As stated in the title I've swapped in the Evo 8 ECU w/MAF and done the initial setup on the Tephra V7 rom: injector scaling/latency, knock multiplier adjustment, fan correction, etc. I've been rolling waste-gate (~10psi) getting everything set: injectors hold within +/- .2 of target AFR at WOT and are within +/- 5% ltft; I've set the knock multipliers per eclipsegsx1736's suggestions of Low 18, Mid 14, High 16, and I'm still seeing a substantial amount of knock (>6 counts), particularly at spool-up.

So I decided I'd throw in some 110 and figure out what's phantom and what isn't; with 110 at waste-gate all my spool-up knock vanished (not what I expected...) and the high-end (5-7k) didn't change at all.

This is where I get confused, I'm moderately convinced that my high-end knock is phantom now, but the spool-up is not. I've read through the Evo 8 swap threads over and over, and I don't see any mention of having to tweak the maps at all; most people are leaning them out up-top if anything. So what's the deal here? How does everyone but me seem to have no issue running the Evo maps as a solid base?

Is everyone having to adjust the timing during spool-up and just not mentioning it? Is everyone running higher knock multipliers? I just can't figure this out; I don't expect tuning to be as simple as plug and play by far, but I don't want to start dropping timing right from the get-go when no one else seems to be. Hortanica even posted that he ran over 23psi with no changes to the stock Evo map, so what's the deal here?

I hope this all doesn't come off as too Newbish, but it's making me crazy! I'd like to have my CEL not flashing at me all the time, and I'd also like my motor to stay in one piece. Sorry for the long post, but I figured it'd be best to be thorough. Thank you all in advance for any help.

have you tried to richen it up or start pulling timing or are you just mad at it for not working with the stock evo maps...?

From your post all i get is that you have dialed in your injectors.

I had to swap out to the evo 8 knock sensor to sort out the phantom knock on my original conversion.

ZombieSix said:

have you tried to richen it up or start pulling timing or are you just mad at it for not working with the stock evo maps...?

From your post all i get is that you have dialed in your injectors.

Click to expand...

My bad, I guess I wasn't near as thorough as I thought, . I previously did some logs and tried richening it up a bit as well as pulling a few degrees of timing from the area; it made no noticeable difference. I then decided to change Roms, to see if it would change anything; it did not.

After reading through the swap thread, I figured at this point it must surely just be phantom knock like so many others experience, so I put a couple gallons of 110 in it and to my surprise, suddenly all the spool-up knock vanished. I later filled up with 93, and the spool-up knock came straight back. This left me at a loss; it seemingly wasn't phantom knock, but adjusting the fuel/timing on 93 didn't seem to do anything either.

At that point I came in nice and frustrated, wrote my venting post and called it good. I wouldn't say I'm mad at it for not working, as in general it works just fine with only this one big nagging issue. However, I'm definitely frustrated that so many people do the swap seemingly trouble-free and I'm sitting here with issues.

It just seems odd to me that I'd be the only one with this problem, when so many others have done this swap seemingly without issue. It seems like everyone just throws in 110, adjusts the multipliers and goes about their business knock-free on the stock maps. Perhaps I'm just missing all the threads of people facing large amounts of spool-up knock that isn't phantom; perhaps they're just not posting about it? With all these cars being pretty similar you'd think I'd see lots of people with the same issue, but I don't, and it sure is strange to me.

I'm tempted to just go back to the 98 ECU as that didn't give me any problems, but there's just so much more support and options with the Evo ECU. I'd like to get map-switching set up and run the E85 patch to really pull what I can get out of my setup. Plus I have an idea to get my Cyclone controlled by the ECU, that I can't do on the 98 ECU. Not to mention development on the 98/99 ECU is completely dead at this point. As it stands, I've got to walk with this new ECU before I can run; I suppose that's just how it goes.

Sitting here with a clear head I believe my next move will be to do some more logging and begin pulling timing and increasing fuel again to see if I just didn't do enough last time. If that doesn't work I'll swap over to the Evo knock sensor and see if that clears it up.

Also, what all did you have to do to get this running smoothly? What Rom is everyone using now? What knock multipliers are you running? Did you have to tune out spool-up knock? Are most people switching to the Evo knock sensor? Most of the swap thread information is vastly out-dated, I'm not really sure if the majority of it is even valuable anymore. I don't mind having to tune this out if that's the case, it's just that I feel like I'm running blind here. Thank you again for all your help, I really do appreciate it.

I swapped to an Evo 8 knock sensor and didn't see any change in the amount of knock detected. From what I can tell, the only difference in the sensor is a longer length of wire to where it connects to the engine harness.

Another possible source may be a difference in the CAS. The Evo's have a different rotor in the CAS. I know someone swapped to one at some point, but I don't remember why or the outcome. That may be messing with the timing just enough. But like I said, I'm not sure, but something to check into.

Post a log if you can.

IIRC, if you use a 9653 ROM, the CAS shouldn't be an issue. The Evo CAS itself is the same as the 2Gb CAS, not sure about the rotor. The Evo has it on the exhaust cam, but it is also located at the angle it should be to give the same signal to the ECU as the DSM head does with it on the intake.

I'm not running the Evo 8 ECU currently, but getting ready to swap back to it. I also had a little annoying knock on spool-up way back when using a V5 ROM. I think lean spool was partly to blame for mine though.

Melbowski- What RPM range do you start seeing knock at?

Also this may help out, and if not it's still a worthwhile read
http://forums.evolutionm.net/ecuflash/382625-knock-control-load-vs-rpm-table-found.html

Don't rule out something mechanical causing the knock. If your mods list is up to date you have a rebuilt shortblock. Maybe its worth your time to just take a close look at it and check the oil if you havent recently.

Alright, sorry for the delay, I've been stuck at work and haven't had any time to poke at the car. I've got some good news for a change! I decided that since I didn't have knock on the DSM stock map I'd interpolate the values from its fuel table to see if it'd make a difference, and it did! The values from the DSM table are over a full point richer in the spool-up region and taper off to be (oddly enough) leaner at the end of the map.

Also interesting to note, the people of 3Si have found that the cross-over point from vacuum to boost is around 80 load on the H8 ECUs, which conveniently is where the stock DSM fuel table switches over from stoich to open-loop afr values. Oddly, the Evo ECU holds stoich until 100 load. Maybe I'm crazy, but I wouldn't think Mitsubishi changed their load calculation formula. Perhaps the newer motors are just more efficient with a better head design that can handle that AFR? Maybe Mitsubishi did change the load calculation? I don't know, I can only speculate.

Anyway, so starting at 80 load on the fuel table I copied the direct values and interpolated the in-between values up until the DSM table became more lean than the Evo, and then I smoothed it in. Took the car for a spin and in 6 pulls, only 1 showed enough knock to make the CEL flash (3 counts)! Not perfect, but a major improvement in comparison to >3 counts on 90% of pulls I dealt with before. If knock hangs around at all, I'll start pulling timing; more logging to be done in the mean-time.

Knock up-top is still prevalent, it was still there even when I ran 110, so it may be phantom. I just did a log going around the block a few times (mind you I live in the boonies, so 1 block is more than 1 mile), so I'm going to try and pull some timing up top to see if that has any effect. It's also running around 9.8:1 AFR at that area, is that enough to cause rich knock? I can't imagine it would, since it's the stock map, but I thought I'd ask.

On a side note, I'm running PTE 780 injectors which I bought from a member here. The part number on them crosses to PTE 780s, but when I scaled them I ended up with a value of 585. I know the Evos scale injectors on the low-side, and PTE are over-rated to begin with, but that's really freaking low. It's also giving me 75% ICD at 10 psi, which seems pretty high as well; maybe I'm just over-thinking things? I just did a Boost Leak Test a couple weeks ago, and came up clear to 18psi (1g BOV isn't crushed yet) so I don't know what to do about that.

I'm sorry to ask so many newb questions, but thank you all so much for your help so far!

knochgoon24 said:

I swapped to an Evo 8 knock sensor and didn't see any change in the amount of knock detected. From what I can tell, the only difference in the sensor is a longer length of wire to where it connects to the engine harness.

Another possible source may be a difference in the CAS. The Evo's have a different rotor in the CAS. I know someone swapped to one at some point, but I don't remember why or the outcome. That may be messing with the timing just enough. But like I said, I'm not sure, but something to check into.

Post a log if you can.

Click to expand...

Gotcha, I've read that the Evo knock sensor is the same and offers no improvement, glad it's been tried out without me buying one. On the thought of the CAS, I'm using a 1g green-top so I wonder if that is having an effect. From what I've read on 3Si, the optical sensors when used on the flashable ECUs increase timing slightly up to ~3 degrees by red-line. Apparently it's due to an uncorrected latency issue that the Hall-effect sensors don't have. I'll look into it more; thanks for the tip. Also, log is attached.

delta448 said:

IIRC, if you use a 9653 ROM, the CAS shouldn't be an issue. The Evo CAS itself is the same as the 2Gb CAS, not sure about the rotor. The Evo has it on the exhaust cam, but it is also located at the angle it should be to give the same signal to the ECU as the DSM head does with it on the intake.

I'm not running the Evo 8 ECU currently, but getting ready to swap back to it. I also had a little annoying knock on spool-up way back when using a V5 ROM. I think lean spool was partly to blame for mine though.

Click to expand...

CAS issue noted above, thank you for the information. I'm currently running the 9653 rom, with lean-spool disabled so hopefully that's not the issue.

PFC_CiarlilloM said:

Melbowski- What RPM range do you start seeing knock at?

Also this may help out, and if not it's still a worthwhile read
Knock control - load vs RPM table found - evolutionm.net

Click to expand...

Spool-up knock was seen from 2500-3500rpm; the high-end knock is rather random, but it's most common around 5k. Thanks for the thread, it is a good read indeed. I thought however that it came down to people opting to change the multipliers as a safer method of adjustment? We do have a slightly different version of the same motor, perhaps the knock control tables actually need adjustment. I don't think at my current skill level that I'm savvy enough for that kind of work though.

bigbird94 said:

Don't rule out something mechanical causing the knock. If your mods list is up to date you have a rebuilt shortblock. Maybe its worth your time to just take a close look at it and check the oil if you havent recently.

Click to expand...

Always a good idea! I just did an oil change a couple days ago; no shavings or anything to indicate a problem. No ticking (3g lifters are a God-send!), tacking, or thumping of any kind. Hopefully it stays that way!

Attached is a log I just took driving around doing some pulls. I obviously let off when I see knock, so there aren't any back-to-back 7kpulls. The log itself is rather long, I hope it's useful. Thank you all again for all of your help, I truly appreciate it.

The log appears to be the wrong one, this one's about 25 seconds of idle only.

FYI I'm now running Knock Multipliers of 19/16/18, which seems to work quite well for the engine in my daily driver. I dialed it in using E85 tuned where I know I will not pick up knock, and lowered the values until I began to see trace amounts of knock where I knew there wasn't any, then raised it up +1 from there. This is how I'd recommend tuning your knock multipliers. You can of course substitute a suitable high octane/leaded gas instead of E85. Again, these values were for MY car, and WILL vary from engine to engine.

Also, it would be useful to learn the sound an engine makes when knocking, which actually changes with RPM and load. Any time I'm full throttle, I've got windows up and everything else off so I can listen to the engine. Low RPM/load knock usually sounds like rattling rocks in a can (you'll often hear newer bone stock cars knock quite a bit as they set off from a stoplight, ECUs nowaways are always shooting for max timing for fuel economy). High RPM/load knock usually sounds like small cracking sounds, similar to the sound a tiny rock makes when it flies up and hits your windshield.. This is how I noticed the multipliers were too high for my engine originally. They make "knock ears" for this purpose too, but as long as your car isn't obnoxiously loud (open dump, etc) you should be able to easily hear it if you know what to listen for.


Beau

Update! So it will suffice to say that I think my problem is pretty much solved! After receiving multiple PMs from other members who *gasp* have the exact same problem, I set out to try again to tune all this knock out. This time however, I have found success! Based on my work and research I believe my problem was two-fold:

1) The Evo maps are not designed to be run at our waste-gate pressure.

Some research on the Evo forums made it quickly apparent that their cars don't have the same waste-gate pressure. The long and short of it is that with the actuator preload they run 12-14psi at waste-gate, whereas we run 8-10psi at waste-gate. I decided to compare our stock maps and see what I could figure out:

2gb DSM Timing:


Evo 8 Timing:


Load to PSI Comparison (Compliments of Ceddy):


You can see that our stock maps are basically shifted to the left in comparison to the stock Evo maps; not very surprising since Evos shouldn't ever really be running in that low of a load range. The fuel maps follow basically the same trend. The Evo timing at WOT at ~10psi just wasn't panning out for what my motor wanted. I decided to interpolate/smooth the DSM timing and fuel into the Evo maps; the results, not surprisingly, were that my knock diminished significantly. A good first step, but there was still some high-end knock left over which leads to my next point.


2) The 1g Optical CAM sensor (Green-top) causes a latency issue with the later ECUs.

Digging around on 3/S/GTO yielded some interesting info on flash-able ECU swaps: using the 1g Optical CAM sensor on an ECU designed for the Hall-effect senor caused a linear timing increase to about 4* by red-line. This problem arises from a latency issue (~80 usec) that the later ECUs are not setup for. Simply put, the timing you see on the table is not what the motor actually receives; it receives that value plus some extra (i.e: + 0* at idle; + 2* at 3.5k; + 4* at 7k; etc.)

Being a 6-bolt swap myself, it's quickly apparent why I was getting so much knock; not only was I running the Evos more aggressive stock timing, but I was gaining additional timing on top of it! More logging and distinctively less timing later, I'm basically knock-free! I don't know if using the black-top CAM sensor will fix this problem, but since it's a Hall-effect type I imagine it will help.

Kind of a long post, but I hope this can help someone out in the future. As a side note, if you're about to do the Evo 8 ECU swap, be aware that this is not a plug in and drive off kind of thing. I spent a good chunk of time trying to figure out what I did wrong and why I was getting so much knock; what I should have been doing was logging and tuning from the get-go! From what I've experienced it takes some work to get things going smoothly, but hang in there! The Evo 8 ECU is a powerful tuning tool, and I completely vouch for it.

Eclipsegsx1736: Thanks for the info and for looking at my log! You're definately right though, I posted the wrong log! I'm running my multipliers at 20/16/18 (low/mid/high) this seems to be the standard that they found in the original swap thread. Seems to work just fine, and knochgoon's testing apparently confirmed that the Evo knock sensor makes no difference. As for listening to knock, thanks to the previous owner my exhaust is obnoxiously loud... When I get a new setup I'll see if I can train my ears, but lets just hope I don't have any knock to hear.

Thank you all again for all your help, I truly appreciate it!

Melbowski said:

2) The 1g Optical CAM sensor (Green-top) causes a latency issue with the later ECUs.

Digging around on 3/S/GTO yielded some interesting info on flash-able ECU swaps: using the 1g Optical CAM sensor on an ECU designed for the Hall-effect senor caused a linear timing increase to about 4* by red-line. This problem arises from a latency issue (~80 usec) that the later ECUs are not setup for. Simply put, the timing you see on the table is not what the motor actually receives; it receives that value plus some extra (i.e: + 0* at idle; + 2* at 3.5k; + 4* at 7k; etc.)

Click to expand...

Glad to hear things are improving! Regarding the notes above, I'd be surprised if this was really the case. However, there's an easy way to verify whether or not this is actually happening. You can either lock timing using Evoscan "IG 5deg BTDC" or make a test flash that has no ignition trims and all ignition maps at 5 degrees flat, and rev the engine up as high as you feel comfortable with (redline?), while pointing the timing light at the engine cover. You should not be able to see any discernible timing drift as RPM increases. If there is a drift, then you'll have to compensate for it like you describe since to my knowledge there isn't a setting to adjust for this on the factory ECU that anyone's uncovered yet.

This pickup delay is a known phenomenon but I haven't heard of it being a result of the optical/hall pickup itself. It's usually more related to ECU processing time etc. In standalone EMS systems, we adjust this using a configurable setting (called Pickup Delay Compensation in AEM EMS) which corrects for small amounts of timing drift as RPM increases. It's normal to only see 1-4 deg max of drift from idle to redline if the pickup delay is incorrect. Any more than this and something else is wrong.

Beau

Melbowski said:

Update! So it will suffice to say that I think my problem is pretty much solved! After receiving multiple PMs from other members who *gasp* have the exact same problem, I set out to try again to tune all this knock out. This time however, I have found success! Based on my work and research I believe my problem was two-fold:

1) The Evo maps are not designed to be run at our waste-gate pressure.

Some research on the Evo forums made it quickly apparent that their cars don't have the same waste-gate pressure. The long and short of it is that with the actuator preload they run 12-14psi at waste-gate, whereas we run 8-10psi at waste-gate. I decided to compare our stock maps and see what I could figure out:

2gb DSM Timing:

You must be logged in to view this image or video.

Evo 8 Timing:

You must be logged in to view this image or video.

Load to PSI Comparison (Compliments of Ceddy):

You must be logged in to view this image or video.

You can see that our stock maps are basically shifted to the left in comparison to the stock Evo maps; not very surprising since Evos shouldn't ever really be running in that low of a load range. The fuel maps follow basically the same trend. The Evo timing at WOT at ~10psi just wasn't panning out for what my motor wanted. I decided to interpolate/smooth the DSM timing and fuel into the Evo maps; the results, not surprisingly, were that my knock diminished significantly. A good first step, but there was still some high-end knock left over which leads to my next point.


2) The 1g Optical CAM sensor (Green-top) causes a latency issue with the later ECUs.

Digging around on 3/S/GTO yielded some interesting info on flash-able ECU swaps: using the 1g Optical CAM sensor on an ECU designed for the Hall-effect senor caused a linear timing increase to about 4* by red-line. This problem arises from a latency issue (~80 usec) that the later ECUs are not setup for. Simply put, the timing you see on the table is not what the motor actually receives; it receives that value plus some extra (i.e: + 0* at idle; + 2* at 3.5k; + 4* at 7k; etc.)

Being a 6-bolt swap myself, it's quickly apparent why I was getting so much knock; not only was I running the Evos more aggressive stock timing, but I was gaining additional timing on top of it! More logging and distinctively less timing later, I'm basically knock-free! I don't know if using the black-top CAM sensor will fix this problem, but since it's a Hall-effect type I imagine it will help.

Kind of a long post, but I hope this can help someone out in the future. As a side note, if you're about to do the Evo 8 ECU swap, be aware that this is not a plug in and drive off kind of thing. I spent a good chunk of time trying to figure out what I did wrong and why I was getting so much knock; what I should have been doing was logging and tuning from the get-go! From what I've experienced it takes some work to get things going smoothly, but hang in there! The Evo 8 ECU is a powerful tuning tool, and I completely vouch for it.

Eclipsegsx1736: Thanks for the info and for looking at my log! You're definately right though, I posted the wrong log! I'm running my multipliers at 20/16/18 (low/mid/high) this seems to be the standard that they found in the original swap thread. Seems to work just fine, and knochgoon's testing apparently confirmed that the Evo knock sensor makes no difference. As for listening to knock, thanks to the previous owner my exhaust is obnoxiously loud... When I get a new setup I'll see if I can train my ears, but lets just hope I don't have any knock to hear.

Thank you all again for all your help, I truly appreciate it!

Click to expand...

Did anybody test this out to see if the 1G Green Top CAS gives you additional timing? I am having an issue with knock. I adding some timing to my WOT and I keep getting weird counts of knock where I don't believe I should be getting knock. I don't believe I am running enough timing to get knock. My AFR is high 10's, low 11's at 20PSI on an E316G.