Bruce_Pi
Posts: 17
Joined: Tue Feb 21, 2017 5:09 pm

(solved)pi 3 B+ killing the I2C shield

Tue May 29, 2018 1:42 pm

Hi,
I have a this i2c shield https://store.ncd.io/product/i2c-shield ... hdmi-port/ in past i have used this shield with Pi 2,3 pi zero and it has been working great.
recenelty i purchased a new Pi3 B+ and when i plug the i2c shield, the pi does not boot. if i unplug the i2c shield the pi starts to boot.
then i try the same shield on pi 3B and it works without any issue.
the pi shield connects to the Pi using these pins
5V, GND,3.3V, SDA and SCL.
i have enabled the hardware I2C in the raspi config section.

any help will be much appreciated.

Thanks
Last edited by Bruce_Pi on Fri Jun 08, 2018 9:25 pm, edited 1 time in total.

wh7qq
Posts: 1315
Joined: Thu Oct 09, 2014 2:50 am

Re: pi 3 B+ I2C shield killing the Pi

Wed May 30, 2018 12:04 am

It sounds as though your power supply is not up to the task. The 3B+ draws quite a bit more current than the 3B or 2B.

Bruce_Pi
Posts: 17
Joined: Tue Feb 21, 2017 5:09 pm

Re: pi 3 B+ I2C shield killing the Pi

Wed May 30, 2018 2:11 pm

the power supply is a BK precision variable power supply and can deliver upto 5A.
I am suspecting the 3.3V regulator on the Pi is not big enough to deliver any current ( the i2c board has a level shifter which consumes 10mA @3.3V)

Update : if i plug the i2c shield after pi boot up if works fine, i am suspecting pi 3.3V could be the issue.

User avatar
FTrevorGowen
Forum Moderator
Forum Moderator
Posts: 5052
Joined: Mon Mar 04, 2013 6:12 pm
Location: Bristol, U.K.
Contact: Website

Re: pi 3 B+ I2C shield killing the Pi

Wed May 30, 2018 3:16 pm

Bruce_Pi wrote:
Wed May 30, 2018 2:11 pm
the power supply is a BK precision variable power supply and can deliver upto 5A.
I am suspecting the 3.3V regulator on the Pi is not big enough to deliver any current ( the i2c board has a level shifter which consumes 10mA @3.3V)
Update : if i plug the i2c shield after pi boot up if works fine, i am suspecting pi 3.3V could be the issue.
What type of level shifter? (FWIW, I've used most of these modules** w/o problems:
http://www.cpmspectrepi.uk/raspberry_pi ... dules.html )
Trev.
** Apart from "G" and "J" which I haven't tested yet. AFAICT only "J" is "ic-based" the others use MOSFETs or similar.
Still running Raspbian Jessie on some older Pi's (an A, B1, B2, B+, P2B, 3xP0, P0W) but Stretch on my 2xP3A+, P3B+, P3B, B+, A+ and a B2. See: https://www.cpmspectrepi.uk/raspberry_pi/raspiidx.htm

Bruce_Pi
Posts: 17
Joined: Tue Feb 21, 2017 5:09 pm

Re: pi 3 B+ I2C shield killing the Pi

Wed May 30, 2018 4:05 pm

level shifter is PCA9306

http://www.ti.com/lit/ds/symlink/pca9306.pdf

as i mentioned earlier this shield works fine with all other Pi models.

User avatar
FTrevorGowen
Forum Moderator
Forum Moderator
Posts: 5052
Joined: Mon Mar 04, 2013 6:12 pm
Location: Bristol, U.K.
Contact: Website

Re: pi 3 B+ I2C shield killing the Pi

Wed May 30, 2018 9:20 pm

Bruce_Pi wrote:
Wed May 30, 2018 4:05 pm
level shifter is PCA9306

http://www.ti.com/lit/ds/symlink/pca9306.pdf

as i mentioned earlier this shield works fine with all other Pi models.
If I'm interpreting that datasheet correctly it appears to be a 3.3V to 1.8V level shifter rather than the 3.3V to 5V modules I've used. Maybe that implies high/higher transient currents (& transient voltage droops) leading to an improper Power On Reset condition?
Trev.
Still running Raspbian Jessie on some older Pi's (an A, B1, B2, B+, P2B, 3xP0, P0W) but Stretch on my 2xP3A+, P3B+, P3B, B+, A+ and a B2. See: https://www.cpmspectrepi.uk/raspberry_pi/raspiidx.htm

Bruce_Pi
Posts: 17
Joined: Tue Feb 21, 2017 5:09 pm

Re: pi 3 B+ I2C shield killing the Pi

Wed May 30, 2018 9:35 pm

it can do
Allows Voltage-Level Translation Between
– 1.2-V VREF1 and 1.8-V, 2.5-V, 3.3-V,
or 5-V VREF2
– 1.8-V VREF1 and 2.5-V, 3.3-V, or 5-V VREF2
– 2.5-V VREF1 and 3.3-V or 5-V VREF2
– 3.3-V VREF1 and 5-V VREF2

anywhere from 1.2V to 5V.

User avatar
FTrevorGowen
Forum Moderator
Forum Moderator
Posts: 5052
Joined: Mon Mar 04, 2013 6:12 pm
Location: Bristol, U.K.
Contact: Website

Re: pi 3 B+ I2C shield killing the Pi

Wed May 30, 2018 10:09 pm

Bruce_Pi wrote:
Wed May 30, 2018 9:35 pm
it can do
Allows Voltage-Level Translation Between
– 1.2-V VREF1 and 1.8-V, 2.5-V, 3.3-V,
or 5-V VREF2
– 1.8-V VREF1 and 2.5-V, 3.3-V, or 5-V VREF2
– 2.5-V VREF1 and 3.3-V or 5-V VREF2
– 3.3-V VREF1 and 5-V VREF2

anywhere from 1.2V to 5V.
Ok. I only quickly read through the datasheet - seems it's more similar to the i.c. used in my module "J" than I thought.
Trev.
Still running Raspbian Jessie on some older Pi's (an A, B1, B2, B+, P2B, 3xP0, P0W) but Stretch on my 2xP3A+, P3B+, P3B, B+, A+ and a B2. See: https://www.cpmspectrepi.uk/raspberry_pi/raspiidx.htm

Bruce_Pi
Posts: 17
Joined: Tue Feb 21, 2017 5:09 pm

Re: pi 3 B+ I2C shield killing the Pi

Mon Jun 04, 2018 2:27 pm

so the issue is with pi 3B+ 3.3V. i used an external regulator to supply 3.3V and it started to boot without any issue.
someone from pi hardware team needs to look into pi3B+ 3.3V power supply ratings.

there are few other people on the forums who are facing the similar issue with new pi 3B+.

dazz100
Posts: 32
Joined: Mon Apr 17, 2017 4:54 am

Re: (solved)pi 3 B+ killing the I2C shield

Fri Jul 27, 2018 9:16 pm

Hi
I am using this XBee to Pi shield found here:
https://store.ncd.io/product/i2c-shield ... xbee-port/

This works perfectly on a 3B but not a 3B+

When the CE XBee HAT is used on a 3B+ the Pi will not boot. The multi-colour splash screen does not show. The green activity led on the Pi does not flash. The screen remains blank. There are no messages at all on the screen.
The red power led on the shield is on.

I have tried multiple CE XBee shields. Same result.
Those same CE XBee shields work perfectly on the 3B.

Their forum blames this problem on the 3.3v regulator:
https://community.ncd.io/t/pr2-2-i2c-sh ... i-3-b/1109

My current work around is to use a Chinese made XBee shield that works perfectly.

This thread on this forum is marked as solved, but I don't see a solution.
I want to produce custom shields with this and other functionality. If there is a limitation with the 3B+, I need to know what it is.

penoud
Posts: 6
Joined: Tue Nov 06, 2018 8:00 pm

Re: (solved)pi 3 B+ killing the I2C shield

Tue Nov 06, 2018 8:13 pm

Same issue. I designed a hat board for accelerating aging. I was using for test phase a 3B. When going in production I used a 3B+ and red was on but no blinking on the green LED. Has anyone found a fix to this issue?

User avatar
FTrevorGowen
Forum Moderator
Forum Moderator
Posts: 5052
Joined: Mon Mar 04, 2013 6:12 pm
Location: Bristol, U.K.
Contact: Website

Re: (solved)pi 3 B+ killing the I2C shield

Tue Nov 06, 2018 10:46 pm

penoud wrote:
Tue Nov 06, 2018 8:13 pm
Same issue. I designed a hat board for accelerating aging. I was using for test phase a 3B. When going in production I used a 3B+ and red was on but no blinking on the green LED. Has anyone found a fix to this issue?
When you designed the hat board what current level were you "expecting" the Pi to be able to provide from its 3.3V supply? I've always assumed that only a few mA's are available for "off-board" use and if more is needed an external regulator, fed from the Pi's 5V (and, ultimately, from the main 5V PSU via the on-board polyfuse etc.), will be needed (sometimes, as indicated in my previous posts, already provided by a level-shifting module).
Trev.
Still running Raspbian Jessie on some older Pi's (an A, B1, B2, B+, P2B, 3xP0, P0W) but Stretch on my 2xP3A+, P3B+, P3B, B+, A+ and a B2. See: https://www.cpmspectrepi.uk/raspberry_pi/raspiidx.htm

penoud
Posts: 6
Joined: Tue Nov 06, 2018 8:00 pm

Re: (solved)pi 3 B+ killing the I2C shield

Wed Nov 07, 2018 7:06 am

Hi,
I only use the 3.3V from the RPi to power a level shifter tpe PCA9306 (on pin VREF1) and two I2C pullup. Then this should not be too much current for the RPi regulator.

penoud
Posts: 6
Joined: Tue Nov 06, 2018 8:00 pm

Re: (solved)pi 3 B+ killing the I2C shield

Wed Nov 07, 2018 7:09 am

I also tried to place a small resistor (10 to 100 Ohm) between RPi 3.3V and my daughter board in order to reduce the start current. This didn't solve my issue.

User avatar
FTrevorGowen
Forum Moderator
Forum Moderator
Posts: 5052
Joined: Mon Mar 04, 2013 6:12 pm
Location: Bristol, U.K.
Contact: Website

Re: (solved)pi 3 B+ killing the I2C shield

Wed Nov 07, 2018 3:54 pm

penoud wrote:
Wed Nov 07, 2018 7:06 am
Hi,
I only use the 3.3V from the RPi to power a level shifter tpe PCA9306 (on pin VREF1) and two I2C pullup. Then this should not be too much current for the RPi regulator.
O.K.
penoud wrote: I also tried to place a small resistor (10 to 100 Ohm) between RPi 3.3V and my daughter board in order to reduce the start current. This didn't solve my issue.
How effective that may or may not be will depend upon the values/type of decoupling capacitors you have (also) added. (I follow "old-style" TTL practice and tend to have several 0.1uF ceramic capacitors + , maybe, a 10 or 100uF electrolytic within my circuits - FWIW there's quite a few examples within my webpages)
Trev.
Still running Raspbian Jessie on some older Pi's (an A, B1, B2, B+, P2B, 3xP0, P0W) but Stretch on my 2xP3A+, P3B+, P3B, B+, A+ and a B2. See: https://www.cpmspectrepi.uk/raspberry_pi/raspiidx.htm

penoud
Posts: 6
Joined: Tue Nov 06, 2018 8:00 pm

Re: (solved)pi 3 B+ killing the I2C shield

Wed Nov 07, 2018 4:02 pm

FTrevorGowen wrote:
Wed Nov 07, 2018 3:54 pm
How effective that may or may not be will depend upon the values/type of decoupling capacitors you have (also) added. (I follow "old-style" TTL practice and tend to have several 0.1uF ceramic capacitors + , maybe, a 10 or 100uF electrolytic within my circuits - FWIW there's quite a few examples within my webpages)
Trev.
Thank for your answer Trev.

I also follow this kind of rules. But in this design I have only one 100nF on the 3.3V no 10 or 100uF. I continue to search the root cause. When I power my daughter borad (not connected to anything) only with 5V I measure 4.2V on the 3.3V. This voltage is created by the PCA9306. This IC maybe contain protection diode that let the 5V power the 3.3V. Then when the power IC on the Pi could be affected by this voltage present since boot time.

aBUGSworstnightmare
Posts: 1480
Joined: Tue Jun 30, 2015 1:35 pm

Re: (solved)pi 3 B+ killing the I2C shield

Wed Nov 07, 2018 4:15 pm

penoud wrote:
FTrevorGowen wrote:
Wed Nov 07, 2018 3:54 pm
How effective that may or may not be will depend upon the values/type of decoupling capacitors you have (also) added. (I follow "old-style" TTL practice and tend to have several 0.1uF ceramic capacitors + , maybe, a 10 or 100uF electrolytic within my circuits - FWIW there's quite a few examples within my webpages)
Trev.
Thank for your answer Trev.

I also follow this kind of rules. But in this design I have only one 100nF on the 3.3V no 10 or 100uF. I continue to search the root cause. When I power my daughter borad (not connected to anything) only with 5V I measure 4.2V on the 3.3V. This voltage is created by the PCA9306. This IC maybe contain protection diode that let the 5V power the 3.3V. Then when the power IC on the Pi could be affected by this voltage present since boot time.
VREF1 needs to be 3.3V (Raspberry Pi side) and VREF2 can be 5V (your external HW side). Have you checked your design for this?

6by9
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 7144
Joined: Wed Dec 04, 2013 11:27 am
Location: ZZ9 Plural Z Alpha, aka just outside Cambridge.

Re: (solved)pi 3 B+ killing the I2C shield

Wed Nov 07, 2018 4:58 pm

penoud wrote:
Wed Nov 07, 2018 4:02 pm
I also follow this kind of rules. But in this design I have only one 100nF on the 3.3V no 10 or 100uF. I continue to search the root cause. When I power my daughter borad (not connected to anything) only with 5V I measure 4.2V on the 3.3V. This voltage is created by the PCA9306. This IC maybe contain protection diode that let the 5V power the 3.3V. Then when the power IC on the Pi could be affected by this voltage present since boot time.
That sounds really bad. The 3.3V line has a specification for 3.3V +/- 5%, so 4.2V would be really quite high.

Your use of the PCA9306 has followed figure 7 of the datasheet? I'm a little surprised if the chip really is letting through those sorts of voltage levels as it really shouldn't.
I wonder what current is available through the chip, as it would cause issues with the PMIC.
Software Engineer at Raspberry Pi Trading. Views expressed are still personal views.
I'm not interested in doing contracts for bespoke functionality - please don't ask.

penoud
Posts: 6
Joined: Tue Nov 06, 2018 8:00 pm

Re: (solved)pi 3 B+ killing the I2C shield

Thu Nov 08, 2018 7:13 am

aBUGSworstnightmare wrote:
Wed Nov 07, 2018 4:15 pm
VREF1 needs to be 3.3V (Raspberry Pi side) and VREF2 can be 5V (your external HW side). Have you checked your design for this?
Yes carrefully.
6by9 wrote:
Wed Nov 07, 2018 4:58 pm
That sounds really bad. The 3.3V line has a specification for 3.3V +/- 5%, so 4.2V would be really quite high.

Your use of the PCA9306 has followed figure 7 of the datasheet? I'm a little surprised if the chip really is letting through those sorts of voltage levels as it really shouldn't.
I wonder what current is available through the chip, as it would cause issues with the PMIC.
I agree regarding the 4.2V is way to high compare 3.3V +/- 5%. I used exactly the figure 7. I didn't have the 200k resistor in my initial schematic. I add this resistor but this modification didn't solve the issue.

aBUGSworstnightmare
Posts: 1480
Joined: Tue Jun 30, 2015 1:35 pm

Re: (solved)pi 3 B+ killing the I2C shield

Thu Nov 08, 2018 8:21 am

6by9 wrote:
Wed Nov 07, 2018 4:58 pm
That sounds really bad. The 3.3V line has a specification for 3.3V +/- 5%, so 4.2V would be really quite high.

Your use of the PCA9306 has followed figure 7 of the datasheet? I'm a little surprised if the chip really is letting through those sorts of voltage levels as it really shouldn't.
I wonder what current is available through the chip, as it would cause issues with the PMIC.
I agree regarding the 4.2V is way to high compare 3.3V +/- 5%. I used exactly the figure 7. I didn't have the 200k resistor in my initial schematic. I add this resistor but this modification didn't solve the issue.
[/quote]
To my understanding of the data sheet the 200kOhms resistor is mandatory.
9.2.2.1 Bidirectional Voltage Translation
For the bidirectional clamping configuration (higher voltage to lower voltage or lower voltage to higher voltage),
the EN input must be connected to VREF2 and both pins pulled to high-side VDPU through a pullup resistor
(typically 200 kΩ). This allows VREF2 to regulate the EN input. A 100-pF filter capacitor connected to VREF2 is
recommended
Have you condsidered replacing the IC (as it might be damaged due to out-of-spec operation)? Can you post the portion of your schematic which deals with the I2C level shifter?

penoud
Posts: 6
Joined: Tue Nov 06, 2018 8:00 pm

Re: (solved)pi 3 B+ killing the I2C shield

Thu Nov 08, 2018 9:11 am

aBUGSworstnightmare wrote:
Thu Nov 08, 2018 8:21 am
To my understanding of the data sheet the 200kOhms resistor is mandatory.
I didn't realise this was a mandatory.
aBUGSworstnightmare wrote:
Thu Nov 08, 2018 8:21 am
Have you condsidered replacing the IC (as it might be damaged due to out-of-spec operation)? Can you post the portion of your schematic which deals with the I2C level shifter?
YES this solve my issue. After IC replacement the 3.3V on the daughter board give me a 3.65V with no load. Then I tried to boot with the daughter board connected to the RPi. It's working well. Thank you for your help.

6by9
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 7144
Joined: Wed Dec 04, 2013 11:27 am
Location: ZZ9 Plural Z Alpha, aka just outside Cambridge.

Re: (solved)pi 3 B+ killing the I2C shield

Thu Nov 08, 2018 9:58 am

penoud wrote:
Thu Nov 08, 2018 9:11 am
aBUGSworstnightmare wrote:
Thu Nov 08, 2018 8:21 am
To my understanding of the data sheet the 200kOhms resistor is mandatory.
I didn't realise this was a mandatory.
9.2.2.1 Bidirectional Voltage Translation
For the bidirectional clamping configuration (higher voltage to lower voltage or lower voltage to higher voltage), the EN input must be connected to V REF2 and both pins pulled to high-side V DPU through a pullup resistor (typically 200 k Ω ). This allows V REF2 to regulate the EN input.
Must = mandatory.
YES this solve my issue. After IC replacement the 3.3V on the daughter board give me a 3.65V with no load. Then I tried to boot with the daughter board connected to the RPi. It's working well. Thank you for your help.
That's good news.
Software Engineer at Raspberry Pi Trading. Views expressed are still personal views.
I'm not interested in doing contracts for bespoke functionality - please don't ask.

Return to “Interfacing (DSI, CSI, I2C, etc.)”