Well, I've considered this a few times and dismissed it each time on the 're-inventing the wheel' basis. Last time I bought an ECU was the link G3 costing £500 incl. VAT including a boost control solenoid - and that was new.
Looking around at the moment, most seem to be approaching the £1K mark and most don't have any sort of knock sensing built in without spending another wedge.
IMO most of the ECU's will have a total component cost below £100 when made in sensible quantity. I guess we're paying for R&D and big mark ups for agents, wholesalers etc.
The only sensibly priced ECU's I'm aware of are the megasquirt and potentially the TGT one if it ever gets released. Megasquirt is very much a DIY/boffin system - the more I try and make head or tail of the website the more confused I get. There doesn't seem to anywhere you can actually go and buy one as a full package with instructions & software. It's a marketing disaster.

With a firend of mine wanting me to develop a programmable electronic ignition unit for the vintage bike market it's got me thinking again about developing a full ECU. Also bearing in mind that with the downturn in my other business I'm going to have more spare time available as we're not going for any expansion this winter.

I see 2 possible routes:

1. Use the megasquirt processor board & source code as a base. Design a custom motherboard for the 3S which will be plug & play, including proper interfaces for our dizzy signals (main failing of megasquirt is you need to fit crank & cam wheels/sensors) and also knock sensor interface. Modify the software to control our cars properly including knock control. Sell as a package ready built, default map loaded, PC software on disk and full instructions.

2. Go the whole hog and develop from scratch.


option 1 has definate technical advantages in that much of the work is already done. It also has legal liability advantages in case of consequential loss claims in that people know they are buying a re-packaged DIY ecu, I am just adding an interface board. Downside is the commercial minefield of dealing with megasuirt and being seen to make any profit from their work (understandable)

Option 2 is possibly my favorite at the moment. The question is whether to go 'open source' with it. This would probably have advantages in terms of attracting technically minded customers, maybe advantages as far as consequential loss claims go. The downside is that it makes it easy for other companies to come in and undercut, although I doubt that will happen as I'm not planning on putting massive markups on - probably a selling price (direct) of £300 ish.

Looking initially at the n/a kit car market as there's less chance of damaging an engine with a bug, then moving back to the gt4 / mr2 in a year or 2 once it's a bit more developed.

What are peoples thoughts ?

_________________
If at first you don't suck seed, try drier grain.

Sounds like a plan chris, I like the idea and although not as technically minded as you I would like to help if I could, finances, doner car etc...

Offers there if you wish.

_________________
1998 ford st24, 186hp 2.5 v6- sold
1997 bmw e36 318ti, mint hellrot red- sold
1992 celica gt4 st185- td04 340hp/330lbs
1998 mercedes E280- sold
1998 pug 1.9 td- mile muncher!
1993 rx7 542hp savage monster!!
2007 BMW 335i- in the shop gettin built!!

My thoughts...what a good idea , if anyone has the electrical/technical understanding needed its you chris. Actually quite shocked at the price component wise of the ecu's currently on offer and if you can develop/manufacture an equivelent at a whole lot less then you are going to be one very popular bunny .
Getting a working prototype then refining into a saleable item sounds like some serious headache but fun at the same time ! Hope you go for it and suceed

_________________
2004 Subaru WRX

seems like a good idea and one you have the basics covered then the hard work is done, converting it to other cars would be relatively easy.

you want to be careful though, you may become a professionally tuner and distributor at this rate!

_________________
FN2 Civic Type R w/Mugen Parts
Toyota Celica GT4 ST205
Toyota Celica GT4 ST185

1) R&D time and cost expected?
2) market potential due to a variety of manufacturers out there already
3) technical, marketing and business skills required
4) cross platform potential eg works on a toyota, mazda, rover, bmw, audi etc; works on S3, S4, S6, V6, V8, diesel, petrol, hybrid etc
5) usability by diy and experienced tuners
6) software interfaces, logging, digital dashes etc
7) support
expected ROI on project and business with expected base unit cost, markup and expected sales etc
9) resource requirements, staff costs, current equity to maintain work on the business

I'll give you a couple of weeks to get back to me with a project and business plan

IMO - its option A. Provides foundation to build upon with support from other MS builders. Once the unit has been built, then you can build from scratch once you know what is working, what doesn't really work and what you need to build which the MS can't do.

I don't expect TGT to come to the party any time soon. And I don't expect any of this to be cheap.

Customers always want cheap - well a plug and play (using adaptor plugs) link, haltech or motec is about as easy as you can get as the installer just chucks in and tunes, gets power, sends customer on their way.

If you have a unit that has to be wired in, no auto tuning capabilities, and requires a tuner to have a degree in electrical engineering then the point has been missed.

If there is a truly viable cheaper option (please build in realistic staff time costs) and there are more than, say, 5 key features which are missing from the top 10 aftermarket ECU manufacturers, then you may have a market to play in.

Also consider that option B is just what megasquirt is for the most part. Sounds like a case of re-inventing the wheel.

I'm just saying.. as a project to play around with, fine. As a business, then you are going up against a number of well known companies and products.

Of note - completed MS units are sold on the local auction website, Trademe, all the time. I wouldn't expect there to be an issue to be honest.

_________________
GT Four Adovansu

"I want to enjoy the powerful and nimble agility behavior."

This is indeed the massive failing of the megasquirt. I do have a degree in electronics, and when I looked at it a few years back, the amount of time I expected to spend getting to grips with what someone else had done in order to make it work on my car, I would expect to spend less time developing from scratch.
I see a massive hole in the current market for a low end ECU which is easy to install and has coherent product backup.

As far as the business and technical side goes, I run my own Electronics business developing stuff far more complex than an ECU. I'm not looking to build a whole business out of it, just fill a hole in the market and get a bit of cash on the side. Costings - I can build a few for the cost of buying another ECU for me. Only leaves my time to develop it, which will be spare time.

_________________
If at first you don't suck seed, try drier grain.

I think the point DeeCee was making is that end-users shouldn't need a degree in electrical engineering. I am sure you would be well capable of designing and building an ECU that would work for you but could your average punter with a modicum of tuning knowledge drive it?

_________________
Don
GT4DC Chairman
1994 Toyota Celica GT-Four ST205WRC JDM 269bhp @ 0.9bar
1994 Toyota Celica GT-Four Special GT 590bhp @ 1.8bar
1989 Van Diemen RF88/89 Formula Ford 1600
2008 Nissan Patrol GU 3.0L ZD30DDTi 154bhp

My intention is to provide a package, including PC software, that will be along similar lines to the other ecu's I've dealt with, specifically link & PFC - hopefully including the best features of each.

Tuners/mappers should be able to use it with a minimum learning curve including DIY tuners. Installers will have full instructions, pinouts, tech. specs. etc. The end user will sit in the driving seat and turn the key, same as always.

There may be companies out there selling ready built megasquirts as a package, but when I went looking I found a jumble of information with very little clarity. Many varients produced by different people, the only centalised resource being a forum where you couldn't see the wood for the trees.
The only thing I did discover is that there isn't a version suitable to run our engines without a load of work fitting cam & crank sensors, and possibly coil on plug or wasted spark ignition.

Don't get me wrong, I'm fully aware of the size of the task. To pay an engineer or sub-contractor to develop it would be a non-starter IMO, that's why I was very surprised to see TGT going for it. I agree with DeeCee that they won't be in the marketplace for a while. Getting the system designed, built, software written and running an engine is the easy bit. testing and refinement is where the months will pass. I would estimate 2 years before I would consider it ready to market fully, but I'm not talking 2 years work, much of that time will be driving (which i do a lot of), collecting logs and doing the odd software/hardware tweak.

_________________
If at first you don't suck seed, try drier grain.

I would more than likely be interested as i see Motec as being superb.... but then it comes with the brand name price tag

This is the crowd I was trying to think of earlier Chris

http://www.extraefi.co.uk/

And a little food for thought
http://www.emulation.com/catalog/off-th ... ckets/qfp/
http://www.emulation.com/catalog/off-th ... erview.cfm
No idea what the implication on BOM would be but that was only a 10 second google atempt

Thanks Steve, looks like there is a 64 pin one that will do the job. Just off out now, will have a better look in the morning.

devices I'm looking at e.g.:

http://www.microchip.com/wwwproducts/De ... e=en552033


Bearing in mind the standard ecu ran a 64HC11 derivative - 8 bit running about 1MIPS, I still think high end pic18 at 16MIPS would eat it for breakfast.

e.g.

http://www.microchip.com/wwwproducts/De ... e=en547757



Regardless, the first step of the programmable ignition will be pic18. At least writing in C it's fairly easy to port the code to a new chip.

_________________
If at first you don't suck seed, try drier grain.

Hardware/software is not my thing TBH - Steve/Lewis and CDB will be much better to comment but from my brief search I believe this type of chip is used by a fair few ECU manufacturer's (e.g. similar specs to DTA / Link / MoTec)
http://www.futureelectronics.com/en/tec ... MAG40.aspx

It is an interesting idea but IMHO it needs to be plug & play unless you intend on offering a lot of technical support or full installation/mapping service. Even then I think you will have to factor in a couple of hours per customer for basic help in the pricing strategy.

I think the biggest hurdle will be the marketing and getting the product into distributors to be offered to customers if that is to be your intended market.

Otherwise you will end up being a halfway house between the Megasquirt (I am sure I have a DVD with that name!) and the DTA ecu which is used more by clubman. The high-end market IMHO is closed to new entry unless you have 20 years experience and a sales/technical support team at races! The geeks market is limited but to be honest probably more where your skills could be best used.

I particularly like the concept of old motorbike ignition systems and I do wonder if there is a market for a super-sleuth ecu conversion for classic and vintage cars using the old carb housings and looking stock. I am dealing with a client at the moment who has a collection of 1940s Bentleys etc and is having issues with emissions expelling into a shared car park so there may be some market here.

The kind of cost you are suggesting is realistic but I assume will not allow for a resellers margin? Take your £300 manufacturing cost including labour time to build and a little bit of profit for you, add shipping and resellers profit and a couple of hours support (£20 + £200 + £70) and then the VAT and you are at a selling price of £710 which is pretty close to Emerald/Link ECu money.

As a part techy the megasquirt (maybe it was a video and is in the loft...) is just too complicated to fathom what options & configuration and software i need.

A budget project specific plug and play item may have some room in the market place - read a new version of a Power FC that can be tweaked from standard map but retains the original safety parameters and just plugs in. It will limit the market to individual car models you have designed & tested but I can't see a place for another ECU manufacturer doing the same thing.

Is it worth a discussion with TGT to see if your product would be of interest to them if they have stalled?

_________________
JP
GT4DC Treasurer

Grey St185 RC
Silver BMW E46 M3
Multicolour yawn Honda VFR400
Silver Honda Civic Type-R

i thaught tgt had released their ecu this year but i havent checked their thread on the oc website. i was looking at megasquirt for when i had my vauxhall and i figured that the reason it was so much cheaper than other stand alone units was that it had to be built up ?

Pretty much spot on with all the points.

The chip you suggest is similar to the more powerful 32bit one I was looking at. Sledgehammer to crack a walnut, but will allow much more room for the future and allow less efficient code that will be easier to adapt to different applications.

As you say, my suggested cost of £300 doesn't include distributer markups. TBH going through distributers would require a selling price of £600-£700 which would render the project a non starter IMO. My plan is to sell direct via website and maybe via ebay.

We'll have a chat about TGT at the curry meet, Nigel may have some knowledge about how theirs is going. Certainly an option, but again I think going through a middle man will kill the project.

_________________
If at first you don't suck seed, try drier grain.

The other thought I've had is that I can build in the cross checking and warning / engine cut that I currently have separate from the ECU - water temp, oil pressure, EGT too high, mixture too weak on boost (if wideband fitted) etc. etc.
I think that would give it a sales advantage, but I bet the other ECU manufacturers would follow suit pretty quick.

_________________
If at first you don't suck seed, try drier grain.