compare the cost of any other ecosystem to the specs of the R5 and you may change your mind.

 

here you go: https://pcmhacking.net/forums/viewtopic.php?f=42&t=8172
i just started using v4 last week so im still getting the details ironed out, but it seems to work well for what it is. its closed source but it wouldnt be hard at all to RE what he has done to start your own branch.

once again, there is no can bus on any car using a p01 or p59 ecu. they use j1850 vpw which is completely different.

 

Sure, but the IO is more or less a solved problem, right? You just integrate the necessary I/O components with whatever microcontroller/etc is at the core of your system. All of these companies already have that stuff working.

Okay, that's new information for me. I'd be curious to understand why/how it uses multiple processors, and why each has separate firmware. But also, flashing new firmware to a CPU or reprogramming an FPGA also fall under the category of solved problems. Furthermore, if that's a significant hurdle, they should just integrate the update process into the system itself, such that the software delivers a package of new firmware+calibrations to the device, and then the device itself takes care of distributing the pieces to its respective internal components. At least, that's how I'd do it if I were trying to build a modern, robust solution.

That sort of thinking is why everything sucks. Seriously -- continuous improvement is the only way to build great stuff. Maximizing ROI at the expense of your product leads to things like when you said "Holley is like drawing with crayons."

If any product decision makers at aftermarket ECU companies are reading this thread, please hear me: you can do better. We, your customers, want you to do better. The aftermarket ECU market has a LOT of room for improvement. Enough that I'm confident that Haltech's head start is not insurmountable. You just need to invest. Put a truly great, modern product out there, put reasonable effort into marketing it, and let it revolutionize the industry. (And, if you're reading this from Haltech -- please keep up the great work, but even you have room for improvement. For example, now that the ECU has wifi, I ought to be able to update my tune from an iPad, or peruse my datalog from an Android phone, or monitor live data from a Mac. There's no reason your software should still be Windows-only.)

 

you dont seem to be understanding this topic, or at least what i mean to be conveying. of course updating firmware is easy and solved. the point was that the nexus uses multiple processors and a FPGA for its function. its an extremely advanced piece of hardware that may not have been possible 5-7 years ago. thinking that hardware design is easy is no different than me thinking you should build me a ecu UI cheap and quick cause 'buttons and input boxes are a solved problem' its far more complicated than that and i think you dont quite understand what you dont understand yet.
its nice to see the interest beyond just throwing a holley at it though.

 

I was trying to stay away from this thread but watching somebody postulate about aftermarket ECMs like they know anything about them just because they are a software engineer is pretty funny.

Do you seriously think that a company is going to release new hardware on a yearly basis? 14-15 of them have been doing nothing for a decade? Holy crap, stick to writing software....

A swear word in GM speak? It's there for a reason and it's hardly a bad thing.

I love posts like this especially coming from somebody who hasn't used any of it, hasn't done anything with it, and is asking such basic questions like what is the best. Anyways, good luck on your quest you are going to need it.

 

It's possible there's something beyond my understanding here, but I don't think so. My computer science curriculum in college included some basics in both electrical and computer engineering, because they serve as foundational principles for the ecosystem in which I work. I am no expert, but I am a good bit beyond a layperson in this department.

I'm not saying hardware design is trivial; I'm saying it's a largely known quantity. We (engineers, collectively) know how to work with processors. We know how to coordinate among multiple processors. We know how to use FPGAs. These are not new things; multiple discrete processors of different types coordinating different tasks has been going on since the 80s if not before. As I said before, for an aftermarket ECU, a team of engineers ought to be able to go from concept to released product in a 2-3 years, give or take. (If you are in some way an authority on this subject, which it seems like you might be, I'd love to learn more about you and what you do, and what you feel I'm missing here.)

I highly doubt that. There may have been practical limitations in terms of cost, but the technical capabilities here are not remotely novel from an engineering standpoint. Given the prices on some of these ECUs, I don't think they're letting cost slow them down either.

To be clear, I don't think hardware design is easy, nor cheap. I think that, just like software, there is a process you go through to develop it. For a lot of the work, it's rote stuff that we all know how to do, and you just have to trudge through and get it done. It takes time, but we know what to do and how to do it. The details of what the rote stuff is differs between software and hardware: for hardware, it's a matter of working your way through each circuit and determining the right values and locations for resistors and capacitors and whatnot, taking the time to understand how your microcontroller is going to behave for a given set of inputs, managing power distribution, etc. For software, it's similar (and I know the details better): we connect input streams and output streams, manipulate collections, perform translations and conversions, establish connections, perform handshakes -- occasionally developing a novel protocol or data format, or reverse-engineering something existing.

In software, once in a while we need a novel algorithm, but that's super rare outside of bleeding edge research. It hardly comes up in my work, and I expect that's true for the electrical and/or computer engineers at Haltech who developed the Nexus line of ECUs and the frequency with which they need to do something truly novel, too.

The hardest thing in software work is the final stage -- you've built everything you think you need to build, and it mostly works, but there are bugs. Some bugs are easy to diagnose and resolve. Those get out of the way quickly. Others are mysterious, intermittent, and difficult to reproduce. Fixing those -- that's the hardest part of my job, and I'll bet most hardware engineers would agree that it's the hardest part of their job, too. Having been in this line of work for the better part of three decades, I have mostly learned where the pitfalls are and developed a good intuition for debugging, and it's relatively rare that I can't figure something out.

BTW, I would love to play a part in building a new UI for engine tuning! (I also want to dig in on the hardware side enough to learn how to build a custom OLED gauge cluster, which I'll then customize with software.)

 

MS3 Pro Ultimate at less than half the price is pretty darn compelling. R5 is clearly better, but 2x? I don’t know…

Okay, whoa. Going to have to learn more about that world! If I’m understanding correctly, this OS can be flashed onto a factory PCM? Does it work on the P01?

Yep, understood. CANbus compatibility is one of the things enticing me toward the aftermarket.

Who said anything about yearly? I’d hope for improvements to the tuning software at least that often, but actual new PCMs with nontrivial improvements should cone every 2-3 years IMO.


I don’t like automatic transmissions, and I definitely lack some knowledge here, but when TM first became a thing, practically all anybody wanted to know was how to disable it.

Is the R5’s torque management more or less the same thing, or is it something else?

I have a FAST XFI 2.0 on another car (‘87 Grand National). My primary impression of it is that it was built by people who didn’t really care or sweat the details. It works well enough, but nothing about it was intuitive or easy. Every second I spent with C-COM was with my internal software-usability conscience griping about how this or that could have been so much better if they would just CARE.

Looking at specs, screenshots, reviews, and YouTube videos for most of the competition, it’s mostly the same. That, combined with the fact that XFI 2.0 cane out in 2013, is the basis for my guess that many of these companies are either resting on their laurels or phoning it in. (Haltech, Link, and MegaSquirt seem to be the exceptions. Maybe AEM too?)

Also, I don’t believe I asked “what is best?” I asked what people are using and what they think of it.

 

Here's another one,
https://www.maxxecu.com/

 

very briefly, the r5 has what amounts to a built in pdm, dual wideband controls and more io. I have one on my wagon and there is literally 2 physical fuses and on relay in the entire car and those are only there for the power windows and locks so they are ‘always on’ rather than ignition. The ms3 is super cool but it’s a bicycle compared to a new Camaro.

An earlier version works on the p01 but you’re better off converting to p59 to use the newer version. I just did that on my 02 pickup and it wasn’t much more than unplugging and plugging in the new one.


What I mean here is the rest of the car speaks vpw so having can bus on your new ecu isn’t useful for interfacing with the stock dash or bcm. I thought you mentioned wanting to keep that stuff but may have misread.

if companies updated the hardware that often they would have to charge waaaaayyyyyy more for them then EOL last years models and not support them. It would be a quickly failing business model in this industry.



sounds like you’re thinking of the gm function that pulls timing on shifts. It seems knee jerk to disable that and think it’s stupid, I’ll admit I used to, but when properly implemented it’s very useful.
haltech uses a torque model and torque managament to control wheel slip/traction


fast ecus are old as ***** so the processor, memory, etc is outdated and very limited to todays standards. It also came from the same guy that kinda started a lot of what we, American v8 owners, know of early standalone efi. Dfi gen7, big stuff, fast, and I think some others all ‘look’ the same for a reason.
earlier when I said ‘not possible 5-7 years ago’ I was referring to the performance/cost/size of what’s possible now. Back when the xfi was designed if you wanted to match todays nexus the box would be the size of a suitcase and cost 50k.

Aem infinity is a rebranded ecu. They just do the software not the hardware.

 

Wow, this sure is an interesting thread. I had NO IDEA there were so many aftermarket ECU alternatives.
I guess I will appear "unimaginative", but I'm happy with my Holley Dominator setup which I bought about 6 years ago.
(I'm still trying to get Traction Control working; it seems my amplifier design for the ABS wheel speed sensor is sub-optimal. <grin>)

Overall the switch to the Dominator was hardly worth it, especially considering the cost and effort to rewire the entire engine compartment.
If I were to do it again now, I would choose something like the "MS3 Pro Ultimate", assuming the software was open-source like the underlying Mega-Squirts. As someone who started working with electronics hardware in the late '60s and then software in the '70s (my career) I like the idea of open-source software and hardware.

To add my 2-cents to the discussion about hardware/software complexity and development, the hardware and software development of an ECU might be reasonably straightforward, but I think the extra work to make it fully programmable for any engine/car and especially the User Interface (UI) and data logging for the tuner/user might be the most time-consuming part of such projects. HP Tuners did this work for factory ECU with years of effort; I think Holley did a reasonable job too.

I'll watch this thread and maybe "upgrade" from my Holley Dominator in the near future.

 

For anyone who wants to know what this thread is really about, the OP has had Kingtalon advising him the usual mix of actual theory and utter nonsense over on the HPT forums.

 

My 2 beer opinion..

My colorcrayon Holley terminator X max started on first crank, it self learns minor adjustments to AFR has good Spark control and , it runs my GM transmission.
I wanted a simple system that I would not be hamstrung to a tuner that locks their tunes or disappears because they got in trouble with the fed for changing the stock ECU's.

I've worked with many aftermarkets and the user interface to program and set up on many is brutal.
To be good with haltech/motec you better be a engine builder AND a tuning specialist.. IT experience is worthless.

Fast and AEM may work in some applications, I have not found one yet...

Most of them,, Can bus compatibility? ROFL... Life is too short. I went Holley glass dash and called it a day.

My needs are,,, simple,,, run my engine decently. Dont cost 4000 dollars to install.
I'm too old to build things from scratch, love the Megasquirt Concept,, , just don't feel like climbing the learning curve.

 

And here my day started out so well........

 

It does not learn spark.

 

Yeah, it's clear he's a bit... off. Don't worry, I'm taking everything he says with a large amount of salt.

 

So he's haunting the HPT forums.. let's see how long that lasts....

 

That does sound really cool. Still can't decide if it's worth an extra two grand though. Maybe!

Cool. Just looked, P59s aren't too expensive either. I may very well go this route! Thanks for the lead.

Yeah, this whole plan, if it comes to fruition, involves a custom dash too. If I went with the Haltech with its PDM, what PCM<->BCM communications would I need to worry about? I know close to nothing about the BCM side of things.

Again, I'm not talking about yearly. New hardware maybe every three years tops.

If they're smart about their designs (modular, architecturally similar where possible), I think they could continue supporting previous models without too much trouble. They could also offer upgrade kits, where you replace the guts of your older model with a newer one. This would require some standardization in the connector department, but I see that as a good thing.

Certainly going from a business model where updates are 10+ years apart to one where they're three years apart would require some changes to the price structure. I don't see a problem here. Introduce new stuff at a much higher price point, and keep selling the older model at a lower price. Kinda like high-end TVs these days. You want a run-of-the-mill LCD, you can get one for <$1000. You want a top-of-the-line OLED with the best possible picture quality? The price is several times higher, but if you wait a few years, you can get that exact same thing for a lot less.

That's the one.

I guess I'll take your word for it.

That does sound potentially more useful. Digging around a bit, it seems that this feature is present on the R3 and the Elite 2500T as well. The R3's price is much more palatable. Compared to the R3, the R5 offers support for dual widebands (nice, but not worth thousands of dollars) and a bunch more I/O (I'd have to count channels I can imagine using to determine whether I care -- but if I did, Haltech sells I/O expander boxes for $300-500 which would almost certainly do what I needed. There'd be a small weight and wiring penalty to go with an R3 + expander vs an R5.

Back when the XFI was designed was 12-15 years ago, not 5-7. The R5 came out three years ago, so I think 5-7 years ago it was entirely within the realm of possibility and probably would have cost just a bit more.

Good to know. Whose hardware are they using?

 

Did my own research and discovered the EngineLab EL129. Now I know!

Their software is apparently built on top of EngineLab's software as well. Nearly identical, but usually a version behind. They hide some things and add some things.

 

the R3 doesnt have any 8a HCOs, only four 25a ones. it would probably be great for an existing car like a camaro. the R5 is designed more for a clean sheet build like a new race car. the IO boxes are neat but not the same as most of the built in IO.
all the software features are the same between both nexus and 2500T

 

Okay, so the R5 costs $1900 more than the R3. If I needed more, I could add 10 8a HCOs and then some with the $1100 PDU. Add the IO box too and now the dual wideband seems like it might be the only remaining difference -- still saving several hundred dollars (although at that point the simplicity and all-in-one nature of the R5 starts to stand out).

This is like some weird kind of bench racing.