Coil Dwell Disassem and Denso Coil Behavior
#1
Coil Dwell Disassem and Denso Coil Behavior
Updates
October 3, 2015: Tables updated to work around the forum display bug for xml/html.
Coil Dwell Disassembly
I finally decided to dig into the coil dwell code. Previous information was mostly correct. Here are the properly interpreted and scaled tables for the 88590015:
As people noted, the RPM-based dwell reduction function is not easily visualized from the tables. The Target Dwell formula is as follows:
Target Dwell = Base Dwell - C*(Reduction Factor)*(1/RPMstart - 1/RPM)
Base Dwell and the Reduction Factor are both dependent on battery voltage.
An important aspect of tweaking the dwell that may have been overlooked is the Upper Bound table that was formerly not fully understood. It limits the Target Dwell to be no larger than a percentage of the time it takes for the engine to do a full revolution. With the wasted spark implementation on the Evo, a coil is firing every engine revolution, so setting the Upper Bound to a percentage of this makes good sense.
To help people visualize changes to the coil settings, I have attached an Excel file that can be used to plot the Target Dwell Time and the Upper Bound Dwell Time as a function of RPM.
Denso Coil Behavior
As with with many people, I am running the SpoolinUp COP that utilizes Denso coils. The reason I jumped into the coil disassembly effort was because I was a bit tired of being in the dark about what these coils can do and whether there is something better out there. I went looking for coil charge data, and amazingly, I found a guy that had recorded the response of a wide range of coils. Bottom line is that the Denso coils are reasonably good for a COP.
The Denso coil is somewhat hotter than the stock Evo 8/9 coil, but interestingly, the Evo 7 coil is just as good as the Denso. The Bosch P100-T is a clearly better COP, but its spendy and requires machining the valve cover to fit. Some of the standalone coils are insanely hot. If a place can be found to put them in the engine bay, and running spark plug wires is acceptable, standalone coils are the way to go.
One performance aspect of the Denso coils that I wanted to know was the time to reach full charge. In the bench testing information that I found, it took only 2.5 ms, presumably at 14 V. However, during my dwell time disassembly work, I needed to verify and test some scalings, so I measured the charge time on my Evo. Turns out that the time to reach saturation was much longer - about 3.3 ms (for 14.1 V at the battery). The cause appears to be the rather small diameter power wires for the coils preventing the coils from seeing full voltage during the entire dwell period. The power wire is the equivalent of 20 ga. During my oscilloscope observations of the power circuit, the voltage was decreasing by several volts during the dwell period.
For stock coil wiring, 3.3 ms should be the target dwell for these coils if you want to get the most out of them. The easiest way to achieve this with the dwell tables is to zero out the dwell reduction factor and raise the Upper Bound value to 50% (or more). Might want to keep a 15 amp fuse handy, as these changes could cause the stock 10 amp coil power fuse to blow.
One last tidbit - I noticed during my oscilloscope observations that the dwell time increases to values greater than the Target Dwell value when engine speed drops below ~1800 rpm. At 875 rpm idle, I was recording a dwell time of ~4.7 ms. I didn't fully analyze the dwell code, so the portion that I didn't analyze in detail must contain the code to allow the dwell to increase at these lower rpms. This higher dwell at low rpm doesn't matter for performance tuning, but decided to mention it for completeness.
October 3, 2015: Tables updated to work around the forum display bug for xml/html.
Coil Dwell Disassembly
I finally decided to dig into the coil dwell code. Previous information was mostly correct. Here are the properly interpreted and scaled tables for the 88590015:
Code:
<scaling name="CoilDwellTime" units="ms" toexpr="x*0.064" frexpr="x/0.064" format="%.2f" min="0" max="16.32" inc="0.064" storagetype="uint8" endian="big"/> <scaling name="CoilDwellTimeUpperBoundPercent" units="% of Eng Revolution Time" toexpr="100*x/256" frexpr="256*x/100" format="%.1f" min="0" max="100" inc="1" storagetype="uint8" endian="big"/> <scaling name="Percent256-8" units="%" toexpr="100*x/256" frexpr="256*x/100" format="%.2f" min="0" max="99.609375" inc="0.00390625" storagetype="uint8" endian="big"/> <scaling name="RPMInverse-16" units="RPM" toexpr="7500000/x" frexpr="7500000/x" format="%.0f" min="0" max="10000" inc="60" storagetype="uint16" endian="big"/> <table name="Ign Coil Base Dwell-Time" category="Ignition - Coil Dwell" address="3cc6" type="2D" level="2" scaling="CoilDwellTime"> <table name="Battery" address="6e0c" type="Y Axis" elements="9" scaling="BatteryVoltage"/> </table> <table name="Ign Coil Dwell-Time Reduction Start RPM" category="Ignition - Coil Dwell" address="1346" type="1D" level="2" scaling="RPMInverse-16"/> <table name="Ign Coil Dwell-Time RPM Reduction Factor" category="Ignition - Coil Dwell" address="3cd6" type="2D" level="2" scaling="Percent256-8"> <table name="Battery" address="6e0c" type="Y Axis" elements="9" scaling="BatteryVoltage"/> </table> <table name="Ign Coil Dwell-Time Upper Bound" category="Ignition - Coil Dwell" address="3ce6" type="2D" level="2" scaling="CoilDwellTimeUpperBoundPercent"> <table name="Battery" address="6e0c" type="Y Axis" elements="9" scaling="BatteryVoltage"/> </table> <table name="Ign Coil Dwell-Time Cranking" category="Ignition - Coil Dwell" address="3cc6" type="1D" level="2" scaling="CoilDwellTime"/> <table name="Ign Coil Dwell-Time Cranking Upper Bound" category="Ignition - Coil Dwell" address="3ce6" type="1D" level="2" scaling="CoilDwellTimeUpperBoundPercent"/>
Target Dwell = Base Dwell - C*(Reduction Factor)*(1/RPMstart - 1/RPM)
Base Dwell and the Reduction Factor are both dependent on battery voltage.
An important aspect of tweaking the dwell that may have been overlooked is the Upper Bound table that was formerly not fully understood. It limits the Target Dwell to be no larger than a percentage of the time it takes for the engine to do a full revolution. With the wasted spark implementation on the Evo, a coil is firing every engine revolution, so setting the Upper Bound to a percentage of this makes good sense.
To help people visualize changes to the coil settings, I have attached an Excel file that can be used to plot the Target Dwell Time and the Upper Bound Dwell Time as a function of RPM.
Denso Coil Behavior
As with with many people, I am running the SpoolinUp COP that utilizes Denso coils. The reason I jumped into the coil disassembly effort was because I was a bit tired of being in the dark about what these coils can do and whether there is something better out there. I went looking for coil charge data, and amazingly, I found a guy that had recorded the response of a wide range of coils. Bottom line is that the Denso coils are reasonably good for a COP.
The Denso coil is somewhat hotter than the stock Evo 8/9 coil, but interestingly, the Evo 7 coil is just as good as the Denso. The Bosch P100-T is a clearly better COP, but its spendy and requires machining the valve cover to fit. Some of the standalone coils are insanely hot. If a place can be found to put them in the engine bay, and running spark plug wires is acceptable, standalone coils are the way to go.
One performance aspect of the Denso coils that I wanted to know was the time to reach full charge. In the bench testing information that I found, it took only 2.5 ms, presumably at 14 V. However, during my dwell time disassembly work, I needed to verify and test some scalings, so I measured the charge time on my Evo. Turns out that the time to reach saturation was much longer - about 3.3 ms (for 14.1 V at the battery). The cause appears to be the rather small diameter power wires for the coils preventing the coils from seeing full voltage during the entire dwell period. The power wire is the equivalent of 20 ga. During my oscilloscope observations of the power circuit, the voltage was decreasing by several volts during the dwell period.
For stock coil wiring, 3.3 ms should be the target dwell for these coils if you want to get the most out of them. The easiest way to achieve this with the dwell tables is to zero out the dwell reduction factor and raise the Upper Bound value to 50% (or more). Might want to keep a 15 amp fuse handy, as these changes could cause the stock 10 amp coil power fuse to blow.
One last tidbit - I noticed during my oscilloscope observations that the dwell time increases to values greater than the Target Dwell value when engine speed drops below ~1800 rpm. At 875 rpm idle, I was recording a dwell time of ~4.7 ms. I didn't fully analyze the dwell code, so the portion that I didn't analyze in detail must contain the code to allow the dwell to increase at these lower rpms. This higher dwell at low rpm doesn't matter for performance tuning, but decided to mention it for completeness.
Last edited by mrfred; Dec 14, 2015 at 01:14 PM.
#2
So to make sure I understand this correctly. Decreasing dwell time helps to give more time to recharge before the discharge cycle? Could this help decrease high RPM knock count possibly by having a greater spark when necessary? Great write up BTW!
#3
http://www.autospeed.com/cms/article...-time&A=113140
#4
Evolving Member
Join Date: Mar 2015
Location: Boring Oregon
Posts: 180
Likes: 0
Received 0 Likes
on
0 Posts
This is seriously interesting stuff
Might have to play with my coils tonight in the bench while My car is down to see just how controlled my actual dwell is
Thanks for this
You have always been a leader in information
Might have to play with my coils tonight in the bench while My car is down to see just how controlled my actual dwell is
Thanks for this
You have always been a leader in information
#5
here's a link to another great article. the autospeed article is an abbreviated version of this one.
http://www.dtec.net.au/Ignition%20Co...alibration.htm
one tidbit worth noting is that the denso coils appear to have built-in current limiting logic circuit that sets the max possible output. the flat top in this coil charge curve shows that current is clamped to about 12 amps. evo 8/9 coil also appears to have current limiting logic set to 7.5 amps. output of the denso coil is mildly higher.
courtesy of jari laine
courtesy of jari laine
http://www.dtec.net.au/Ignition%20Co...alibration.htm
one tidbit worth noting is that the denso coils appear to have built-in current limiting logic circuit that sets the max possible output. the flat top in this coil charge curve shows that current is clamped to about 12 amps. evo 8/9 coil also appears to have current limiting logic set to 7.5 amps. output of the denso coil is mildly higher.
courtesy of jari laine
courtesy of jari laine
Last edited by mrfred; Jun 2, 2015 at 08:00 AM.
#6
Evolving Member
Join Date: Mar 2015
Location: Boring Oregon
Posts: 180
Likes: 0
Received 0 Likes
on
0 Posts
So.. if.. I'm correct and in basic summary
the adjustment ifs used because even though the factory settings allow it to work..... the denso still isn't being used to its full potential and you could in turn eek a few more hp with supporting adjustments.
The logic data you posted is amazing
Makes me want to steal the oscilloscope from school.
the adjustment ifs used because even though the factory settings allow it to work..... the denso still isn't being used to its full potential and you could in turn eek a few more hp with supporting adjustments.
The logic data you posted is amazing
Makes me want to steal the oscilloscope from school.
#7
So.. if.. I'm correct and in basic summary
the adjustment ifs used because even though the factory settings allow it to work..... the denso still isn't being used to its full potential and you could in turn eek a few more hp with supporting adjustments.
The logic data you posted is amazing
Makes me want to steal the oscilloscope from school.
the adjustment ifs used because even though the factory settings allow it to work..... the denso still isn't being used to its full potential and you could in turn eek a few more hp with supporting adjustments.
The logic data you posted is amazing
Makes me want to steal the oscilloscope from school.
btw - those traces are not mine. they belong to a guy named jari laine. a boat-load of traces are on his picassa album page.
https://picasaweb.google.com/106058819169295752360
my crappy $500 oscilloscope produced the same primary winding charge curve for the denso coil. i was also able to monitor the square wave trigger signal.
Trending Topics
#8
Evolving Member
Join Date: Mar 2015
Location: Boring Oregon
Posts: 180
Likes: 0
Received 0 Likes
on
0 Posts
Well sir
I will input the script and try to tinker!
Do you think I should remove my cdi for more accurate findings?
This is definitely something I'm interested in
Also I think I must be blinded by the schools equipment. After receiving a mechanical engineering degree and building race motors for 4 years I thought it would be good to go back for computer science since ive been having so much fun tuning. I had no idea the cheaper oscilloscopes had gotten so good
Also I see your NW
I'm in PDX! Go rain!
I will input the script and try to tinker!
Do you think I should remove my cdi for more accurate findings?
This is definitely something I'm interested in
Also I think I must be blinded by the schools equipment. After receiving a mechanical engineering degree and building race motors for 4 years I thought it would be good to go back for computer science since ive been having so much fun tuning. I had no idea the cheaper oscilloscopes had gotten so good
Also I see your NW
I'm in PDX! Go rain!
#9
Well sir
I will input the script and try to tinker!
Do you think I should remove my cdi for more accurate findings?
This is definitely something I'm interested in
Also I think I must be blinded by the schools equipment. After receiving a mechanical engineering degree and building race motors for 4 years I thought it would be good to go back for computer science since ive been having so much fun tuning. I had no idea the cheaper oscilloscopes had gotten so good
Also I see your NW
I'm in PDX! Go rain!
I will input the script and try to tinker!
Do you think I should remove my cdi for more accurate findings?
This is definitely something I'm interested in
Also I think I must be blinded by the schools equipment. After receiving a mechanical engineering degree and building race motors for 4 years I thought it would be good to go back for computer science since ive been having so much fun tuning. I had no idea the cheaper oscilloscopes had gotten so good
Also I see your NW
I'm in PDX! Go rain!
i grew up in socal - still learning to appreciate the rain... we should meet up sometime.
#10
Evolving Member
Join Date: Mar 2015
Location: Boring Oregon
Posts: 180
Likes: 0
Received 0 Likes
on
0 Posts
I only use a plug in Hyperforce..
My idea was to see if I could get the same results with my COP unplugged
From the cdi by using the data posted and toying with other variables in the ECU
I was born in England lived there till I was 9 then moved to so cal as well haha.
I moved to Oregon because the weather is what I missed from England
And yes pm me
We can gather details
My idea was to see if I could get the same results with my COP unplugged
From the cdi by using the data posted and toying with other variables in the ECU
I was born in England lived there till I was 9 then moved to so cal as well haha.
I moved to Oregon because the weather is what I missed from England
And yes pm me
We can gather details
#13
Evolving Member
Join Date: Mar 2015
Location: Boring Oregon
Posts: 180
Likes: 0
Received 0 Likes
on
0 Posts
Also I'm sure there are benefits to doing this with any COP system
But to truly say what any of its capabilities are is somewhat unknown
MrFred did an excellent job of explaining the theoretical application and graciously decided to provide a patch to test everything and discover what modifying the parameters will do