FrankenFerrari - V12 Ferrari 308

Tonight I realized 2 things can be true at the same time, I had low oil pressure and the sensor was dead.  but now the dash gauge it displaying EXACTLY what the ECU tells it to...first time its ever been right I'm pretty sure

I'm assuming it's better in the car, but i really had to look at those pictures to see the needle in the gauge.

Ian F (Forum Supporter) said:

I'm assuming it's better in the car, but i really had to look at those pictures to see the needle in the gauge.

Easier....but the gauges are for sure more stylish that functional.....they are hard to read and not even close to calibrated.  The needles on the speedo and tach would be much easier to see if they weren't hidden by the steering wheel most of the time. I have the ECU driving the tach too of course....and added a calibration function similar to what I just did on the oil pressure so it also reads the correct rpm rather than just a good estimate as it used to. Its stuff like this that I love about ECU, I can just add a feature that no other ecu has because only me and 3 other people on the plant would ever want or need it  

....all the little things that add "character" and make the car so lovable  

bentwrench said:

The worst part about those TBI kits is the limitations of TBI and they still need an EFI fuel supply, so you have the work of installing a proper fuel system and don't get the benefit of port injection. 

They make it sound as if you can simply bolt on a carb like device and have EFI. In the real world there are way too many factors to consider and TBI introduces some of it's own. It is next to impossible to create something that will self tune on anything you bolt it to.

Marketing.....

There was an episode of Engine Masters (I think) where they were comparing 2 engines, or some combination of cams and exhausts on one engine - and they were using a Holley TBI setup. They actually ended up ditching the EFI for one of the combinations because "the EFI's computer couldn't quite make sense of that combo". I think maybe Holley was a sponsor. And IMO that one little line and the change to throw a carb on there should have been the focus of the whole episode - WAIT, they're charging $1500+ for this setup and it can have irresolvable tuning issues on an American V8? Like, their whole value proposition is just a total joke?

It wasn't like some oddball setup that they were trying to fool the EFI system with, either. I wish I could find it now.

Looked at the fuel gauge...another bad sending unit. Best I can tell its the exact same gauge as the oil pressure with a different printed face...same 8V signal line, 0-300ohm sender......I'm a little temped to install a cheap common sending unit in the tank as the replacement is $200 and let the ECU run the gauge...hmmmmm

Edit: Fuel gauge is fixed. Opened the sending unit and the wipers are copper and had a bit if oxide, and both ground through just a dot of a contact point to the metal cover...a little 600 grit to polish everything up and it's good as new. Took under 20 minutes including drilling the rivers and finding screws with nyloc nuts to replace them. Thought I snapped pic but it turns out no, sorry

Tonight was the big night, I connected the CAN expander, fired it up....yeah, nothing.  Email out begging for help....there is a reason I waited to do this last.....

In reply to mke :

Dang. That's a bummer- but I'm sure you'll get though it. 

mke said:

Tonight was the big night, I connected the CAN expander, fired it up....yeah, nothing.  Email out begging for help....there is a reason I waited to do this last.....

Everyone reading this knows you're gonna get it sorted out.

You guy have more confidence than I do I promise you that!  I did get the expander spitting out something, I can see it on a scope now.  The ECU is still ignoring it at the moment.  I cheated a little on the wiring so I will get the exact resistor its supposed to have and correct that before proceeding.

In reply to Syscrush :

That is the the thing isn't it?  Carbs have a real elegance about them, especially the slide type....they are analog computers really.  The venturi or area under the slide is  a mass flow sensor that outputs a vacuum signal rad by the fuel jets and while the vacuum signal grows as the square or the mass flow, the fuel mass flow meters (jets) add a restriction proportional to the mass fuel flow that when done right gives a constant air/fuel mixture at all air mass flue points.....they just work, on most engines anyway.

EFI is the brute force approach to the problem.  You can correct for anything and everything, just keep adding sensors and tables until you get it.  OEMs spend months if not years tuning.  A guy i know used to be part of the tuning teams, they'd drive for weeks, though mountains, desserts, everywhere logging everything the whole time.  and that is after the dyno teams do their thing for months.  The quality of tune the OEMs produce are amazing to me honestly.....-10F outside, turn the key and it starts then you pull away....gone are the days of hoping it starts, spraying ether or pumping the gas pedal then trying to t keep it running...let it warm up before you dare to drive.  Amazing.

DIY EFI installs are in between somewhere with most, if we're being honest, closer to carb the OEM.  I've personally never seen a single install that is OEM quality.....even saying we're talking say mid 90s oem quality., mid 80s quality installs I've seen.  Most setups these days will run and drive pretty well once they are warmed up much like a good carb setup, but maybe have a few added features so overall better than carbs....but everyone I've ever messed with there are just little things I just can't quite fix....like hot start after 5 or more minutes is great, but at 10 seconds you need to open the throttle or I can have good stable idle with a little stumble off idle or I can have no stumble with an unstable idle.....stuff like that.  Not a big deal really but the project never feels done really because you know its not OEM quality...its good but......frikin OCD......

Way back I damaged a couple clips inside the main ECU connector so today was replace the connector day father than start adding more wires to the broken one, still there were about 50 to move 

Probably a good chunk of the the rest of the day will be new wires.  The expander needs its 12 MAP signal lines for starters and I'm not even sure what all the wires in the cabin are all about , clearly I was plotting something 5 years ago.

I am a little proud of myself for planning the CAN expander.....there was a bundle with all the power, ground signal all run and labeled.  

All the wire I used everywhere is aircraft grade and all the analog signal lines are shielded, which slows things down a bit.  Its a bit harder to strip because its fine strand and each strand is cad plated so you need to be really careful not to nick them....I use a thermal stripper

Then on the shield stuff the shielding is braided so it needs to be unbraided and either twisted and run to ground (I ground the ecu end, never ground both ends) or they make these cool connectors that heatshink and then solder themselves when you get them hot enough, which only works with tefzel aircraft wire, pvc auto wire would catch fire.  The solder area changes from red to clear when the solder melts so you know its done and the heatshink seals so moisture can't get in....pretty cool but it take time to do.

Finished wiring in the MAP wires yesterday but couldn't get the damn thing working....played with every setting....moved it to CAN B......went inside and had  drink.  This morning I remembered that early on I'd mistakenly type the CAN addresses in HEX and the letter part was accepted...hmmm...I wonder....yeah, changed the addresses from decimal to HEX and I have 12 MAP readings.  The way the board works the 12 readings are not truly independent even thought there are 12 sensors, but its easy to see the 0-10% error mistuned TBs would make and larger errors also show just not in a simple linear way, and because this is a board thing the old way I was getting the signals would have the same issue. 

A couple little things to fix in the model and display but its mostly working now....at least with a brake bleed pump hook up to pull vacuum,  still need to see what it does on a running engine.

Real men use binary.

LOL, yes, and no I really don't speack hex or binary 

What's that they say?

There are 10 kinds of people:

    those that understand binary 

     those that don't

On that whole 2 steps forward 1 step back path getting the cylinder reading working made me realize the multiMAP itself was really working right.  When I saw the output didn't match the lowest cylinder I quickly told myself and all of you that's normal....no its not.

The blue is MAP1 voltage, which I confirmed with a meter, its right.  The purple is the multiMAP output, it doesn't track right and never goes below 1.3V, but by 1.4V its not longer tracking right at all, so while it could read to about 37kPa, 40 kPa it had really lost any hope of calibration.  

hmmmmmm........Pulled the wire out of the ECU connector and check with a meter...same  answer

hmmmm......where's the circuit drawing, this is normally where I get myself into trouble because, yeah, I suck at electronics, but not today.  The analog inputs on the ECU all have a 100k 5V pullup so when they are not connected to anything they read 5V which while often very helpful its buggering my NBO2 sensor reading and I thought maybe this too......but no.  What's buggering this is the isolation amp, its apparently not setup right and doesn't have the correct range..  But, this thing is a pretty dead simple analog circuit, I just moved the output wire to bypass the amp and its pretty much fixed.  Now I'm pretty sure the 5V pullup is causing voltage reading not to drop quite as low as it should, but that I can and did calibrate and now the MAP is about 1.5kPa above the cylinder reading at atmosphere, and about 1kPa high at the most vacuum I could pull....plenty close enough (red and orange lines)

Trying to work through some of these wires. 

I connecter the rear axle speed....it reads the speedo sensor and that end was wired, now it goes to the ecu.

Looking up front I see a couple long 3 conductor wires...must be wheel speed, easy.  there are 3 2 conductor....I'll make one CAN to make future work easy, you know, now that I'm a CAN system master :)

 my wiring pinout shows white/yellow as the "slow down" light

I show a clutch and brake pedal, I probably thought they would be good for traction and/or launch control....a couple of the orange wires I guess.

Then another 3 wire labeled steering so it looks like I was planning to add a steering angle/position sensor for logging I guess I was thinking.  anyone have a good (aka low cost) solution for steering position?

This stuff takes forever....

On a more serious note, I've been enjoying following this project.  I consider myself lucky if I know something you do not.

Spent a bit more time on MAP sensors after weather.com told me the correct baro reading for my area was 101.5 not the 99.5 may sensors displayed....101 ish now.  On the other end the sensors on the multiMAP have a min output of 20kPa while the baro goes down to 10 which let me confirm my brake bleeder pump was pulling under 20 and I could use the lower flatline as a second cal point.  That all seems close enough now and much better than it all was.

Syscrush said:

bentwrench said:

The worst part about those TBI kits is the limitations of TBI and they still need an EFI fuel supply, so you have the work of installing a proper fuel system and don't get the benefit of port injection. 

They make it sound as if you can simply bolt on a carb like device and have EFI. In the real world there are way too many factors to consider and TBI introduces some of it's own. It is next to impossible to create something that will self tune on anything you bolt it to.

Marketing.....

There was an episode of Engine Masters (I think) where they were comparing 2 engines, or some combination of cams and exhausts on one engine - and they were using a Holley TBI setup. They actually ended up ditching the EFI for one of the combinations because "the EFI's computer couldn't quite make sense of that combo". I think maybe Holley was a sponsor. And IMO that one little line and the change to throw a carb on there should have been the focus of the whole episode - WAIT, they're charging $1500+ for this setup and it can have irresolvable tuning issues on an American V8? Like, their whole value proposition is just a total joke?

It wasn't like some oddball setup that they were trying to fool the EFI system with, either. I wish I could find it now.

I actually watched that episode and still kept trying cause I had already bought the TBI and installed it. Dual plane vs Single plane intakes is a big controversy because the TBI sellers say it does not matter, when in reality single plane is the only option and it is counter-intuitive for a street car.

I am convinced that the TBI systems CAN work, but it is going to rely on equal parts luck and EFI knowledge/skills  if it is to do so. I actually learned enough about EFI to make it worthwhile, because a big part of what I learned is that no way, no how, would it ever be as reliable as OEM EFI or an analog carb. With that learned, I have a carb on the car now. Fitech and Holley VASTLY oversell the TBI.

Bottom line, if you had to fly in a home made plane with a carb or a TBI,( or any other hobe-brewed EFI)  which would you choose?

mke said:

Spent a bit more time on MAP sensors after weather.com told me the correct baro reading for my area was 101.5 not the 99.5 may sensors displayed....101 ish now.  On the other end the sensors on the multiMAP have a min output of 20kPa while the baro goes down to 10 which let me confirm my brake bleeder pump was pulling under 20 and I could use the lower flatline as a second cal point.  That all seems close enough now and much better than it all was.

The baro reading reported by weather.com is corrected for elevation.  You're ECU is likely expecting an absolute reading.

MegaSquirt

APEowner said:

The baro reading reported by weather.com is corrected for elevation.  You're ECU is likely expecting an absolute reading.

It shouldn't be but I'll see if I can find another data source just to be sure. We're only at about 600ft around here but that is 2kPa-ish relative to sea level so I'll see what I can find.

Edit..oh G..d D..m it you're right.  I need to re-do the calibration. 

bentwrench said:

MegaSquirt

I've already got plenty of problems to sort, I don't need more that have no solutions 

Nohome - I am convinced that TBI systems will retain all the TBI short comings.

There are many reasons that the OEMs stopped using TBI, IMHO it is a waste of time to try and make one work.

mke said:

APEowner said:

The baro reading reported by weather.com is corrected for elevation.  You're ECU is likely expecting an absolute reading.

It shouldn't be but I'll see if I can find another data source just to be sure. We're only at about 600ft around here but that is 2kPa-ish relative to sea level so I'll see what I can find.

 

Edit..oh G..d D..m it you're right.  I need to re-do the calibration. 

Ok, that's fixed.