From what I've read about Pilot Injection in other diesel engines, it can produce more power and more efficiency at higher RPMs. As I understand it, PI promotes more complete combustion.

 

Nope - Pilot injetion is disabled on tip-in...
Well for those of us that have had the flashes, it's just disabled :-)

 

Yep, if it means anything..

 

My truck is an Aug03 build. I drove it off the lot with PI enabled. My brother bought the same truck a week later with PI enabled. I received a recall letter in the mail for the Dec03 flash, which disabled PI. The reason the letter stated was due to California emissions, it said nothing about problems with the truck. While getting service, the dealer flashed the truck and disabled PI. The truck is louder and mileage went downhill. I advised my brother not to go near the dealer.. which he hasn't, and his truck is much quieter and his mileage is much higher. So, it's my opinion that Ford must've had some problems with EPA or something.. and disabled the PI for that reason...

 

There is a guy in houston Texas that makes and sells a tuner called a Tweecer for EECIV Mustangs, trucks and cars. It would be a small thing to get him to build one for a PSD
The fuel injection control module would have to be tackled seperately

 

The 6.0L is a completely different beast. It doesn't use the EEC-IV/EEC-V type computer that the Tweecer is designed for.

 

I bought Autotap some time ago for my now gone Suburban. When I tried to use it on the PSD it would not talk to it. Called Autotap and was advised that the software does not support the PSD because Ford went to a CAN? bus system on the 2003, and they are still 6 months away from a product that interfaces with it. All mfgs have to be on this system by 2007 but Ford went early with the PSD and everyone is scrambling to catch up. I suspect that is what took Diablo so long to get a product on the market.

Software is not going to fix the current PSD FI pressured injection problem. The fix lies in redesign of the injector or switching to a pump pressure rail system.

Has anyone ever seen a printout of the code associated with the 6.0? I have looked at a 90s OBD1 calibration/code for a Chevy product and it is 100+ pages of uncommented machine code and it was not reverse engineered until someone with inside knowledge 'hepled'. Even then it took 2 years of work to apply that knowledge to a specific year for a specific vehicle.

A friend of mine is a calibration engineer that worked for Lotus on the corvette LT5 DOHC in the 80s and he spent years doing dyno and road testing to develop the calibration/code for that motor. He now works for Toyota doing the development calibration for thier F1 engines. I once asked him how hard it would be to hack/develop current OBDII systems. He basically said it could not be done quickly without considerable expense (or inside information) and some very talented engineers and you can bet those are the people Diablo et all have hired to do the calibrations in the preditor and this is just the calibration tables. The control code is a different problem. If you want to see how just a simple DIY open source for fuel control is done then go here.

http://www.bgsoflex.com/megasquirt.html

This effort has been ongoing for 2 years and would be similuar to what you propose but the level of expertese would have to be increased at least 10 fold and it is just not going to happen. If it was that easy Ford would have done it.

Tyler

 

Well... I've said all along that we'd need some smart engineers and that this would cost us an uncommercial amount of time. That's the advantage of open source. Multiple actors can contribute for their own benifit without a single entity having to foot the whole bill.

Now... first, I want to thank you ... I've bookmarked that site. We can note that they're building a FICM for older gas engines, not a PCM/FICM combination. In fact, the majority of the engines they discuss would have been carbourated origionally... so it's a different type of project.

They're also doing this with 100x less horsepower. They're using a 68HC11 (I believe it's a variant with ram and flash) which is an 8-bit processor clocked at less than 1Mhz. It's certainly adequate for thier purposes, but not for ours.

We wouldn't be referencing the Ford code at all. When you write the source code, understanding a hex dump is not an issue.

Ford doesn't have any reason to be open with it's code. I don't know why Ford's actions would figure in your reply.

It's interesting that you talk about some hardware redesign to restore pilot injection. That's been a thought too. However, to be able to change injectors or even experiment with the problem, we need a platform that works.

I'm beginning to think (from the accumulated weight of posts) that PI working may just be the use of a specific (much smaller) list of oil for replacement in the engine.

 

I agree. The issue is a very complicated one, especially to people like me who require interfaces with no blank screens. Just changing the fuel map on the Tweecer was above the level I cared to invest the time to learn and it is relatively simple compared to what you are talking about. Sure someone could do it, but is the effort worth it?

 

There's a lot of conjecture here... let me give some info:

Diablo has had hardware talking to the CAN system for a long time (remember, the 02 Explorer has a hybrid!). They were tuning the 6.0 Powerstroke in the early summer of last year. There were a lot of issues with the truck then and they didn't want to touch it.... learned a lesson from Hypertech in the 90s with crappy Chevy engines that would self-destruct and I'm sure Hypertech got blamed for a lot of problems not of their doing.

I have no doubt this system you're discussing can be made. However, I think the limited number of people who will actually participate and/or use it is going to relegate it to a time-consuming project for a small handful of hobbyists. You see, most people don't want that much control, its simply too complicated and too prone to making mistakes that can result in a blown engine. The Tweecer is a prime example. Its a great system but it has a limited audience of hard-core users.

 

Exactly, but my answer runs up against the word police again!

 

Wow. I must say that the megasquirt page is full of information. They have 'Diesel' as a choice in their fuel tables, but none of the documentation I came across mentioned any of the issues that I've been reading about here for a few months.

However, it assures me of a few things:

• We have ample power. My design allows us to use virtually off-the-shelf parts which will make the unit fairly inexpensive and not require any electronic assembly. Our goals are different and our design is differnet, but we'll easily be able to run the engine on my design.
• The use of a simulator is valuable. They call it the "stimulator" ... but essentially, they found it very valuable.
• The use of logs for tuning and simulator use are valuable.
• That reasonable starting values for the system can be wrung out of physics calculations. I was somewhat worried about this ... there's a large distance between "ideal" and "real world" ... but they start with calculated values.
• That someone else has done this type of thing and that it's "doable."

Really, my initial estimate that in total we'd exhaust somewhere around 20 person-years of effort and produce a design that would be able to be used to control almost any diesel engine designed around similar principles seems about right. No corporation could pay for this, but this is where the concept of "open source" or "free software" comes in.

From a company perspective, 20 person-years is $2M. That's probably more money than you could make back in a few years. Someone (like the engineer quoted second hand above) who is familliar with the business case for doing or not doing something would immediately find it infeasible (Unless you were planning to ship 200k units). Note that $2M doesn't include facilities.

Open source blows those assumptions ... just like open source is giving Micro$oft a run for it's money, open source changes the economics of a project drastically. If 50-odd people work on the project (and there seems to be a good supply of odd people here ) each person is donating some fraction of a person year. Some more than others, but everyone is enjoying their endevour.

In the end, the participants have achieved something significant and a few of those individuals might even go on to commercialize some of the effort. Everyone benifits and the cost of creation is bourne by the creators ... paid by their satisfaction, use of the end result and, in some cases, some commercial gain.

People create things naturally. It isn't true that people require a monopoly on their work to have incentive to create. That's one model that works well for a certain class of things. Open source is another model. I don't know how much the average car guy has been exposed to open source (beyond reading about linux in the paper these days), but it works like this...

Bob knows a lot about cars. Bob's also a nice guy. Fred is an architect... and has outfitted his toolbox well, but hasn't had much time to use them. Fred's also a decent guy.

Joe, their neighbour, arrives home one night with a car problem. Joe asks Bob to help and they start to diagnose it. When Fred arrives home, Bob and Joe have decided they need a tool that neither has. Fred is only too happy to help and lends his tools.

Joe gets his car fixed. Bob is happy because he can share some knowledge with two friends. Fred is happy because he helped fix something, and learned how to use one of his tools to fix a real problem.

The internet has made us all neighbours. We've been talking over the back fences on this site now for a few years. Now I'm proposing that we should build something. Maybe it's a deck. Maybe it's more than that ... a community center or somesuch. I dunno.

I feel strongly that we can produce a PCM/FICM replacement if we had the right team. I think we'd need a couple more computer guys on the team... there's lots of work there... but we also need some diesel engine guys, too. Someone with a shop might be useful. It may be useful, for instance, to have an engine somewhere on a stand (connected to a load) that we could work with.

I digress. I'm even more convinced that it's doable. Here's the ring. A few more people have to throw their hats in.

...
(edited, corrected typo ... )

 

Keep in mind that you can't produce a baby in a month with nine pregnant women.

 

Already factored in the estimate. I suppose, to be fair, if you were a business paying for this, you'd narrow it down to 6 months of work with 4 or 5 people. Open source is generally less efficient in total overall effort, but it gets things done whilst occupying the time that is otherwise leisure.

 

Send me a finished product and I will dyno and drag test it