LikesAChallenge
Posts: 32
Joined: Fri Dec 05, 2014 3:39 pm

I2C and 5V - Definitive answer?

Wed Mar 09, 2016 12:46 pm

Background:

I'm preparing to deploy a number of humidity/temp sensors and these will be positioned towards the top end of what I2C can deal with length wise (~2m, potentially 5m+).

I've tested up to 1.5m with 3.3V with no obvious degradation in performance.

Would using 5v (the sensors support 5V) enable me to push the line length higher?

Some posts suggest the RPI is not tolerant to 5V on SDA/SCL but others say it is fine as it is only a sink to GND. Also they say as the RPI has 1.8k? pullup resistors to 3.3V I don't need external 2.2k resistors?

(https://www.element14.com/community/thr ... hread=true
& https://learn.adafruit.com/adding-a-rea ... ng-the-rtc)

So my questions are:

1. Will 5V rather than 3.3V allow my line lengths to be higher? My logic would suggest it would (larger difference between logical high and logical low).
2. Do I need external pull-up resistors if using 3.3V? (I assume I would for 5V as the Pi resistors pull up from 3.3V? - Using 2.2k at the moment for 3.3V trials.)

3. If 1 is true, will I harm the Pi using 5V?
4. If 3 is true, is it worth using a level shifter?

Thanks for your help :D

User avatar
RaTTuS
Posts: 10531
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK
Contact: Twitter YouTube

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 12:47 pm

5V will kill your RPi
How To ask Questions :- http://www.catb.org/esr/faqs/smart-questions.html
WARNING - some parts of this post may be erroneous YMMV

1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX
Covfefe

User avatar
joan
Posts: 14686
Joined: Thu Jul 05, 2012 5:09 pm
Location: UK

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 12:51 pm

You should not feed 5V into a Pi GPIO.

It should be safe to power the sensors from 5V provided they do not have pull-ups to their input voltage fitted.

Power the sensors and then measure the voltage at SDA and SCL on the sensor.

scotty101
Posts: 3803
Joined: Fri Jun 08, 2012 6:03 pm

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 1:24 pm

I have seen examples where people have used logic level shifters on both ends.

At the end with the sensor they step up the voltage and then at the receiving end they step it down again. Likely to work if the speed of your signal is low enough or your logic level shifters don't introduce much latency. Using a higher voltage does improve the signal to noise ratio.
Another option is to have a small microcontroller right beside the sensor and then communicate back to the pi by another suitable protocol like RS485.

I2C was designed to be an inter-circuit protocol i.e. between one side of a PCB to the other rather than for external sensors. Doing anything else you accept the risk of it not working.
Electronic and Computer Engineer
Pi Interests: Home Automation, IOT, Python and Tkinter

User avatar
RaTTuS
Posts: 10531
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK
Contact: Twitter YouTube

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 1:26 pm

Level shifters will work
How To ask Questions :- http://www.catb.org/esr/faqs/smart-questions.html
WARNING - some parts of this post may be erroneous YMMV

1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX
Covfefe

hippy
Posts: 6743
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 4:56 pm

RaTTuS wrote:5V will kill your RPi
Yet others say it won't when appropriately current limited, and same say they are successfully using a Pi in such a configuration.

It is fair to say it's recommended to not connect 5V to a Pi's GPIO pin, and certainly never without current limiting. As to what current limiting is required is anyone's guess.

There is conflicting opinion and information on what injection currents are within specified limits for a Pi's GPIO input which will hopefully be cleared up when the GPIO electrical specifications are released.

In the case of I2C it's more complicated than a simple voltage signal at 5V. Output to the I2C bus should be open-collector pull-down so nothing should ever be putting 5V onto the bus. But there may be some leakage path from a +5V power rail through the I2C device to the i2C bus line via its input circuit. Would that be enough to damage the Pi ? I would doubt it but honestly don't know.

Would taking the i2C bus pull-ups to 5V cause damage ? That's more likely, and increasingly so as pull-up resistance is reduced.

Probably the easiest way to extend I2C bus length and keep things safe is to stick to 3V3 and simply run the bus slower.

Massi
Posts: 1691
Joined: Fri May 02, 2014 1:52 pm
Location: Italy

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 5:08 pm

well, powering a sensor with 5V does not imply that it will output 5V
this is often true for a lot of sensor (think to all the open collector outputs), also (and probably more) for i2c
i2c bus is "terminated" by a pullup on sda and scl
this termination commonly is done by the master (the pi, in our case)
in fact, pi i2c-1 bus has phisical 1.8k pullups.

I this is the case, a sensor powered with 5V will not output 5V on the i2c bus (since it only pulls down lines when needed) and should not destroy the pi

BUT

there's another problem. Sensor side, this time. Probably the sensor will accept as a valid logic HIGH level, a voltage related to the power supply (something like 0.8 * Vcc)
If this is the case, the Pi 3,3 V isn't enough to reach this voltage and so you can't be sure of the proper working of sensor.

All that said, but a ton of level converter and live happily in the world of 3,3V :)

Edit: damn, i was clearly late.. So to add something, i'm using my i2c bus something like 10m away from the pi using range extenders that work quite good. These
http://sandboxelectronics.com/?product= ... 715-module

User avatar
GTR2Fan
Posts: 1601
Joined: Sun Feb 23, 2014 9:20 pm
Location: South East UK

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 5:11 pm

hippy wrote:
RaTTuS wrote:5V will kill your RPi
Yet others say it won't when appropriately current limited, and same say they are successfully using a Pi in such a configuration.
Shoving 5V into a 3V3 logic pin via a current-limiting resistor and hoping that the internal protection diode will save you is an awful way of doing it.

That would be relying on the current-handling capacity (along with many other potentially strange characteristics) of internal protection diodes that are there primarily to handle static discharge. They're not there for the purpose of correcting bad circuit design in the outside world and should never be relied upon to do so.

That would be very poor practice.
Last edited by GTR2Fan on Wed Mar 09, 2016 5:27 pm, edited 2 times in total.
Pi2B Mini-PC/Media Centre: ARM=1GHz (+3), Core=500MHz, v3d=500MHz, h264=333MHz, RAM=DDR2-1200 (+6/+4/+4+schmoo). Sandisk Ultra HC-I 32GB microSD card on '50=100' OCed slot (42MB/s read) running Raspbian/KODI16, Seagate 3.5" 1.5TB HDD mass storage.

LikesAChallenge
Posts: 32
Joined: Fri Dec 05, 2014 3:39 pm

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 5:13 pm

Thanks all.

I shall look at the output of the SDA and SCL lines when the sensor is plugged into 5V.

Also, the sensor already works perfectly at 3.3V, so it must have internal logic to work that out (20% low, 80% high).

I'll look at level changers at the Pi end.

Thanks again.

hippy
Posts: 6743
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 5:41 pm

GTR2Fan wrote:Shoving 5V into a 3V3 logic pin via a resistor and hoping that the internal protection diode will save you is an awful way of doing it.
Indeed, "hope" is not a good approach to circuit design. But if something is designed in a manner which is compliant with the electrical specification of the device that is something else.

In the absence of full electrical specs for the Pi's GPIO it would be recommended not to do it.

User avatar
GTR2Fan
Posts: 1601
Joined: Sun Feb 23, 2014 9:20 pm
Location: South East UK

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 5:49 pm

hippy wrote:In the absence of full electrical specs for the Pi's GPIO it would be recommended not to do it.
Either with or without full electrical specs, it should never be done, ever! That's one of the basic golden rules that any qualified EE should already be fully aware of. I'd have a hard time trusting any of their circuit designs if they weren't aware of the basics.

This is why some of the old guard EEs around here (including myself) are getting so wound up about the urgency with which some people are practically begging for full electrical specs. Design your circuits properly based around the already well documented requirements for conventional 3V3 logic and they're totally unnecessary.
Pi2B Mini-PC/Media Centre: ARM=1GHz (+3), Core=500MHz, v3d=500MHz, h264=333MHz, RAM=DDR2-1200 (+6/+4/+4+schmoo). Sandisk Ultra HC-I 32GB microSD card on '50=100' OCed slot (42MB/s read) running Raspbian/KODI16, Seagate 3.5" 1.5TB HDD mass storage.

hippy
Posts: 6743
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 6:56 pm

GTR2Fan wrote:
hippy wrote:In the absence of full electrical specs for the Pi's GPIO it would be recommended not to do it.
Either with or without full electrical specs, it should never be done, ever!
Why should something within electrical specs not be done ?

For example; if a product specifies that 20mA can be safely injected into an I/O pin and through clamping diodes then what is the problem with injecting 10mA ?

User avatar
GTR2Fan
Posts: 1601
Joined: Sun Feb 23, 2014 9:20 pm
Location: South East UK

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 7:39 pm

hippy wrote:Why should something within electrical specs not be done ?
As I said, diodes designed to be protection diodes often have strange characteristics, so they can't always be expected to behave exactly like a normal diode under all conditions once you push them in to conduction. Bear in mind that diodes also exhibit capacitance that varies with current, so you really wouldn't want to be forcing one in to conduction if there was any fast or clocked data involved anyway.

Remember that we're talking about microscopically tiny diodes here that are invisible to the naked eye, not something anywhere near the size of, let's say, a 1N4148 signal diode that's typically used in low frequency small signal clamping circuits and can still only sink around 50mA continuously while getting slightly toasty.

The diode is there solely as the device's last line of defence against the outside world. Would you really want to use it for a purpose it was never designed for in the knowledge that, if it fails, it's almost certainly goodbye Pi? Using the internal diodes in this way is creating a highly non-ideal solution to a problem that will never exist in the first place if something is designed by a competent design engineer.
For example; if a product specifies that 20mA can be safely injected into an I/O pin and through clamping diodes then what is the problem with injecting 10mA ?
It is just possible that a similar figure could be specced, but it would almost certainly come under the "absolute maximum safe limits" section, ie, the stuff you want to be staying well clear of at all times in any good circuit design.

Whatever, it certainly won't be advocating ever deliberately pushing the internal protection diodes in to conduction as I've never seen this advocated on any datasheet for anything in the past 35 years. For perfectly sound reasons, it just isn't done.
Last edited by GTR2Fan on Wed Mar 09, 2016 8:27 pm, edited 1 time in total.
Pi2B Mini-PC/Media Centre: ARM=1GHz (+3), Core=500MHz, v3d=500MHz, h264=333MHz, RAM=DDR2-1200 (+6/+4/+4+schmoo). Sandisk Ultra HC-I 32GB microSD card on '50=100' OCed slot (42MB/s read) running Raspbian/KODI16, Seagate 3.5" 1.5TB HDD mass storage.

hippy
Posts: 6743
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 8:24 pm

I suppose it does come down to what one considers "good circuit design" or acceptable. Much like some would never go parachuting because there is risk intrinsic in that while others are happy to take the chance.

Microchip were at one time happy to have an application note showing how to connect I/O pins of their PIC devices direct to mains via suitable resistors but have backed off that as chip technology has changed, but they haven't convinced everyone it's not acceptable to keep doing that.

For interfacing it probably greatly depends on which side of the fence one sits; the don't do it unless it's guaranteed safe to do, or, it's okay to do it unless proven to be detrimental.

User avatar
GTR2Fan
Posts: 1601
Joined: Sun Feb 23, 2014 9:20 pm
Location: South East UK

Re: I2C and 5V - Definitive answer?

Wed Mar 09, 2016 8:48 pm

hippy wrote:I suppose it does come down to what one considers "good circuit design" or acceptable. Much like some would never go parachuting because there is risk intrinsic in that while others are happy to take the chance.
I'd like to be able to agree with you, but good circuit design always means doing nothing that a datasheet hasn't explicitly told you is safe to do. I just stick by their rules because they know their device better than I ever will by guessing.
Microchip were at one time happy to have an application note showing how to connect I/O pins of their PIC devices direct to mains via suitable resistors but have backed off that as chip technology has changed, but they haven't convinced everyone it's not acceptable to keep doing that.
You've actually jogged my memory into recalling the only time I can remember a manufacturer ever recommending this. If it's the device I'm thinking of (something to do with zero-crossing detection?), I believe they recommended a 5Meg resistor in series limiting the current to around 20uA, so that doesn't come as any great surprise to me really. A few tens of microamps is unlikely to do any harm, but I still wouldn't be doing it without a manufacturer's guarantee that it was entirely safe to do so permanently.

It's worth remembering that latch-up of any I/O pin on the Pi due to an internal protection diode failure can also lead to a catastrophic cascade effect that wipes out the entire SOC, so it's certainly not a risk I'll be taking even if it's sanctioned. I wouldn't mind losing a cheap PIC over it, but not an entire Pi. I would have mentioned this earlier, but it's difficult to judge the technical level of members without knowing them. I'm pretty sure that this particular point is one you'll understand having dealt with PICs and the like, so hopefully you won't dismiss this as utter nonsense.

I wonder if it was the smaller fab size since then that's stopped Microchip recommending it. Internal protection diodes are scarily miniscule at 40nm and below. I don't know what size fab 4000 series CMOS logic was on originally (something huge I'd imagine), but the protection diodes in those were so incredibly fragile that they couldn't withstand any physical handling at all out-of-circuit if you wanted to guarantee not dramatically shortening their lifetime. I worked with those for nearly 20 years.
For interfacing it probably greatly depends on which side of the fence one sits; the don't do it unless it's guaranteed safe to do, or, it's okay to do it unless proven to be detrimental.
Yes, and the designer whose designs will work the most reliably and for the longest will be the one who sticks by the letter of whatever datasheet the manufacturer makes available without making any assumptions at all. I call that competent designing. Going back to the topic title, I think the answer is: Assume that you can't until you're told with no uncertainty by the manufacturer that you definitely can. ;)
Pi2B Mini-PC/Media Centre: ARM=1GHz (+3), Core=500MHz, v3d=500MHz, h264=333MHz, RAM=DDR2-1200 (+6/+4/+4+schmoo). Sandisk Ultra HC-I 32GB microSD card on '50=100' OCed slot (42MB/s read) running Raspbian/KODI16, Seagate 3.5" 1.5TB HDD mass storage.

geoffr
Posts: 89
Joined: Wed Aug 22, 2012 11:25 am
Location: Melbourne, VIC

Re: I2C and 5V - Definitive answer?

Thu Mar 10, 2016 4:40 am

To clarify a couple of points:
  • You don't need external pull-ups on the 3.3v I2C bus - the Pi already has pull-ups on the I2C
  • It is probably worth looking at using logic level converters to get your I2C to 5V, especially if the I2C devices you want to use are designed to run on 5V. The only case where I have been running a I2C device at 3.3V is where I am using a DS3231 RTC, and I am interfacing the INT pin on the RTC back to the RUN pin on the Broadcom SOC, and everything has to run at a 3.3V level.
You can build a 3.3V to 5V logic level converter for I2C quite easily using 2 MOSFETs and a few resistors.

hippy
Posts: 6743
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: I2C and 5V - Definitive answer?

Thu Mar 10, 2016 1:01 pm

GTR2Fan wrote:It's worth remembering that latch-up of any I/O pin on the Pi due to an internal protection diode failure can also lead to a catastrophic cascade effect that wipes out the entire SOC, so it's certainly not a risk I'll be taking even if it's sanctioned. I wouldn't mind losing a cheap PIC over it, but not an entire Pi. I would have mentioned this earlier, but it's difficult to judge the technical level of members without knowing them. I'm pretty sure that this particular point is one you'll understand having dealt with PICs and the like, so hopefully you won't dismiss this as utter nonsense.
I absolutely wouldn't dismiss it as nonsense, it's a perfectly sensible 'zero risk' approach. But others are prepared to take greater risks, especially when it appears to be minimal risk. It's not something to be done with a commercial product, certainly not anything safety critical, even mission or business critical, but can be acceptable outside those.

I wouldn't risk a $35 Pi either, but I am more comfortable risking a $5 Zero on a non mission critical, throw-away, project. Especially if that minimal risk comes out better against the cost of having done it properly. I would always advocate that a risk assessment is done.

User avatar
GTR2Fan
Posts: 1601
Joined: Sun Feb 23, 2014 9:20 pm
Location: South East UK

Re: I2C and 5V - Definitive answer?

Thu Mar 10, 2016 1:49 pm

hippy wrote:I absolutely wouldn't dismiss it as nonsense, it's a perfectly sensible 'zero risk' approach. But others are prepared to take greater risks, especially when it appears to be minimal risk. It's not something to be done with a commercial product, certainly not anything safety critical, even mission or business critical, but can be acceptable outside those.
I absolutely agree here, except that it shouldn't ever be promoted on public forums that anything other than a 'zero risk' approach is an acceptable way of doing things, because those less knowledgeable as to why it shouldn't be done will continue spreading the word that it's absolutely fine to be doing it (as happened with Microchip ;)). I've always called this kind of situation a CWOD, a Chinese Whisper Of Destruction. :D
I wouldn't risk a $35 Pi either, but I am more comfortable risking a $5 Zero on a non mission critical, throw-away, project. Especially if that minimal risk comes out better against the cost of having done it properly. I would always advocate that a risk assessment is done.
Applied to our own hardware, that makes perfect sense as we know the risks involved. I just wouldn't suggest to anyone else that they do something that contained any element of risk whatsoever to their hardware as they may not have the knowledge required to assess for themselves whether or not an element of risk considered acceptable by someone else is acceptable to them. Zero risk is acceptable to everyone, or at least should be if it can be logically proven.

If we're lucky enough to be told the forward voltage drop of the internal protection diodes and it's absolutely guaranteed to be higher than that of a readily available small-signal Schottky diode, I may even start suggesting placing an external Schottky strapped from the relevant I/O pin to the appropriate rail so that the internal diode is never pushed in to conduction and isn't being relied upon in circuits where it can't be guaranteed that an input signal won't ever go out of spec.

It's clearly a precaution that isn't necessary at all if we know that the input signal is always going to be within already known design voltage constraints for the I/O pins, but it's not always possible to guarantee this on someone else's workbench when we can't see exactly what they're doing and have no direct control over it. Amateurs are bound to make silly mistakes from time to time, and I consider it our job here to minimise the risk of potential damage when these mistakes are made. :)
Pi2B Mini-PC/Media Centre: ARM=1GHz (+3), Core=500MHz, v3d=500MHz, h264=333MHz, RAM=DDR2-1200 (+6/+4/+4+schmoo). Sandisk Ultra HC-I 32GB microSD card on '50=100' OCed slot (42MB/s read) running Raspbian/KODI16, Seagate 3.5" 1.5TB HDD mass storage.

hippy
Posts: 6743
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: I2C and 5V - Definitive answer?

Thu Mar 10, 2016 6:14 pm

I am extremely grateful for your taking the time to explain your position and you've explained it to me clearer than anyone else has ever managed to. I have to say I think you are right and make good and fair points.

So now I'm going to have think how my willingness to embrace 'informed recklessness' fits in with that. I'm not adverse to that but do agree that we should not encourage anyone to do anything where the potentially adverse consequences are not understood.

User avatar
rpdom
Posts: 16110
Joined: Sun May 06, 2012 5:17 am
Location: Chelmsford, Essex, UK

Re: I2C and 5V - Definitive answer?

Thu Mar 10, 2016 6:42 pm

hippy wrote:So now I'm going to have think how my willingness to embrace 'informed recklessness' fits in with that. I'm not adverse to that but do agree that we should not encourage anyone to do anything where the potentially adverse consequences are not understood.
Well said Sir :)

User avatar
dentex
Posts: 180
Joined: Sun May 20, 2012 7:58 am
Location: Italy
Contact: Website

Re: I2C and 5V - Definitive answer?

Fri Sep 09, 2016 7:58 pm

Hello!
I'm reviving this post because I was searching for a "Definitive answer" on this matter, too.

I'm sorry to say that probably the thread in the end went more on the "good circuit design" topic; as a result, it's not perfectly clear to me if it's safe or not to use a I2C device, with the RPi, and power it from 5V.

Searching the Internet for answers, I found this resource:
https://www.element14.com/community/mes ... -i2c#59441
...where it seems that "it depends on the device".

Also Massi's answer seems to go in the same direction:
Massi wrote:well, powering a sensor with 5V does not imply that it will output 5V
this is often true for a lot of sensor (think to all the open collector outputs), also (and probably more) for i2c
i2c bus is "terminated" by a pullup on sda and scl
this termination commonly is done by the master (the pi, in our case)
in fact, pi i2c-1 bus has phisical 1.8k pullups.

I this is the case, a sensor powered with 5V will not output 5V on the i2c bus (since it only pulls down lines when needed) and should not destroy the pi

BUT

there's another problem. Sensor side, this time. Probably the sensor will accept as a valid logic HIGH level, a voltage related to the power supply (something like 0.8 * Vcc)
If this is the case, the Pi 3,3 V isn't enough to reach this voltage and so you can't be sure of the proper working of sensor.
[...]
What I was trying to understand, it's if is safe to use one of those LCD with I2C backpack with no level shifting:
http://www.sainsmart.com/sainsmart-iic- ... no-r3.html
(actually it's a Chinese version, but it's really identical in every aspect, at least looking at it).

If the answer would be "yes, it's possible to use the I2C-LCD without level shifting", would it be an additional problem to use, i.e., an RTC clock linked to same I2C bus and powering via the 3v3 pin?

[Disclaimer: I've been having such a configuration, but with a 16x2 I2C LCD, on my PiAC project (see signature) for more than a year... no problem so far; but I understand I might have been just lucky. Unfortunately, only now I'm realizing there are possible drawbacks]

Thanks for any answer.
One of my projects with the Raspberry Pi:
--------------------------------------------------------
(Raspberry) Pi Aquarium Controller: https://github.com/dentex/piac
...and its dashboard: https://freeboard.io/board/O17ACD

User avatar
joan
Posts: 14686
Joined: Thu Jul 05, 2012 5:09 pm
Location: UK

Re: I2C and 5V - Definitive answer?

Fri Sep 09, 2016 8:17 pm

Power the backpack from a Pi 5V and ground pin.

Then measure the voltage on the backpack's SDA and SCL pins.

If it's 5V then nobody is going to assure you it is safe.

User avatar
dentex
Posts: 180
Joined: Sun May 20, 2012 7:58 am
Location: Italy
Contact: Website

Re: I2C and 5V - Definitive answer?

Fri Sep 09, 2016 8:20 pm

joan wrote:Power the backpack from a Pi 5V and ground pin.

Then measure the voltage on the backpack's SDA and SCL pins.

If it's 5V then nobody is going to assure you it is safe.
Thanks for this "practical" answer. Will be the 1st thing I'll do tomorrow when I'll go back home.
One of my projects with the Raspberry Pi:
--------------------------------------------------------
(Raspberry) Pi Aquarium Controller: https://github.com/dentex/piac
...and its dashboard: https://freeboard.io/board/O17ACD

User avatar
dentex
Posts: 180
Joined: Sun May 20, 2012 7:58 am
Location: Italy
Contact: Website

Re: I2C and 5V - Definitive answer?

Sat Sep 10, 2016 8:34 am

joan wrote:Power the backpack from a Pi 5V and ground pin.

Then measure the voltage on the backpack's SDA and SCL pins.

If it's 5V then nobody is going to assure you it is safe.
It's zero. :shock:
One of my projects with the Raspberry Pi:
--------------------------------------------------------
(Raspberry) Pi Aquarium Controller: https://github.com/dentex/piac
...and its dashboard: https://freeboard.io/board/O17ACD

User avatar
joan
Posts: 14686
Joined: Thu Jul 05, 2012 5:09 pm
Location: UK

Re: I2C and 5V - Definitive answer?

Sat Sep 10, 2016 8:43 am

dentex wrote: ...
It's zero. :shock:
When measured between Pi ground and backpack SDA/SCL when powered from Pi ground and Pi 5V?

That's fortunate. Apparently it's not like many of the backpacks which have their own pull-ups to VCC.

That should be fine to use on the Pi.

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