mke
mke SuperDork
11/3/24 7:54 a.m.
Pete. (l33t FS) said:

...it has 2/3rds the displacement but 1/3rd the weight?

The comment confused me at first but I think what he was trying to describe is just how the low end feels.  His toy car is an S2000, about the same size as the 308 and about the same peak torque and hp rpm as my car so as it revs it kind of feels the same, faster but similarish.  But at 2k rpm an S2000 isn't happy about anything but his pickup is so I think he was trying to process how it could be that rpm didn't seem to matter.  

And you got it right or 2 accounts.  1st is there is not replacement for displacement and 5.4liter pulls a lot harder down low than 2liter no matter how its tuned. 2nd when you said ported.  People love to hop up honda engines because they make all kinds of top end, but they do it with monster ports  that are death to lowend performance and torque in general.  Google says that car makes 153ft-lbs @ 7500 and then 238hp@8300, only 1000rpm between the peaks is kind on the definition of a peaky engine.  My engine has much smaller ports giving higher velocities so I get a lower torque peak @7k and higher rpm peak@9k that is only 9k because I'm intentionally choking the exhaust, it wants to be 10.5k where I fear it would explode. So much wider and flatter, and that flatter torque curve carries most all the way down to idle so at 2k rpm it should be making like 250-300 ft-lb torque, more than double the S2000's peak number and reminding more of his pickup than his sports car.

Its kind of hard to describe.  It just drives nice everywhere and acts quite civilized, until it doesn't devil

 

Your tuning oops reminds me of the time I was in the middle of a road tune and, for whatever reason, cycled the key while rolling down the road in gear.  My computer runs speed-density and takes a barometric reading when first powered up... from the MAP sensor.  There are limits to how far it will go, but I ended up tuning the car with an assumed baro reading of 80kpa.  Oops.

Gotta love megasuirt.  I'm pretty sure they are the only ones who do that because its boarderline retar....not wise 

JoeTR6
JoeTR6 SuperDork
11/3/24 10:26 a.m.

It's easy enough to run a dedicated baro sensor with Megasquirt.  That was a requirement for setting up the TR6 in Colorado.

Perhaps worse is what Mazda did for the Mazdaspeed Miata.  They used a single pressure sensor to read manifold and baro pressure that is switched by a solenoid controlled vacuum switch.  Since the solenoid switch is exposed to atmosphere and the intake, it accumulates dirt and oil vapor.  It's a matter of time before it sticks and causes an incorrect pressure reading.  The infamous MSM "bog".

Pete. (l33t FS)
Pete. (l33t FS) MegaDork
11/3/24 10:44 a.m.

In reply to JoeTR6 :

Chrysler did that for many years, too, with a solenoid to switch from manifold source to atmospheric source to double check the baro reading when decelerating in gear.

 

 

mke
mke SuperDork
11/3/24 10:48 a.m.
JoeTR6 said:

It's easy enough to run a dedicated baro sensor with Megasquirt.  That was a requirement for setting up the TR6 in Colorado.

I'm sure that's true and i need to lean not to kick the MS bear, I know its quite popular and mostly not too bad theses days.  My point was normally when you selector no baro the ECU doesn't use baro because while nice its not needed for speed-density tuning.  MAP means Manifold Absolute Pressure, the MAP reading is correct where ever you are and so is the tuning for the most part.  What does happen though is the what you maybe tuned as mid throttle and 0.95 lambda becomes WOT where you maybe would want 0.9lambda.  Not perfect but fine.

The MS scheme uses baro in the math whether you have one or not with the record at start hack.  So if you change elevation the tuning is flat buggered meaning the baro sensor is always required.   Its fine to require a baro and a better setup, if they were just honest about that and I think the resellers are all recommending it these days.

The mazda issue sounds like they know they needed to add a sensor but didn't ave another input pin in the ECU.  Clever but its always the mechanical parts that let clever systems the system down.

mke
mke SuperDork
11/3/24 10:51 a.m.

Walked into the shop and what the...

Oh my.....at least now the source should be obvious

Hot start was getting sad so I figured Saturday's no fuel pressure drive maybe fouled something but the 1st 3 plugs out say I need to look a bit deeper.  I had the cylinder fuel correction off and a quick look at sync suggest it would have made the issue worse with lean 4 the lowest and rich 6 highest MAP.  I have a handheld manometer with differential mode so I think I'll redo sync using that once i find the oil problem.


 

JoeTR6
JoeTR6 SuperDork
11/3/24 7:26 p.m.

No worries about the MS criticism.  I like mine, but the tuning can be a pain and I realize there are better ECUs out there.

I hope that coolant leak is easy to track down and fix.  This is coming along very well.

mke
mke SuperDork
11/3/24 11:02 p.m.
JoeTR6 said:

No worries about the MS criticism.  I like mine, but the tuning can be a pain and I realize there are better ECUs out there.

I hope that coolant leak is easy to track down and fix.  This is coming along very well.

Better is subjective.  Many cars run just fine with the various MS versions.  The baro thing though I always found weird is all.

Sadly that puddle is motor oil.. and I'm struggling to find a source.  I took the car on several runs Saturday, the last with my buddy and I pulled straight in and shut down.  On that run we both saw the oil light just blink on a shift...rpm dropped and it was hot so low rpm, low pressure, thats what i thought in the moment.....but the oil leak makes me wonder if maybe it could be be something more like the pickups not keeping up so the sump was full and the tank empty and it hadn't fully recovered when I shut down so oil was sitting where it doesn't normally sit? That or maybe the frame is leaking motor oil....some kind of ferrari corrosion protection system?  The tank level seems low, I need to fire it up and see where it settles and then maybe I'll drain the sump to see how much is in there....

 

mke
mke SuperDork
11/5/24 12:12 p.m.

Most all the velocity stacks broke....its like the screws tightened themselves while it was running.  

I'm reprinting them now and ordered some metal sleeves...when they get here I'll machine the prints to accept them and that should prevent a repeat.

I was planning to fire it back up this afternoon, but the new WB controller should be here today so I'm thinking I'll get those on-line and plan to run Wed or Thursday.  My general plan is ignore the oil leak for the moment and focus on the other issues.  

For the oil, it about to get pretty cold so I will just plan on pulling the engine.  That will lets me fix the leaking front cam cover, reseal the timing cover and probably rework the whole feed/return oil pump part that makes installing such a bugger. and if I'm doing that I'm thinking lower the return pump and hoses....I know the setup is suboptimal and strongly suspect I'm seeing the oil level in the sump vary significantly so exactly when I shut down its full or empty and leaks or doesn't so 2 problems.  That and fix the power steering are the big must do winter items and I guess suspension geometry will move to if there's time list.

APEowner
APEowner UltraDork
11/5/24 1:38 p.m.
mke said:

Most all the velocity stacks broke....its like the screws tightened themselves while it was running.  

I'm reprinting them now and ordered some metal sleeves...when they get here I'll machine the prints to accept them and that should prevent a repeat.

I was planning to fire it back up this afternoon, but the new WB controller should be here today so I'm thinking I'll get those on-line and plan to run Wed or Thursday.  My general plan is ignore the oil leak for the moment and focus on the other issues.  

For the oil, it about to get pretty cold so I will just plan on pulling the engine.  That will lets me fix the leaking front cam cover, reseal the timing cover and probably rework the whole feed/return oil pump part that makes installing such a bugger. and if I'm doing that I'm thinking lower the return pump and hoses....I know the setup is suboptimal and strongly suspect I'm seeing the oil level in the sump vary significantly so exactly when I shut down its full or empty and leaks or doesn't so 2 problems.  That and fix the power steering are the big must do winter items and I guess suspension geometry will move to if there's time list.

i wonder what the CTE is on the stack material vs the fasteners.

mke
mke SuperDork
11/5/24 2:25 p.m.
APEowner said: i wonder what the CTE is on the stack material vs the fasteners.

That could be part of it, but if you look at the pic its like the screws are driven through the part by about 1/8" in some cases and sitting much deeper than the the non-broken spots. When I called my son out to fire up his printer he thought I put them on with the screw gun and broke them all...crack, crack, crack....I had to ask if he truly believed I was stupid enough to not stop after the first one rather than break them all one by one if that was how they broke?   I've never seen anything like it.    With the new sleeves, the stack will basically float with little to no load on it...hopefully that fixes it.  The next step i guess would be make metal stacks and that is a lot more work for me so I'd like to avoid it.

mke
mke SuperDork
11/6/24 10:00 a.m.
Paul_VR6 (Forum Supporter) said:

look to see if there is an idle timing table. I like to use the bottom row for decel advance coming off fuel cut, and it would be hard with a single table. Other way to "fix" timing is just put 6 everywhere for mow like its 1998

Is there a particular ECU you've used and really liked for this? Looking for ideas to borrow. as I have not been loving what I've been able to do with idle timing.  The better it runs the hard it has been to get it to drop into idle and the higher rpm I had to start pulling out timing...last weekend I was pulling timing  at 2k and its become kind of surgy.  So, last night I created an idle timing table (1D using CLT as the input but I may need to make it 2D for rpm and/or add additional conditions to make decel right).....but I just hate table switching so what I wrote is below 2% throttle pedal its idle table, above 5% throttle pedal is normal, and between the timing is interpolated so a pseudo 4D timing table.  I also added a cranking position for the throttle so hopefully tonight I can see how it all works.

 

mke
mke SuperDork
11/7/24 9:03 p.m.

I put the bushes in the resin printed parts and can now torque to my heart's content.

A buddy offered to try filament printing with a high temp polycarb so if these fail, that's next I guess.


and while it feels like I step back as much as forward, this feels like forward

 

I added some new code so cranking throttle is in crank position and timing is normal so it fires up nice, then drops to idle settings.  Coming off idle there is a smooth transition in the timing which should make it a bit easier to get rolling form a stop, I'll test that bit tomorrow.

The new WB controllers are in and working and I still saw a little difference bank to bank so I settled on sync the TBs, adjust the bank linkage to get matching lambdas then resync down the banks.  The new lambda values are very close to what I was getting from my rear bank sensor and kind of confirmed the front bank so pointing in the right general direction but was not to be trusted which is kind of what I thought.  It seemed good tonight, hopefully its all still good tomorrow.

A little public service message because I know there are a bunch of MS ecu people, the calibration voltages the controller outputs are very helpful, but  spreadsheet 14point7 publishes for helping calibrate the analog output only works right at 14.7AFR. If you plug in 14.1 for E10 or in my case 1 because I use lambda not AFR, the calibration numbers are gibberish.  I sent an email with the required correction so hopefully they'll fix it.

Paul_VR6 (Forum Supporter)
Paul_VR6 (Forum Supporter) UltraDork
11/8/24 9:50 a.m.

If you do it in lambda it all works out though. 
 

For the idle even the simple way ms does it with the rpm/timing curve in ms2 or the closed loop idle timing on ms3 is good enough. You need need to make sure fuel is good before it works well though. 

mke
mke SuperDork
11/8/24 11:02 a.m.
Paul_VR6 (Forum Supporter) said:

If you do it in lambda it all works out though. 

I do work and lambda and that is why the issue popped so clearly because the cal values were nonsense.  A little digging found =10*C9+C10 and =20*C9+C10 in the output where the 10 and 20 are the AFR at 14.7.  Replacing them with =0.68*C3*9+C10 and =1.36*C3*C9+C10 where C3 is the AFR conversion factor corrects it.  The rest is all good as its already done in lambda.

 

For the idle even the simple way ms does it with the rpm/timing curve in ms2 or the closed loop idle timing on ms3 is good enough. You need need to make sure fuel is good before it works well though. 

I'll take a look at what they do.  My issue is probably a little specific to me and not one I've deal with in the past.  The TBs are just big, needed for the top end but big.  The engine is also on the big side displacement wise if not cylinder displacement wise andf together idle timing is 6 degrees and 2k running is 20 degrees....idle throttle position is 2000-2500 with normal timing so I failed on the just use the main timing table path.  Going to an idle timing table gets it to drop into idle no problem, but then it explodes out, just touch the throttle and the tach goes to 2500....at least it did before I added the interpolation region which parked was giving a nice smooth transition where I could find and hold any point in between 1000 and 2500....I'll road test it later today but I'm hopeful.

I think once I get it behaving nicely open loop I'll see what I can do to closed loop idle control so the rpm is stable with time.

wawazat
wawazat SuperDork
11/8/24 2:37 p.m.

If these newly printed velocity stacks crack, print them in nylon not PC.  Better chemical resistance and is the same base material used in injection molded intake manifolds by the OEM's.

wawazat
wawazat SuperDork
11/8/24 2:37 p.m.

If these newly printed velocity stacks crack, print them in nylon not PC.  Better chemical resistance and is the same base material used in injection molded intake manifolds by the OEM's.

mke
mke SuperDork
11/8/24 6:28 p.m.

Today I turned the key with the engine at 65F and it started and settled to idle without me touching anything, really nice!

....but it just didn't sound quite right.  Coil output 1 is not working correctly and I run waste spark so cylinders 1&6 are dead.  I do love this ECU but I'm not sure they are still actually in business and I know its not in the budget.  Its always something.

mke
mke SuperDork
11/9/24 7:19 p.m.

I spent some time looking at the ecumaster pro16 and its pretty ok overall.  Its got a bazion I/O pins but not a lot of ways to use them that I could figure out.  The list of pre-defined outputs is fairly short and there are only 5 custom outputs which are pretty nice but its only 5 so no way to use more outputs, at least not that I saw.  Simple stuff like there is a defined channel for check engine light, but the only way I could see to use it other than to select it in a custom output.  Kind of the same in inputs, LOTS of pins available but how to use them?  Its a maybe at $2544

The haltech Nexus 5 seems quite good but more expensive .  But it allows 20 generic inputs and 20 generic outputs and can definitely do almost everything I currently do other than trim fuel from 12 MAPs, but it can read and display them.   Overall its the nicest I looked at today, but $4500 though, nearly double the price of their 8 cyl unit and the other options on the list sad

The prices on the motec M150 have dropped but other than it is 12fuel/12spark its still pretty sad until you add a bunch of upgrades and they've removed the generic examples that I know were there once upon a time so there is no way to actually see if it can do what I want to do.  I think it mostly can?  I know if I add the $2400 development license i can make it do everything my enginelab could do (except log, that;s another $400 for basic, $1500 for  good),  but there does not appear to be anyway to know what lesser licenses do other than a generic feature list.   But plus side it looks like if I order from Oz an M150 with the basic GPA license can be had for $2500+ shipping & connectors and from what I could see in the 1 example it will probably mostly work?

The enginelab website will take my order for a replacement, just plug it in and I'm done. $1900

There are no manuals anymore so I have to downloads the software and go through it with whatever in application help there is so its slow and maybe I just mis what I was looking for not really knowing what they call it or where it might be found.

oh and no oil puddle Friday after running Thursday, but a huge puddle today after running Friday. No idea what's up with that.

mke
mke SuperDork
11/9/24 9:45 p.m.

I lied.  I got a drink and spent a couple hours looking at what I could do with a rusEFI Huge for $825 with the connectors

rusEFI huge

I have an open question on the current rating for the H-bridges....so can it drive my 7A actuator.  

Its open source so a little lite on support but it looks like it can do everything except the MAP cylinder trim and TB sync if I learn to work the scripting that lets you add custom features.   I MIGHT be able to make the TB sync display work with the 2 custom gauges that are available and CAN info for the reading and a CAN selector to tell it which reading to display.  It doesn't look bad if the H-bridge has the current I need.

Paul_VR6 (Forum Supporter)
Paul_VR6 (Forum Supporter) UltraDork
11/10/24 8:07 p.m.

I have an ee degree and efi savvy and couldnt get a rusefi working in a reasbable period of time. Its like a bunch of programmers that never tuned a car just had an orgy. 

not an ecu masters fan at all, there are some features either dont work or arent documented in a way where they can be figured out. 

haltech is generally decent. 

maxxecu would be another to look at. 

I've gotten somewhere around 700 cars running on ms but its not for everyone

guy with a ferarri should have motec money 😅

APEowner
APEowner UltraDork
11/10/24 8:57 p.m.

Have you thought about fixing the ECU you have?  The driver circuit should be pretty straight forward and you've got functional ones to compare to.

mke
mke SuperDork
11/10/24 9:35 p.m.

In reply to Paul_VR6 (Forum Supporter) :

Funny because my today rusEFI forum post was:

"Ok, I've invested about 6 hours searching and reading and sadly I think its time for me to tap out. Its clear a ton of effort has gone in, but if I'm being honest with myself there is just no way I could get this working with the documentation available. I programmer I am not :("

The basic ECU has come a long way and Im sure I could get it working and the control logic has gotten to where I honestly think its pretty ok.  

The problem for me is the number of selections.."iF this revision then that" with the then that part not documented.  The effort in the last few years has clearly gone into Lua which is a scripting language (and I fickin hate scripting languages) to access the the full functionality...where to find it, how to load it it, none of that is documented anywhere I could find it and I just don't enjoy those kinds of battles.  So as capable as it appears to be and as much as I wan tot save money its just not for me.

 

maxxECU is on my to-look list.  maybe download the software tonight.

haltech was the first ECU I installed on this car 24 years ago so I have a soft spot for it....but the latest version looks quite good.

motec - I had a motec M800+ignition expander I almost installed but sold.  Then I had an M150 I almost installed but sold.  Their reputation is great but I honestly can't understand why, all I've ever seen is money grubbing ashholes telling me I should be happy to pay more for less features and support.  Bosch owns them now and something feels a bit off....the M150 is $1000 less than it was,  $2-3k less ordered from Oz with licenses, that's never happened.  But also as I posted earlier there is no longer any way to know what any or the ECUs actually do and they no longer offer any training on the development package, I believe because that competes directly with bosch motorsport stuff.  I'm pretty sure they are pushing the ECUs to general market or dropping them and keeping all the accessories much as holley did with AEM.  The motec come with no data logging....there is an FPGA and I think 512M storage on board that is capable of 2000 channles at 1000hz you can't use.  You get 30 sec storage when the PC is connected.  For $400 you get 200channels at 200hz, then for another $1400 (after you pay the $400 im pretty sure) you get everything the HW is able to do.  Its the same with the firmware "packages".  The ECU comes with nothing, completely nonfunctional.  $800 get you basic functionality but it only reads $400/channel motec lambda controllers, $1200 adds traction control, $1500(?"?)adds paddle shift, $2400 on top of the $3k hardware and its a less good enginelab fully programmable ECU (logging is still extra) so I'm pretty sure I'd roll the dice on another enginelab ECU before giving motec my money.

MS...700 installs?  Wow!  The only one I ever messed with was to remove it for a guy 15 or 20 years ago.  I know its WAY better now but its still 16 bit and sort of weird left over for the old days.  I just couldn't.  I had lunch with Bruce once, pretty nice guy.  The MS3 was originally designed to use a MPC5634 processor and at the time time I was involved with an open source ECU project using the same family...in hindsight I probably should have made a deal with him because my project folded as most open efforts do. And they put on a new version of the old processor and launched a decent product.\

My problem now is budget.  I agreed to hold car spending to $250/month, I need a lot of months to pay for a new ECU.

mke
mke SuperDork
11/10/24 9:42 p.m.

In reply to APEowner :

I have....but that is not really my thing.  It's a 2 sided board and the connectors need to be unsoldered to remove it.  I asked about another dead driven and they told me 50/50 chance.  Maybe for this i send it.

mke
mke SuperDork
11/11/24 6:26 p.m.
Paul_VR6 (Forum Supporter) said:

maxxecu would be another to look at. 

I've been through the software and manual.  There are many things I like and I can get it to do all the think I currently do other than the cylinder trims,  The CAN support looks top shelf, I could probably put that to good use.

..... but I have no idea how its calculating the fuel.  Its asking questions in odd ways so I really wonder why.  It honestly feels like the actual engine control is an after thought. If I buy 1 of these I'd need ot run it on the bench first I think to understand what its doing so I know how to setup everything. 

mke
mke SuperDork
11/14/24 9:02 p.m.

I do really like a lot of things about the maxxecu pro, the software is just nice. The actual engine control is state of the art 2000 but the software lets me alter and add in enough places that I'm sure I can get it working nicely.  The h-bridge outputs (throttle drivers) are just sad at 2A when my throttle (BMW E46) actuator draws 7A so I'd need to keep the external driver I currently use.....but Lana did not approve a $2500 car budget advance.  Maybe for the spring.

The ecu I have has 10 spark outputs, #1 is fried.  With many newer ECU it would be a non-issue as i could just pick a different output but ecu automatically assigns 1-6.  Tonight though I had all 12 cylinders running again.  The software lets me select 12 cylinder coil on plug as if there are 12 outputs because early HW did have 12...mine actually does have 12 too but the softer can't access 11&12 correctly.   But that wouldn't help anyway, output 1 is dead....but it also supports 2-stroke and allowed me to select 12 cylinder 2-stroke so now I can waste spark off any 6 driver I want. The spark is fixed

The down side is I now have 2 fuel pulses per cycle.   I still have 12 outputs so cylinder trim still works.  I set the global fuel trim to 50% but it was still pig rich and idle VE went from 50 to 35...I may put it back to 50 and mess with the global trim more.  It wasn't running great and I realized after I shut down the injection timing table at 360 on a 360 degree cycle it not right.  Also I can't set 2 cylinders to the same crank position, I can but the ECU gets confused and cuts spark.   so I set the pairs (1/6, 2/5, ect) 0.5 degree apart....I think I'll increase it a bit though as I don't know what the ECU is actually capable of and injection angle isn't all that critical so no need to find the ECUs limit.

But it runs again.  

 

 

You'll need to log in to post.

Our Preferred Partners
CifMk6qJjwmmARSDDOWtrC6UtJE31Or3A42D96vx238h2MmWbxOZ959IEqUQdruR