Forums

Why are IGBT's used in ignitions so much?

Started by Joerg April 28, 2018
lørdag den 28. april 2018 kl. 22.22.54 UTC+2 skrev Joerg:
> Considering the typical saturation voltages of >1V it doesn't look > attractive yet they are popular: > > https://www.fairchildsemi.com/datasheets/FG/FGD3040G2_F085.pdf > > Why? > > There are plenty of cheap MOSFETs that are in the low tens of milliohms > and thus show only a fraction of the losses while the coil current ramps > up. You'd just have to add a zener from drain to gate in case the spark > plug goes bad, the wire comes off or the controller lets the dwell time > run out of hand. Plus some other gate protection but that's no problem. >
http://www.st.com/content/ccc/resource/technical/document/application_note/8d/46/36/a7/b6/f2/45/10/CD00003911.pdf/files/CD00003911.pdf/jcr:content/translations/en.CD00003911.pdf
On 28/04/2018 21:23, Joerg wrote:
> Considering the typical saturation voltages of >1V it doesn't look > attractive yet they are popular: > > https://www.fairchildsemi.com/datasheets/FG/FGD3040G2_F085.pdf > > Why? > > There are plenty of cheap MOSFETs that are in the low tens of milliohms > and thus show only a fraction of the losses while the coil current ramps > up. You'd just have to add a zener from drain to gate in case the spark > plug goes bad, the wire comes off or the controller lets the dwell time > run out of hand. Plus some other gate protection but that's no problem. >
I'd guess the answer is ruggedness. Automotive engineering is one of the most conservative around. Possibly once in the past someone did use a MOSFET and it failed and they've been averse ever since. Modern avalanche rated FETs may be getting tough enough but auto designers may feel comfortable with a solution they know and trust? piglet
On Sun, 29 Apr 2018 12:18:32 -0700 (PDT), Lasse Langwadt Christensen
<langwadt@fonz.dk> wrote:

>s&#2013266168;ndag den 29. april 2018 kl. 19.23.11 UTC+2 skrev Jim Thompson: >> On Sun, 29 Apr 2018 09:52:21 -0700, Joerg <news@analogconsultants.com> >> wrote: >> >> >On 2018-04-28 15:16, Tim Williams wrote: >> >> "Joerg" <news@analogconsultants.com> wrote in message >> >> news:fkk3gpFmm06U1@mid.individual.net... >> >>> There are plenty of cheap MOSFETs that are in the low tens of >> >>> milliohms and thus show only a fraction of the losses while the coil >> >>> current ramps >> >>> up. >> >> >> >> At what Vds(max)? ;-) >> >> >> > >> >You don't need much above 400V for an ignition. Easy to do with either kind. >> >> 350V/5A is the general requirement, so I'd probably margin a wee-bit >> higher than 400V, maybe 600V. > >most ignition igbt have an active clamp around 350-400V to protect the coil >if there's a loose plug wire >
I don't do "most", I roll my own, based on "CD-Ignition-Basic.pdf" on the S.E.D/Schematics Page of my website. I was running experimental Ford Thunderbirds with "cruise control" back in the late '60's using that ignition (at Philco-Ford, Santa Clara; later embellished with multi-firing while I was at Dickson Electronics, Scottsdale). Be a piece-a-cake now with an appropriate NMOS. And it doesn't need a plug to survive... might toast the coil, though >:-} ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | STV, Queen Creek, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | Thinking outside the box... producing elegant solutions, by understanding what nature is hiding. "It is not in doing what you like, but in liking what you do that is the secret of happiness." -James Barrie
s&oslash;ndag den 29. april 2018 kl. 22.00.39 UTC+2 skrev Jim Thompson:
> On Sun, 29 Apr 2018 12:18:32 -0700 (PDT), Lasse Langwadt Christensen > <langwadt@fonz.dk> wrote: > > >s&oslash;ndag den 29. april 2018 kl. 19.23.11 UTC+2 skrev Jim Thompson: > >> On Sun, 29 Apr 2018 09:52:21 -0700, Joerg <news@analogconsultants.com> > >> wrote: > >> > >> >On 2018-04-28 15:16, Tim Williams wrote: > >> >> "Joerg" <news@analogconsultants.com> wrote in message > >> >> news:fkk3gpFmm06U1@mid.individual.net... > >> >>> There are plenty of cheap MOSFETs that are in the low tens of > >> >>> milliohms and thus show only a fraction of the losses while the coil > >> >>> current ramps > >> >>> up. > >> >> > >> >> At what Vds(max)? ;-) > >> >> > >> > > >> >You don't need much above 400V for an ignition. Easy to do with either kind. > >> > >> 350V/5A is the general requirement, so I'd probably margin a wee-bit > >> higher than 400V, maybe 600V. > > > >most ignition igbt have an active clamp around 350-400V to protect the coil > >if there's a loose plug wire > > > > I don't do "most", I roll my own, based on "CD-Ignition-Basic.pdf" on > the S.E.D/Schematics Page of my website. I was running experimental > Ford Thunderbirds with "cruise control" back in the late '60's using > that ignition (at Philco-Ford, Santa Clara; later embellished with > multi-firing while I was at Dickson Electronics, Scottsdale). Be a > piece-a-cake now with an appropriate NMOS. And it doesn't need a plug > to survive... might toast the coil, though >:-}
protecting the coil is the point of the clamp
On Saturday, April 28, 2018 at 1:22:54 PM UTC-7, Joerg wrote:
> Considering the typical saturation voltages of >1V it doesn't look > attractive yet they are popular: > > https://www.fairchildsemi.com/datasheets/FG/FGD3040G2_F085.pdf > > Why?
Aggressive marketing, obviously. Take the typical IROC (right-out-of-college) engineer, assign it to make an ignition, and Google will pull him into a maze of IGBT datasheets. I'm fond of the old Mark10 design (Heathkit, too) with capacitive discharge, which works fine with forty-year-old SCRs. My car has a potted mystery module and was designed with a LONG harness connection to the coil and reluctor sensor. There was no active intelligence used in the parts of the design that I can see.
On Sun, 29 Apr 2018 13:17:05 -0700 (PDT), Lasse Langwadt Christensen
<langwadt@fonz.dk> wrote:

>s&#2013266168;ndag den 29. april 2018 kl. 22.00.39 UTC+2 skrev Jim Thompson: >> On Sun, 29 Apr 2018 12:18:32 -0700 (PDT), Lasse Langwadt Christensen >> <langwadt@fonz.dk> wrote: >> >> >s&#2013266168;ndag den 29. april 2018 kl. 19.23.11 UTC+2 skrev Jim Thompson: >> >> On Sun, 29 Apr 2018 09:52:21 -0700, Joerg <news@analogconsultants.com> >> >> wrote: >> >> >> >> >On 2018-04-28 15:16, Tim Williams wrote: >> >> >> "Joerg" <news@analogconsultants.com> wrote in message >> >> >> news:fkk3gpFmm06U1@mid.individual.net... >> >> >>> There are plenty of cheap MOSFETs that are in the low tens of >> >> >>> milliohms and thus show only a fraction of the losses while the coil >> >> >>> current ramps >> >> >>> up. >> >> >> >> >> >> At what Vds(max)? ;-) >> >> >> >> >> > >> >> >You don't need much above 400V for an ignition. Easy to do with either kind. >> >> >> >> 350V/5A is the general requirement, so I'd probably margin a wee-bit >> >> higher than 400V, maybe 600V. >> > >> >most ignition igbt have an active clamp around 350-400V to protect the coil >> >if there's a loose plug wire >> > >> >> I don't do "most", I roll my own, based on "CD-Ignition-Basic.pdf" on >> the S.E.D/Schematics Page of my website. I was running experimental >> Ford Thunderbirds with "cruise control" back in the late '60's using >> that ignition (at Philco-Ford, Santa Clara; later embellished with >> multi-firing while I was at Dickson Electronics, Scottsdale). Be a >> piece-a-cake now with an appropriate NMOS. And it doesn't need a plug >> to survive... might toast the coil, though >:-} > >protecting the coil is the point of the clamp > >
Actually I was thinking dissipation in the coil primary with no load. My scheme will apply 350V to the primary, period, secondary loaded or not; and the switch device only sees 350V, irrespective of load. ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | STV, Queen Creek, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | Thinking outside the box... producing elegant solutions, by understanding what nature is hiding. "It is not in doing what you like, but in liking what you do that is the secret of happiness." -James Barrie
On 2018-04-29 15:08, Jim Thompson wrote:
> On Sun, 29 Apr 2018 13:17:05 -0700 (PDT), Lasse Langwadt Christensen > <langwadt@fonz.dk> wrote: > >> s&#2013266168;ndag den 29. april 2018 kl. 22.00.39 UTC+2 skrev Jim Thompson: >>> On Sun, 29 Apr 2018 12:18:32 -0700 (PDT), Lasse Langwadt Christensen >>> <langwadt@fonz.dk> wrote: >>> >>>> s&#2013266168;ndag den 29. april 2018 kl. 19.23.11 UTC+2 skrev Jim Thompson: >>>>> On Sun, 29 Apr 2018 09:52:21 -0700, Joerg <news@analogconsultants.com> >>>>> wrote: >>>>> >>>>>> On 2018-04-28 15:16, Tim Williams wrote: >>>>>>> "Joerg" <news@analogconsultants.com> wrote in message >>>>>>> news:fkk3gpFmm06U1@mid.individual.net... >>>>>>>> There are plenty of cheap MOSFETs that are in the low tens of >>>>>>>> milliohms and thus show only a fraction of the losses while the coil >>>>>>>> current ramps >>>>>>>> up. >>>>>>> >>>>>>> At what Vds(max)? ;-) >>>>>>> >>>>>> >>>>>> You don't need much above 400V for an ignition. Easy to do with either kind. >>>>> >>>>> 350V/5A is the general requirement, so I'd probably margin a wee-bit >>>>> higher than 400V, maybe 600V. >>>>
I'd do that as well although the usual zener from collector to base or drain to gate will rein in anything that wants to go past zener voltage plus Vbe or gate threshold.
>>>> most ignition igbt have an active clamp around 350-400V to protect the coil >>>> if there's a loose plug wire >>>> >>> >>> I don't do "most", I roll my own, based on "CD-Ignition-Basic.pdf" on >>> the S.E.D/Schematics Page of my website. I was running experimental >>> Ford Thunderbirds with "cruise control" back in the late '60's using >>> that ignition (at Philco-Ford, Santa Clara; later embellished with >>> multi-firing while I was at Dickson Electronics, Scottsdale). Be a >>> piece-a-cake now with an appropriate NMOS. And it doesn't need a plug >>> to survive... might toast the coil, though >:-} >> >> protecting the coil is the point of the clamp >>
There also needs to be a current limit for coil protection against melt-down.
> > Actually I was thinking dissipation in the coil primary with no load. > My scheme will apply 350V to the primary, period, secondary loaded or > not; and the switch device only sees 350V, irrespective of load. >
That is what IIRC most capacitive discharge ignition circuits do. AFAIK that comes at the price of a shorter spark versus inductive. -- Regards, Joerg http://www.analogconsultants.com/
On 2018-04-29 12:19, Lasse Langwadt Christensen wrote:
> l&oslash;rdag den 28. april 2018 kl. 22.22.54 UTC+2 skrev Joerg: >> Considering the typical saturation voltages of >1V it doesn't look >> attractive yet they are popular: >> >> https://www.fairchildsemi.com/datasheets/FG/FGD3040G2_F085.pdf >> >> Why? >> >> There are plenty of cheap MOSFETs that are in the low tens of milliohms >> and thus show only a fraction of the losses while the coil current ramps >> up. You'd just have to add a zener from drain to gate in case the spark >> plug goes bad, the wire comes off or the controller lets the dwell time >> run out of hand. Plus some other gate protection but that's no problem. >> > > http://www.st.com/content/ccc/resource/technical/document/application_note/8d/46/36/a7/b6/f2/45/10/CD00003911.pdf/files/CD00003911.pdf/jcr:content/translations/en.CD00003911.pdf > >
Thanks, I have read papers like that. On of the things that puzzles me is evident in figure 10, over 1V in Vcesat. A FET can run circles around that and run with much less cooling, meaning an overall smaller module, less weight. -- Regards, Joerg http://www.analogconsultants.com/
On 2018-04-29 13:24, whit3rd wrote:
> On Saturday, April 28, 2018 at 1:22:54 PM UTC-7, Joerg wrote: >> Considering the typical saturation voltages of >1V it doesn't look >> attractive yet they are popular: >> >> https://www.fairchildsemi.com/datasheets/FG/FGD3040G2_F085.pdf >> >> Why? > > Aggressive marketing, obviously. Take the typical IROC (right-out-of-college) > engineer, assign it to make an ignition, and Google will > pull him into a maze of IGBT datasheets. >
I remember the days when IGBTs came out and lots of less experienced used them for just about anything. Their designs worked but were expensive in production.
> I'm fond of the old Mark10 design (Heathkit, too) with capacitive discharge, > which works fine with forty-year-old SCRs. > > My car has a potted mystery module and was designed with a > LONG harness connection to the coil and reluctor sensor. > > There was no active intelligence used in the parts of the design that I can see. >
My Citroen had a really cool system. The coil's secondary fed both plugs in its two-cylinder engine (wasted spark method) which meant no distributor. It also meant no distributor could go bad. However, it could make for interesting effects when hand-cranking it on a cold day. You had to make sure no white or particularly clean car war parked behind and nobody stood near the exhaust. Once the bang was so bad that the 2nd muffler pot fell off and started rolling down the street. Embarrassing. -- Regards, Joerg http://www.analogconsultants.com/
On Sun, 29 Apr 2018 13:24:39 -0700 (PDT), whit3rd <whit3rd@gmail.com>
wrote:

>Aggressive marketing, obviously. Take the typical IROC (right-out-of-college) >engineer, assign it to make an ignition, and Google will >pull him into a maze of IGBT datasheets.
You could not be more wrong. I'm still an engineering consultant to Ford's Dearborn engine development lab so I know a thing or two about how things are done. Ford (or more likely their captive contractor Delco) doesn't go out to the marketplace and buy something that is close to what they need. When tens of millions of a part are involved, they develop a functional specification for a part that is EXACTLY what they need and put that out for bid. The respondents are first evaluated as to their ability to handle the volume. Then Delco invites chip makers to submit a bid and some sample parts. The sample parts are tested to standards that few other industries require. With millions of parts involved, even a 0.0001% failure rate would cause a massive recall and massive bad publicity. They'll shave pennies in other areas. Eliminating packaging and using a COB die is now common. The finished assembly, say a coil-on-plug is also massively tested under extraordinary conditions, both in the lab and on the road. Ford has a proving ground facility in Greenland inside the Arctic circle for cold weather testing. A similar facility is located at a desert equatorial location (I forget where) for hot weather testing. After the assembly is accepted for production, the finished product is statistically sampled continuously to verify that performance remains the same throughout the part's life. The chip foundry, being geared up for very high production, realizes that the part can be useful for other things and so markets it outside the industry and at a low price. And BTW, a newly graduated engineer is never assigned "go design a widget" project. Newly hired engineers undergo an apprenticeship that lasts about 4 years. During that time they learn to do things the automotive way. Only seniority and accomplishments leads to more complex projects. One of my very good MechE friends at Ford is about to retire. He has told me that he was in his 30s before he was assigned a significant design project. John John DeArmond http://www.neon-john.com http://www.tnduction.com Tellico Plains, Occupied TN See website for email address