winter39

Hmm...anyone consider the Rtek 2.0 in the lower end of this catagory? Before everyone trashes me I know it does not do everything that the full stand alones do but it has a nice plug and play package for those that do want to run in the area of 275-300hp. For guys wanting the big 400+ hp a full standalone is the way to go.
The Rtek is created from your original ECU. You can go to their site and read up on it (no rewiring required at all buy you). I know I would consider it for a hybrid turbo application for sure. It also has some nice features like it uses a palm PDA to interface and make changes, it displays all error codes for the RX7, etc. One drawback would be it is only for the S4TII's right now with other models becoming available soon.
Here is the link if you are interested: http://www.pocketlogger.com/?pid=rtek7v2feat

coldfire

iTrader: (1)

if plug-n-play is really that important, they have made a pNp MegaSquirt (only S4 right now).
https://www.rx7club.com/group-buy-center-69/zeal-engineering-plugnplay-megasquirt-group-buy-533404/#post5574295

$525 is a decent price for a unit that you can just plug in (you don't have to build), and will allow you to do quite a bit.

i will be making my own pNp adapter for my S5, since i don't want to rip out the stock harness just yet.
keep in mind though, when done right, a new harness is the way to go, because you never know what could be worn out/defective on the stock harness.

Tofuball

MAX HP does not decide what ECU to use.

Dragoon

I also would recommend microtech. Just installed a LT8s in a race FB with a NA S5 motor with porting. Started imediatly, runs great, very flexable. Although Mictrotech could have a better fine tuning matrix, but for the money its great. If its gonna be a daily driver id recommend installing a GM Throttle position sensor, as any FC's TPS's are more of a switch then a accurate sensor.

All you need to make it run is external FC coils (plus all sensors on the motor of coarse)

PowerFC's are pretty nice as well, but do do any real tuning you need to have or know a friend with a data logit just to access the main tuning options. I have seen 2 now installed in FC's with a simple harness adapter and software changes

nik

features I find key in ECU are:
1. closed loop O2 control
2. datalogging to disk/laptop
3. some way to feed air fuel readout into the ecu's datalogs(or feed ecu parameters into a datalogger)

I would not consider any ECU that doesn't support all of these.

As for megasquirt, the microcontroller on the megasquirt-1 is ancient technology and I believe this is the only one at this point that supports rotary. The megasquirt-II is a much better chip but I don't believe there is full leading/trail ignition support yet. I would stay away from these until they mature further and become a great EMS.

I have personally used e6x and powerFC and they are both good, but powerFC depends on stock wiring harness and most of ours are in **** condition...

coldfire

iTrader: (1)

Calling the megasquirt-1 "ancient" i think is a bit harsh. it is no more ancient than some of the older aftermarket ECUs, and consider that both the MS-I and Haltech E6K run 8-bit processors. both ECUs have been proven to work on many applications.

the main advantage of most of the aftermarket ECUs over the megasquirt is RESOLUTION. some may find the 12x12 maps restrictive, but for others this might be fine.
performance/feature wise it is near the bottom of anything mentioned here (far from the "best").
it really depends on application and preference (DIY aspect with MS). so $300 for a MS-I might be adequate for one person, whereas $1000+ for an E6X or similar might be adequate for another...

B6T

iTrader: (2)

I think if I were to go with a "standalone" in my GSL-SE, I'd get that plug and play Megasquirt for the S4 FC, and an S4 engine wiring harness. Then it's just a matter of untaping the harness and removing unnecessary wires (for the air meter, etc), which wires to add (general purpose outputs) and then figuring out which wires need a connection to the body harness (12v constant, 12v switched, fuel pump, etc)... I think this would be the most efficient solution.

winter39

I dont agree, the ECU system will dictate what can safely be run on it. For the RTEK 2.0 (Which is what I was refering to) it uses the stock boost sensor. Unless you want to run some insanely large turbo and have it run only to 12-14 psi then Rtek would be fine and more then 300 HP would be possible. Also, you are limited by the offset of the fuel correction map (+/- 15% over stock settings) with the RTEK. This means that you can't command, lets say 200% more fuel then the stock map at any one operating position. This would dictate the max sized safe injectors and the max power you could get from your car. In the case of the RTEK the ECU would dicate "safe operation HP". I beleive this is the same for other systems. If the MS is limited by its resolution then you would not be wise to have an engine setup with high HP that could not be properly supported by the ECU. I am sure the full standalones have a threshold that is "much" higher before you would consider it the limiting factor.

Aaron Cake

iTrader: (1)

Microtech doesn't have closed loop (at least the LT-8 doesn't) but does have a "Cruise" map which it switches to when it detects steady throttle. So you get the same effect basically.

Haltech, Wolf and Megasquirt all do closed loop.

If you can do the DIY work, I think you would be hard pressed to beat the Megasquirt in terms of functions. What other $140 ECU offers boost control, nitrous control, traction control, dual fuel maps, staged injectors, rotary support, closed loop, etc. etc.

nik

Aaron, Haltech e6k doesn't use 8 bit microcontrollers, i don't know where you got your info but it's wrong. 8 bit microcontrollers with no hardware timer = **** precision for event timing. For a turbo car 12x12 map is not sufficient and will affect driveability/fuel economy.

The only combination that I think is acceptable is to take a megasquirt -II, which has a far better microcontroller and adapt the code for rotary. Although I am pretty sure you could do no trailing without much recoding.

Black91n/a

One thing with the Megasquirt is that it may "only" have a 12x12 map, but it interpolates between the setpoints, making for a smooth transitions between the points on the map. AFIK not all ECU's do this, so that's an advantage it has there.

RacerJason

iTrader: (5)

For Professional Motorsports 1. Pectel 2. Motec (i.e. M800) 3. Magnetti Marelli (uncommon)

For Motorsports/Lappers 1. Motec 2. Haltech 3. Microtech/Wolf 3D/AEM

For the budget minded enthusiast 1. Megasquirt

As Aaron mentioned always research your area for stand alone resources, tuning familiarity, and credibility/reputation.

In World Challenge the teams opt to use one of a few possible custom loom makers. We're talking $8000 USD installs not including Motec ADL. If you come to Mosport labour day weekend have a look under as many hoods as you can. The looms are works of art using Raychem and other aerospace components.

nik

I have not seen an ECU that didn't do this interpolation within the map. The problem is that you start to split load bands up too much with boost. For a 3 bar map setup, you have 4 vac bands, and 8 boost bands. This is really not enough mapping in the vacuum areas. For a 1 bar n/a only setup it's not as bad. Same applies to RPM, but 12 is enough for most people.

One must remember that megasquirt-1's hardware was designed as an affordable fuel-only computer. However, its nice that the thing is so damn cheap, and the upgrade path to MS-II is there, so there's no reason not to buy it. At some point my car will have a megasquirt in it, just not yet.

Tofuball

First off, the MegaSquirt IS a standalone ECU. You can take that stock ECU and throw it away, along with (among other things) it's AFM and MAP sensor.

The MegaSquirt is not limited by res, and even people with ancient low res ECUs can run very high HP numbers safely. The reason for this is that WOT tuning at top RPM (what gets you those peak HP numbers) is the LEAST effected by the ECU's res.

You could run a carburator and dizzy for + or - 5 HP on a 700HP car. Same hp for a Haltec or a MegaSquirt or maybe even an SDS.

Motec, and the MegaSquirt 2, are different stories. You can gain some very small amount of HP on a high HP setup (especially if it is very high revving, we're talking 10K to 12KRPM) by the fact that it holds very accurate to a decimal of timing even at very high HP.

Most of them dont.

What are you talking about? You can MOVE the bins on the MegaSquirt, the idea is to move the bins so that they control all the non liniar parts of the map. If your map has more then 12x12 such points, go ahead and get a different ECU.

Rotary engines (even boosted ones) are usually VERY linear, except at low RPM around idle.

So, by moving the bins, if you dont like having 8 'boost bands' then just change the number to the left of the table (kPa) and bang, you have 7 'boost bands' and 5 'vac bands'

most other ECUs do not offer movable bins.

muythaibxr

The megasquirt 1 chip has 2 hardware timers with 2 output compare channels each...

I've personally never seen more than about 50-100 microseconds, which is about .5 - 1 degree at 8000 rpms... and it always jitters on the retarded side, so it's no less safe. Dwell moves around a bit more, but as long as you have it set to around 3 ms it's fine.

Many people forget that the megasquirt also interpolates between the bins on the table, and allows you to move all the rpm and MAP bins... I'm pretty sure haltech doesn't do either of those, which is why it needs such a large table. A 12x12 interpolated table with movable bins is every bit as good as what haltech offers.

I have just finished taking the ms1 hires code and making staging work with it.... that means 100x the old MS code's output resolution (down to the microsecond) for rx7 owners....

I have actually started writing the wasted-spark wheel decoder code for ms2/extra. James and I are basically going to add as many of the msns-extra features as we can to the ms2. I'm actually finished the wheel decoder, and wasted spark/COP ignition for it, and have started working on rotary outputs. Next on the list will be staging.

The main disadvantage of the ms2 chip is going to be that it has far fewer general purpose outputs available to the user... so making it run a rotary will be easy enough, but then you'll be out of outputs for other things.

The ms2's advantages are going to be things like:
1) almost no ignition jitter
2) calculates fuel at 1 rpm intervals instead of 100 rpm
3) x-tau accel enrichment algorithm will make emissions easier
4) PID predictor closed loop. You tell it how long a change in throttle takes to affect AFR, and how long it takes the sensor to respond, and it runs closed loop and makes a *FLAT* AFR... it just predicts the amount of fuel needed and gets very close without overshooting or oscillating when tuned right.

Almost none of these things affect maximum power, but instead they affect drivability, and even without these things, the ms1/extra code with the hires additions does a VERY good job if tuned right of getting good drivability

I know for sure that the e6x does not support 3, or 4, and possibly 2, and I'm pretty sure that people are afraid to run 0 split on the haltech because they can't guaruntee that trailing will not fire before leading.

The MS1 has successfully run many high-power engines with no problem, with very large injectors and a LOT of boost... running ignition as well. The accuracy has not been a problem in those installs... you forget that all the code for the ms1 is hand-coded in assembly, which makes for some very tight code. So your argument that the ms1 chip is not worth buying is untrue.

nik

Wow, you guys sure love your megasquirts And I do admit it has evolved quite a bit since the early days. I guess you guys don't see that I believe the megasquirt is a great idea, just because I don't think it's as evolved as the
aftermarket ECUs of today.

Does megasquirt-1 support full trailing ignition as well as leading right now? At a quick glance that would require 3 output compares plus one for injectors if you are batching them all together and then some switching for the staging. You would need at least one trigger input as well, and I am guessing this is where you run out of hardware timer channels?

The haltech does interpolation as far as I can tell from experimenting, same with powerFC. The e6x does not have the variable resolution, which is a very nice feature. The powerFC has a 20x20 map and variable resolution. I still believe the 12x12 map is a limitation for half bridgeport engines like mine(not linear), although the variable res helps the problem.

You are right about the MS-2 having too few outputs, but that is because they chose a small footprint package for the microcontroller. The features you are bringing up are future implementations, and it is not reasonable to compare what the megasquirt will become, to what it is now. Right now you can only bring up megasquirt-1 currently implemented features for rotary. With time(and hardware upgrades), the megasquirt will be the best option for aftermarket ECU. Right now, I believe the (MS1)chip is too limited and has reached it's maximum capacity. People have taken it much farther than it was intented for, which is a good sign... Where were you guys when efi332 was around?

Finally, I clearly said if someone is interested there is no point in NOT buying the megasquirt, since it is so cheap and has possible upgrade to MS-II.

muythaibxr

Yes, it supports full leading and trailing, and doing so only requires 1 output compare. We are clever coders. jsmcortina over at msefi.com is responsible for most of it, but I got trailing working, and it works without a hitch. We're using the 1 output compare and bitbanging our outputs in the ISRs to get up to 6 channels for ignition... yes we support 6 cylinder COP

We also support wheel decoding up to 60-2, and second trigger input, so we can use a stock, unmodded CAS.

The 2 injector channels are 2 more output compare channels.

staging is based on RPMS or MAP, or a combination of the 2, and I've implemented a gradual staging trasition as well, so it doesn't just immediately drop the primaries to the lower pulse width while bringing the secondaries online... it gradually transitions them in over a user-selectable number of tach events. I don't think ANYONE does that.... or if they do they don't advertise it.

For the haltech not doing interpolation, I was going on hearsay... so I could very well be wrong, but I disagree that 12x12 is too little... Right now, we could support 16x16 or more on ms2 and there are no plans to go higher because nobody actually using the MS has complained.... However, I've been able to add code to msns-extra that allows you to use 2 tables instead of 1, giving you 24x12... you just have to overlap the two tables a little to keep the interpolation going.... so it's effectively maybe 22x12?

Every time we think the ms1 chip has reached capacity, someone goes and does something new with it like the hires (usec resolution) code.

I understand that, but even that isn't necessary to say... right now the ms2 isn't really an upgrade because it won't run our cars.... ms2/extra will do that, and *very* soon. For now msns-extra on an ms1 with the v3 board is every bit as good as many other ECUs costing 2-3x more.

The features I'm bringing up for ms2 are already there... the only thing I'm adding right now are features from msns-extra, like the wasted-spark/COP wheel decoder, rotary ignition, etc...

X-tau, PID predictor loop, etc.. are already out on the standard ms2 code... so it IS a valid comparison to say that nobody else does these things.

As for the MS2 and its outputs.... the GPIO board is coming out soon, as is the router board... those give the user COP and sequential injection, and all the outputs you can think of using CAN bus to communicate between modules...

My point is not only does the MS1 have the features to compete with ECUs that cost more than double the MS, MS2 will soon *BEAT* those ecus with ms2/extra, and future development is being done (think ms3 and XGATE) that will absolutely blow the other guys away.

coldfire

iTrader: (1)

the info is from my head, so could be wrong. i know the MS uses the 8-bit Motorola MC68HC, and i was under the impression the e6k ran an 8-bit chip also...or maybe i read 8MHz bus speed. not quite sure. guess it doesn't really matter, the point i was kind of trying to make is, it may be considered "old" technology, but can still run most engines fine.
i mean, the EFI332 was a 32-bit chip that was actually quite a modern design at that time, but that complexity is actually kind of what prevented it from taking off as much as the much simpler 8-bit.

on another note, the quickness of Ken (muythaibxr) putting info into this thread prooves there is support for the MS! haha