Post by thread:[PICLIST] [PIC] kill a PIC

Chis, > The best way to unexpectedly kill a PIC through I/O that I have seen, > although I have not been stupid enough to try it myself - is to attempt to > implement the MC ap note on triac control, where it suggests (and shows a > drawing of) connecting the AC line directly to I/O for zero-cross > detection - relying on the pin's internal protection diodes. > That was just a ridculous idea to try and implement. I can see no problem with this, as long as the series resistor is large enough to limit the maximum current to the manufactures specification. .......................... Zim -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email spam_OUTlistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

> Chis, > > > The best way to unexpectedly kill a PIC through I/O that I have seen, > > although I have not been stupid enough to try it myself - is to attempt to {Quote hidden}> > implement the MC ap note on triac control, where it suggests (and shows a > > drawing of) connecting the AC line directly to I/O for zero-cross > > detection - relying on the pin's internal protection diodes. > > > That was just a ridculous idea to try and implement. > > I can see no problem with this, as long as the series resistor > is large enough to limit the maximum current to the manufactures > specification. This is what I like to refer to as TV set engineering. ;-D (ducking for cover now) IMHO, TV set design and antenna theory (vs. reality) are indistinguishable from black magic. I tend to eventually "let the smoke out" of TV's when I try to tinker with them. So now I stick with moderately low-voltage (120V or less) equipment with somewhat understandable schematics. You know.....not like TV's. ;-) I respect the ability of people to connect high-voltage directly to pic by not exceeding the current limits, but it still seems a little risky to me. I tend to expect that touching a pin on an IC is not going to shock me. That hot chassis stuff 'gets' me every time. ;-D michael -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email .....listservKILLspam@spam@mitvma.mit.edu with SET PICList DIGEST in the body

Hi, > > The best way to unexpectedly kill a PIC through I/O that I have seen, > > although I have not been stupid enough to try it myself - is to attempt to {Quote hidden}> > implement the MC ap note on triac control, where it suggests (and shows a > > drawing of) connecting the AC line directly to I/O for zero-cross > > detection - relying on the pin's internal protection diodes. > > > That was just a ridculous idea to try and implement. > > I can see no problem with this, as long as the series resistor > is large enough to limit the maximum current to the manufactures > specification. It is almost impossible to keep it under the maximum current allowed all the time !!! Remember that you have many high voltage spikes going on any regular power line. Even more if you are using a dimmer that generates lots of noise. Imagine that switching a heavy inductive load such as a halogen lamp transformer ! The internal diodes cannot protect from these spikes. You have to use external protection to make sure you do not violate the specs. That application note has always amased me because it is just pure bad enginnering. I would never use that to make a real product. Best regards, Alexandre Guimaraes -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email listservKILLspammitvma.mit.edu with SET PICList DIGEST in the body

At 12:12 PM 2/24/02 +1100, you wrote: >I can see no problem with this, as long as the series resistor >is large enough to limit the maximum current to the manufactures >specification. I don't see any *inherent* problem with it, but trying it with a standard 0.25W resistor is practically guaranteeing trouble, and I doubt it would pass safety certification in any developed country. Instead use an approved resistor rated for a few thousand volts, which might cost 10-20 cents rather than << 1 cent. Philips' VR37 series meets the requirements, and can be bent to fit 0.5" pitch pads. The value should be high enough that enough current to cause latchup will not flow *even under transient conditions*. Best regards, >.......................... Zim > >-- >http://www.piclist.com#nomail Going offline? Don't AutoReply us! >email .....listservKILLspam.....mitvma.mit.edu with SET PICList DIGEST in the body Spehro Pefhany --"it's the network..." "The Journey is the reward" EraseMEspeffspam_OUTTakeThisOuTinterlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com 9/11 United we Stand -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email listservspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body

At 07:00 AM 2/25/02 +0200, you wrote: >I know this should be OK in theory as it is in the manufacturer's specs, but >in practice I would never design a curcuit for mass production like that, or >do anythink that could take any pin outside the supply voltage range, even >using schottkey diodes to clamp the pins could be dangerous. There is a >parasitic scr in every chip waiting to be fired! Junction-isolated CMOS chips (which is most of them), yes. If you guarantee the current will never get to the trigger current (which is tested) then you'll be safe on that account. That's why you need a resistor that won't arc over for even an instant, with say 3-4,000 volts applied. Remember that similar voltages are created by walking across a carpet, so pins that are brought out need similar protection. P.S. I agree with distrusting application notes. I have been burned in the distant past by using application note circuits (from RCA), bought samples and wasted several days before I realized that the app note circuit could *never* work reliably. Now I always do the calculations from the min/max figures on the data sheet, and don't trust the people who wrote the app notes (who seem to be young and inexperienced engineers trying to be clever and sell chips, and don't put the care in there to make a commercial conservative design that will always work under a wide set of conditions). In a way, it's a bit like suckering you in.. here buy our part and you can make this product with only 17 parts. If they showed you the 34 you really need to do a good job, you'd shrug and flip the page. Unfortunately, some are suckered to the point of actually trying to produce the "design" with the 17 parts and get burned. I've just been struggling with such a PIC "design" that someone else did. 8-( Best regards, Spehro Pefhany --"it's the network..." "The Journey is the reward" @spam@speffKILLspaminterlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com 9/11 United we Stand -- http://www.piclist.com hint: To leave the PICList KILLspampiclist-unsubscribe-requestKILLspammitvma.mit.edu

> The internal diodes cannot protect from these spikes. > ...... > That application note has always amased me because it is just pure bad > enginnering. Let's see, the PIC input diodes can take about 20mA (I'm guessing). With a 100K resistor, that's 2KV input With 1Meg that's 20 KV . And BTW, that's continuous current. For narrow pulses, the PIC could handle proportionally more current. My point is that the resistor will die long before the PIC ! (a 1 Meg resistor with 20 KV across it is dissipating 400 Watts !!) And of course you would have a small cap at the PIC to filter out the HF junk. This cap would also greatly limit any spikes. The weakest link would be the voltage rating of the resistor, you would need a couple in series at least. And yes, I personally would put a MOV or something at the junction of the resistors. Just pointing out that the App note isn't crazy when you actually do the sums... ..................... Zim -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu

At 01:50 AM 2/26/02 +1100, you wrote: >And BTW, that's continuous current. Absolute maximum clamp current is +/-20mA. Absolute maximum. >For narrow pulses, the PIC could handle proportionally more current. Unh, no, not a good assumption. >My point is that the resistor will die long before the PIC ! >(a 1 Meg resistor with 20 KV across it is dissipating 400 Watts !!) More likely it is arcing and dissipating 40kW ;-) >And of course you would have a small cap at the PIC to filter out the HF >junk. >This cap would also greatly limit any spikes. Really? Without doing the math, I'd certainly think that any such capacitor would cause a lot of phase shift before it did much good. Best regards, Spehro Pefhany --"it's the network..." "The Journey is the reward" spamBeGonespeffspamBeGoneinterlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com 9/11 United we Stand -- http://www.piclist.com hint: To leave the PICList TakeThisOuTpiclist-unsubscribe-requestEraseMEspam_OUTmitvma.mit.edu

{Quote hidden}> -----Original Message----- > From: Graeme Zimmer [SMTP:RemoveMEgzimmerTakeThisOuTOPTUSHOME.COM.AU] > Sent: Monday, February 25, 2002 2:51 PM > To: PICLISTEraseME.....MITVMA.MIT.EDU > Subject: Re: [PIC] kill a PIC > > > The internal diodes cannot protect from these spikes. > > ...... > > That application note has always amased me because it is just pure bad > > enginnering. > > Let's see, the PIC input diodes can take about 20mA (I'm guessing). > > With a 100K resistor, that's 2KV input > With 1Meg that's 20 KV . > > And BTW, that's continuous current. > For narrow pulses, the PIC could handle proportionally more current. > > My point is that the resistor will die long before the PIC ! > (a 1 Meg resistor with 20 KV across it is dissipating 400 Watts !!) > > And of course you would have a small cap at the PIC to filter out the HF > junk. > This cap would also greatly limit any spikes. > > The weakest link would be the voltage rating of the resistor, you would > need > a couple in series at least. > And yes, I personally would put a MOV or something at the junction of the > resistors. > > Just pointing out that the App note isn't crazy when you actually do the > sums... > This topic has been discussed numerous times on the list. IIRC, the general feeling was that injecting currents into the protection diodes caused unpredictable behaviour and was best avoided. The magnitude of current was well below that which could cause any kind of SCR latchup as well. I'll have a scan through the archives if I get 5 minutes. Mike -- http://www.piclist.com hint: To leave the PICList EraseMEpiclist-unsubscribe-requestmitvma.mit.edu

> > The internal diodes cannot protect from these spikes. > > ...... > > That application note has always amased me because it is just pure bad > > enginnering. > > Let's see, the PIC input diodes can take about 20mA (I'm guessing). 20 ma is the correct number. > With a 100K resistor, that's 2KV input > With 1Meg that's 20 KV . > > And BTW, that's continuous current. > For narrow pulses, the PIC could handle proportionally more current. Not at all. That is instantaneous. The latch-up will occur very fast as the semi conductor areas are very small. It is not triggered by temperature rise, as fas as I know. > My point is that the resistor will die long before the PIC ! > (a 1 Meg resistor with 20 KV across it is dissipating 400 Watts !!) The chip will enter scr latch up before the resistor even fells the problem. The spikes will arc over the resistor and even if they dont they will not harm the resistor because they are very fast. > And of course you would have a small cap at the PIC to filter out the HF > junk. > This cap would also greatly limit any spikes. The cap will not help much. It will not act fast enough ! > The weakest link would be the voltage rating of the resistor, you would need > a couple in series at least. > And yes, I personally would put a MOV or something at the junction of the > resistors. The MOV, a very fast one, would make the design more reliable. I would use 2 resistors in series and the MOV in the junction of the resistors. That should be enough to keep the high voltage out. > Just pointing out that the App note isn't crazy when you actually do the > sums... Sorry but I would never use that without the MOV's at a product of mine. Best regards, Alexandre Guimaraes -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestEraseMEEraseMEmitvma.mit.edu

At 09:22 PM 2/24/02 -0300, Alexandre Guimarães wrote: > It is almost impossible to keep it under the maximum current allowed all >the time !!! Remember that you have many high voltage spikes going on any >regular power line. Even more if you are using a dimmer that generates lots >of noise. Imagine that switching a heavy inductive load such as a halogen >lamp transformer ! The internal diodes cannot protect from these spikes. You >have to use external protection to make sure you do not violate the specs. >That application note has always amased me because it is just pure bad >enginnering. I would never use that to make a real product. At 03:15 PM 2/25/02 +0000, Michael Rigby-Jones wrote: >This topic has been discussed numerous times on the list. IIRC, the general >feeling was that injecting currents into the protection diodes caused >unpredictable behaviour and was best avoided. The magnitude of current was >well below that which could cause any kind of SCR latchup as well. I can comment on both the above statements: I *do* use a PIC input pin to sample zero-cross in several products. There are 2 - 1 Meg resistors (0.25W, 400V) in series to sample a 120 Vac input. Some boards have 1n4148 clamps to Vdd & Gnd, some boards have a 4.7V zener clamp to Gnd, some boards have no external diode clamps. I don't use a capacitor filter at the pin if the intent is to sample zero crossing (too much phase shift) but do use a capacitor if the requirement is only use the AC mains as a timebase. I have had *NO* failures related to this out of several thousand boards - even those without external diode clamps. But I did a couple of things ahead of time: I qualified the resistors I use by testing several hundred pieces with a pulsed 1200 Vac supply to ensure that they do not break down or arc *and* I have a MOV at the front end of the board - where the 120 Vac enters the card. I feel that 2 resistors in series, each resistor having a 10x test margin, is adequate. I *always* use clamp diodes if the PIC has a/d input that I am using. I have measured significant a/d errors on the 12c671, 16c71, 16c73 parts if there is *ANY* substrate current resulting from the ESD protect diodes conducting because some other input pin was above Vdd or below Gnd. I don't know about the comparitor inputs on the '62x family of parts - I'll find out when I do a project that actually uses those those features. On the other hand, minor substrate currents (0.1 mA or less) appear to have no effect on the purely digital inputs on the PICs I have used to date. Finally, this (2 resistors + MOV) seems to work for me with a 120 Vac mains supply. All bets are off at higher voltages. dwayne Dwayne Reid <RemoveMEdwaynerspam_OUTKILLspamplanet.eon.net> Trinity Electronics Systems Ltd Edmonton, AB, CANADA (780) 489-3199 voice (780) 487-6397 fax Celebrating 18 years of Engineering Innovation (1984 - 2002) .-. .-. .-. .-. .-. .-. .-. .-. .-. .- `-' `-' `-' `-' `-' `-' `-' `-' `-' Do NOT send unsolicited commercial email to this email address. This message neither grants consent to receive unsolicited commercial email nor is intended to solicit commercial email. -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestTakeThisOuTspammitvma.mit.edu

Hi, Dwayne > I *do* use a PIC input pin to sample zero-cross in several products. There > are 2 - 1 Meg resistors (0.25W, 400V) in series to sample a 120 Vac > input. Some boards have 1n4148 clamps to Vdd & Gnd, some boards have a > 4.7V zener clamp to Gnd, some boards have no external diode clamps. I > don't use a capacitor filter at the pin if the intent is to sample zero > crossing (too much phase shift) but do use a capacitor if the requirement > is only use the AC mains as a timebase. It is ok to have the pin connected that way, you are protecting it. > Finally, this (2 resistors + MOV) seems to work for me with a 120 Vac mains > supply. All bets are off at higher voltages. The MOV is giving you the reliability you need ! You can be pretty sure that 120 Vac is not 2Kvolts. I do not like diode clamps to VCC or GND at all. You inject all the noise into your power supply and may get many components in trouble. Small voltage surge arrestors are available and are not too expensive. That way you just dissipate the spikes as heat and do not propagate them trough all the board at the power supply rails. Best regards, Alexandre Guimaraes -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads