Strange Results of Injector Change - Miata Turbo Forum - Boost cars, acquire cats.

Welcome to Miataturbo.net   Members
 


MEGAsquirt A place to collectively sort out this megasquirt gizmo

Reply
 
 
 
LinkBack Thread Tools
Old 02-23-2014, 09:13 AM   #1
Moderator
Thread Starter
iTrader: (11)
 
sixshooter's Avatar
 
Join Date: Nov 2008
Location: Tampa, Florida
Posts: 15,669
Total Cats: 1,561
Default Strange Results of Injector Change

Cross posted from my build thread to attract more answers:

I swapped from RX7 460cc EV1 injectors to Deatschwerks 700cc EV6 injectors. They didn't like the adjustment of the req_fuel or maybe the injector opening time settings. Please take a look at the info and tell me what I'm missing. Did DW perhaps send me the wrong injectors (mislabeled) or am I doing something wrong?

Quote:
Originally Posted by 18psi View Post
460/700=.657 which is roughly what you should be multiplying req_fuel by

Quote:
Originally Posted by sixshooter View Post
Strange things are afoot at the Circle K.

I swapped in the DW700cc's in place of the RX7 460cc's. I changed the req_fuel as discussed above (7.4 to 4.9), took a tenth of a millisecond out of the injector opening time (1.2 to 1.0), the battery correction voltage from 0.1 to 0.2, and the car would barely run. I ended up mucking with those values to get it from 1.0 to 1.05 on the opening time, and it ran better. I gradually increased the req_fuel until it reached 7.3 and my AFR at idle and while free revving seemed pretty good there. WTF? Are my dead times that off or are the injectors nearly the same size as the ones that came out of the car? Strange.

I looked at injector opening times in many places online tonight and came up with these tables comparing an EV1 like the one that came out of the car to the EV6 I put in it.




The discussion came from here:
Megasquirt MSEXTRA / MS3EFI • Injector Open Time (View topic)

I'm curious what the more experienced might have to say.

I road tuned for a few minutes, and it went well, but didn't think to datalog until after I got home. I know datalogging would tell me a bit more.
Quote:
Originally Posted by sixshooter View Post
A few new notes after sleeping on the problem.

I had done a fresh air recalibration on the O2 sensor last week since the car had been sitting since September. It seems to be relatively accurate in relation to the performance characteristics at idle.

I have read that the req_fuel is somewhat arbitrary and that the VE scaling is the important factor in making the engine run. As it sat yesterday, I had dropped the req_fuel from 7.4 to 7.3 and the idle AFRs were similar to the idle prior to the change. Then, the cells around the 101kpa and up seemed to want about 20 to 25 points less VE value (roughly around 120 down to 100) to get back to target AFRs. This tells me that if I left the req_fuel at 7.4 the car may have been pretty close in the middle and top cells. The idle cells were a touch rich at 7.4. This is telling me the injectors are acting too similar to the size that came out. They are supposed to be ~60% larger.

The car is at 1/4 tank of fuel and it is old fuel from September. I will be topping the car off with fresh fuel and retuning to eliminate that as a variable.

FWIW, I could adjust the req_fuel up or down while idling and watch the idle AFRs ease upward or downward, and likewise, adjusting the opening time setting would effect it similarly.
sixshooter is offline   Reply With Quote
Old 02-23-2014, 11:08 AM   #2
Elite Member
 
DNMakinson's Avatar
 
Join Date: Jan 2013
Location: Seneca, SC
Posts: 2,701
Total Cats: 269
Default

If most of the VE is still acting correctly then the REQ_FUEL is right. If low pulse width areas ( idle and <40kpa) are lean, add dead time. If rich, decrease dead time. This assumes the dead times were correct old inj. If the low kpa rows suddenly rise rather than continuing to decrease from the rows above, then your old dead times were wrong (too short), and this is adding to the errors.

Also, if you believe the curves, you could correct for low PW nonlinearity in MS3.
DNMakinson is offline   Reply With Quote
Old 02-23-2014, 11:09 AM   #3
Boost Czar
iTrader: (61)
 
Braineack's Avatar
 
Join Date: May 2005
Location: Chantilly, VA
Posts: 72,872
Total Cats: 1,790
Default

From my experience, the larger DW injectors (they used to sell, unsure about now) required large deadtimes and PW at idle and would not idle smoothly unless richer than 13.0:1 at idle because they couldn't cycle fast enough.
Braineack is offline   Reply With Quote
Old 02-23-2014, 11:40 AM   #4
Elite Member
 
DNMakinson's Avatar
 
Join Date: Jan 2013
Location: Seneca, SC
Posts: 2,701
Total Cats: 269
Default

And, according to the graphs, dead time should be about 0.8mS (where the black line stops at zero flow).
DNMakinson is offline   Reply With Quote
Old 02-23-2014, 12:33 PM   #5
Boost Czar
iTrader: (61)
 
Braineack's Avatar
 
Join Date: May 2005
Location: Chantilly, VA
Posts: 72,872
Total Cats: 1,790
Default

yea but also looks like the lowest possible PW is 1.5ms
Braineack is offline   Reply With Quote
Old 02-23-2014, 02:21 PM   #6
Moderator
Thread Starter
iTrader: (11)
 
sixshooter's Avatar
 
Join Date: Nov 2008
Location: Tampa, Florida
Posts: 15,669
Total Cats: 1,561
Default

Quote:
Originally Posted by DNMakinson View Post
If low pulse width areas ( idle and <40kpa) are lean, add dead time. If rich, decrease dead time. This assumes the dead times were correct old inj. If the low kpa rows suddenly rise rather than continuing to decrease from the rows above, then your old dead times were wrong (too short), and this is adding to the errors.
It seemed to be pulling fuel pretty evenly across the board as it autotuned.

Quote:
Originally Posted by Braineack View Post
yea but also looks like the lowest possible PW is 1.5ms
Yeah, I was estimating 1.4 from looking at the chart last night but was using a setting of 1.05 yesterday. The info with the DWs said it was around 1.1ms. And the old EV1s should have been around 1.6 according to the other chart above but it was set at 1.2 on my previous tune (was my basemap setting that came with the MS1).

Quote:
Originally Posted by DNMakinson View Post
If most of the VE is still acting correctly then the REQ_FUEL is right.
But it shouldn't be right if I went to 60% larger injectors than before. This is why I'm wondering if they mislabeled a set of their 450cc as 700cc at DW. That would explain them not needing a big change in req_fuel to run right. The req_fuel is almost identical.
sixshooter is offline   Reply With Quote
Old 02-23-2014, 03:31 PM   #7
Elite Member
 
DNMakinson's Avatar
 
Join Date: Jan 2013
Location: Seneca, SC
Posts: 2,701
Total Cats: 269
Default

Yes, VE table should be right, as the change you made in true REQ_FUEL should compensate across the entire table.

The thoughts on dead time is that it is part of the linear portion of the curve and occurs at every pulse regardless of length. The non-linear portions should be dealt with separately. I admit I have no experience there, but do know that the functionality exists.

We need to hear from someone who has done it, or Six can be the first. Seems that with the curves in hand, one should be able to simply key them in.
DNMakinson is offline   Reply With Quote
Old 02-23-2014, 03:44 PM   #8
Elite Member
iTrader: (2)
 
triple88a's Avatar
 
Join Date: Apr 2006
Location: Chicago, IL
Posts: 9,131
Total Cats: 598
Default

Quote:
Originally Posted by Braineack View Post
From my experience, the larger DW injectors (they used to sell, unsure about now) required large deadtimes and PW at idle and would not idle smoothly unless richer than 13.0:1 at idle because they couldn't cycle fast enough.
I idle my 800cc dws (820 flow tested) at 15afr all day long. 700s shouldnt be a problem.
triple88a is offline   Reply With Quote
Old 02-23-2014, 07:42 PM   #9
Elite Member
 
DNMakinson's Avatar
 
Join Date: Jan 2013
Location: Seneca, SC
Posts: 2,701
Total Cats: 269
Default

Quote:
Originally Posted by sixshooter View Post
Cross posted from my build thread to attract more answers:

I swapped from RX7 460cc EV1 injectors to Deatschwerks 700cc EV6 injectors. They didn't like the adjustment of the req_fuel or maybe the injector opening time settings. Please take a look at the info and tell me what I'm missing. Did DW perhaps send me the wrong injectors (mislabeled) or am I doing something wrong?
Six. I'm sorry, I missed this part. I thought the problem revolved solely around the idle area. Indeed, it does seem that things do not make sense.

You put in new injectors, inversely changed the REQ_FUEL, and went lean everywhere. Well, sort of, except that you changed RWQ_FUEL back up to 7.3 and then AutoTune started pulling fuel.

This gets back to looking like the issue resides at low Pulse Width only (idle and free rev are both low fueling parts of the table). So, it seems to me that the evidence is still that the injectors are correct, just not so good at the low pulses (a condition made worse by their larger size).

In looking at the curve for the injector, it would appear that the data exists for correcting for Non-Linearity. This goes into tables in TS under Fuel Settings / Injector small pulse widths. I can walk you through the steps to create the table, but the problem is that I don't know the nomenclature of New ms and Original (ms). If you want to try, the only issue is that if we are backwards then the condition will worsen. Then just swap the columns in the table(s) to make the curve correct. But I think this will be the correct columns.

Using the second chart, follow the black line. This says that dead time should be 0.6ms. This is the NEW ZERO / ZERO, and should also be what you enter into TS as dead time.

Start with a third column, the Flow.
The black line would represent the Original ms (as what would have been called for based on Linear performance).
The red line would represent the New ms (what needs to be called for before dead time is added).

So for 0 flow, Orr = 0 and new = 0
for 10 flow, Orr = .20 mS and new = .65 ms
for 20 flow, Orr= .35 ms and new = .70 ms
for 30 flow, Orr= .50 ms and new = .73 ms
for 40 flow, Orr= .60 ms and new = .76 ms
for 50 flow, Orr = .73 ms and new = .80 ms

That gives the 6 data points that will describe the curve for the NL area of the injector. Put those in for each of the 4 injectors, and Enable.

Also set the dead time to the .6ms

After that, things are linear.

This should add a lot of fuel for the low pulse widths (any below what are now .2 before dead time, or with your present dead time of 1.05, anything between 1.25 and 1.78). NOTE, if you are not seeing any values this low, then what I am describing will not make any difference. However, unless your fuel pressure is low at idle (my 99 is return-less) then I suspect you are operating in that region.

Even if you choose not to try this approach, I would highly suggest:
1) use the correctly calculated REQ_FUEL
2) use Include AFR_Target
3) don't let idle drive your table

DNM
DNMakinson is offline   Reply With Quote
Old 02-23-2014, 08:41 PM   #10
Elite Member
 
DNMakinson's Avatar
 
Join Date: Jan 2013
Location: Seneca, SC
Posts: 2,701
Total Cats: 269
Default

Further looking says that 1st and 6th point set should be equal. So I would skip the "for 30 Flow…" and add a last line of Orr .85 ms and New .85 ms
DNMakinson is offline   Reply With Quote
Old 02-23-2014, 09:57 PM   #11
Moderator
Thread Starter
iTrader: (11)
 
sixshooter's Avatar
 
Join Date: Nov 2008
Location: Tampa, Florida
Posts: 15,669
Total Cats: 1,561
Default

I read the posts from earlier today and did a couple more hours of research online in various places. Things I was reading about the EV6 injectors kept pointing to increasing the opening time, lag time, or deadtimes depending who is using the term. My problem was that I was expecting shorter lag times than my previous settings.

As a diagnostic measure I went back to my pre-injector swap map. I increased the "opening time" way up to 1.40ms as it appears in the above chart (1.4 is where it seems to really catch up to the line) and changed the req_fuel to the 4.8 number calculated above. It idled pretty well. That's what it wanted all along. Long deadtimes, lagtimes, opening times, whatever you want to call them.

I took the car for a little fresh fuel and then checked my settings by letting TunerStudio VE Analyze Live do some autotuning. The settings in my VE table were close, as the numbers were alternating adding and removing a little bit of fuel in the cells as it tuned. It was very close. So I guess it is close to right.

I verified what was going on by datalogging. The same approximate boost pressure and RPMs that gave me 100% duty cycle with the 460cc RX7 injectors is giving me around 60% duty cycle on the DW 700cc's, which is close to the percentage difference in size between them. So it must be close to right.

I put in a help desk request with Deatschwerks so I will see what they can tell me about recommended settings.


Thank you to David for going into such depth regarding my issue. I'm sure my MS1 could benefit some code upgrades but I'm happy for a simple enough if not perfect solution at this point. One day I will jump ahead to MS3 but for now...

I think all is well. Maybe this will help others some day. And maybe Lars will clear out his VM inbox or respond to my text from yesterday.
sixshooter is offline   Reply With Quote
Old 02-23-2014, 10:08 PM   #12
Elite Member
 
DNMakinson's Avatar
 
Join Date: Jan 2013
Location: Seneca, SC
Posts: 2,701
Total Cats: 269
Default

Sure.

It seemed that this feature was added, but no one has tried it. If this is really how the injectors perform then shorter dead time and NL adjustment would be the best solution. I misread that you were running MS3.

In the end, for the most part, adding extra dead time is a good (similar) fix.

And, as often happens, Brain's short answer was what you needed.
DNMakinson is offline   Reply With Quote
Old 02-23-2014, 10:15 PM   #13
Moderator
Thread Starter
iTrader: (11)
 
sixshooter's Avatar
 
Join Date: Nov 2008
Location: Tampa, Florida
Posts: 15,669
Total Cats: 1,561
Default

Quote:
Originally Posted by DNMakinson View Post
Sure.

It seemed that this feature was added, but no one has tried it. If this is really how the injectors perform then shorter dead time and NL adjustment would be the best solution. I misread that you were running MS3.

In the end, for the most part, adding extra dead time is a good (similar) fix.

And, as often happens, Brain's short answer was what you needed.
And I sincerely appreciate everyone's willingness to help.
I'll get by with a little help from my friends.
sixshooter is offline   Reply With Quote
Old 02-24-2014, 12:29 PM   #14
Junior Member
iTrader: (3)
 
Join Date: Aug 2010
Location: Oklahoma City
Posts: 76
Total Cats: 1
Default

Hey Steve,

Figured it might be easier to answer here as I figured the community would have already gotten involved from time you sent the inquiry through our site and the time I read it this Monday morning.

What part number do you have? As far as I know, we don't have an EV6 injector for the Miata. It would either be a Denso injector or an EV14.
LowFlyin' is offline   Reply With Quote
Old 02-24-2014, 12:38 PM   #15
Moderator
Thread Starter
iTrader: (11)
 
sixshooter's Avatar
 
Join Date: Nov 2008
Location: Tampa, Florida
Posts: 15,669
Total Cats: 1,561
Default

Quote:
Originally Posted by LowFlyin' View Post
Hey Steve,

Figured it might be easier to answer here as I figured the community would have already gotten involved from time you sent the inquiry through our site and the time I read it this Monday morning.

What part number do you have? As far as I know, we don't have an EV6 injector for the Miata. It would either be a Denso injector or an EV14.
It is a 22U part number. I guess that would be the Denso. I thought it was an EV6 design from outside discussion, so that's an error on my part.
sixshooter is offline   Reply With Quote
Old 02-24-2014, 12:51 PM   #16
Junior Member
iTrader: (3)
 
Join Date: Aug 2010
Location: Oklahoma City
Posts: 76
Total Cats: 1
Default

Yeah, the 22U would be a Denso.

That injector was phased out before we started collecting our new data. It might be similar enough to the 22S-700 we have now that you could use that info as a starting point.
LowFlyin' is offline   Reply With Quote
Old 03-06-2014, 09:14 AM   #17
Newb
 
Join Date: Mar 2014
Posts: 5
Total Cats: 0
Default

Quote:
Originally Posted by sixshooter View Post
I increased the "opening time" way up to 1.40ms as it appears in the above chart (1.4 is where it seems to really catch up to the line) and changed the req_fuel to the 4.8 number calculated above. It idled pretty well. That's what it wanted all along. Long deadtimes, lagtimes, opening times, whatever you want to call them.
Long time lurker. I too have recently purchased these injectors. Also had similar troubles running the recommended deadtimes from DW.

I also increased deadtime to ~1.5ms and the car started to behave before I read this thread. I'll try 1.4ish this coming weekend.

Just wondering what if you had any other news/responses?
frizzles is offline   Reply With Quote
Old 03-06-2014, 09:38 AM   #18
Newb
 
Join Date: Mar 2014
Posts: 5
Total Cats: 0
Default

Well, just found this.
DW must've been using their old measurement technique when they sent this datasheet out.

Matches what we all have seen:

Attached Thumbnails
Strange Results of Injector Change-img_20130330_112808.jpg  
frizzles is offline   Reply With Quote
Old 03-06-2014, 12:16 PM   #19
Junior Member
iTrader: (3)
 
Join Date: Aug 2010
Location: Oklahoma City
Posts: 76
Total Cats: 1
Default

Yeah, that's the old data sheet and data.

Keep in mind, the values shown are raw data values using the equipment we have in house, an oscilloscope and direct power source for a trigger.

Since the signal has to pass through the longer wires and other components of the harness before it gets to the injector, the values plugged into the ECU are going to be a little different. The shape of the curve should remain the same, though.
LowFlyin' is offline   Reply With Quote
Old 03-06-2014, 07:04 PM   #20
Newb
 
Join Date: Mar 2014
Posts: 5
Total Cats: 0
Default

Thanks Chris for the quick reply. Would you mind telling us how the values were measured before?
frizzles is offline   Reply With Quote
 
 
Reply

Related Topics
Thread Thread Starter Forum Replies Last Post
Project Gemini - Turbo Civic on the Cheap Full_Tilt_Boogie Build Threads 58 Yesterday 10:04 PM
Changing from 460cc to flow force 610cc in diypnp 90 Turbo MEGAsquirt 19 10-19-2015 04:23 PM
Going back to stock. Need some 1.6 parts. Trent WTB 2 10-01-2015 01:15 PM
Time to start learning and play with tuning The Gleas MEGAsquirt 3 10-01-2015 10:30 AM


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -4. The time now is 11:13 PM.