rusefi.com

So, someone is editing my posts?

I don't mind if you do it, in so much as you let me know you are doing it.

abecedarian wrote:So, someone is editing my posts?

I don't mind if you do it, in so much as you let me know you are doing it.

edited_posts.png

Yes, I did correct your typo after asking you if that was a typo. As a matter of fact, I was a bit surprised you did not edit your own post before I did.

russian wrote:

abecedarian wrote:So, someone is editing my posts?

I don't mind if you do it, in so much as you let me know you are doing it.

edited_posts.png

Yes, I did correct your typo after asking you if that was a typo. As a matter of fact, I was a bit surprised you did not edit your own post before I did.

You didn't ask me if it was a typo.
I only realized the typo after you telling me Google didn't know that part number.


Can we move along now?

abecedarian wrote:Can we move along now?

We should

So why include many IGBT's on the board for someone who...
... likely has a distributor and likely only has one ignition coil because of that?


Me thinks trying to solve TOO MANY things with one solution only convolutes, and dilutes, what that solution can be.

Perhaps:
- one RUSEFI board for Kettering / single igniter + non-EFI
- one RUSEFI board for Kettering / single igniter + EFI
- one ....

have I made my point?


If there was really one ring to rule them all, no one would have problems, no?

abecedarian wrote:So why include many IGBT's on the board for someone who...
... likely has a distributor and likely only has one ignition coil because of that?

Me thinks trying to solve TOO MANY things with one solution only convolutes, and dilutes, what that solution can be.

We might be on the same page. I am envisioning that each IGBT would be on a tiny board which is soldered on the prototyping area. EGT chips go away from Frankenso and they are either replaced by an EGT solder-on board or external EGT2CAN device. MAP is a tiny solder-on board. No idea if that would stay this way forever, but for that's probably the best way to go. All the random stuff just does not fit on the same board, regardless of the size.

PS: No idea what Kettering is.

I'm picturing an add on PCB that goes above those W jumpers. This would allow you to choose an add on module to customize as required for your application.

Ah I see D12573190 doesn't exist, then 12573190 is a reasonably common LS1 coil near plug device with internal ignitor.

kb1gtt wrote:I'm picturing an add on PCB that goes above those W jumpers. This would allow you to choose an add on module to customize as required for your application.

We cannot drive an IGBT directly from ARM, right? So an IGBT always runs with a hi-side driver, right? Does it mean the hi-side driver should be on the same extension board?

How small can we go with the hi-side driver which drives the IGBT - can we use SOT23?

IGBT's have high side drive and low side drive just like MOSFET's. The low side with a +V logic level are the most common, just like a MOSFET. Also just like a MOSFET, they do have benefits of being driven hard to +V and GND. The short way to explain this, is that an IGBT is basically a MOSFET that drives a transistor. This is why a MOSFET and IGBT are very similar.

A key feature of an IGBT is that it offers better isolation than a MOSFET, which saves your STM32 from getting blown away. The trade off is that IGBT's typically make more heat, and they are typically slower acting. From the software side of things, you can generally think of a MOSFET and IGBT as being the same.

Any how, these mentioned parts are driven with a 0V to 5V signal. Items like the J701 could be different inside.

kb1gtt wrote:Any how, these mentioned parts are driven with a 0V to 5V signal

Which stm32 cannot offer. In my test I was driving the IGBT via a TC4427 with VP=12v
So in our case we need 5V thus we need something to make 5V out of 3V, some SO23 thingy?

On state voltage for FGB3440G2_F085 is a max of 1.75V. The on state voltage for AUIRGS30B60K is a max of 2.95. So both of those chips should be able to be driven by 3V.

I think you are thinking of the J701, which has a IGBT inside, as well it has other stuff. The other stuff is what I think is causing the bumps in the carpet.

kb1gtt wrote:IGBT's have high side drive and low side drive just like MOSFET's. The low side with a +V logic level are the most common, just like a MOSFET. Also just like a MOSFET, they do have benefits of being driven hard to +V and GND. The short way to explain this, is that an IGBT is basically a MOSFET that drives a transistor. This is why a MOSFET and IGBT are very similar.

A key feature of an IGBT is that it offers better isolation than a MOSFET, which saves your STM32 from getting blown away. The trade off is that IGBT's typically make more heat, and they are typically slower acting. From the software side of things, you can generally think of a MOSFET and IGBT as being the same.

Any how, these mentioned parts are driven with a 0V to 5V signal. Items like the J701 could be different inside.

The LSx coils take a ground and 12v via relay, preferably charged by the ECU- they do not draw much current until the "trigger" is pulled high since the contain THEIR OWN IGBT / IGNITER.

When a TTL logic HIGH is applied to the trigger, that signal causes the coil to charge, and when the trigger goes TTL LOW, the coil fires- due to the size of the coil and such, it only takes around 5-6mS to charge the coil for firing, and if you hold the trigger high more than around 7.5-8ms, it will self-discharge, and that 'duration' is somewhat random so you can't just take the coil high and wait for it to spark.

Last I heard, the LSx coils will fire with a 3v3 signal on their trigger, which is in compliance with the other things I heard about TTL signal levels being able to trigger them, so no logic level translation should be required... probably because they need to trigger under some low-temp/low-volt situations.

If you really need translation, a TXS0102(two channel)/TXS0104(4 channel)/TXS0108(8 channel) translator would probably work, and would also allow you to pull signals externally to some safe range, or tri-state inputs, so as to prevent things from actuating.

These LSx style coils are the original reason why the Frank boards have the TC4427 HI/LO drivers. That chip allows you to drive to 12V and GND, which works well for external ignitors like this, or if you have an old school coil, you can also drive things like the J701. I haven't seen an external ignitor we can't drive. I believe that also covers automotive style CDI.

I agree that this LCx coils are good for when you an do the coil near plug approach. Good fire power and good dwell times.

kb1gtt wrote:These LSx style coils are the original reason why the Frank boards have the TC4427 HI/LO drivers. That chip allows you to drive to 12V and GND, which works well for external ignitors like this, or if you have an old school coil, you can also drive things like the J701. I haven't seen an external ignitor we can't drive. I believe that also covers automotive style CDI.

I agree that this LCx coils are good for when you an do the coil near plug approach. Good fire power and good dwell times.

You don't drive LSx coils with 12v.
LSx coils typically have 4 connections (not really in any particular order as different coils may have different functions at terminal locations):
1- 12v, switched.
2- trigger signal.
3- ground, common with ECU.
4- ground, common to engine block / battery.

The trigger is TTL IIRC, so should be compatible with either 5v or 3v3 MCU logic levels. The MCU pulls "HIGH" to start the coil "charge", then when the "HIGH" is released, fire the coil; time spent "HIGH" sets the dwell, but coils also self-limit and fire after ~8ms "HIGH"... probably a good idea for a weak pull-down to ground to keep current-sink through the MCU to a minimum, i.e. the MCU can pull "HIGH" to charge the coil, then go low, with the pull-downs sinking current to ground.


CDI is a totally different thing compared to inductive discharge ignition, so I won't comment regarding the efficacy of what's already been done.

The Frankenso board allows you to select the + rail. There is a jumper which makes it reasonably easy to select if you are driving to 5V or 12V. It also can be connector to another V source with a jumper wire. Perhaps it was the HEI's that wanted the 12V drive. I forget which is which. I just know that for external ignitors, we have a flexibility that allows for almost any external igntor to be used.

Been a while since I've really delved into the ignition systems on GM, Ford or Chrysler for that matter.
To the best of my recollection, none of them drove the (+) 12v to the coil; they all switched the (-) ground or provided a logic-type level signal to trigger an igniter.


My '89 Celica, '91 Toyota pickup, '88 4Runner, '72 Firebird, '78 and '81 VW Scirocco's, '78 Type II Transporter, '81 CX500TC ... all switch the coil ground to fire the coil.
I'd have to do some more research but IIRC even the '73 Coronet (318 CID), '66 Mustang (200 CID), '72 Dart (225 CID) and '78 HiLux (2189 CC) I had access to growing up switched the ground.

So, why provide something like IGBT or MOSFET giving 12v out? ... or am I missing something?

abecedarian wrote:So, why provide something like ... MOSFET giving 12v out?

12v out is used to drive external logic-level ignition modules (not coils, but devices before the coil): Mazda Miata (integrated with the coil), Ford Festiva (J701), Mazda Protege (J701) Honda Accord (inside the distributor module)

Why is it so hard to find protected MOSFET's?

I've been looking for a P-Channel, thermally protected, 10amp or more, enhanced mode MOSFET, DPAK or D2PAK, preferably with a voltage clamp protection. However so far no dice.

russian wrote:

abecedarian wrote:So, why provide something like ... MOSFET giving 12v out?

12v out is used to drive external logic-level ignition modules (not coils, but devices before the coil): Mazda Miata (integrated with the coil), Ford Festiva (J701), Mazda Protege (J701) Honda Accord (inside the distributor module)

I'm still not getting it.

kb1gtt wrote:Why is it so hard to find protected MOSFET's?

I've been looking for a P-Channel, thermally protected, 10amp or more, enhanced mode MOSFET, DPAK or D2PAK, preferably with a voltage clamp protection. However so far no dice.

NTF6P02? IRF9540? ... Oh, wait. You said thermally protected too, so maybe not.

What system in any vehicle, other than powering the starter motor or fuel pump directly, requires 10A current?

abecedarian wrote:I'm still not getting it.

provided a logic-type level signal to trigger an igniter.

J701 is a popular igniter which is triggered by +12v.

russian wrote:

abecedarian wrote:I'm still not getting it.

provided a logic-type level signal to trigger an igniter.

J701 is a popular igniter which is triggered by +12v.

I guess that would make it a "not logic level" triggered device?
But just a quick check on my side suggests it's powered by 12v on the black / yellow and white / blue wires (one to coil, other to igniter) and switched ground to trigger on the white wire- the only wire going back to the ECU... at least on a Honda. http://www.tegger.com/hondafaq/igniter-operation/igniter-on-car-test.pdf

abecedarian wrote:I guess that would make it a "not logic level" triggered device?

Absolutely. I was just trying to provide a closest similar.

Can you check 1994 Mazda Miata? The igniter is inside the coil pack.

russian wrote:

abecedarian wrote:I guess that would make it a "not logic level" triggered device?

Absolutely. I was just trying to provide a closest similar.

No worries. Like I said, I may be missing something.

Can you check 1994 Mazda Miata? The igniter is inside the coil pack.

Have one I can borrow for a while?

There are people out there that need 12V drive for external ignitors, and we have it. abecedarian is not one of them, so I don't see much reason for continuing the discussion. We are looking to drive an ignitor that's internal to rusEFI, as many existing wiring harness do it that way. Can we pull the discussion over to doing this with an internal ignitor?

kb1gtt wrote:Why is it so hard to find protected MOSFET's?

I've been looking for a P-Channel, thermally protected, 10amp or more, enhanced mode MOSFET, DPAK or D2PAK, preferably with a voltage clamp protection. However so far no dice.

You should not choose a discrete MOSFET as a driver for a signal or source going out from an ECU, that is considered potentially dangerous in the automotive industry since a fault in a component or the harness must not cause a potential fire-hazard which a discrete driver has no ability to comply (a consideration in this case when people, most likely, are going to hack into their own wiring harness). So the reason you having difficulties finding such a device is most likely because there are device-alternatives matching the robustness and compliance criteria you need. Consider using Automotive-graded HSS/LSS for this, they are designed for this exact purpose makes your life easier, have HW-detected current cut-off and gives the possibility of SW diagnostics as well (most likely overkill here, but if one should want to do that in the future...).

Example of ST's HSS-devices (you should find devices up to 100Amp without problems):
http://www.st.com/web/en/catalog/sense_power/FM1965/SC1037

If anyone is interested: The implementation in an OEM ECU is to relay on the internal current cut-off of the HSS/LSS for electrical issues, mainly to save the driver itself or other intra-ECU HW. But in addition to this you have a SW current characteristic profile (see it as the expected behavior of the actuator) this is continuously compared with the HSS-internal current measured feedback. If the true value goes outside the expected, the µC triggers the SW fail safe routine which is usually enough to prevent hazards such as harness damage caused by heat build up before the HW-cut-out is triggered. But both of these triggering conditions can of course be individually detected by the µC which sets diagnostics and possibly triggers fault recovery or limp-home state.

Thanks for the link and info. I have looked over those and posted my technical reply in this thread http://rusefi.com/forum/viewtopic.php?f=4&t=747&start=49 I was a bit miss-placed when I replied in this thread, as we have a high side drive that works for external ignitors, I was in the wrong forum when I posted that reply about high current high side drive. Oops.

Also I have added a note to the "general suggested environments" wiki found here http://rusefi.com/wiki/index.php?title=Manual:Hardware#General_suggested_environment

russian wrote:Infineon BTS2140 is Obsolete, but still available for about $1.20 from China.

BTS2140-1B compatible

russian wrote:in the middle on the back we have two coil drivers mounted on the prototyping area.

Its better to use an external igniter, the coil signal is very noisy, a typical ignition IGBT clamps at around 400-450V. I don't know of any OEM ECUs that have the igniters internal.

(I've moved to what I believe is a more relevant topic)

Rhinoman wrote:I don't know of any OEM ECUs that have the igniters internal.

My point is to use stock Dodge Neon 2003 ecu harness completely stock - I believe OEM has igniters inside ECU, see http://rusefi.com/wiki/index.php?title=Vehicle:Dodge_Neon_2003

Same story with my long gone http://rusefi.com/wiki/index.php?title=Vehicle:Mini_Cooper_2003

I believe BTS2140 was originally used by Bosch ECU like this one: