ADC Chanel not working?

[reaperarch]

hi all,

im working on an accelerometer project, but when i read the voltage on the y pin with a multimeter i get 1.6 (half supply voltage). when i read it with adc10
i always get 8.(other channels are fine).

is it possible that by soldering it i damaged the picaxe?
has anyone had similar experiences?

thanks.

 

[BeanieBots]

Need more data from you.
Which PICAXE?
What is the ACTUAL voltage on each input pin?
What is the ACTUAL returned reading from the ADC?
What is your power rail voltage?
Good idea to post your code as well.

 

[reaperarch]

hi, thanks for the swift reply.
picaxe is 08m
voltage varies with acceleration (0v - 3.3v) from -1.5g to +1.5g
the adc on x returns about 500 (level)
z returns about 800 (1g)
y returns 8

power rail is at 3.3v
and code is just the standard readadc10 pin,wordvar

thanks

 

[BeanieBots]

Yes, but what is the voltage on the ADC pin when you get those values?
Does it all work ok if for example you use a POT as a voltage source?
Also, is it a very old 08M?
What is the firmware version?

 

[reaperarch]

the voltage is about 1.5v on the y pin, so adc should be about 500
i can get it up to about 11 on the adc with a pot, so lets say the adc channel as a greatly reduced range.
its a fairly new axe. got it 2 days ago.

 

[BeanieBots]

OK, somethings wrong there!
Please confirm the firmware version anyway, just in case.

Disconnect the "non working" input. (which one is it?).
Write a simple loop which tests it and use debug to monitor.
Connect to Vcc via a 10k resistor.
Then connect to 0v via 10k resistor.
What are the two readings?

 

[reaperarch]

firmware is 9.2
0v = 8
3.3v = 12

 

[BeanieBots]

V9.2 is OK, so not that issue
Sounds like it's connected to something it shouldn't be.
However, not to 0v or Vcc but maybe a 'low' output.
Which physical pin are you using?
Is there any "history" such as miss-connection to sometime in the past?

 

[reaperarch]

i picked through the entire circuit with a cont. tester. all was ok.
my question is really can the heat generated by a soldering iron damage a pic?
(obviously using the soldering iron on the legs, not intentionally trying to damage it)

 

[hippy]

One trick to make sure that a pin isn't inadvertently connected to something it shouldn't be is to take it out the socket, bend its leg at right angles and plug it back in. It'll be waving its little leggy in the air disconnected from everything. Take care when bending the pins and if you go this route be prepared for the leg breaking off.

A safer trick which avoids leg bending is to use an intermediate socket with the leg of that pulled out but there can be problems with contact between the sockets.

Before anything else though - Check you're reading the right pin and haven't confused leg with pin. Check the input goes to the correct leg / pin.

You can damage chips by overheating them but in my experience you have to do really well to do damage. At this point in time I'd suspect something other than over-heating. If you have a spare 08M drop that in its place as it won't have been heated at all ( I presume you are using sockets ).

 

[reaperarch]

lol, unfortunately not.
its boxed in by the accel. at the top and a transmitter on the side.
when im feeling strong, ill desolder everything and replace the chip.

thanks for all the advice

 

[Dippy]

As he asked about soldering damaging it then I presume he didn't use a socket. Oh dear.

Can you post your schematic and code please.

I wrecked an ADC input on a dsPIC once. Rest of chip OK, so I'm not saying it's impossible. But, equally, I'm not saying definitely that this has happened with your application.
One other thing I am saying, however, is that when prototyping/experimenting use a socket.

PS. And as you have now discovered, it's best to test it on the bench.. in pieces

 

[BeanieBots]

Shoving 1/2A into an input can certainly "weld" it low. Damage caused by solder heat more often than not 'melts' the entire substraight and hence kills the entire chip.
The values returned suggest a short to neigbouring pin which is configured as ouput and held low.
Of course, knowing the actual leg used and the actual code used would confirm/dismiss all of that speculation instantly. Along with any leg/pin confusion. (or even word/byte errors).

 

[Technical]

Are you try to use adc on input 3....? (doesn't exist!)

 

[nbw]

That's a good point. On a less serious note, ADC Chanel Number 3 on an 08M might smell like smoke, rather than the more pleasant Chanel Number 5 on the x2 parts................................

 

[Dippy]

A VERY good point. And I'm surprised it wasn't noticed during pre-assembly tests.

But without code sample and a schematic we'll never know... RIP.