Detect whether a pin is connected to GND

Question

Hi everyone!

Recently I've started working on my own IoT project. An Airco that has no connectivity to the world (except for a remote, that I never use), to connect it to the internet.

Note: The Arduino Uno is connected with an NodeMCU ESP-12F via serial. The Arduino Uno communicates with the AC circuitry.

However, I'm facing a serious problem that stops me from continuing with the project.

The AC has got three modes, Cool, Dry and Fan. Microprocessor controls the specialised equipment inside the AC (such as the refrig., the pump etc) so there is no line going to the other control board that is high when one of the three modes is active.

The MIP (microprocessor) controls three leds to let the user know which mode is active. (here comes the, for me, very difficult part)

The GND connections of the LED's are connected to the MIP (pulls, I guess) and they are being connected to GND when a mode is active.

The challenge for me, with my Arduino Uno, is to check which GND connection of the three LED's is connected to the GND line of the external circuitry.

I've made a beginning, for each line I've used this setup:

My selfmade and NOT complete schematic of the AC circuit

The values I get are usable, but they require taking loads of samples, and taking the average of all the samples, to check which line has got the lowest value (and thus is connected to GND). However, the values are fluctuating very much, and are not very accurate.

As required, the GND of the Arduino is connected to the GND of the AC circuitry.

If you want to see a few pictures of the AC circuitry, check out my Google Drive here.

The project is a hobby project, next to my school projects, so there is no deadline or something

Thank you in advance for your help,

Tim

Message was edited by: Tim Koers Added the AC circuitry schematic

Added the AC circuitry schematic

jc2048 · Answer

If you're going to do all that, I think it would be worth your while sitting down with large sheet of paper and drawing out the circuit. It should be easy to do - you can see all the tracks, and the conventional components are simple to recognise. It will make it much more straightforward to plan your modifications. You'll also get more answers here because many of us can read schematics quickly but don't have the time (or inclination) to trace out a circuit from photographs.

jc2048 · Answer

Are you sure the drive is direct. [It's difficult to make out from your photographs because of the large blob of silicone over the capacitor that they've added.] The processor (assuming it is a processor) is controlling the two 7-segment displays, multiple switches, and multiple LEDs and it's possible that the indicator LEDs you are interested in are combined with the rest of the multiplexing. If they are, the readings you get will be very confusing. If it is multiplexed, you should still be able to come up with a way to read it with an Arduino. It would just be a matter of reading at the right time, triggered by one of the multiplex signals. In fact it wouldn't be too difficult to read the whole lot, including the value on the display and the switch states, if you wanted to. Add some analogue switches across the ones on the board and you could even control it. Hold off on the filter until you know more about the circuit. One of your caps is the wrong way round. And if the drive is switched you have a differentiator that will generate negative going pulses (though the static protection diodes will probably save you from damaging the Arduino input).

Jan Cumps · Answer

It is possible that the circuit that drives the LEDs is of an 'open drain' type. in that case, the output floats when the LED pin isn't pulled to ground, and you can't make a useful measurement. can you remove your added circuitry, and check the voltage of all 3 LED kathodes when they are turned off, with a volt meter? Does that give you a value you can work with? Is it stable?

jc2048 · Answer

Good start on the schematic.

It's more usual to use R as an identifier for a resistor - L usually indicates an inductor (coil).

The HC164 has CMOS logic levels at the inputs, the HCT164 has TTL logic levels. It would be good to have the HC part on the schematic if you can find one (I don't know Eagle, so I can't help you locate one - there's a fairly active Eagle contingent here, but you'll probably need to post questions in their area of the site to get their attention).

Are you sure that the CLR input connects to GND? That would hold the output in the clear state permanently (the clear input is active low and, from memory - though it's a long, long time ago - I think it overrides everything else)?

Does the QA output really connect directly to the e segment? Are you sure that there isn't a resistor along the way? The way you've got it, there wouldn't be any current limiting for the LEDs that form the segment.

rsc · Answer

Here's another option. Depending on the available drive current of the micro, you could replace the LEDs with optocouplers (or add them in parallel to the LEDs) That would give you clean signals without disturbing the control micro. Scott

Jan Cumps · Answer

Depending how fast your photo resistor is, it may show the same variances as the LED. It's our eye that makes the led look stable, but an electronic sensor may very well be faster and follow the fluctuations of the signal going into the led.

Jan Cumps · Answer

Tim Koers wrote: Do you mean de-solder the wires? The wires are jumper cables ... Don't desolder them please. Measure with everything in place. If you have an oscilloscope at school, it might be good to probe these points too. That will immediately show if there's multiplexing involved with the LEDs.

tim687 · Answer

I wasn't even aware of the L and R identifiers. I searched for resistor in the Eagle Library, and since some parts are very hard to find, I used the first one that I saw (the one with the L as the identifier) I'll edit that when I get back working on it at the end of this weer or something.

I'll check the connection of CLR to GND to make sure that it's connected as it's supposed to. Furthermore, the segment display is common annode, so the shift register should output a GND signal to make a letter visisble. This means that there should be current limiting resistors on the annode lines of the display. (and as far as I can remember, they are there)

rsc The project is a hobby project, and I don't have them laying around, so I don't think that that is a good option. However, I'll keep the optocoupler in mind when designing a project in the near future!

Thanks for your input though!

Tim

Detect whether a pin is connected to GND

Top Replies