Vin affecting analogRead(sensorPin)?

LeafLabs Garden » Support » Maple Hardware Support

Vin affecting analogRead(sensorPin)?

(6 posts) (2 voices)

Started 3 years ago by pico
Latest reply from pico

pico
Member

Strange problem, but perhaps this is a known and understood phenomenon.

After a recent rush of blood to the head, I purchased a Maple 5 clone (Iteadmaple v1.0) from Iteadstudio (which is one of my regular haunts for Arduino compatible hardware).

I've been playing with the board using the maple-ide v0.0.12, and everything is largely working fine. I'm reading a thermistor value on analog pin 0 (pin 15) using the standard voltage divider approach: a 10K resistor from the 3.3v pin to pin 15, and then connecting the 10K/25C thermistor between pin 15 and ground.

Works a treat reading the values when the supply voltage is 7.5V, but starts reading very erratically whhen using a 12V supply. At first I thought it must be a noisy power suppply the boards filtering didn't like, but I've used several power supplies now, and have convinced myself that the higher voltage is itself the problem.

I was reading that the genuine Maple 5 board prefer to be powered at lower supply voltages than 12V, which is their max. rated input, as there is more headroom available for current output from the board that way (I don't really understand why this would be so, but that's what I believe I read.)

One thing about the Iteadmaple clone was that is supposed to have a greater tolerance for supply voltage input, up to 35V as I recall. So I was surprised to see the effect. Basically, the sensor pin voltage reading seems to be jumping all over the place when using a 12V, but settles down to a very solid and reliable reading when using a 7.5V supply.

Using a 7.5V adaptor (I have a few, and it works robustly with all of them) is a reasonable practical work-around, but it has me scratching my head as to "why?". Is this an issue the clone inherited from the reference design, or is this something peculiar to this board?

Any insights into what might be going on is appreciated.

Posted 3 years ago #
mikep
Member

I didn't see a schematic for the ITeadMaple, so I am guessing here a little bit;
- it looks like the board has a switching power supply (unlike the Maple) so that allows higher voltages. But it could also create some "noise" due to the switching frequency.
- when powered at 7.5V the power supply works at a different frequency than at 12V or at a different on/off ratio, both of these could affect the level of noise generated

I would try this:
- use a small capacitor across your thermistor to filter out any high-frequencies on the analog pin (something like 10 uF should work, it's not critical.)

- I assume you tried a different analog pin already

- also maybe the analog reference voltage pin has some noise on it, maybe a capacitor on there could help, hard to say without the schematic of the board

Hope that helps,
Mike

Posted 3 years ago #
pico
Member

Thanks Mike, all sensible suggestions of things to try, and your hypothesis makes sense, too. I will investigate further and report back.

Posted 3 years ago #
pico
Member

Well, the mystery is still not solved, but I have learnt a few things.

Mikep's post got me wondering about AREF. Now, the genuine Maple 5 does not have an AREF, and neither does the Iteadmaple, or so it would seem (which is unsurprising since this is a limitiation of the chip). But, upon further reading, I see the Maple 5 does have the vdda 3.3 pin where the 5V pin is situated on the Arduino. (This could be better documented for newbies -- I suggest an entry explaing this difference here: http://leaflabs.com/docs/arduino-compatibility.html#header-numbering-and-incompatibilities would be an appropriate documentation improvement.)

Now, the Iteadmaple 1.0 actually supplies 5V on this pin -- making it more compatible with the Arduino, but less so with the Maple 5. Neither is this difference documented anywhere in the docs on the Iteadstudio site as far as I can see. The board silkscreening is actually blank for this header, which doesn't help either! And since, as Mikep also pointed out, the board schematic is a dead link at the moment, it's hard to know the exact nature of whatever other differences remain.

Indeed, it's unclear whether the Iteadmaple 1.0 actually provides a vdda pin. There is a set of header pads roughly where the power supply select header is located on the Maple 5, of which one pad is suggestively labled "vdd 3.3" on the silkscreen. So I used this as a reference to A0/D15 pin rather instead of the 3.3 pin on the power header, which I had been using. However, the symptoms remains unchanged -- with a 7.5V input supply I get stable readings, with a 12V supply, very unstable readings (and I'm only looking at 10 bits of resolution.)

Disappointing. So I'm inclined to think that the Iteadmaple 1.0 doesn't provide a vdda pin after all, which is a significant shortcoming, if true. I may even try to follow up with Iteadstudio to see if they can provide a schematic or otherwise shed light on the issue. But for now, the 7.5V workaround is still my only practical answer. I did try a 0.1uF cap across the 3.3V header pin and ground to try smooth things a bit, but to no discernable effect.

BTW, to correct my earlier post: The rated supply input for the Iteadmaple 1.0 is 6-23V, not 35V as I posted above.

Posted 3 years ago #
mikep
Member

Hi pico,
some good investigative work there.
Interesting that the IteadMaple provides actually 5V on the respective pin, that should make it more compatible with Arduino shields that need 5V (e.g Ethernet shield).
Did you try also try a cap across the thermistor, i.e from the Analog pin to ground? I would definitely try 10uF or so, as close to the analog pin and ground as possible, 0.1uF might not be big enough to filter out the ripple.

Posted 3 years ago #
pico
Member

I emailed Iteadstudio last night regarding the dead link to the schematic, and they've already fixed it (which is a pretty prompt response, I think.) So hopefully that may resolve a few questions.

Looking at the schematic seems to confirm that there is no seperate breakout for a vdda pin -- everything 3v3 looks tied to a common "vdd33" line. I haven't looked at the Maple 5 schematic to compare, though, so I might be missing something.

The power rectification is in two stages Vin -> 5v0 -> 3v3. There is a lot of filtering going from Vin to 5v0, but much less going from 5v0 to 3v3, which is probably reasonable. Not clear to me why increasing Vin from 7v5 to 12v0 would affect anything much at all, looking at the schematic.

I may yet try a larger cap across the thermister, but I'll probably continue just to use the lower voltage supply to keep the ADC under control. More than anything else I'd just like to understand what the underlying problem is! Frustrating.

Posted 3 years ago #

RSS feed for this topic

Reply

You must log in to post.