Post by thread:What is a good capacitive value with a 2.2k resis

> Could anyone recommend a good capacitive value used with a 2.2K resister > to filter out noise. I have some limits switches which are NO and > connected to a 2.2K pullup resistor. However the moment the limit > switches are closed to ground, they cause my PIC processor to reset. I > am thinking that the limit switches are somehow picking up some high > frequency noice, which is not appearing on my 20MHz scope. I know that > the pins are set for input mode, so it is not shorting the system. What you describe is rather strange and suggests some other problem. If you are in a "normal" environment this would be an extremely unusual occurrence. Check carefully that eg switches really go from pins to ground, ground return is same as PIC ground etc. Have you got any power supply decoupling on the PIC eg 10 uF or 100 uF Vdd to ground capacitor? What is your power supply. IF the switches are some distance from the processor on long wiring runs then voltage pickup is more likely. If so, firstly repeat the test with switches with very short leads adjacent to the PIC. If this works OK then you noise theory sounds more likely. Assuming remote switches and long leads. First ensure switches have a 2 wire connection back to PIC ie do NOT use eg equipment ground as earth return. Connect via a twisted pair of wires (this balances induced noise in each leg so that they cancel). The time constant of an RC circuit is T = RC so C = T/R In this case it's not the 2k2 that you want to use as your R but a resistance in the noise supplying circuit. Try this: - Place a 1k resistor in EACH leg of the wires going to the limit switch - resistors located at the PIC. end. - Place a 0.1 uF capacitor across the PIC to ground connection (where the two resistors reach the PIC and ground). - Use a 100k pullup resistor. Time constant is now 2k x 0.1 uF = 200 uS to discharge the cap and 100k x 0.1 uF = 10 ms to charge the cap. You could use values which are closer for charge and discharge R's but have to ensure that the input circuit properly pulls the circuit to a logic low. A larger capacitor within reason won't hurt. eg a 1 uF or even a 10 uF for test purposes. The 10 uF/100k will have a 1 second time constant so will be no good in "real" use (and the PIC prefers not to have such a slow transition) BUT this will show you if your problem is noise. BUT FIRST - do perform the test with a switch right at the PIC and no long leads. If this fails then you are doing something else wrong that you haven't described here. Russell McMahon -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

Be careful if putting big caps directly across a small switch. The cap will charge up when the switch is open & then discharge with high current through the switch when it is closed. The current can easily reach several hundred amps (momentarily) and weld or otherwise damage the switch contacts. Placing a 10 - 100 ohm resistor on series with the cap. will normally prevent damage but still provide enough filtering. More elaborate circuits using more resistors & diodes can also be used. Richard P > Could anyone recommend a good capacitive value used with a 2.2K resister > to filter out noise. I have some limits switches which are NO and > connected to a 2.2K pullup resistor. However the moment the limit > switches are closed to ground, they cause my PIC processor to reset. I > am thinking that the limit switches are somehow picking up some high > frequency noice, which is not appearing on my 20MHz scope. I know that > the pins are set for input mode, so it is not shorting the system. <snipped> A larger capacitor within reason won't hurt. eg a 1 uF or even a 10 uF for test purposes. The 10 uF/100k will have a 1 second time constant so will be no good in "real" use (and the PIC prefers not to have such a slow transition) BUT this will show you if your problem is noise. BUT FIRST - do perform the test with a switch right at the PIC and no long leads. If this fails then you are doing something else wrong that you haven't described here. Russell McMahon -- -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

That should have been a shunt capacitor... Sorry about the dumb error. Rich ----- Original Message ----- From: "Richard Graziano" <spam_OUTrgrazia1TakeThisOuTrochester.rr.com> To: <.....PICLISTKILLspam@spam@MITVMA.MIT.EDU> Sent: Sunday, February 02, 2003 9:32 PM Subject: Re: What is a good capacitive value with a 2.2k resister to filter out high frequency noice? > For a Butterworth configuration LPF, the cut-off frequency is equal to > 2PiRC. That is a resistor in series with the signal and a shunt resistor > from the load end of the resistor to signal return. > > Rich > ----- Original Message ----- > From: "Russell McMahon" <apptechKILLspamPARADISE.NET.NZ> > To: <.....PICLISTKILLspam.....MITVMA.MIT.EDU> > Sent: Sunday, February 02, 2003 3:54 PM > Subject: Re: What is a good capacitive value with a 2.2k resister to filter > out high frequency noice? > > > > > Could anyone recommend a good capacitive value used with a 2.2K resister > > > to filter out noise. I have some limits switches which are NO and > > > connected to a 2.2K pullup resistor. However the moment the limit > > > switches are closed to ground, they cause my PIC processor to reset. I > > > am thinking that the limit switches are somehow picking up some high > > > frequency noice, which is not appearing on my 20MHz scope. I know that > > > the pins are set for input mode, so it is not shorting the system. > > > > What you describe is rather strange and suggests some other problem. > > If you are in a "normal" environment this would be an extremely unusual > > occurrence. Check carefully that eg switches really go from pins to > ground, > > ground return is same as PIC ground etc. Have you got any power supply > > decoupling on the PIC eg 10 uF or 100 uF Vdd to ground capacitor? What is > > your power supply. > > > > IF the switches are some distance from the processor on long wiring runs > > then voltage pickup is more likely. If so, firstly repeat the test with > > switches with very short leads adjacent to the PIC. If this works OK then > > you noise theory sounds more likely. > > > > Assuming remote switches and long leads. First ensure switches have a 2 > wire > > connection back to PIC ie do NOT use eg equipment ground as earth return. {Quote hidden}> > Connect via a twisted pair of wires (this balances induced noise in each > leg > > so that they cancel). > > > > The time constant of an RC circuit is T = RC > > so C = T/R > > In this case it's not the 2k2 that you want to use as your R but a > > resistance in the noise supplying circuit. > > Try this: > > - Place a 1k resistor in EACH leg of the wires going to the limit > > switch - resistors located at the PIC. end. > > - Place a 0.1 uF capacitor across the PIC to ground connection (where {Quote hidden}> > the two resistors reach the PIC and ground). > > - Use a 100k pullup resistor. > > Time constant is now 2k x 0.1 uF = 200 uS to discharge the cap > > and > > 100k x 0.1 uF = 10 ms to charge the cap. > > > > You could use values which are closer for charge and discharge R's but > have > > to ensure that the input circuit properly pulls the circuit to a logic > low. > > > > A larger capacitor within reason won't hurt. eg a 1 uF or even a 10 uF for > > test purposes. The 10 uF/100k will have a 1 second time constant so will > be > > no good in "real" use (and the PIC prefers not to have such a slow > > transition) BUT this will show you if your problem is noise. > > > > BUT FIRST - do perform the test with a switch right at the PIC and no long {Quote hidden}> > leads. If this fails then you are doing something else wrong that you > > haven't described here. > > > > > > > > Russell McMahon > > > > -- > > http://www.piclist.com hint: The list server can filter out subtopics > > (like ads or off topics) for you. See http://www.piclist.com/#topics > > > > -- > http://www.piclist.com hint: The list server can filter out subtopics > (like ads or off topics) for you. See http://www.piclist.com/#topics > -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

> Could anyone recommend a good capacitive value used with a 2.2K resister > to filter out noise. That of course depends completely on the frequency you want to start filtering at. The rolloff frequency for an RC filter is: F = 1 / (2 * Pi * R * C) where F is the frequency in hertz, R the resistance in ohms, and C the capacitance in farads. I keep the constant 10**6 / (2 * Pi) = 159155 permanently in a variable in my calculator. It can be divided by any two of the resistance in ohms, the frequency in hertz, or the capacitance in microfarads to yield the third. Very handy. > I have some limits switches which are NO and > connected to a 2.2K pullup resistor. However the moment the limit > switches are closed to ground, they cause my PIC processor to reset. I > am thinking that the limit switches are somehow picking up some high > frequency noice, which is not appearing on my 20MHz scope. I know that > the pins are set for input mode, so it is not shorting the system. There is something else going on, like inductance in the line causing a kickback. If the switch is external to the board, then you should protect the PIC inputs. After the 2.2Kohm pullup, put a 1Kohm resistor in series with diodes to ground and supply. After that put another 10Kohm resistor in series to the PIC pin. That should protect from over and under voltage spikes to a few tens of volts. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.