Tuning the NSX 3.0L Using the OEM ECU

Hi guys, just got done reading through this whole thread and sad to see no updates in 2 years. Id love a cheap basic way to tune my bolt on na1, and this seems to fit the bill, so hopefully development is still ongoing.

Anyway, when I went to look up the demon2, it appears to be discontinued, with this newer product being offered “Snake emu”

Anyone try this out yet? Would it still be compatible with the same files and software as the demon?
 
Hi guys, just got done reading through this whole thread and sad to see no updates in 2 years. Id love a cheap basic way to tune my bolt on na1, and this seems to fit the bill, so hopefully development is still ongoing.

Anyway, when I went to look up the demon2, it appears to be discontinued, with this newer product being offered “Snake emu”

Anyone try this out yet? Would it still be compatible with the same files and software as the demon?
The Snake is definitely the one to go with, but IIRC must be modified to accept the NSX ECU inputs. I think John can explain better, but it will work with Tunerpro and accept our data files.
 
I don’t plan on supporting anything besides the Moates Demon 2 which seems to be available again. I can chip the ECU for $500 including the Demon board and get it set up for logging. Or you could probably find someone else to do it cheaper or DIY isn’t too bad if you have the tools for thru hole reworks.

Also boostednw has been problematic lately according to posts in the Honda tuning groups so I would steer clear of them.
 
If the Demon 2 is available again, that's big news. Anyone who wants to mod the OEM ECU and not deal with aftermarket should grab one while you can get them!
 
I rarely use TunerPro for anything, and admittedly I'm a bit rusty here with it, but I receive a verification failure when using the "alpha" file unless I open my stock PR7-A03 and write that to the Demon first.

Perhaps I'm overlooking something simple here. Any insight as to why this may be?

Thank you!
 
Which file are you trying to write, exactly, to the Demon?
Screenshot 2025-03-03 162530.png

I plan to use this as a starting point, but so far, the only way to get that to successfully verify is to write the stock ROM, then open that, modify whatever, and continue on. If I disconnect from the Demon, and attempt writing again, it succeeds, but fails verification. When I download the file from the Demon, it shows some corrupt values in spots as well, which is likely why it fails the verification?
 
How is the Demon chip wired up? You need the +5V and ground wires in addition to the 28 pin connection or it can be flaky. Also writes are sketchy if the Demon is in an ECU but the ECU isn't powered up (via bench power supply or from turning the ignition on).
Well, that's indeed the scenario here right now (Demon in ECU, ECU on desk). :)

Current install:
PXL_20250228_204710039.MP.jpg
 
Ah. Yeah flashing the demon like that with the ECU un-powered does not work well if at all. You'll need to power the ECU with a 12V supply to bench flash, or take the Demon completely out of the ECU. Or just put the ECU in the car.

No problem; I did notice that if I removed the board from the ECU it would write and verify essentially every single time I tried, so it was driving me nuts over the weekend. I'll toss this thing in and throw it on the dyno soon then.

Thanks for the confirmation!
 
Let us know how it goes! So far I think I'm the only person whos dyno'd their car with a Demon setup so far. What mods do you have?

It's a super basic setup.

'91 model with 65,000 miles
Pride long tubes
Unresonated test pipes
Pride V2 exhaust
Carbon intake scoop

Sheetz brand 93 octane from central PA!

The 100% stock baseline was 236hp / 183tq, and my dyno wideband showed pretty lean fueling on the baseline (14.0-14.2:1) at WOT, so this is mainly to richen her up a touch and ensure it's happy.

23b9ca1e-1ad3-4a48-b8a9-a9f98c759e34-1_all_264.jpg

I have a bunch of experience with the four cylinder OBD1 stuff from years and years ago, but I haven't touched a stock ECU of this vintage in longer than I can remember. I deal exclusively with the newer Honda/Acura stuff these days, thankfully.
 
That looks like a pretty normal/typical NSX dyno- nice! Open loop fueling on the stock tune does go into the low 14s. Since you've only got headers/exhaust, there might be enough room left in the injectors to bump fuel a little bit.
 
Just wanted to pop in and update.

I got a chance to toss the car on the dyno this afternoon.

I had *just* enough injector duty to richen it up to my liking. It made the most power at 12.3-12.5:1, and funny enough, it didn't want any additional ignition advance over the stock table values. It actually ended up with less advance at peak torque (and the 93 octane I'm using is pretty good stuff, based on my experience with it on a wide range of engines).

I'm very pleased with the results, and the car runs great.

255hp / 200tq is where it ended up.

Solid lines are with all parts installed and tune dialed in, dotted lines are stock.

1000016631.png
1000016626.jpg
1000016627.jpg
1000016620.jpg
 
Last edited:
Solid results and an actually realistic dyno! Did you find any more torque playing with the tune compared to the stock mapping or were the numbers pretty much the same?
It picked up torque everywhere from start to finish via enrichment alone. That was the single biggest difference.

I yanked 3deg from 2000-8000rpm as a test, and it lost a fair amount everywhere except 5800-6500rpm. I ended up pulling 2deg in that range for the final run (compared to the stock ignition mapping).

I noticed I wasn't able to input a value higher than 8000rpm for the 'final' rev limit, not realizing it has a clamped maximum value in the XDF I have open. The curve looks like it could benefit from having just a little more RPM, so I'm going to bump it to 8,200rpm.
 
It picked up torque everywhere from start to finish via enrichment alone. That was the single biggest difference.

I yanked 3deg from 2000-8000rpm as a test, and it lost a fair amount everywhere except 5800-6500rpm. I ended up pulling 2deg in that range for the final run (compared to the stock ignition mapping).

I noticed I wasn't able to input a value higher than 8000rpm for the 'final' rev limit, not realizing it has a clamped maximum value in the XDF I have open. The curve looks like it could benefit from having just a little more RPM, so I'm going to bump it to 8,200rpm.
First, it's so nice to see a good tuner working with the NSX ECU! I'm also going to do some timing tuning this spring, though John and I have speculated that pulling timing a little might yield gains. Thus, it's interesting to see that pulling timing seemed to work better for you, which is similar to feedback from past NSX tuners.

Next, I would think twice about raising the rev limit. The NSX has an issue with harmonic resonance above 8,100 rpm that will shatter the oil pump gear during extended high rpm operation.

images.jpg

Conventional wisdom is to keep the limiter stock unless you invest in a billet gear. It also appears that Honda corrected the flaw for the 1997+ cars and you can also replace with the -J21 pump. Decades ago, the "hot" mod was to get a chip with a 8,300 rpm limiter- many blown up oil pumps as a result.
 
Last edited:
First, it's so nice to see a good tuner working with the NSX ECU! I'm also going to do some timing tuning this spring, though John and I have speculated that advancing a little might yield gains. Thus, it's interesting to see that pulling timing seemed to work better for you, which is similar to feedback from past NSX tuners.

Next, I would think twice about raising the rev limit. The NSX has an issue with harmonic resonance above 8,100 rpm that will shatter the oil pump gear during extended high rpm operation.

View attachment 198266

Conventional wisdom is to keep the limiter stock unless you invest in a billet gear. It also appears that Honda corrected the flaw for the 1997+ cars and you can also replace with the -J21 pump. Decades ago, the "hot" mod was to get a chip with a 8,300 rpm limiter- many blown up oil pumps as a result.
That's a bit of a downer. I'll reconsider the raise, then, I suppose. I saw the note, but figured it was more of a cover-your-butt kind of warning for the casual user. 😅
 
For the RDX injectors I think we discussed retarding timing since better atomization should theoretically result in a faster burn.

@DRobIMW if you ever get the itch to change something, RDX injectors (or other modern injector) could be interesting. I never got to test them back to back with the stock ones but 10-15 years ago there was a vendor making dubious claims of 30+ lbft with RDX injector/chip kits.

The safest option for raising your redline is the Toda oil pump gear. The later J21 pump seems to have an updated rotor but I don't think anyone has tried testing it's limits. I remember a while back @Honcho and I were comparing the rotors side by side of an old pump and a J21 and IIRC the J21 rotor set was heavier by a noticeable amount so it seems like the material or production process was changed somehow for the inner rotor. (overall weight, inner rotor, outer rotor)

But yeah the limit in the XDF file was sort of a "don't do this unless you know what you're doing" type of deal. I got myself pretty good with it though, a while back I was testing the ECUs capability to run at 9000+ and it kept going into fuel cut at 8000, took me a while to realize the XDF was limiting it. (with an engine simulator board, not in my car)

Old rotor set: 244.1g / 122.4 / 121.7
J21 rotor set: 250.2g / 128.3 / 121.9
 
Last edited:
I have JDM 2/3/4 gears and NSX-R final drive to swap in here, maybe I'll just pull it all out and do a 3.2 and upgrade the oil pump gear while I'm working on it, because I have a mental problem where I can't leave anything alone.

I find it very hard to believe +30lb-ft exists from sticking in RDX injectors, too. Did someone really make such a claim? 🙄. There might be a few here and there on the table from injector phase adjustment, but... I bet it's nothing revelatory here.
 
Back
Top