Any insight for Automotive Sensor Inputs/ Front-End Protection.
in [Design]
|
From: Joerg on 28 Apr 2010 21:45 Jim Thompson wrote: > On Wed, 28 Apr 2010 18:26:16 -0700, Joerg <invalid(a)invalid.invalid> > wrote: > >> Jim Thompson wrote: >>> On Wed, 28 Apr 2010 17:50:17 -0700, Joerg <invalid(a)invalid.invalid> >>> wrote: >>> >>>> Jim Thompson wrote: >>>>> On Wed, 28 Apr 2010 17:11:18 -0700, Joerg <invalid(a)invalid.invalid> >>>>> wrote: >>>>> >>>>>> Jim Thompson wrote: >>>>> [snip] >>>>>>> I was designing for aerospace whilst you were still in knickers :-) >>>>>>> >>>>>>> Check the facts before you burp. >>>>>>> >>>>>> Then you should be intimately familiar with DO160. Yet you haven't >>>>>> answered the question above, how you would solve that situation for 15c >>>>>> or less without a TVS. >>>>>> >>>>>> How did the Romans say? Hic Rhodos, hic salta. >>>>> That'd be _way_ after my time in aerospace. >>>>> >>>> The Romans? Man, you _are_ old :-) >>> Yep. >>> >>>> Ok, then: The tougher category is 80V for 100msec, followed by 48V for >>>> one second. Three times at 10sec intervals. Depending on what it is the >>>> circuit can pause, but ideally it should just continue to work through >>>> these episodes. Throw in 600V spikes as well but they are brief and thus >>>> easy. So? >>>> >>>> Remember, you've got 15 pennies to play with, no more. And if you use >>>> lots of parts then you are charged a pick-and-place tax. >>> I first encountered load dump around 1964. I'll not tell you how I >>> solved it, I'll make you read my patents and see if you can discern >>> how I did it... by clever circuit design... no TVS's... I don't even >>> think they existed then :-) >>>>> What is it you have running in any space vehicle? >>>>> >>>> One electronic board but not at liberty to tell. >>>> >>>> >>>>> What is it you have running in any car electronics? >>>>> >>>> Nothing, assuming big rigs don't count here. However, some of my 12V >>>> designs are allowed to be run on vehicles and on ships (and do so nicely). >>>> >>>> >>>>> Perhaps you have something in an implant. I do. >>>>> >>>> I've got tons designs of intra-body stuff. But all this doesn't matter. >>>> You said using TVS is like making simple problems difficult so let's not >>>> deviate here. How do you solve the problem you snipped away, sans TVS, >>>> for less than 15c in parts? >>> Clever design, read my patents and learn from a (very old) master :-) >>> >> Hey, you are deviating now. How would you solve the problem I presented? >> "Spend the next two days reading through umpteen patents" ain't the >> answer. The answer would be a circuit sketch. So? > > YOUR problem, as a designer, is that you must use off-the-shelf stuff > that is not designed to withstand abuse, and ADD stuff to it to make > it survive. > Yes. So why then did you say that TVSes make simple problems difficult? That's plain wrong, and unsurprisingly you haven't been able to come up with the simpler solution to the (quite common) problem I presented. Because there is no simpler or lower cost solution than a TVS. > MY problem, when designing a chip, is to make the circuit survive on > its own, in the first place. Over the years I've become most clever > at that skill :-) > > MY '60's designs survive load dump. Wonder how ?:-) > That wasn't the issue in this thread. We (and that includes the OP) have to live with what's available at Arrow, Digikey and so on. Clients do not have half a million bucks and one year of patience to wait for a better chip that then saves 15c in discretes, and maybe amortizes over a decade, or never. They want the solution here, now, and cheap. And they get that solution. That's life. -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
From: Jim Thompson on 28 Apr 2010 21:53 On Wed, 28 Apr 2010 18:45:51 -0700, Joerg <invalid(a)invalid.invalid> wrote: >Jim Thompson wrote: [snip] >> >> YOUR problem, as a designer, is that you must use off-the-shelf stuff >> that is not designed to withstand abuse, and ADD stuff to it to make >> it survive. >> > >Yes. So why then did you say that TVSes make simple problems difficult? >That's plain wrong, and unsurprisingly you haven't been able to come up >with the simpler solution to the (quite common) problem I presented. >Because there is no simpler or lower cost solution than a TVS. > > >> MY problem, when designing a chip, is to make the circuit survive on >> its own, in the first place. Over the years I've become most clever >> at that skill :-) >> >> MY '60's designs survive load dump. Wonder how ?:-) >> > >That wasn't the issue in this thread. We (and that includes the OP) have >to live with what's available at Arrow, Digikey and so on. Clients do >not have half a million bucks and one year of patience to wait for a >better chip that then saves 15c in discretes, and maybe amortizes over a >decade, or never. They want the solution here, now, and cheap. And they >get that solution. That's life. I was just arguing that there's usually a clever approach that doesn't even need the expense of a TVS. "Life" is good design, not patches. ...Jim Thompson -- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | The only thing bipartisan in this country is hypocrisy
From: John Devereux on 29 Apr 2010 02:30 Joerg <invalid(a)invalid.invalid> writes: > Jim Thompson wrote: >> On Wed, 28 Apr 2010 18:09:18 +0100, John Devereux >> <john(a)devereux.me.uk> wrote: >> >>> Johnny5 <dirtylogicdesigns(a)gmail.com> writes: >>> >>>> On Apr 26, 6:35 pm, Jim Thompson <To-Email-Use-The-Envelope-I...(a)On-My- >>>> Web-Site.com> wrote: >>>>> On Mon, 26 Apr 2010 16:18:05 -0700, Joerg <inva...(a)invalid.invalid> >>>>> wrote: >>>>> >>>>>> Jamie wrote: >>>>>>> Johnny5 wrote: >>>>>>>> I was curious if anyone had any insight into protecting an Opto- >>>>>>>> isolator LED input from load dump, Jumpstarts, and many of the other >>>>>>>> signals in the auto environment. FETS,CLIPPERS,Clamps? >>>>>>>> Thank you in advance, >>>>>>>> It is an open drain SSR. >>>>>>> TVS diodes. Transorbs is another name for them. >>>>>> Also Transzorb. And ST calls them Transil. It's trade names, like >>>>>> Kleenex or Kim Wipes. >>>>>> [...] >>>>> No canned answers. It depends on the circuitry. >>>>> >>> [...] >>> >>>> Thank you all for the input, a little more on the circuit . Sensors >>>> are pressure switches/Temp that are connected to Vbatt. Pressure High= >>>> relay that switches Vbatt on/off. No current Limit. So I was thinking >>>> it would be best use opto. HSR312 SSR, Vr=7v, Vf=1.6V , control >>>> current approx 2mA. >>>> >>>> >>>> Automotive opto front end for Vbatt switched with no current limit >>>> >>>> Bav99/ TVS with a current limiter suffice? ? >>>> polarity protection p channel fet as well? any choice on fet, dmos >>>> only? >>>> thks >>> Do you actually need to do anything? LEDS will survive brief overcurrent >>> just fine. If you size the LED resistor for 12V operation near the 2mA, >>> you will have to get up to 100V or so to even get to its rated 25mA >>> maximum. Any such pulses will be very brief, the LED won't mind. >>> >>> I would just use a series resistor and connect a signal diode in >>> anti-parallel across the LED. >> >> There you go! >> >> So many posters here want to make simple problems difficult... TVS's, >> etc :-] >> > > Those posters tend to design reliable electronics :-) > > Figure 1, for a dose of reality: > http://www.automotivedesignline.com/howto/205101011 Hi Jeorg, That LTC part will likely blow up before the simple resistor+opto solution. The Ford load dump pulse is only 60V peak with 150ms decay time constant. The resistor+opto will survive 100V indefinitely. Furthermore is is extremely likely to survive pulses of much higher voltage, while the LTC part is extremely likely to fail. Yes, of course you can make it still more robust with a R+TVS+R feeding the LED. > Usually an LED survives that. But sometimes, especially when dumps > happen in rapid succession, it doesn't. I don't have too much > experience with automotive stuff for passenger cars but I do with > electronics for larger vehicles such as aircraft. There, a non-TVS > design is most likely going to hit the fan before it ever sees its > fuselage mounting spot. At the RTCA test lab. -- John Devereux
From: John Devereux on 29 Apr 2010 09:24 Jamie <jamie_ka1lpa_not_valid_after_ka1lpa_(a)charter.net> writes: > Jim Thompson wrote: > >> On Wed, 28 Apr 2010 18:09:18 +0100, John Devereux >> <john(a)devereux.me.uk> wrote: >> >> >>>Johnny5 <dirtylogicdesigns(a)gmail.com> writes: >>> >>> [...] >>>>Thank you all for the input, a little more on the circuit . Sensors >>>>are pressure switches/Temp that are connected to Vbatt. Pressure High= >>>>relay that switches Vbatt on/off. No current Limit. So I was thinking >>>>it would be best use opto. HSR312 SSR, Vr=7v, Vf=1.6V , control >>>>current approx 2mA. >>>> >>>> >>>>Automotive opto front end for Vbatt switched with no current limit >>>> >>>> Bav99/ TVS with a current limiter suffice? ? >>>> polarity protection p channel fet as well? any choice on fet, dmos >>>>only? >>>>thks >>> >>>Do you actually need to do anything? LEDS will survive brief overcurrent >>>just fine. If you size the LED resistor for 12V operation near the 2mA, >>>you will have to get up to 100V or so to even get to its rated 25mA >>>maximum. Any such pulses will be very brief, the LED won't mind. >>> >>>I would just use a series resistor and connect a signal diode in >>>anti-parallel across the LED. >> >> >> There you go! >> >> So many posters here want to make simple problems difficult... TVS's, >> etc :-] >> >> ...Jim Thompson > Lets see. > > Series resistor and diode = 2 components > > TVS diode = 1 component and they come in bi or unidirectional. No, you need the resistor (unless you are going to use an unreasonably large transorb). Load dump impedance is something like 2 ohms. -- John Devereux
From: Jim Thompson on 29 Apr 2010 10:27
On Thu, 29 Apr 2010 07:30:11 +0100, John Devereux <john(a)devereux.me.uk> wrote: >Joerg <invalid(a)invalid.invalid> writes: > [snip] >> >> Figure 1, for a dose of reality: >> http://www.automotivedesignline.com/howto/205101011 > >Hi Jeorg, > >That LTC part will likely blow up before the simple resistor+opto >solution. The Ford load dump pulse is only 60V peak with 150ms decay >time constant. [snip] Alternators have been improved over the years. In my day (mid '60's) you could see _400V_! So, above 20V, my field driver devices turned off and went into BVCER mode and rode out load dump. The control chip itself was low current, so a series carbon resistor plus an active "zener" on-chip sufficed for protection. (Everything bipolar.) ...Jim Thompson -- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | The only thing bipartisan in this country is hypocrisy |