EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: t1d on August 09, 2021, 06:34:31 am
-
I Bbd'd this project: http://embedded-lab.com/blog/making-a-mini-led-christmas-tree/. (http://embedded-lab.com/blog/making-a-mini-led-christmas-tree/.) Five pins on a PIC12F683 driving 20 individual LEDs = kewl.
I verified all connections on the board, pin-to-pin. All good.
As a matter of course, I tested for a dead short across Vsupply and Ground with my Siglent SDM3065X in the continuity function. I got an odd glow across the LED grid and it appeared that the circuit was trying to operate from just the amperage provided by the DMM. It was changing the glow patterns.
I changed the DMM setting to Ohms and read across Vsupply and Ground. There was an extremely high resistance, so I knew there wasn't a short and that the circuit must have been trying to fire up. Wow! Something to know/remember. Maybe you have a good idea of what was happening?
So, I moved on. Full power on. It worked! No smoke. Woot!
Vsupply was 5 volts. As the patterns changed, the current draw on the PSU was very low... I did not watch it forever, but the high I remember was only 0.05A.
I don't like to code, so I was just working with the HEX. If the coder used PWM, maybe 0.05 was reasonable. Hmm... Only 4 LEDs are on at any one moment and if there is a duty cycle that scales down the power... The LEDs were bright, but not extremely bright.
I think I will power it up, again, and confirm the current draw. I think that I will investigate the current output on the DMM in continuity mode, too.
I see CharliePlexing being used to do some very powerful things, especially when PWM and RGB LEDs are thrown into the mix.
Here is the DropBox link to a short video of the blinking (just white) lights:
https://www.dropbox.com/s/8tjn3tajvdt5y1f/IMG_0074.MOV?dl=0 (https://www.dropbox.com/s/8tjn3tajvdt5y1f/IMG_0074.MOV?dl=0)
Thank you for your help.
-
PICs from that era could typically operate from as low as 2.0V. The project in question has the PIC12F683 internal brownout reset (BOR) circuit enabled which has a threshold in the range 2.0V to 2.2V (Datasheet param. 35). At 2.2V white LEDs with a typical Vf of around 3V would draw very little current, and the PIC itself with its oscillator configured for 4MHz would only draw a few hundred uA. (Datasheet param D016: typ. 340uA, max. 450uA).
It therefore seems likely that your multimeter outputs >2V in continuity test mode, but without much current. I would certainly be cautious using it on low voltage LSI logic circuits and absolutely would not use its continuity mode on a board with any sort of battery backup (or otherwise powered) as it appears that it could turn on gates and forward bias semiconductor junctions.
0.05A is rather odd - charlieplexing *NORMALLY* only turns on one LED at a time and 8 bit PICs of that era are typically only rated to source or sink 25mA from any I/O pin (and in many cases can't even reach that due to the I/O's relatively high output resistance) They generally do a bit better sinking than sourcing so its *POSSIBLE* its driving two LEDs at at time, overloading the cathode side I/O pin. However is it possible that you failed to accurately measure and record the circuit's current consumption?
-
Thank you, Ian, for your reply.
I would certainly be cautious using it on low voltage LSI logic circuits and absolutely would not use its continuity mode on a board with any sort of battery backup (or otherwise powered) as it appears that it could turn on gates and forward bias semiconductor junctions.
Good info!
However is it possible that you failed to accurately measure and record the circuit's current consumption?
I did run the circuit again and closely watched the PSU's (Siglent SPD3303X-E) current display. The draw was even less than I thought; just 0.01A, when all the LEDs are illuminated. Of course, a current reading from a PSU on this type of circuit is fraught with inaccuracies.
-
Under 20mA would be what I would expect for a single direct charlieplex on a PIC so 0.01A on the PSU's meter would be reasonable.
-
Note that modern LEDs are extremely efficient, and visible light can be produced with only a few uA of current.