Offline
Newbie
Karma: 1
Posts: 28
|
 |
« on: October 06, 2012, 10:11:13 am » |
Hi all, I have 4 Arduino linked together, sometime one of this can be power off but on its gpio can be a voltage that come from other Arduino. This can damage the chip. So I must to isolate all the digital input with optocoupler, but i don't know what is the best solution for the analog input, can you help me?
|
|
|
|
|
Logged
|
|
|
|
|
Massachusetts, USA
Offline
Tesla Member
Karma: 98
Posts: 6389
|
 |
« Reply #1 on: October 06, 2012, 02:52:49 pm » |
Why do yo have four Arduinos linked together? How are they linked together?
|
|
|
|
|
Logged
|
|
|
|
|
|
|
Offline
Edison Member
Karma: 114
Posts: 2205
|
 |
« Reply #3 on: October 06, 2012, 04:03:31 pm » |
This can damage the chip. It can, but that doesn't mean that it will. So I must to isolate all the digital input with optocoupler, but i don't know what is the best solution for the analog input, You don't have to isolate. But if you do isolate, the best solution will depend on your application, your parts availability, your tolerance for complexity, your bank account, etc. can you help me? Sure but not enough to go on.
|
|
|
|
|
Logged
|
|
|
|
|
Offline
Edison Member
Karma: 3
Posts: 1713
|
 |
« Reply #4 on: October 06, 2012, 04:33:23 pm » |
Like they said above, you my be able to get away with something simple like a 1k resistor protecing the pin insted of optocoupling
|
|
|
|
|
Logged
|
|
|
|
|
Left Coast, CA (USA)
Offline
Brattain Member
Karma: 279
Posts: 15340
Measurement changes behavior
|
 |
« Reply #5 on: October 06, 2012, 07:01:57 pm » |
Yes, I would give simple series resistors a shot. However I'm not sure 1K is the right value, you want to prevent current in excess of what the clamping protection diodes are rated for. I suspect 10K or more is more suitable, maybe more? The datasheet should give clamping diode ratings somewhere I would think?
Lefty
|
|
|
|
|
Logged
|
|
|
|
|
Offline
Edison Member
Karma: 114
Posts: 2205
|
 |
« Reply #6 on: October 06, 2012, 07:13:05 pm » |
Those diodes are rated 20ma. It would take over 200v to kill those diodes + 10k resistor.
Not to mention the low-pass filtered you got there.
|
|
|
|
|
Logged
|
|
|
|
|
0
Offline
Tesla Member
Karma: 73
Posts: 6638
Arduino rocks
|
 |
« Reply #7 on: October 06, 2012, 09:05:00 pm » |
The protection diodes are rated at 0.5V max forward bias - that's rather less that 20mA I think - or have I missed the right part of the specs?
Actually specifying the voltage is a rather poor way of doing this as forward voltage varies significantly across the operating temperature range.
|
|
|
|
|
Logged
|
|
|
|
|
Anaheim CA.
Offline
Edison Member
Karma: 31
Posts: 2314
Experienced old Whitebeard with a Full head of Hair...
|
 |
« Reply #8 on: October 06, 2012, 09:17:24 pm » |
Yeah they make good thermometers, easy to calibrate but do require a 2 point cal for <5% accuracy.
Bob
|
|
|
|
|
Logged
|
“The solution of every problem is another problem.” -Johann Wolfgang von Goethe
|
|
|
|
Offline
Edison Member
Karma: 114
Posts: 2205
|
 |
« Reply #9 on: October 06, 2012, 09:22:08 pm » |
as forward voltage varies significantly across the operating temperature range. -2mv/c, pretty much. How many degrees of temperature do you consider "significant"?
|
|
|
|
|
Logged
|
|
|
|
|
Left Coast, CA (USA)
Offline
Brattain Member
Karma: 279
Posts: 15340
Measurement changes behavior
|
 |
« Reply #10 on: October 07, 2012, 12:27:00 am » |
The protection diodes are rated at 0.5V max forward bias - that's rather less that 20mA I think - or have I missed the right part of the specs?
Vf and Imax are two different specification for the internal clamping diodes. The first tells when it starts to conduct, the second how much current can conduct before damage.
Actually specifying the voltage is a rather poor way of doing this as forward voltage varies significantly across the operating temperature range.
Yea, the Vf spec is of just passing interest, we know that a 5 volt signal will fully turn on the positive clamping protection diode on a input digital pin on a 3.3 volt powered board, what we want to know is how much continous current the diode can withstand, that is the datasheet spec of interest.
Lefty
|
|
|
|
|
Logged
|
|
|
|
|
Offline
Newbie
Karma: 1
Posts: 28
|
 |
« Reply #11 on: October 07, 2012, 02:32:58 am » |
So can I protect the analog input only with something such as 10K series resistor? If true is a very good info. In alternative i think to use analog switch but i need something that when is power off go in hight impedance. Do you know some ic like this?
|
|
|
|
|
Logged
|
|
|
|
|
0
Offline
Tesla Member
Karma: 73
Posts: 6638
Arduino rocks
|
 |
« Reply #12 on: October 07, 2012, 07:02:55 am » |
Many things can be protected with a large resistor - so long as no appreciable current flow is needed. A digital input in normal operation takes < 1uA so even a 1M resistor would allow it to function (but much slower than normal - 10k is more reasonable a compromise).
|
|
|
|
|
Logged
|
|
|
|
|
Offline
Newbie
Karma: 1
Posts: 28
|
 |
« Reply #13 on: October 07, 2012, 08:02:36 am » |
And is also ok for and analog input?
|
|
|
|
|
Logged
|
|
|
|
|
Offline
Edison Member
Karma: 3
Posts: 1713
|
 |
« Reply #14 on: October 07, 2012, 09:37:49 am » |
For an analog input don't go any higher than a 10k resistor but it should work to protect it the same, if you go higher it will negatively affect how fast you can do an analog read, in the datasheet it says best with source impedance <10k ohms
|
|
|
|
|
Logged
|
|
|
|
|
|