Page 3 of 4

Re: Good non surface mount MOSFET to drive injectors?

Posted: Thu Jan 09, 2014 5:27 am
by nicholasjanssen
http://www.megamanual.com/ms2/pickups.htm
Image
A Hall effect sensor is an "active", magnetic field presence sensor. It is based on the Hall effect. The principle was discovered in 1879. When Edwin H. Hall applied an current to a piece of metal inserted between two magnets, he found it created a secondary voltage in the metal at a right angle to the applied voltage. In the sensors we are concerned with, the Hall effect is used to change of resistance in a semiconductor in a magnetic field, and this is then used to switch an output voltage from high to low, or vice versa.

The Hall effect sensor consists of semiconductor material which will conduct current when the material is subject to a magnetic field. These types of sensors require a "flying magnet", wheel. Instead of teeth on the wheel, as in a variable reluctor sensor, you must have small magnet and a shutter wheel.

Re: Good non surface mount MOSFET to drive injectors?

Posted: Thu Jan 09, 2014 5:28 am
by nicholasjanssen
Fred wrote:I already explained it as much as I'm going to. It's not a problem at all. Your signal type is not supported...

You could hack it, but even if you do, it'll never run well. Only mediocre running is possible without significant changes/work. You already have a huge job ahead of you to make it work at all. Hence why bother? Fit a crank wheel to it and get it running.

ok I understand. thanks very much though you have been very helpful. I will continue working and investigate further.

Re: Good non surface mount MOSFET to drive injectors?

Posted: Thu Jan 09, 2014 5:41 am
by Fred
Your link is broken.

Re: Good non surface mount MOSFET to drive injectors?

Posted: Thu Jan 09, 2014 5:52 am
by nicholasjanssen
Fred wrote:Your link is broken.
it is broken.... very strange.. let me test again

http://www.megamanual.com/ms2/pickups.htm
it might be possible they are blocking any incomming hits from our website?

anyhow i was trying to find "even teeth in the source code"

Code: Select all

                if(KeyUserDebugs.decoderFlags & CONFIGURED_SYNC){
                        KeyUserDebugs.currentEvent++;
                        if(KeyUserDebugs.currentEvent == numberOfRealEvents){
                                resetToNonRunningState(COUNT_OF_EVENTS_IMPOSSIBLY_HIGH_NOISE);
                                return;
aparently there must be a sync if some sort to prevent ?? flawed random noise that will cause the engine to be impossible to tune??

thus the sync is a safety feature that prevents noise on an -even- tooth - no missing tooth- hall effect sensor from being misread at any and all RPM levels..

?


This "even tooth without sync" issue might be fixed by me upgrading to a 1993 *REAL EFI VERSION* distributor for the exact sane engine size and displacement. I will try to research the 1993 distributor as a posible solution.. ( if any of my guesswork above is correct)

Re: Good non surface mount MOSFET to drive injectors?

Posted: Thu Jan 09, 2014 6:19 am
by nicholasjanssen
* This decoder is for any 4 tooth/slot hall or optical cam speed sensor and to
* be used for distributor and/or 4 shot batch injection only.

HallOrOptical-Distributor-XofY.c

i will try to modify this file to work someday :(

Re: Good non surface mount MOSFET to drive injectors?

Posted: Thu Jan 09, 2014 3:34 pm
by malcom2073
nicholasjanssen wrote:Actually, I am very curious about the idea that it cannot work currently. Is it a code or hardware problem? I am not very knowledgeable as to exactly what is wrong with the signal it provides. I ordered the ecu board already, and it should arrive soon. I will fix it, because i am certain if a 25+ year old ecu can read the signal so will a 2014 board, even if it requires hackery.

worse case I will use a 1995 distributor which will have a "different better signal"
It's a code problem. Your old ECU is capable of reading it because it's designed to. FreeEMS firmware is not designed to support that style of pickup, it's designed for more modern means (missing tooth wheel).
nicholasjanssen wrote:aparently there must be a sync if some sort to prevent ?? flawed random noise that will cause the engine to be impossible to tune??

thus the sync is a safety feature that prevents noise on an -even- tooth - no missing tooth- hall effect sensor from being misread at any and all RPM levels..

?

This "even tooth without sync" issue might be fixed by me upgrading to a 1993 *REAL EFI VERSION* distributor for the exact sane engine size and displacement. I will try to research the 1993 distributor as a posible solution.. ( if any of my guesswork above is correct)
"Sync" with reference to the crank means to know if cylinder 1 is TDC (on either compression or exhaust, doesn't matter which).
"Sync" with reference to the cam means to know if cylinder 1 is TDC on the compression stroke.

You have a very similar setup to my car which has a chevy distributor providing 8 pulses per engine cycle (4 per crank revolution). What this means is that the ECU doesn't actually know when cylinder 1 is TDC, it just knows that ONE of the cylinders is TDC, the distributor handles spark, and injection doens't care, it fires batch fire.

With current decoders, you need some sort of sync. Either a cam or crank. This is typically done via a missing tooth in the crank wheel. The reason why your stock ECU could do this, is that it didn't worry about "syncing", that is it didn't know which cylinder was TDC, and it didn't care. It is possible to wire injectors to batch/bank fire in such a way that it does not matter (fire the TDC compression and TDC exhaust cylinders together), but most batch/bank fire fuel injection setups fire left/right rather than that, and will provide different fueling based on how you start the car (It could be normal one start, and 90 degrees out of phase the next).

In addition to the issues that a lack of sync brings up, providing 8 pulses per cycle means that there are 90 degrees of crank rotation between pulses. This means that timing is mediocre at best, whereas a 24 tooth wheel would give you 15 degrees between pulses, providing for a much more stable timing base for ignition/injection events.

nicholasjanssen wrote:HallOrOptical-Distributor-XofY.c

i will try to modify this file to work someday :(
I've modified a missing tooth decoder to work with my engine (syncless), but it is in no way optimal. Do yourself a favor and get pickup with some sort of sync ability (missing tooth wheel, or distributor with a cam sync). That will make life much easier for you. Check into the 1993 distributor see if it has something like this. If not, fitting a toothed wheel isn't really that difficult.

Re: Good non surface mount MOSFET to drive injectors?

Posted: Thu Jan 09, 2014 6:02 pm
by nicholasjanssen
malcom2073 wrote:
nicholasjanssen wrote:Actually, I am very curious about the idea that it cannot work currently. Is it a code or hardware problem? I am not very knowledgeable as to exactly what is wrong with the signal it provides. I ordered the ecu board already, and it should arrive soon. I will fix it, because i am certain if a 25+ year old ecu can read the signal so will a 2014 board, even if it requires hackery.

worse case I will use a 1995 distributor which will have a "different better signal"
It's a code problem. Your old ECU is capable of reading it because it's designed to. FreeEMS firmware is not designed to support that style of pickup, it's designed for more modern means (missing tooth wheel).
nicholasjanssen wrote:aparently there must be a sync if some sort to prevent ?? flawed random noise that will cause the engine to be impossible to tune??

thus the sync is a safety feature that prevents noise on an -even- tooth - no missing tooth- hall effect sensor from being misread at any and all RPM levels..

?

This "even tooth without sync" issue might be fixed by me upgrading to a 1993 *REAL EFI VERSION* distributor for the exact sane engine size and displacement. I will try to research the 1993 distributor as a posible solution.. ( if any of my guesswork above is correct)
"Sync" with reference to the crank means to know if cylinder 1 is TDC (on either compression or exhaust, doesn't matter which).
"Sync" with reference to the cam means to know if cylinder 1 is TDC on the compression stroke.

You have a very similar setup to my car which has a chevy distributor providing 8 pulses per engine cycle (4 per crank revolution). What this means is that the ECU doesn't actually know when cylinder 1 is TDC, it just knows that ONE of the cylinders is TDC, the distributor handles spark, and injection doens't care, it fires batch fire.

With current decoders, you need some sort of sync. Either a cam or crank. This is typically done via a missing tooth in the crank wheel. The reason why your stock ECU could do this, is that it didn't worry about "syncing", that is it didn't know which cylinder was TDC, and it didn't care. It is possible to wire injectors to batch/bank fire in such a way that it does not matter (fire the TDC compression and TDC exhaust cylinders together), but most batch/bank fire fuel injection setups fire left/right rather than that, and will provide different fueling based on how you start the car (It could be normal one start, and 90 degrees out of phase the next).

In addition to the issues that a lack of sync brings up, providing 8 pulses per cycle means that there are 90 degrees of crank rotation between pulses. This means that timing is mediocre at best, whereas a 24 tooth wheel would give you 15 degrees between pulses, providing for a much more stable timing base for ignition/injection events.

nicholasjanssen wrote:HallOrOptical-Distributor-XofY.c

i will try to modify this file to work someday :(
I've modified a missing tooth decoder to work with my engine (syncless), but it is in no way optimal. Do yourself a favor and get pickup with some sort of sync ability (missing tooth wheel, or distributor with a cam sync). That will make life much easier for you. Check into the 1993 distributor see if it has something like this. If not, fitting a toothed wheel isn't really that difficult.
Thanks, I <3 more information. From what I gather, the 25yo ecu can detect a "sync" to cylinder 1 because it can measure the "smaller" tooth out hf the 6 of the teeth. I understand the firmware must be written to accomplish this task, and I have no problem having a go at it.

I am more curious if it is possible for the AD9S12XDP512M2 to have the ability to time events. (code suggests yes). The other factor I am also curious about is the quality and strength of the signal. From the schematics posted. This PIP from EEC-IV isn't a direct feed from the HALL sensor, It is actually an amplified signal sent to the very old school ECU. If this amplified signal is as strong as I hope, and If it is as clean as I hope, I will use it as is. I will run a coax wire directly from the PIP output to a voltage divider/capactior(to clean the signal??) and then read those pulses without sync in the ecu, hopefully being able to log tons of data at the same time.

If the data logged is good, I will then attempt to time the short tooth/vane. I will gather max times for the vane, how it works at 4000rpm ( redline for the engine) and min times for the vane. If the signal is clean and strong, I very much hope to be able to accurately detect the "thin" vane.

thats my plan at least

Re: Good non surface mount MOSFET to drive injectors?

Posted: Thu Jan 09, 2014 9:23 pm
by malcom2073
nicholasjanssen wrote:Thanks, I <3 more information. From what I gather, the 25yo ecu can detect a "sync" to cylinder 1 because it can measure the "smaller" tooth out hf the 6 of the teeth. I understand the firmware must be written to accomplish this task, and I have no problem having a go at it.
Oh excellent, I totally missed that in the picture above, that's way better, and that hardware should work for you if you write a decoder to support it
nicholasjanssen wrote:I am more curious if it is possible for the AD9S12XDP512M2 to have the ability to time events. (code suggests yes). The other factor I am also curious about is the quality and strength of the signal. From the schematics posted. This PIP from EEC-IV isn't a direct feed from the HALL sensor, It is actually an amplified signal sent to the very old school ECU. If this amplified signal is as strong as I hope, and If it is as clean as I hope, I will use it as is. I will run a coax wire directly from the PIP output to a voltage divider/capactior(to clean the signal??) and then read those pulses without sync in the ecu, hopefully being able to log tons of data at the same time.
The decoders are based on interrupts that trigger on either rising edge, falling edge, or both. Being that there are only 8 pulses, I'm sure there's enough time for you to calculate the length of the tooth, and use that to figure out which tooth is the TDC tooth.

Re: Good non surface mount MOSFET to drive injectors?

Posted: Thu Jan 09, 2014 10:43 pm
by Fred
There is no such thing as "syncless" in FreeEMS!! There is a CylinderSync that knows where within a cylinder's slice it is, and not which cylinder that is. If you remove the concept of sync, do not post a thread in the vehicles section, as it's not FreeEMS anymore at that point. FreeEMS is/will be a mark of quality, and a syncless setup is no such thing. Mike, your hacked decoder still has sync! Which type of sync you set, and whether it's correct and accurate or not, is up to your hacking. Likely you left it wrong. Missing tooth is wrong on cam applications too, FYI.

You can't use XofY because you don't have X of Y you have X-1 of Y and 1 of Z. Yes you could write a decoder for this, but it's not as easy as you think it is. Time based stuff like this requires a very distinct pattern to have good confidence. The scope trace above shows a very subtle difference with no way of knowing which is which until AFTER it's running. Thus the ECU is injecting fuel with the dizzy in control and then switching from batch to sequential later. I've done schedule changing on the fly, but it's non-trivial and fraught with danger.

Unless the later distributor has a sophisticated pattern within, you're likely wasting your time installing it.

Re: Good non surface mount MOSFET to drive injectors?

Posted: Thu Jan 09, 2014 11:11 pm
by nicholasjanssen
Fred wrote:There is no such thing as "syncless" in FreeEMS!! There is a CylinderSync that knows where within a cylinder's slice it is, and not which cylinder that is. If you remove the concept of sync, do not post a thread in the vehicles section, as it's not FreeEMS anymore at that point. FreeEMS is/will be a mark of quality, and a syncless setup is no such thing. Mike, your hacked decoder still has sync! Which type of sync you set, and whether it's correct and accurate or not, is up to your hacking. Likely you left it wrong. Missing tooth is wrong on cam applications too, FYI.

You can't use XofY because you don't have X of Y you have X-1 of Y and 1 of Z. Yes you could write a decoder for this, but it's not as easy as you think it is. Time based stuff like this requires a very distinct pattern to have good confidence. The scope trace above shows a very subtle difference with no way of knowing which is which until AFTER it's running. Thus the ECU is injecting fuel with the dizzy in control and then switching from batch to sequential later. I've done schedule changing on the fly, but it's non-trivial and fraught with danger.

Unless the later distributor has a sophisticated pattern within, you're likely wasting your time installing it.
Fred again you show why you are the best. I looked through the code, and it is all not only amazing, but organized also!

I very much want to do it correctly.
Would it be possible to fabricate more teeth into a spare decoder by cutting notches?
(i realize this will require a switch to COIL pack or even coil on plugs)?

Is more teeth better, or would just maybe the "thin" tooth removed?