FMII and MS3: What fuel economy should I expect?
 

As some of you know, I've had an FMII and a Rev-built MS3 installed in my '95 with original engine by a local tuner shop that recommended the MS, and I'd appreciate feedback on what kind of fuel economy I should expect. Would you please share what you're seeing if you have a similar setup?

I had an old JR M45 and a Link before installing the turbo and I typically saw 21-24 mpg combined. From talking with Brandon at FM, he's got the same hardware run by a Link and he's seeing 26-28 in town and 28-30 on the highway driving conservatively. Brandon suggested I should see similar fuel consumption with perhaps a 5-10% bump because of Megasquirt's improved efficiency over the Link.

When I first picked up the car after the installation, I wanted to baseline fuel consumption with the new setup. In particular, I wanted to see how it compared to the JR/Link setup. To that end, I've been driving the car very conservatively - mostly out of boost since an engine is usually running that way - and generally boosting to 7psi or less. I expected a substantial improvement in off-boost consumption because the turbo avoids the parasitic drag from the supercharger.

But instead of an improvement, I'm typically seeing consumption of 15-19 mpg with a fair combination of street and highway driving. The tuner and I are both seeing that the AFR target table appears correct and the VE table is showing mixtures that are hitting the targets. We replaced the WBO2 sensor with a new Innovate unit and that hasn't improved anything. We're also running with sequential fuel and ignition to RC 550 injectors and the FAB9 CoP system.

The car is actually running great and making good power except for the fuel. My tuner is saying everything should be working, he's done about everything he can, and that 15 mpg could be about the best we can get. That doesn't seem right since I was seeing better with the supercharger/Link and the new setup should be *more* efficient when staying out of boost. At the very least, a 1.8L engine in a 2400 pound car shouldn't be getting mid-to-high teens in fuel consumption. That sounds more like a big pickup truck.

My tuner says the next step is to "force" the ECU to burn less fuel by playing with timing and leaning out the mixture. He's the expert so I need to defer to his opinion, but that doesn't feel quite right to me.

Alternatively, he suggested the fuel pressure might be too low to make my RC injectors happy. As part of this build, I had them install a used Vishnu fuel rail that has a fuel pressure regulator attached. Since it's used, I suppose it's possible that the regulator could be bad but that seems unlikely.

Any thoughts about what could be going on?

When you're cruising at a constant speed what is your wideband showing? With normal driving my gas mileage has been pretty consistent between a turbo and a stock car. You do seem lower than I'd expect it to be.

I usually lean out cruise afr to 15.5:1. That seems to pick up a little bit that gets offset when you get in the throttle. I see people claiming +30mph pretty regularly but that hasn't been my experience. I typically get 21-24mpg on 93 octane, my record tank was 29mpg. Even on E85 I get 17-20mpg.

Another thing to look at could be overly aggressive accel enrichment. This is the time when you should be posting logs (and pictures of logs) of different run states.

I typically see 19-20mpg with my EFR6258 + Basic MS3. My old MS1'd NA was less - typically 16-17 in town. I saw as low as 12.5mpg on two separate tanks (one 100% town tank, one 100% highway tank at entirely unmentionable speeds), and as high as 30mpg (65mpg, some hills, no boost, 100% highway).

IOW, it depends almost entirely on how you drive. If you beat it like a rented mule, it will return mid to high teens. If you drive it like your grandmother is riding shotgun, it will return mid to high 20s.

I get over 30 on the highway.

Post your tune and logs. We can't help you without them.

I drive mostly off boost... 20-24mpg when I ran RC injectors, I average almost 28-30mpg on newer five-o 460cc ev14

Stock ecu with begi reflash + xede though. Cruising AFR bounces between 14.7-15.3

Thanks for the responses, guys.

@pdexta, I'm seeing 14.7 at cruise, which is what the Rev's base AFR target table is set to.

@chiburbian, I'm happy to post logs if you all are willing to look at them. I'd sure appreciate the help. Is there anything in particular I should be logging?

@savington, I'm driving it like my grandmother is riding shotgun in an attempt to optimize my baseline fuel consumption. I'd be happy with low to mid 20s since that's what I was getting with the supercharger/Link.

Log all sorts of different driving conditions.

@aidandj, my tuner doesn't want me to post the tune because he considers that his intellectual property. I appreciate that he's left it unlocked so I'd like to respect his request, even if it's inconvenient. I can look into any settings that you suggest though.

@girz0r, my tuner suggested the RC 500s may also be a problem because of an inefficient spray pattern. I'm open to getting other injectors if they'll improve efficiency.

OK, I'll go for a drive and post the logs when I come back. Sure appreciate the help. Thanks, guys.

We probably can't help you then. Talk to your tuner then, sounds like its his problem.

Yes, it's his problem to fix. I'm trying to understand what the heck is going on. Hard for the engineer in me to keep out of this, especially since we've been wrestling with it for a couple months.

I thought of an additional bit of information to share. Up until recently, the shop has been unable to duplicate the same poor fuel consumption that I've been seeing. We talked about it in greater detail and realized that they've been driving the car really hard with a lot of boost, in comparison with the grandmotherly way that I've been driving it. So whereas I've been seeing 15-19, they've been seeing 19-24. So it seems like the problem is in the non-boosted areas of the tune.

I'll go log a drive and get back to you. Thanks again.

OK, folks. I drove around for some errands and took a log file. I'd say this is a fair representation of the kind of driving I've been doing. I'd sure appreciate it if you could see if anything looks out of the ordinary.

That looks like a composite log. Not a datalog.

Back in 2007-2010 I could easily get 32mpg on the interstate with a Begi S1, rx7 440cc injectors, and MS1 as long as I did not floor it to much. I would try to cruise around 15.5-16.0:1.

Oh, I suppose it is a composite log. I didn't know there were different kinds of logs since this is all new to me.

I don't see an option for "data log" on the logging tab in TunerStudio. I searched around online and in the help file and I can't find anything there either. Would you please clue me in on how to get you a log you can use?

Oh, I think I see it now. It's under the Data Logging menu instead of the Diagnostics & High Speed Loggers tab. Is that the one?

Yeah, click the dropdown menu and hit start log. Go drive, then hit stop log before you turn off the car.

OK, I got a datalog this time. Would you all please take a look and let me know what you think?

I don't think it's necessarily the root of your problem, but you should definitely investigate getting better injectors. The RCs are pretty junky compared to the current batch of affordable Bosch ev14 injectors.

Flowforce, Injector Dynamics, etc

@elfini~fc3s, thanks for the feedback on the injectors. I agree that it's unlikely the injectors are causing the whole problem but now I'm appreciating that they're not helping.

I don't want to distract us from discussing the core issue of reviewing my tune, but is there a consensus on which injectors would be my best choice? It appears all the companies are using the same basic injector that they modify a bit. Flow Force seems least expensive, Deatchworks is in the middle, and Injector Dynamics is the most expensive. Is there a reason I wouldn't go for the least expensive since they're basically the same?

Flowforce doesn't modify them at all, which is why they are the least expensive. All Nigel does is buy in bulk to reduce cost, then flow-test, dynamically match, and include everything needed to drop them in. He's local, so we helped with some fitment foibles that have always plagued the ID1000s, so the FF610s now fit better than those.

For stock bottom end or light builds (sub-325whp), the FF610s give up nothing.

Flow Force 610cc EV14 Injectors

Thanks for the feedback. I'm still running my stock engine and the 250 hp it's generating honestly feels like more than I want. I might dial it back. So the 610s look like they'd work fine and are about 10% less expensive than the Deatchworks. Always nice.

In the meantime, I'm looking forward to getting feedback on the log I posted yesterday afternoon.

Ace, can you go for a 30+ minute drive and post the log of that? I am messing around with scatter plots to maybe diagnose the problem but I need more data points.

It does look like your AFR target never goes above 14.7 in cruise. Plus, your AFRs aren't hitting 14.7 reliably in cruise. Do you have EFR/EGO control turned on?

Oh, so you need a longer drive so there's more data? Sure, I can do that.

I wouldn't expect the AFRs to go much over 14.7 since those AFR targets are set to not go much leaner than that. I've attached that table so you can take a look.

If I understand your question properly, I believe I've got EFR/EGO turned on. The Megasquirt is running in closed loop mode and is making adjustments based on inputs from the O2 sensor. But I don't want to assume we're talking about the same setting. Can you point me to where I'd check in TunerStudio to make sure?

Below is part of my AFR table. I have mine a bit tailored due to my wideband reading a bit off, but should give you an idea of what I aim for when cruising.
https://www.miataturbo.net/attachmen...ine=1450895592

I am at work and can't look at tunerstudio but I believe it's:

Fuel>AFR/EGO Control>

I am by far not the best tuner here so I will be happy to hear other opinions, but it seems like to me that your AFRs are jumping around a lot in cruise if I remember your log correctly.

Also, your AFRs in cruise don't have to be 14.7:1. My map is closer to Shuiend's.

@shuiend, OK, I see you've leaned out the AFR targets in that midrange zone. The table I'm working from is the base configuration Rev placed on the ECU and I don't *think* my tuner changed any of that. But if you're not getting any knock then I figure your AFRs would be OK and would result in less consumption. Out of curiosity, why do you suddenly go rich at the higher RPMs?

@chiburbian, I'm attaching a screenshot of the AFR/EGO screen. It looks correct based on my knowledge.

I'm going to go for a 30+ minute drive as requested. If you tell me stuff looks generally OK in the log then maybe we can play with the AFR targets, autotune a bit, and then see where we are.

If your tuner actually tuned cruise and stuff by hand then autotune will screw that all up. If he didn't even change the basemap and then autotune I would find a new tuner

He definitely tuned by hand on a dyno and I didn't figure that any autotuning I did would be a final configuration. I wanted to see what kind of effect it had on fuel consumption and then revert it when I take the car back to him. But maybe it's better to share what we learn here and leave the changes to him.

And honestly, I can't remember what the base AFR table looked like from Rev so I don't know what changes might have been made.

I logged a good long ride on my favorite mountain roads and then did some stop-and-go driving for good measure. Please let me know if this will work for you.

I wanted to bounce something else off you too. We've been talking about leaning out the mixture in the non-boosted part of the fuel table, but I'm thinking back to the AFR targets with my Link. I've attached that table from Flyin Miata for reference. The Link was generally running at 14.8 - 14.0, which is about what the Megasquirt is running, and had lower fuel consumption than the Megasquirt.

So on the one hand, it seems reasonable to lean out the non-boosted mixture if that's what's needed, but I'm struggling to reconcile that against my experience with the Link. All things being about equal, the Megasquirt seems like it should be returning at least as good fuel economy as the Link at the same A/F mixture.

What do you all think?

Is your AFR gauge matching your indicated AFRs in tunerstudio? (in other words, is it calibrated on the megasquirt?)

It's hard to tell if they're saying *precisely* the same thing on the gauge and in TunerStudio because the numbers move so quickly. But in general, yes, it looks like the numbers match.

You can setup in TS in the VE Analyzer live section to only autotune certain sections of your VE table. You could set min and max rpm then use the custom filter to only do between certain KPA's I believe. Get that setup, lean out the center section of your AFR table like I have mine, then go drive on the interstate at constant speeds between 60-90mph. Also you don't need to worry about that lean and knock. I have ran as lean as 18:1 with no knocking in that area, I don't suggest it, but it is doable. If you want to dig deeper you can search some of Hustler's old posts about his super mileage white DD where he was working on tuning for MPG.

@shuiend, you said that your AFR targets are tweaked because your O2 sensor is a bit off. Would you suggest that I use the 15.4 that you're running or should I try something else?

Hi, folks. I took advantage of my day off to do a bit of experimentation based on your feedback. I tweaked the AFR target table a little differently than shuiend suggested, but I believe I got the general idea deciding to make it a little more conservative (rich) in the higher RPMs. I figure I can lean things out even more if things look like they're going well.

I then autotuned everything in the non-boosted part of the table, i.e fuel load <100 kpa. The car seemed to run just as well at 15.4 as at the 14.7 I was running before. Would you please take a look at that updated table and let me know what you think?

I also took a bit of a datalog and would appreciate your feedback on that.

Finally, I checked if the O2 gauge reads the same as in TunerStudio as chiburbian suggested. It turns out that the gauge is reading about .3 lower than TS. So if the gauge is at 14.7 then the software is showing 15.0. It seems like that could account for excessive fuel consumption since the car would be running richer than the MS thinks.

I went into Tools/Calibrate AFR Table, selected the Innovate LC-1/LC-2, and wrote it to the ECU just to be sure. I didn't see anything change in TunerStudio though. Do you have a suggestion for how to make them read the same?

Agreed.


OP, I looked at the first log you posted. At 1772 seconds into it, it shows 4,000 RPMS, 4.5 PSI boost, 12.4 AFRs, and 37 degrees of ignition advance.

Just so you know, that's wayyyyyyyyyyyyyyyyyyyyyyy too much timing, and your detonating. RPMs drop while in boost, which could be a misfire but likely it's from detonation.

At 1919 seconds on the log, you hit max boost. SEVERE detonation is shown, to the point that RPMs were dropping when you gave it gas. At this point you were running 8 PSI boost, 12.0 AFRs, and 36 degrees of timing!!!!!!!!!!

Start over on your spark table, and do a compression test and pray you haven't nuked a piston. And new plugs. And whoever did you spark table, don't ever let them touch your car again.

Oh, wow. Thanks for the feedback. That's sure not what I wanted to hear since I had the ECU professionally tuned to avoid this kind of trouble.

I'm confused about what you're saying because I'm not hearing any detonation at all and the car feels really strong. Also, when I look at that log, I see RPMs and boost gaining until I let off the throttle. Soon afterwards I see the boost drop first, followed quickly by RPMs. I also figure that an AFR of 12.0 is about right as well. Isn't that what I should expect?

Based on your suggestion, I'd expect to see RPMs drop before boost. No?

I know the tuner messed around with timing so that the numbers in the table are offset by an equal number of degrees in the base timing. So basically, he set the base timing offset from TDC where you'd otherwise expect and then added that offset to the timing table. Based on my own experiments, I believe his offset is about 10*. I don't fully understand his logic, but perhaps that accounts for the weird timing number?

Oh, I see. You're talking about the bit that starts at 1919 seconds and goes to 1921. I thought you were talking about the curves that finished at 1919 seconds.

Even so, it appears to me that during those 3 seconds, I hit the gas, the boost went up, and then RPMs went up with a little lag. I'm still not seeing where RPMs were dropping where I was giving the car some gas so I'm not able to see where the detonation would be.

That's a shitload of timing still. Basemap is like 16 degrees there

Check your timing with a timing light. It's easy to do and it's covered here many times. Do that, make sure the numbers match, adjust if necessary.

This is what i was referring to.

https://www.miataturbo.net/attachmen...ine=1451005690

Best case scenario is a terrible misfire. But if you're running anything close to the timing your log shows, it's detonation city.

I run 20 degrees at that spot on E85 with 9:1 compression and a built motor. About 16-17 on pump gas.

This is a pic of my car going from 1st through 4th. I used launch control on the start, so ignore the RPMs jumping around there as it's normal for launch control. But notice how the RPM trace is butter-smooth and only climbing up? That's how it's supposed to be.

https://www.miataturbo.net/attachmen...ine=1451006132

Thanks very much for taking the time to point out the problem. I wouldn't have seen that myself so I really appreciate your helping me learn.

Needless to say, this is very upsetting news since I paid a lot of money to avoid this situation. I'll share this feedback with my tuner when he's back in the shop on Monday. He's said that he's doing stuff in his tune that only makes sense if you know what's going on in his head. As an example, I put the timing light on the car at idle and was seeing the crank pulley ticks way off the TDC marks. In the Trigger Wizard, the Ignition Offset Angle is set to 6*. I had to change the offset to -4* to get the marks to line up. That's when he told me about his deliberate manipulation of those numbers in both the base timing and the timing table. So I know he's doing something but I don't understand what it is.

Regardless, I'll ask about the timing and what's up with the detonation. By itself that would tell us that the timing is incorrect even with the offset calculations, right?

If I don't get a decent answer then I'll likely tune it myself.

Based on what you just wrote, you're actually running 10* less timing that what your logs say (6 - -4 = 10). That means the "36" degrees in your log was actually 26. Less bad, but 26 is still wayyyy too much timing. I run 20 there on E85, 16-17 when on 93 Octane.

Pull the plugs and get a 10x magnification jewelers loupe and look at the porcelain of the plug. If it's detonated, you will see black specs on the white porcelain. If it's severe, you'll see little balls of aluminum too. If you can see any of this without a 10x magnification your hammering the motor with detonation.

Also as I said, the RPMs dropping in boost is one of two things, detonation, or a misfire. Either is bad, though detonation is worse as it's much more destructive. Reading the plugs will show you if the motor has detonated. If plugs are clean under 10x magnification and good lighting I'll be shocked. If it's a misfire it could be several things causing it, and it's not uncommon in the miata world as these cars have crap ignition coils from the factory.

Yeah, that's what it sounds like with the timing. I asked about this, and like I said, I was told that he's doing stuff that can only make sense if I understand everything he's doing, which he didn't want to tell me.

But it sounds like the 10* offset might have been calculated incorrectly. If the timing is supposed to be 16* and we're seeing 36* in the log then perhaps he put a 20* offset in the ignition map instead of the 10* it should have been. Am I thinking about this right?

I just went outside and checked the #1 and #2 plugs. Both of them looked perfect, which is a relief. I'm running the FAB9 CoPs and we had some trouble with misfires that I thought had been fixed by switching to sequential ignition. I guess not, which means we've still got some work there. As you said, that's better than detonation, and we still appear to have a timing issue to fix.

If you didn't use a 10x magnification and good lighting then you need to. Wrote all that out for a reason.

Yes, I looked at them with a 10X magnifying glass in daylight. Thanks for making sure though.

Pat has a true crank sensor, and you (likely) read off the timing belt still.

So, take what pat says about the smoothness of the RPM plot with a massive grain of salt.

Your tuner still sucks massive peen though, if he didn't set the trigger wheel he missed step 1. Whatever is going on in his head is irrelevant, it's built on bullshit. He likely won't admit it, he either doesn't know or doesn't care.

Your RPMs don't jump around like that from a cam angle sensor driven by a timing belt. If that sensor's RPM jumped around THAT Much on its own (sensor or timing belt inaccuracy), it wouldn't be accurate enough to predict timing of the crankshaft to within a degree or two.

Mine do, my car does not detonate or misfire in boost. There's a reason I want a crank trigger wheel.

I also run 26degrees around 4k and 4psi, but I don't often see that cell and I'm on e85...

Thanks for the input, guys. Yes, I'm running all the timing off the CAS and don't have a crank trigger wheel so I'm grateful for another explanation. For the record, my tuner warned about problems running timing off the CAS as opposed to the crank. You may recall from another thread that we determined the CAS was surely less accurate than the crank, but should be sufficient since many NAs run that way.

I'm not in a position to refute Deezum's opinion on my tuner since I'm not sure what he's doing. He's been working on my car for several months and I'm still seeing issues. That's sure not good. This thread has been very useful in helping me understand what might be going on. I'll bounce this off him and see what I make of his responses. Then I'll decide how to proceed.

In the meantime, can you guys think of why TunerStudio would be reporting AFRs .3 higher than the gauge is saying? Recall that if the gauge is showing 14.7 then TS is typically showing 15.0. I checked the calibration in TS and it appears most O2 sensors use the same voltages to correspond to AFRs so it seems like there shouldn't be much need for fiddling here.

It's not your CAS, it's impossible for your RPMs to drop by 100 RPMs WHILE ACCELERATING when under load due to belt stretch/CAS sensor inaccuracy, but at the same time the CAS be accurate enough to predict ignition timing withing a degree or two.

Fab9 coils are pretty but a terrible weak setup. Those coils are designed to be driven by a CDI (400+ Volts to coils) but are instead driven by an ignitior (12V to coils), that's your misfire problem. Stock coils are stronger than the Fab9 setup. Get a CDI ($$$) or a real ignition system upgrade. LS coils, FM big spark, IGN1A coils, etc.

Well, that's disappointing since I got the Fab9 setup to avoid misfire under boost. But it's not a huge surprise considering some of the issues we've had with it. Bryan from Fab9 suggested playing with the dwell and duration a bit but my tuner found the suggested settings increased heat in the coils to an unacceptable level. His website also says the kit should support ~19psi with the included module. I guess that's not true, eh? I'd thought we'd fixed the misfires by switching to sequential spark but that appears not to be the case.

Either way, thanks for letting me know. I take it you're not a fan of the Toyota CoPs then?

Fab9 suggested an AEM Twin Fire module if more spark is needed from the system. It appears this would be the item in question. https://www.acceleratedperformance.c...odules-30-2821. So yes, pricey, but potentially less expensive than the FM kit since I already have the Fab9. On the other hand, it also appears to be discontinued from AEM, although I can't tell what they replaced it with.

Either way, it seems like this is a secondary problem to the fuel consumption at this moment. I'm seeing 3 issues that need to get resolved pertaining to that:
1. Understand why the AFR readings in TS are reading .3 leaner than the gauge.
2. Verify what's going on with the timing.
3. Retune with a leaner mixture in the 15.1 - 15.4 range as suggested.

Then I'll focus on the ignition and avoiding misfires.

GL getting a CDI for under 600. That AEM one is impossible to get buy new, maybe you could find a used one though. Truth is nobody uses those anymore outside of bike engines.

I don't know much about the toyota COPs. I know the ones I listed are hot nasty coils that work- they can put out 100 mJ all day with 3ms dwell and 200+ with lots of dwell. I run mine at 3-4ms at cruise, 6ms at idle and very high load. 30 thousands plug gap. Might up that to 40, I tested it earlier this year and it took 5.5ms to fire a 40 plug gap with absolutely no misfires in boost at 28 PSI.

It's possible your car has a misfire at cruise. If so, that would absolutely affect your mileage in a negative way. So I wouldn't write that off as a secondary issue just yet. What spark plug gap do you run? What plugs?

I get 15-19 mixed driving on E85 with an extremely lossy suprecharger setup dragging my fuel economy down. And this is my AFR table.

https://www.miataturbo.net/attachmen...ine=1451167102

EDIT: I'm only running 4.5ms of dwell at high load with .030 gap. Next set of plug I may go back to 40 and 6ms dwell up top, we'll see.

I tried the fab9 coils with there old module. It did blow out spark above 6 psi. Fab9 sent me a new module to try. How old is your kit? Do you have there newest ignition module?

Double post oops

@patsmx5, I did find a couple of websites offering the AEM module for $400-$450 so maybe they're offering old stock. In any case, what are people using as replacements for these CDI modules?

Wouldn't I notice if I had a misfire at cruise? I sure noticed when I had the misfire issue with the Fab9 CoPs before and it seems like I'd notice even more under acceleration. More to the point, wouldn't misfires show up on the datalogs with squiggly RPM lines like you pointed out?

Right now, the MS is configured for 3.5ms dwell and 1.5ms duration, which is consistent with Bryan's suggestions. This is up from 3ms dwell before. I recall I'm running a .030 gap, although I can't remember for sure. The gap had been smaller before too.

@90_Turbo, I verified with Bryan that I have the newer module so I'm all good there.

Motec makes a 4 cylinder CDI box but it's $$$. I would go with something cheaper like the coils I suggested before I did a CDI.

You posted a log here with a misfire in it and you didn't know it. So I can't say you would notice it at cruise.

Look at the pic of the log I posted. See how the RPMs are smooth, and the AFR trace (second graph, white trace) is dead-flat smooth too? That's because I don't have a misfire. Open your log and look at yours and compare. RPM and AFR jumping around is the common sign of a misfire. And shit mileage. Could be something else but you've got 3 out of 3 on obvious signs of a misfire.

If you have stock coils try them. Or try gapping your plugs to .020 and see if it's any better. Or just put some better coils on there. Or run more fuel in cruise, it's easier to light off a 14.7 mixture vs a leaner one thus less chance for a misfire.