Post by thread:[PIC:] 18f452, problems with interference

Hi! I been trying to solve this for quite a few days now but just can't find a solution. Short story: Etched 100 PCBs and deliverd them. Buyer calls back saying the system crasches sporadicly. The system measures tempratures and depending on several thing switches a few relays (220v) on or off. The problem seems to be the switching of 220v on and off. If I just switch an ordinary table lamp on/off nearby my cicuit I can get it to lock up. Sometimes after just 10 or 20 switches but on rare occasions I have to switch it 15 000 times. I tried running it on batteries and that made no difference, so it's not a dip in the power supply. I put an 18f452, a led, a reset circuit (standard) and a 3.68MHz crystal on my labdeck. Loaded it with a Blink a LED and can get it to freak out by switching a lamp. I get the feeling the something shocks my PIC from the outside, not through the power line. I have activated the BOR and tried it to make sure it works. The symptoms are several different, the PIC runs faster or slower, stops executing, stops executing but timers still run, or it resets. The WDT does not kick in when the PIC stops executing. I must have missed something but can't figure out what. tia /Fredrik, going crazy any day now... -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email spam_OUTlistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

frax wrote: > Hi! > > I been trying to solve this for quite a few days now but just can't find a > solution. > > Short story: > Etched 100 PCBs and deliverd them. Buyer calls back saying the system > crasches sporadicly. > The system measures tempratures and depending on several thing switches a > few relays (220v) on or off. Welcome to the real world. Up to this point, your job has been a cakewalk. NOW you earn your pay. > The problem seems to be the switching of 220v on and off. If I just switch > an ordinary table lamp on/off nearby my cicuit I can get it to lock up. > Sometimes after just 10 or 20 switches but on rare occasions I have to > switch it 15 000 times. You have the classic symptoms of radiation (RF pulses, noise pulses from power, relay noise, etc) being imposed onto your equipment. First, you must stop outside conductors from bringing the noise into the PCB from outside. To do this, every wire going to the PCB has to be FILTERED to remove the problem. If the wire is a signal wire, sometimes just a series resistor (220 ohm) and a small cap (1000pF) will do the job. If it is power or relay wiring, you will need to use RF chokes and these will have to be able to carry the current. Next, look at the relay circuits carefully. The collapsing coils of a relay generate a noise spike of incredibly high energy. You MUST use a Transorber or varistor- NOT a diode or Zener diode- to stop the relay coil pulse, because diodes are NOT fast enough to crush that strong signal. (I know there are many app notes indicating that diodes are recommended, but these are WRONG by being outdated- they are from a time when some noise could be tolerated- and those days are gone. If you've done the two paragraphs above religiously, your system is probably now working at this point. But here are some other pointers to nail electrical noise: 1. Make sure the PCB has a GOOD ground system. Ideally it means a ground plane, because that simple act stops RF from passing through the PCB from top to bottom. But if your product can't afford an inner plane, create a GND copper pour on both top and bottom. You will be amazed how well the system can reject noise with a good ground. 2. Try to design your product into a metal box or a metallized plastic box. These metal coatings that look like copper on the inside of the plastic case work almost as good as metal in stopping external noise sources. 3. Finally, PROVE that your design is now fixed. I found that the best way to do this is with a 110V surplus power relay; wire it up with a 10' cord, without any coil suppressors, so that it acts like a buzzer, by turning on and off hundreds of times a second. Wrap that 10' cord near your product. If it works NOW, its fixed. --Bob {Quote hidden}> I tried running it on batteries and that made no difference, so it's not a > dip in the power supply. > > I put an 18f452, a led, a reset circuit (standard) and a 3.68MHz crystal on > my labdeck. > Loaded it with a Blink a LED and can get it to freak out by switching a > lamp. > > I get the feeling the something shocks my PIC from the outside, not through > the power line. I have activated the BOR and tried it to make sure it works. > > The symptoms are several different, the PIC runs faster or slower, stops > executing, stops executing but timers still run, or it resets. The WDT does > not kick in when the PIC stops executing. > > I must have missed something but can't figure out what. > > tia > /Fredrik, going crazy any day now... > > -- > http://www.piclist.com#nomail Going offline? Don't AutoReply us! > email listservKILLspammitvma.mit.edu with SET PICList DIGEST in the body > > -- Replier: Most attachments rejected -------------- Bob Axtell PIC Hardware & Firmware Dev http://beam.to/baxtell 1-520-219-2363 -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email .....listservKILLspam.....mitvma.mit.edu with SET PICList DIGEST in the body

>-----Original Message----- >From: Bob Axtell [EraseMEengineerspam_OUTTakeThisOuTCOTSE.NET] >Sent: 13 July 2004 09:55 >To: PICLISTspam_OUTMITVMA.MIT.EDU >Subject: Re: [PIC:] 18f452, problems with interference >Next, look at the relay circuits carefully. The collapsing >coils of a relay generate a noise spike of incredibly high >energy. You MUST use a Transorber or varistor- NOT a diode or >Zener diode- to stop the relay coil pulse, because diodes are >NOT fast enough to crush that strong signal. (I know there >are many app notes indicating that diodes are recommended, but >these are WRONG by being outdated- they are from a time when >some noise could be tolerated- and those days are gone. Is this really true? Whilst the generic 1N400x recitifiers are slow, the cheap 1N4148/1N914 type diodes are very fast, low capacitance switching diodes and are robust enough for clamping most relays. Regards Mike ======================================================================= This e-mail is intended for the person it is addressed to only. The information contained in it may be confidential and/or protected by law. If you are not the intended recipient of this message, you must not make any use of this information, or copy or show it to any person. Please contact us immediately to tell us that you have received this e-mail, and return the original to us. Any use, forwarding, printing or copying of this message is strictly prohibited. No part of this message can be considered a request for goods or services. ======================================================================= -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email @spam@listservKILLspammitvma.mit.edu with SET PICList DIGEST in the body

At 09:45 AM 7/13/2004 +0200, you wrote: >Hi! > >I been trying to solve this for quite a few days now but just can't find a >solution. Okay, so switching a LOAD is what does it, right? Take the load off and the relays switch fine, right? So.. I think you've essentially got problems with the noise created by the contacts of the relays. It can be getting into your circuit in a number of places. Your layout is one thing to suspect. Make sure you've got NO wires running from a PIC pin off the board. Preferably a ground plane under everything to do with the PIC. SHORT wires to the crystal or resonator. Different paths for the relay ground and the PIC ground back to the power supply.. and many other things I can't begin to guess at. Try to visualize the CURRENTs and the LOOPS. You want the currents to follow benign paths to to MINIMIZE LOOP AREA (ideally with power planes, next best is traces overtop of each other on a double-sided board, worst case would be single-sided traces running around the outside of the board in a big loop like a deliberately created antenna. Good luck. There is *probably* a way to hack a fix for these 100 units without replacing the boards, but it's best to be conservative to start with, IMHO. I like to use ground planes, shielding, filtering, series impedances, opto-isolation, shunt capacitance, ferrite beads, common-mode chokes, copper pour, minimum inductance layout and other techniques up- front to minimize the risk of disaster, which is what you have right now, unfortunately. Best regards, Spehro Pefhany --"it's the network..." "The Journey is the reward" KILLspamspeffKILLspaminterlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email RemoveMElistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

Michael Rigby-Jones wrote: {Quote hidden}>>-----Original Message----- >>From: Bob Axtell [spamBeGoneengineerspamBeGoneCOTSE.NET] >>Sent: 13 July 2004 09:55 >>To: TakeThisOuTPICLISTEraseMEspam_OUTMITVMA.MIT.EDU >>Subject: Re: [PIC:] 18f452, problems with interference > > >>Next, look at the relay circuits carefully. The collapsing >>coils of a relay generate a noise spike of incredibly high >>energy. You MUST use a Transorber or varistor- NOT a diode or >>Zener diode- to stop the relay coil pulse, because diodes are >>NOT fast enough to crush that strong signal. (I know there >>are many app notes indicating that diodes are recommended, but >>these are WRONG by being outdated- they are from a time when >>some noise could be tolerated- and those days are gone. > > > Is this really true? Whilst the generic 1N400x recitifiers are slow, the > cheap 1N4148/1N914 type diodes are very fast, low capacitance switching > diodes and are robust enough for clamping most relays. Alas, while they are significantly faster than the venerable 400X-series, the delay in switching time is significant enough to allow a lot of energy to get through. You can google for supporting specs; look at the IR (Power MOSFET) website, if I recall, it had the best documents on protecting power MOSFETs. --Bob {Quote hidden}> > Regards > > Mike > > ======================================================================= > This e-mail is intended for the person it is addressed to only. The > information contained in it may be confidential and/or protected by > law. If you are not the intended recipient of this message, you must > not make any use of this information, or copy or show it to any > person. Please contact us immediately to tell us that you have > received this e-mail, and return the original to us. Any use, > forwarding, printing or copying of this message is strictly prohibited. > No part of this message can be considered a request for goods or > services. > ======================================================================= > > -- > http://www.piclist.com#nomail Going offline? Don't AutoReply us! > email RemoveMElistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body > > -- Replier: Most attachments rejected -------------- Bob Axtell PIC Hardware & Firmware Dev http://beam.to/baxtell 1-520-219-2363 -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email listservEraseME.....mitvma.mit.edu with SET PICList DIGEST in the body

On Tue, 2004-07-13 at 03:45, frax wrote: > Hi! > > I been trying to solve this for quite a few days now but just can't find a > solution. > > Short story: > Etched 100 PCBs and deliverd them. Buyer calls back saying the system > crasches sporadicly. > The system measures tempratures and depending on several thing switches a > few relays (220v) on or off. > > The problem seems to be the switching of 220v on and off. If I just switch > an ordinary table lamp on/off nearby my cicuit I can get it to lock up. > Sometimes after just 10 or 20 switches but on rare occasions I have to > switch it 15 000 times. > I tried running it on batteries and that made no difference, so it's not a > dip in the power supply. Stupid small thing, but you do have a reversed diode across the relay coils right? What are you switching the coils with? TTYL ----------------------------- Herbert's PIC Stuff: http://repatch.dyndns.org:8383/pic_stuff/ -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email EraseMElistservmitvma.mit.edu with SET PICList DIGEST in the body

> (I know there >>are many app notes indicating that diodes are recommended, but >>these are WRONG by being outdated- they are from a time when >>some noise could be tolerated- and those days are gone. > >Is this really true? Whilst the generic 1N400x recitifiers are slow, the cheap 1N4148/1N914 type diodes are very fast, low capacitance switching diodes and are robust enough for clamping most relays. I don't know what he's on about. . The only thing you need to do is provide a harmless path for the current induced by the collapsing magnetic field. When you stop the current through the inductor, the collapsing field attempts to induce this current, and the voltage at the switch rises as high as necessary to get the current to flow. A diode across the coil (preferred) or a zener across the switch, will do that job. -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email RemoveMElistservEraseMEEraseMEmitvma.mit.edu with SET PICList DIGEST in the body

David VanHorn wrote: >>>are many app notes indicating that diodes are recommended, but >>>these are WRONG by being outdated- they are from a time when >>>some noise could be tolerated- and those days are gone. >> >>Is this really true? Whilst the generic 1N400x recitifiers are slow, the cheap 1N4148/1N914 type diodes are very fast, low capacitance switching diodes and are robust enough for clamping most relays. > > I don't know what he's on about. . > The only thing you need to do is provide a harmless path for the current induced by the > collapsing magnetic field. When you stop the current through the inductor, the collapsing field > attempts to induce this current, and the voltage at the switch rises as high as necessary to get > the current to flow. A diode across the coil (preferred) or a zener across the switch, will do that job. No, the diode is too slow, and the zener is even slower than that. The diode across the coil serves two purposes: 1. Prevent the inductive kick from damaging the driving device, whether IC, MOSFET, or transistor. 2. Reduce the RF generated by the kick from being propagated into other circuitry. One problem with the diode method is that diodes have a finite switching time, i.e. a delay before turning on. The spike before the diode conducts will radiate energy and the spike may still damage the driving device. The second problem with simple diodes is that the PN junction must dissipate the energy spike. This sudden heat may cause the junction to fail, usually by spot overheating (short). This tendency makes diodes unacceptable for good engineering design. Transorbers, Varistors and other similar devices switch in a 1 ns or so, and they all have the ability to handle large pulses of energy without failing. The second problem has to do with the PCB layout, and where the protective device is located. The protective device must be located as close to the relay coil as possible, since the heavy current spike that circulates in the PCB traces to the coil will be a RADIATION source to other comnponents. -- Replier: Most attachments rejected -------------- Bob Axtell PIC Hardware & Firmware Dev http://beam.to/baxtell 1-520-219-2363 -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email RemoveMElistservspam_OUTKILLspammitvma.mit.edu with SET PICList DIGEST in the body