanbob
Posts: 6898
Joined: Fri May 31, 2013 9:28 pm
Location: Mansfield UK

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Fri Aug 16, 2019 8:06 am

No you cant repair your pi , the only option is replacement.
We want information… information… information........................no information no help
The use of crystal balls & mind reading are not supported

User avatar
JohnBeardmore
Posts: 211
Joined: Thu Nov 15, 2012 11:03 pm
Location: Derbyshire UK.
Contact: Website

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Fri Aug 16, 2019 2:50 pm

So to ask the obvious question again,

"Has this issue been overcome in the Pi4 ?".

If not, I'm still sticking with the nonplussed Pi3B if I can still get them.


Cheers, J/.
Author of oBeMS open source Building energy Management System.
Automatic Meter Reading (AMR), Building Management System (BMS),
Building Energy Management System (BEMS), Infrastructure Control System (ICS).
See: http://t4sustainability.co.uk/oBeMS/

User avatar
rpdom
Posts: 15168
Joined: Sun May 06, 2012 5:17 am
Location: Chelmsford, Essex, UK

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Fri Aug 16, 2019 3:38 pm

The Pi 4B has a different power management circuit which is less susceptible to being killed by shorting the GPIO pins.

drgeoff
Posts: 9803
Joined: Wed Jan 25, 2012 6:39 pm

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Fri Aug 16, 2019 3:52 pm

rpdom wrote:
Fri Aug 16, 2019 3:38 pm
The Pi 4B has a different power management circuit which is less susceptible to being killed by shorting the GPIO pins.
1. Note that "less" is not a synonym for "not". :)

2. The statement is only applicable to the power pins on the GPIO header. The RPi4B GPIOs are no less susceptible to electrical abuse than those on previous models.

User avatar
rpdom
Posts: 15168
Joined: Sun May 06, 2012 5:17 am
Location: Chelmsford, Essex, UK

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Fri Aug 16, 2019 4:39 pm

drgeoff wrote:
Fri Aug 16, 2019 3:52 pm
rpdom wrote:
Fri Aug 16, 2019 3:38 pm
The Pi 4B has a different power management circuit which is less susceptible to being killed by shorting the GPIO pins.
1. Note that "less" is not a synonym for "not". :)

2. The statement is only applicable to the power pins on the GPIO header. The RPi4B GPIOs are no less susceptible to electrical abuse than those on previous models.
Indeed. The Pis are not built to withstand electrical (or any other sort of) abuse.

User avatar
JohnBeardmore
Posts: 211
Joined: Thu Nov 15, 2012 11:03 pm
Location: Derbyshire UK.
Contact: Website

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Fri Aug 16, 2019 5:35 pm

Don't think any of my failed Pis were abused or had their 3V3 rails shorted. Most were in IP65 boxes in a chicken shed measurig air quality, not on the work bench having uninsulated DMM probes poked at the pins. The only other one was in a similarly protective box in a nice warm chemistry lab sniffing the gases coming of an air cleaning catalyst.

They certainly don't seem to need to be shorted to fail.

We do use power from the 3V3 rail though, variously running SPI AtoD and DtoAs and a watchdog timer (555 and 4017). Between them these have a few uF of decoupling capacitors. Would the switch on suge from those caps be enough to break the 3B+ regulator ?

Sice the early 7805 type linear regulators of the 1970s, most regulators I've encountered have been basically short circuit proof. It seems unconscionable to me, to use a regulator for the amateur 'my first computer' market, which is so easily destroyed or fails spontaneously.

Glad that the 4B PSU is less susceptible. Might try one then.

Cheers, J/.
Author of oBeMS open source Building energy Management System.
Automatic Meter Reading (AMR), Building Management System (BMS),
Building Energy Management System (BEMS), Infrastructure Control System (ICS).
See: http://t4sustainability.co.uk/oBeMS/

mattmiller
Posts: 2106
Joined: Thu Feb 05, 2015 11:25 pm

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Fri Aug 16, 2019 8:02 pm

The Pi 4B has a different power management circuit which is less susceptible to being killed by shorting the GPIO pins
Have you got a source for that info?

PeterBell
Posts: 9
Joined: Mon Nov 17, 2014 4:46 am
Location: Philippines (since 2007)

Re: GPIO only has current at 5V pins (none at 3.3V pins)

Sat Aug 17, 2019 9:08 am

JohnBeardmore wrote:
Fri Aug 16, 2019 5:35 pm
Don't think any of my failed Pis were abused or had their 3V3 rails shorted.
Indeed! As I stated, the one that failed on me had no peripheral devices connected, other than to the ethernet, HDMI, microSD slot and USB power port. There was nothing near the GPIO pins.

Return to “Troubleshooting”