Is MS still the answer?

[elbekko]

Still vapour at the moment, but have been mulling this over and doing a little bit of research.

I've been thinking about MS for my P38 for a long time. Integration itself should be relatively straightforward, we're doing that for @Escape's install.

On my car, I have the extra complication of having the electronic gearbox, that talks over CAN to the ECM. In the past I've thought about making a little module that interfaces with MS and talks CAN to the gearbox, but honestly, that sounds like a lot of R&D, and I find myself not wanting to do more software development in the evening when I'm already doing it all day long.

So on to the vapour part... why not something else? There are a couple of good aftermarket systems around, like Haltech, Holley, ...
From watching a bunch of car YouTube channels, the Haltech seems like a very good option. And they have support in Europe.

A wishlist of features I have, in order of importance:

• Petrol control with direct coil drive
• Gearbox control
• LPG control (low impedance injectors)
• E85 support would be nice (Scandinavia has much more E85 than LPG)
• Drive-by-wire with cruise control (I dislike the P38's vacuum system)

Looking at the Haltech offering, the Elite 2500 looks nice, if not a little bit overkill: https://www.haltech.com/product/ht-151302-elite-2500/
But it has everything I want/need:

• Both low and high impedance injector control.
• Works with any trigger wheel.
• Fully configurable gearbox control.
• DBW with cruise control support.
• Support for flex fuel sensor for E85.
• A bunch of programmable inputs and outputs.
• Easily extendable over CAN with a bunch of available modules.
• Easy tuning.
• Good support.

And by the time you get all of that in MS as add-ons (if you even can, nobody wants to risk cruise control), you're sailing well past the price of an Elite 2500. And you have to somehow make it all work together.

So... tell me why it's a bad idea

[ThreePointFive]

The only reason I've come up with when thinking through the same 'problem' is the support. These companies may fold and they invariably are using some proprietary code that cannot be unpicked by anyone else/competitors. I do wonder if problems develop, how will anyone be able to lend a hand diagnosing what's going on if no one understands the core functioning of some long-defunct company's product. At least with MS, it is very unlkely to die outright and it can all be easily understood even if it was no longer being developed. MS1 is 'dead' to all developmental intents and purposes, yet it's far from out-of-support.

Certainly MS is not a winner on cost grounds alone, which was a surprise when taking it on for the first time.

The other reason applicable for most of us is that we're using engines based on technology one-up from hitting one rock with another rock to create sparks, so the complexity of these systems and the efficiencies and performance they can extract is lost on our applications.

[elbekko]

11 minutes ago, ThreePointFive said:

The only reason I've come up with when thinking through the same 'problem' is the support. These companies may fold and they invariably are using some proprietary code that cannot be unpicked by anyone else/competitors. I do wonder if problems develop, how will anyone be able to lend a hand diagnosing what's going on if no one understands the core functioning of some long-defunct company's product. At least with MS, it is very unlkely to die outright and it can all be easily understood even if it was no longer being developed. MS1 is 'dead' to all developmental intents and purposes, yet it's far from out-of-support.

That's a fair point, but it's no worse than the stock system in that regard, and at least the software is out there and available. The software on the ECU is closed source, but is anyone realistically going to modify or dive into it anyway? I've tried looking through the MS1 code, it looks to be a very unstable house of cards with little room for changes/improvements, and wouldn't be all that easy to go debugging in.

15 minutes ago, ThreePointFive said:

The other reason applicable for most of us is that we're using engines based on technology one-up from hitting one rock with another rock to create sparks, so the complexity of these systems and the efficiencies and performance they can extract is lost on our applications.

That's true in a way, and I don't expect to get 100hp extra out of it. But more flexibility, and being able to upgrade to more modern things in some regards (like the throttle body), sounds pretty nice to me. Sure, no RV8 is DBW, but the Thor uses a run-of-the-mill Bosch throttle body, which should be trivial to swap out.

[ThreePointFive]

In all honesty I was trying to poke holes in the idea and really couldn't come up with much. When deciding how to do the 4.0 it made sense for me to go with something that was readily understood and supported here specifically, as I knew I was going to have many, many problems and people on this forum are helpful.

I'm not likely to be doing a project like this again, but can't see why I would stick to MS as opposed to any other systems. I've watched waaaay too much MotorTrend not to at least look at MSD or something similar.

[Bowie69]

If I wanted to do it quickly, I'd go with a Microsquirt, as I know it, and not too costly. 

If I wanted complete configurability, great support from a company that has been doing it since the 1980s (and showing no signs of slowing down), I would go with Haltech.

If I wanted a project, I would go with Speeduino.

If I wanted to get the most out of whatever system I fitted, I would go with what the local rolling road chappie says he would be happy to tune.

 

Megashift exists for MS, but gets a little pricey when I priced it up, IIRC you need a separate ECU To run things, so you in effect would end up running two Microsquirts, connected together with CAN(?), so that is £600+ before you start. The 8ft looms are nice though. Alternatively if Haltech can do the trans (assume by interfacing with stock TCU?) then you are laughing, honestly.

 

 

 

 

 

[elbekko]

3 hours ago, Bowie69 said:

Alternatively if Haltech can do the trans (assume by interfacing with stock TCU?) then you are laughing, honestly.

They can. Directly. Line pressure sensor support and everything from what I've seen.

I should send the guys an email at some point to go over my silly ideas, but from what I've seen in their (good) documentation, it should be fairly straightforward.

[Dave W]

MS is still a good option and people are doing a lot with the built in CAN comms these days which wouldn't take a lot of work to bridge to a gearbox controller, even if you had to covert MS CAN signals into different ones using an Arduino or similar. I've just built something similar for my van using an ESP32 to bridge CAN devices and broadcast the CAN data on Bluetooth so I can view it on my phone. ESP32 has a built in CAN controller, WiFi, BLE and costs around £3 each from China !

That said, MS is not what it used to be either in support or ethos. The original spirit of open source collaboration which resulted in hardware, firmware and software advances got lost somewhere along the way with commercial offerings taking precedence and costs increasing despite costs of manufacture and components reducing. The MS forums are dead so support is nowhere near as good as it used to be although we do still have good resources for it here.

Speeduino is probably the current equivalent of where MS used to be in terms of having an active community and open source development, I'm looking to replace the MS on my LS with a Speeduino at some point simply because it's cheaper than an MS upgrade (I'm still running MS1), faster and more flexible. I'll also be looking at controlling the gearbox with it. The Speeduino forums are still active, you have a choice of processors, the firmware is written in a modern(ish) language and there's even documentation to help developers customise the firmware !

I honestly think that, these days, Speeduino is the community built, open source engine and gearbox control system. MS is now very much a commercial offering with high costs which means it needs to be compared, both by feature and price with other commercial offerings and it pretty much falls down when compared to other commercial system. Having closed firmware isn't an issue as long as it works and you can get access to it for diagnostics and tuning/adjustment.
 

[FridgeFreezer]

As Dave says, MS has slipped from the original open ethos and has also (in MS3) suffered the curse of trying to be all things to all people.

For basic EFI I'd still choose MS1 or MS2 as it's a fairly simple, robust, and well proven system and can be done fairly cheaply.

If you want more bells & whistles / integration with other systems I guess the commercial offerings may sway it as MS3 is just an utter dog's breakfast despite ticking lots of boxes on paper it's ended up being the worst of all worlds.

I've not yet seen a Speeduino ECU that doesn't look like a Maplins project, not sure I'd want to fit it to a Land Rover and jiggle it round for 1000's of miles on & off road. Also it's suffering a bit of feature-creep and is already more expensive/complex than it needs to be.

[=jon=]

I've been following the MS vs. other ECUs a bit, not so much for the landy, but for my other pet project which is a turbo'd MX5 - currently it's running a derivative of MS3 (a licenced plug and play mainboard wired for the mazda ECU plugs with an MS3 processor card). It works, and there's good support for getting it properly tuned, but some things are not great compared to the newer competition from people like Haltech. Even more so when you look at the non MS3Pro units which still use crappy DB37 connectors rather than proper automotive waterproof connectors, but are close to the same cost as the lower end alternatives. 

On the US MX5 forums where megasquirt is/was the standard, a lot of people are looking to move away as the MS development seems to have stagnated, or 'common' features like DBW and long term trim tables have been in beta forever, or firmware niggles are never resolved (or massively broken between releases), and the source code isn't open to let the community fix them any more. 

The common alternatives are Haltech - although from what I understand their CAN implementation is closed to third party integration which may limit options, Link, who have been around forever or MaxxECU who seem to be fairly new but can also do transmission control, including the BMW DSG gearboxes which are gaining popularity for swaps. There's also people like Syvecs, Motec who are more motorsport biased.. 

[Hybrid_From_Hell]

We tinkered with MS3 and decided quickly to drop it and aviod like the plague.

if you open the ECU it has PCB on PCB on PCB and awful posts to keep apart connectors and as a grown up said "Its bloody awful design"

 

MS2 should be more than enough although with the CAN system lots of additional work.

Check the USA forums on MS there maybe someone who has been there sorted it done it and can help - THATS the key difference with MS over any other systems is that its a community and peeps can help !

Nige

[Ed Poore]

I've heard decent thinks about the Link ECUs as well and they seem reasonably priced for a "professional" system like the Holly and Haltech ECUs. There are a fair few distributors in the UK so I'd imagine there are a few on the continent as well.

[FridgeFreezer]

I think if you want a ton of CANbus controls and bells and whistles you may as well go for a "professional" one as at least in theory you can phone them up and shout at them when it doesn't work.

MS1/2 is great for what it was intended to be - a basic DIYable EFI system. For sparks & fuel it's absolutely great, the more junk you try and bolt on after that the messier it gets and you're into diminishing returns compared to just running whatever the factory ECU was.

For ripping crusty old carbs + dizzy or Lucas EFI off and replacing with something better it's just what you need.

[MECCANO]

On 8/5/2021 at 11:01 PM, FridgeFreezer said:

've not yet seen a Speeduino ECU that doesn't look like a Maplins project, not sure I'd want to fit it to a Land Rover and jiggle it round for 1000's of miles on & off road. Also it's suffering a bit of feature-creep and is already more expensive/complex than it needs to be.

I'll let you know, ill be using it to replace the Mems on the t16.  As for expense. i think speeduino has a long way to go before it gets to the price of MS. the kit for the board was only about £80.  Admittedly it was some time ago.

[ThreePointFive]

On 8/3/2021 at 1:01 PM, Bowie69 said:

If I wanted to get the most out of whatever system I fitted, I would go with what the local rolling road chappie says he would be happy to tune.

Reviving this as weirdly it came up today. I have been looking at how to get my car to be the nicest to drive that it can be and a rolling road in Somerset was highly recommended (PVE tuning). As the website said they do work with MS, I called them but it turns out they dont. They were not kind about MS and stated there are always problems with it.

It's an interesting one as the best environment for tuning has to be a rolling road with the proper sensors and tools but it's off-limits because I'm on MS.

I've made some assumptions about the cost and complexity of a new kit - whatever that might be - I can't see a change being justifiable but that's resigning myself to a 'good enough' tune for the life of the vehicle. I'm also not sure how compatible most kits would be with such rudimentary engines. Further research required.

[Bowie69]

11 minutes ago, ThreePointFive said:

They were not kind about MS and stated there are always problems with it.

This is often the perception, it is however muddied by the fact that anyone and everyone can 'have a go' and end up with a substandard installation that causes massive issues -the hardware is good, so is the software.

There's a lot more flexibility in the software of other aftermarket ECUs, but you are looking at twice what you paid for MS, and you still have to do the wiring yourself or pay someone else to do it -(tip: auto elec-chickens are NOT cheap).

Really don't need a rolling road to get it nice to drive, but it will take time and driving around and lots of tweaking. At the end of it you know how to tweak the engine to run differently, and can fix stuff yourself.

If you rolling road it and it isn't right still then you have to go back, and back, and back again.... which is again really not cheap.

Rolling road is great if you re chasing the last HP and maximum timing etc, but not actually that good for driveability -for that you can set up most of the part throttle stuff, but AE and WUE still have to happen on the road and through many cold starts to get it right.

It is most definitely not the silver bullet, though of course has its place.

[smallfry]

This is true. Years ago when I was into Minis and the A series engine, I used a long gone company here in Kent for rolling road tuning. The old (as he seemed to me then) guy said that although you can tune an engine on a rolling road to get the figures, you cant beat doing it on the road, as every vehicle is different.

[ThreePointFive]

Fair comments, I was taken aback by the response towards the system. Also some annoyance that it was an outright no rather than letting me decide based on the likelihood of a sub-optimal result. I did like the idea of paying someone rather than constantly relying on kindness and favours, too.

MS is still the best system for me thanks to the support around it, but again, that relies on this forum and the patience of people.

 

[Bowie69]

To be fair, after the 20th person to turn up with an MS installation held together with insulting tape and scotchlocks I'd probably be a bit wary too.

[landroversforever]

15 hours ago, ThreePointFive said:

Fair comments, I was taken aback by the response towards the system. Also some annoyance that it was an outright no rather than letting me decide based on the likelihood of a sub-optimal result. I did like the idea of paying someone rather than constantly relying on kindness and favours, too.

MS is still the best system for me thanks to the support around it, but again, that relies on this forum and the patience of people.

 

I'd also add that I helped a mate convert his Zetec powered Robin hood to Microsquirt from one of the aftermarket options. I'm not 100% sure on which one so I won't mention, but it was one of the common ones. It was setup in a previous build professionally which was then transferred to the kticar and retuned. Again professionally and again on a rolling road. 

Starting with one of the base tunes, getting it running and only ~20-30mins into the test drive running the autotuning function he commented that it was driving 10x nicer than it ever did. 

[elbekko]

Reviving this idea...

I've been in contact with Haltech support to talk about this. According to them, none of my requirements should be an issue, BUT

Apparently tuning the gearbox control isn't easy (guy said 2-3x harder than fuel map tuning), and really should be done by an experience tuner on a rolling road.

Is it a terrible idea to say "how hard can it be?" and do it anyway?

The other option is to make something that speaks CAN, and talks to the stock gearbox ECU, which sounds like a pain in the butt (but on the other hand, should be fairly straightforward once you know what needs to be said).

[=jon=]

Have they said it will drive low impedance LPG injectors in addition to high impedance petrol injectors, with switchable maps between them?

This is one of the options I've been looking at - I've got MS3 + an LPG ECU, it all works but it's not 'great' - there's various things that could be done better if everything was driven from one unit... 

[simonr]

Someone at work was talking about implementing Speeduino.  He had previously run MS2 and some commercial implementation (don't know which nor in what order he tried them).  Se said Speeduino was like a breath of fresh air.  I've not looked at it at all - but his enthusiasm for it was infectious.

[elbekko]

22 minutes ago, =jon= said:

Have they said it will drive low impedance LPG injectors in addition to high impedance petrol injectors, with switchable maps between them?

This is one of the options I've been looking at - I've got MS3 + an LPG ECU, it all works but it's not 'great' - there's various things that could be done better if everything was driven from one unit... 

I explicitly asked about it in my e-mail, and they didn't say no to it anyway And according to their specs, they can do both out of the box.

[=jon=]

Thanks - I read through their specs and it says they can run high/low/peak hold, but I never got as far as asking if you could do a mix - run half the injectors with one and the other half with the other, or swap based on an external input if you are switching between the two via relay or similar... I'd be interested to see how you get on

[FridgeFreezer]

I looked at Speeduino and it sounds great but the hardware seemed quite "maplins kit"...

It would be possible to re-design it as a single sensible PCB that fits in a nice sensible case of course.