Board logo

Megasquirt
02GF74 - 8/1/19 at 06:03 PM

I've just bought a raspberry pi 3, cost £34, and it got me wondering about megasquirt.

A quick search show unassembled megasquirt costs £350, HOW MUCH!!!!?

Seems an awful lot as I'm sure the additional components required for engine management would not cost more than £50 in addition to the pi.

Of course the downside is getting the software. And looking at arduino, which is about £ 5?, then surely megasquirt is hugely over priced and its days are numbered.

[Edited on 8/1/19 by 02GF74]


coyoteboy - 8/1/19 at 06:51 PM

The BOM cost of a megasquirt 3 is about 200 quid I think, ms1 is about 90. Don't underestimate the cost of specific parts like MAP sensors. However the BOM cost isn't really what you're paying for, you're paying for a decade of software support, free for life ongoing support, experience and skills.

Arduino isn't up to the job, there are ECUs based on Arduino already, but they are really just hobby boards. Rpi is cheap, but also not designed for the task and sold in the millions not hundreds.

There's dozens of MS clones out there selling cheaper, but they don't get the market because the support isn't there.


Angel Acevedo - 8/1/19 at 07:00 PM

quote:
Originally posted by coyoteboy
The BOM cost of a megasquirt 3 is about 200 quid I think, ms1 is about 90. Don't underestimate the cost of specific parts like MAP sensors. However the BOM cost isn't really what you're paying for, you're paying for a decade of software support, free for life ongoing support, experience and skills.

Arduino isn't up to the job, there are ECUs based on Arduino already, but they are really just hobby boards. Rpi is cheap, but also not designed for the task and sold in the millions not hundreds.

There's dozens of MS clones out there selling cheaper, but they don't get the market because the support isn't there.


I found my MS invoice a few weeks ago and almost fainted at what I paid for it.
Next build if ever I´d go for Stock ECU...
Although some Stock ECU´s are not mod friendly....

But regarding support, I can´t complain.


coyoteboy - 8/1/19 at 07:11 PM

I'll be honest, I've considered designing my own a stack of times but the software side and having Access to knowledge of all the sensor systems (mainly timing pickup) is hard to come by.


rf900rush - 8/1/19 at 08:04 PM

I built a MS1 quite a few years ago for just over £100.

There was good support back then including MSextra.

£100 for a basic EFI system that did every thing I needed was a reasonable price.
£350 is what one should be paying for a high end one, not a DIY system.

The only thing to watch out for with the alternative micro's is a lot are not really for automotive use.

Motorola now NXP Freescale do make micros automotive use.


scudderfish - 8/1/19 at 08:58 PM

Through MSDroid work I've come across (and started supporting) Speeduino https://speeduino.com/wiki/index.php/Speeduino which seems to be a spiritual successor to MS1/2Extra. Quite a lot of people are successfully running their cars on it. If my MS1 ever goes pop it'll be replaced by one of these.


snapper - 8/1/19 at 09:42 PM

You need an ECU you can tune
MS has a couple of autotune programs but they don’t do the whole job.
Your now into either a huge amount of learning to understand how to get the best out of it.
Another problem is not many rolling roads I know of will touch MS


coyoteboy - 9/1/19 at 12:31 AM

quote:
Originally posted by snapper
You need an ECU you can tune
MS has a couple of autotune programs but they don’t do the whole job.
Your now into either a huge amount of learning to understand how to get the best out of it.
Another problem is not many rolling roads I know of will touch MS


MS is pretty easy to tune manually, if you know engines fairly well. Plus there's a few MS tuners in the UK now and after all that work, most people will be happy to drop a base tune on and drive a hundred miles to a specialist. I'm not sure the RR support argument stands anymore.


MikeR - 9/1/19 at 09:22 AM

I bought the parts to assemble a speeduino for under a hundred quid.

I soldered it up, tested it and the magic (lucas derrived) smoke broke free.....

As lucas magic smoke is as rare as rocking horse poo these days & no one in China is selling a generic rip off I bought a fully assembled speeduino for under 200 quid with modifications

However there is now a movement to push a new board (seawolf) which fully assembled is around 250. Its also got an easier to use wiring connector. I didn't pick it as it was brand new and I was worried about being the first to find problems. Hindsight I should have waited 6 / 12 months and then bought seawolf.


coyoteboy - 9/1/19 at 01:11 PM

Were you just looking for a challenge MikeR? Seems like at 250 quid you're 3/4 of the way to an MS1, which is well supported and works a treat. With regards the earlier questions of tuning availability - does *anyone* in the tuning world know speeduinos?


MikeRJ - 9/1/19 at 08:02 PM

quote:
Originally posted by rf900rush
I built a MS1 quite a few years ago for just over £100.

There was good support back then including MSextra.

£100 for a basic EFI system that did every thing I needed was a reasonable price.
£350 is what one should be paying for a high end one, not a DIY system.



You won't get much for £350. A entry level Omex 600 or Emerald K6 is around £700 though obviously a significant chunk of that will be paying for the software development rather than hardware.

In terms of hardware the microcontroller in an ECU is one of the cheaper parts. The cost of the power electronics to drive coils and injectors will be considerably more, as will the case and connector and loom.

A basic ATmega 328 based Arduino is not up to the job of driving a four cylinder engine, but you can use the Speeduino firmware on a Arduino Mega 2560, which gives you something of similar functionality to the VEMS system, though somewhat more 'experimental'. If you just want to get your engine up and running quickly rather than spending most of your time playing with ECUs it's probably not the way to go.

A Raspberry Pi is totally inappropriate for this application. It uses a system on chip designed for running relatively heavy weight operating systems and graphically intensive user interfaces rather than running time critical processes.