| Author |
Message |
seyyah
Joined: 07 Oct 2001 Posts: 578 Helped: 1
|
21 Apr 2004 15:57 pic mclr |
|
|
|
|
| This pin is very sensitive. Even i touch this pin with a screwdriver, pic resets. In my circuit i connect this pin to 5V via a 10K resistor. I also connect a 100n capacitor parallel to this pin. But if something happens in ground (for example turning off a relay), pic resets. How can this be possible. When i take out the capacitor there is no problem. Can anyone explain me how mclr pin goes to zero while a capacitor is connected? what can i do to prevent this? Is it possible or suitable to use pic without this capacitor connected to mclr pin. Is it safe? Are other mcus are as sensitive as pic to the effects like this one?
|
|
| Back to top |
|
 |
E-design
Joined: 01 Jun 2002 Posts: 947 Helped: 68
|
21 Apr 2004 16:02 mclr connected 44 pin |
|
|
|
|
The newer pic's like the 16F628A have an option to use the internal power-up timer (all pic's have one according to a uchip note) and thus no need for the external circuitry like in the older 16F84A.
I remember a friend of mine once having a similar problem when switching relays and by putting an additional resistor (few kOhms) in series with the MCLR pin to the cap was able to overcome this problem.
Here is a document about reset and causes which might not address your problem by name, but still good reading
http://www.microchip.com/download/lit/suppdoc/toots/reset.pdf
Last edited by E-design on 21 Apr 2004 16:55; edited 2 times in total |
|
| Back to top |
|
 |
tgq
Joined: 15 Jul 2002 Posts: 305 Helped: 2
|
21 Apr 2004 16:38 pic mclr cap |
|
|
|
|
| maybe you have bad (too close) PCB tracks ?
|
|
| Back to top |
|
 |
hugo
Joined: 01 Jan 1970 Posts: 286 Helped: 27 Location: canada
|
21 Apr 2004 17:22 mclr pic |
|
|
|
|
Hi,
What do you mean by : "I also connect a 100n capacitor parallel to this pin."
There is no need to have a capacitor to MCLR pin!!!
You have to connect 100n capacitors parallel with power supply only!!!
Hope it help!
Bye
Last edited by hugo on 21 Apr 2004 21:33; edited 2 times in total |
|
| Back to top |
|
 |
kcmohr
Joined: 21 Apr 2004 Posts: 2
|
21 Apr 2004 18:59 microchip mclr |
|
|
|
|
| I agree with Hugo. I use a 10k pull up w/ no cap. Set the fuses for PUT (power up timer) and BROWNOUT. Else, you can use a uprocessor supervisor and be sure to NOT USE, PUT and BROWNOUT.
|
|
| Back to top |
|
 |
mrcube_ns
Joined: 10 Apr 2002 Posts: 429 Helped: 11 Location: Dark side of the Moon
|
21 Apr 2004 20:24 sensitive mclr pin |
|
|
|
|
First: Try to put that capacitor (usualy 100n ceramic) as close as you can to supply pins of pic.
Second: if there are sparks when reley go off you have to make distance between rele and mcu (or made some kind of metal rooms in box where all project are mounted).
Regards.
Mr.Cube
|
|
| Back to top |
|
 |
seyyah
Joined: 07 Oct 2001 Posts: 578 Helped: 1
|
21 Apr 2004 21:11 does mclr need pullup? |
|
|
|
|
To hugo: I mean i connected a cap to mclr pin. One of its leg is connected to mclr and the other is to ground. I see this configuration everywhere. This ensures that the pic won't be resetted due to the transients, i think.
But it made the opposite effect which causes a reset. I wonder how this mechanism works?...
I take the ground side leg of the capacitor out and it worked good until i touched with a metal. When i touch this leg (unconnected leg of the cap)with a metal it resets again.
Then i connected this leg to the VSS of pic. I also connected a small cap between vss and vdd and changed 10k resistor with a 4.7k resistor.
And it doesn't reset now.
So do you have any comment about the reason of this event? I rather wonder, what happens when i touch the unconnected pin of the capacitor that causes the voltage level at the mclr pin goes down and results in a reset?
|
|
| Back to top |
|
 |
wessoms
Joined: 15 Jan 2002 Posts: 17
|
21 Apr 2004 23:44 microchip pic mclr pin esd |
|
|
|
|
| The reason why is because the 100nF cap will pass the negative transient spike from the ground to MCLR when the relay changes states. The current thru the 10K resistor from VDD to the MCLR pin is not enough to keep the transient spike from pulling the pin LOW. There for a resest happens. When you increased the current to the MCLR pin by lowering the resistance to 4.7K the problem went away. Hope this answers the question.
|
|
| Back to top |
|
 |
papyaki
Joined: 13 Apr 2002 Posts: 564 Helped: 19 Location: A small village somewhere in Gaul
|
22 Apr 2004 7:58 mclr supervisory circuit |
|
|
|
|
Hi
As said E-Design in his previous post, Microchip recommand to put a serial resistor between the capacitor and MCLR pin. The recommanded value is comprised between 100ohm and 1kohm to avoid ESD/EOS value.
See the diagram page 26 of the following document :
h**p://www.microchip.com/download/lit/pline/picmicro/families/16f8x/35007b.pdf
They also provide app note AN607 :
h**p://www.microchip.com/download/appnote/pic16/00607b.pdf
|
|
| Back to top |
|
 |
btbass
Joined: 20 Jul 2001 Posts: 1187 Helped: 113 Location: Oberon
|
22 Apr 2004 9:25 mclr decoupling capacitor |
|
|
|
|
| Sounds like you have bad decoupling on your supply rail. Have you got a 100nF cap right next to the pic supply pin, and good decoupling on your relay supplies?
|
|
| Back to top |
|
 |
sam_manchali
Joined: 14 Jun 2003 Posts: 105 Helped: 3 Location: sharjah
|
22 Apr 2004 15:19 filters for mclr |
|
|
|
|
hi,
i have also experienced this problem, but i think u need to have a free-wheeling diode across the coil of the relay.
i am quite sure this will work. The pic resets as the coil, when s/w on and off causes a back emf , which disturbs the gnd. u dont need a good pcb or a 10k with cap on th eMCLR pin, i normallyu pull it directly to Vcc.
sam
|
|
| Back to top |
|
 |
lguancho
Joined: 07 Jan 2004 Posts: 266 Helped: 5 Location: Singapore
|
23 Apr 2004 9:23 filtering the mclr pin |
|
|
|
|
| Hi. Did you try to power your pic and relay separately? Also, does your pic device has "BROWNOUT" mechanisim? Maybe you could try to deactivate it.
|
|
| Back to top |
|
 |
Google AdSense

|
23 Apr 2004 9:23 Ads |
|
|
|
|
|
|
| Back to top |
|
 |
simce
Joined: 09 Jan 2004 Posts: 195
|
23 Apr 2004 10:40 caps on mclr should |
|
|
|
|
Here is my experience that solves most of these problems:
1. You must use "star" topology for power distribution across pcb that has MCU and relays.
2. You must use inverse polarized diode connected parallel to relays coils.
3. If contacts of relay passes high current it is very advisable to use snubber circuit parallel to contacts and to put relays far from PCB on which MCU is mounted.
4. Use separate galvanic isolated supply for MCU and relays.
5. Use 100ohm-1k pull-up resistor on MCLR pin.
6. If you made good PCB use 100nF capacitor between MCLR and GND.
7. Use inductors, and pi filters in power supply filters, and use more 100nF capacitors parallel to first filter capacitor after the input diode bridge.....
These are basic precautions that has to be obeyed, there are also more things to take care of, but if you obey this things you have 90% chance to make good MCU driven relay system.
Simce
|
|
| Back to top |
|
 |
seyyah
Joined: 07 Oct 2001 Posts: 578 Helped: 1
|
23 Apr 2004 17:55 microchip mclr resistor capacitor |
|
|
|
|
To simce: i've done all of these except 4 and 7 and it didn't solved the problem.
To all: The only thing solved the problem: When i connect the capacitor's ground pin to pic's ground pin, it didn't reset again.
Relays diodes are connected. I tried to connect serial resistor(220 ohm) to cap of mclr it didn't work. I redesigned the ground to put mcu and relays far away from each other. I cancelled the brownout. I connected more capacitors and small capacitors and tantalum capacitors. I tried all of the your recommendations. May be they have some help but if the cap is connected but its ground isn't connected to pic's ground then there is problem. If cap isn't connected than there is no problem.
|
|
| Back to top |
|
 |
simce
Joined: 09 Jan 2004 Posts: 195
|
23 Apr 2004 19:05 pic mclr capacitor |
|
|
|
|
Have you tried to increase capacitor that follows diode bridge after transformer. It has to be at least 4700uF!.
With 2x4700uF i controlled 10 relay for contact current of 10A each.
Anyhow you solved your problem! But turning Brownout is not a good idea!
Simce
|
|
| Back to top |
|
 |