[Success Story] 2004 Mazda MX5 1.8 VVT #50

Your chance to introduce yourself and your vehicle
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2004 Mazda MX5 1.8 VVT #50

Post by AndreyB »

Now that I am thinking about it I an not 100% sure if I got MIL to work on NB2 :( I know it's just a trivial low side driver on NA6 but no idea what's the deal with NB2 :( Need better record keeping :(
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

Mine works but is very very shy, the light is very weak.
It's actually on 5v I think, maybe 12v is the answer but I don't want to burn it...
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2004 Mazda MX5 1.8 VVT #50

Post by AndreyB »

What does the wiring diagram say? I would expect it to be low-side driven, not high side driven?
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

On the 2003 Miata page low side driver is specified, but on the 2003 alt which is working for me it's stated that high side driver should be used. So as I see it it's just a question of switching the wiring to an unused low side driver.

Will try that tomorrow eventually and give you some feedback on it. It would be great to change the wiki page for that
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2004 Mazda MX5 1.8 VVT #50

Post by AndreyB »

Crazy Striker wrote:
Sat Aug 31, 2019 8:59 pm
On the 2003 Miata page low side driver is specified, but on the 2003 alt which is working for me it's stated that high side driver should be used. So as I see it it's just a question of switching the wiring to an unused low side driver.

Will try that tomorrow eventually and give you some feedback on it. It would be great to change the wiki page for that
In my opinion this is a vague post. Is this about MIL? Alternator? What are the URLs of the pages you are referring to? Are you trying to report an issue? Please edit your post and provide a bit more details.
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

https://rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_2003 is the one stating low side should be used for CEL.

https://rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_2003_alt is the one referring to high side driver for CEL. I used the wiring diagram from this one to build my harness and most of functions are correctly wires, but obviously not for the CEL. It glows very poorly following this wiring diagram.

I will try the low side driver tomorrow as stated by the first link and report results

Gwendal
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

So I tried the low side driver for the CEL and it is not working. Seems it should be driven using +12V as +5V seem to barely light the bulb.
As I'm dur to a 1000km trip next friday to get hoem I will not be able to test it anymore until mid September.

I had one or two questions still.
Is your engine able to idle without the IAC with the stock Throttle body idle screw settings ? I need like 30% of duty cycle for the RPM to stay at 800rpm approx, and on the stock ECU the idle is kinda instable. I'm starting to think my throttle body bypass screw is clogged up...
Otherwise, for the auto idle, do anyone have approximative PID ans such values that would work on a NB2 miata ? I tried but it was pretty instable with what probably was a positive feedback loop. I studied PID quite a lot so I understand what had to be changed but I couldn't get it to stabilize...

But so far I'm so happy of how everything turns out, I'm considering building another rusefi one day so that I can make a clean one using a reflow oven.
@russian what method do you use to place and solder components for your kits ? SMD stencil, solder paste and oven ?

Gwendal
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2004 Mazda MX5 1.8 VVT #50

Post by AndreyB »

Something is weird with that MIL, you are confirming my vague recollection :(

At the moment there are some discussions of better idle control in the 93 miata thread and in Slack chat.

I order assembled Frankenso from China, I hate soldering.
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

Oh so you use pcb assembly services, what website are you using for that purpose ? I was considering getting into microrusefi but I'm not sure about the number of available low side outputs if I want to get into weird things like water/meth injection. I could as well use an Arduino over CAN but I would prefer it to be controlled by rusEFI 😅 The other option is to design a new pcb suited for my needs which considering the amount of work you've already put in validation of used technologies should be quite easy too but time consuming.

I definitely need to get into that slack thing, I'll look for that

I think it's just some lack of usage which led to the MIL issue. It would be easy to test with a paper clip and the stock harness. You plug the pin to gnd, then to 5v then to 12 v and we should have our answer. Unfortunately I can't work on that this week but as I saw you are working on the VVT test mule 😉

Cheers

Gwendal
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2004 Mazda MX5 1.8 VVT #50

Post by AndreyB »

Just checked: check engine light is dim if connected to +5v and bright if connected to GND. This tells me it's a +12v bulb with LOW side control.
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

Hi all !

I'm slowly preparing the "beast", starting by the fabrication of mechanical parts. After receiving the piping I cut steel for supercharger flanges and started welding today. Considering I discovered welding this morning the result is pretty good. It was done using stick (MMA) then TIG to solve small issues. I just need to cut and Weld flanges for the BMW ETB that I received today and the supercharger bypass, as well as the supercharger bracket.
One of the machinists is gonna turn my pulley so that it fits the supercharger.

For the software part I need to configure the ETB control, I received the VSH2SP30 H bridge modules. Are they driven through the PWM input ?

Cheers

Gwendal
Attachments
IMG_20190911_180344.jpg
IMG_20190911_180344.jpg (2.85 MiB) Viewed 18777 times
IMG_20190911_180228.jpg
IMG_20190911_180228.jpg (4.62 MiB) Viewed 18777 times
IMG_20190911_171251.jpg
IMG_20190911_171251.jpg (2.98 MiB) Viewed 18777 times
IMG_20190911_180237.jpg
IMG_20190911_180237.jpg (3.69 MiB) Viewed 18777 times
IMG_20190911_210858.jpg
IMG_20190911_210858.jpg (2.6 MiB) Viewed 18777 times
User avatar
AndreyB
Site Admin
Posts: 14292
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: 2004 Mazda MX5 1.8 VVT #50

Post by AndreyB »

Did you decide to keep mechanical throttle body? We shall chat about this. I am afraid if the struggle would be to pass all the knowledge related to ETB tinning, we have next to zero documentation on how to get a nice bias curve and nice PID settings
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

Yup I'll leave the mechanical TB on the stock position, so it can eventually act like a failsafe.

For the TB I'm ready to try it out, I can try to tune the PID using an arduino before so that I can write a code quickly

Gwendal
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

Hi all !

Today I had the opportunity tu try rusEFI on another car and I'm quite pleased about how it worked out. A friend of mine with a 2002 1.6 Miata has a possible alternator issue, the regulator doesn't regulate well and it sends 17v in the battery when running over 5k rpm...

So to sort out if it was his alternator or his ECU I plugged in the rusEFI without MAP or any computer and the car started right away, running a bit rich because of the 1.8 tune but overall it was fine. And considering I was still seeing 16.5V on the battery terminals I think the issue has been ruled out...

So very pleased to see it's so much reliable 😅

For now I can't use rusEFI on my car because I'm already struggling with the incoming inspection and emissions issues (no cat for now)... But as soon as I pass this part I swap my injectors for yellow RX8 and plug rusEFI in. Auto-tune will do the rest.

Gwendal
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

Hi everyone !

Lately I haven't had the opportunity to keep working on RUSEFI because of all the work needed on the chassis (Miata NB cursed chassis rail...) and pass my biannual inspection, but soon I shoul be able to make progress as I'm switching to Proteus ECU (https://rusefi.com/forum/viewtopic.php?f=4&t=1646). I had the opportunity to remove power steering so I'm now running depowered (with welded pinion) rack.

I ordered the boards as well as the components, and cut the harness from my kind of ghetto frankenso implementation to crimp the AMPSEAL connectors for the stock harness adapter. The connectors were obtained for free through TE sample system, thus greatly reducing the board cost.

For the supercharger install, I now have all pulleys machined, have all the piping material, the IC, the uprated chinesium radiator, the dual fan setup, the pulley tensioner, GM IAT, uprated RX8 injectors (550cc blue ones), fuel pressure regulator, fuel pump upgraded wiring, and new arduino AFT/BOOST/everything gauge with can implementation.
Still need an uprated fuel pump (walbro 255 or AEM) and maybe a coolant reroute.
I'll keep this updated as soon as I solder the boards.

Gwendal
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

Hi everyone !

After butchering the Proteus soldering I was able to plug it into my car. My harness adapter kind of worked well, I was able to get crank and cam signal and RPM sync, as well as all digital inputs working. Vbatt wasn't working because I did not solder the 82k sensing resistor. I could get stock MAP (baro) value close enough to be realistic (93kpa on a sunny day). Injector output #4 as well as ignition output #3 works well, but not the other. This has to be 100% because of my poor soldering skills... Not aving a good iron handy I won't be able to sort it easily soon. I also need to replace my makeshift resistor arrays with real ones.

Attached to this post you can find the .csv high speed trigger logger that TS produced during cold cranking.

I'll soon order 90% assembled proteus boards as well as a PnP adapter board, so I may have a working board sooner than expected.

cheers

Gwendal
Attachments
2020-04-22_18.31.17.csv
(1.23 MiB) Downloaded 326 times
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

Hi everyone !
I'm back with good news of Proteus and Pnp adapter both working.
Running using MAP sensor and speed density, VE and ignition are good enough for daily driving.

So attached is two logg I wasn't able to post on RE Online because it's too heavy, picturing a 10 and 15 min drive with some semi spirited driving.

Will report back with more result soon

Gwendal
Attachments
DataLogs.zip
(6.32 MiB) Downloaded 288 times
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

Hi everyone ! After a few weeks daily driving my NB2 on Proteus I'm now con fident enough to drive it 100% rusEFI.
So I decided to swap my trigger plate in favor of a more precise one, coming from Mazda Protege.
This wheel is 100% bolt on crankshaft pulley, just have to swap the old one out by removing the damper pulley and swap the new one back in, checking the crank angle sensor clearance.
So with this 36-1 wheel you get 10 degrees of timing accurcy instead of the 90 degrees of the original one, which helps a lot especially during cranking.
VVT is till working as intended, and I'm still running Wasted spark with fully sequential injection.
My last tune on rusEFI Online (https://rusefi.com/online/view.php?msq=71) is dialed in correctly for a NB2 running this wheel.

Gwendal
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

Hi folks !

I now run my NB2 on straight E85. We have that chance in France to have E85 stations popping everywhere, last time it was 3 km (2 freedom length units) from home.
Modified my injector rating to accomodate for lower energy density and the car runs really great. Timing was a bit too high before so now thanks to better octane rating I have a real increase in low end torque compared to when running on stock ECU and 98 octane.

Next step will be to clean my future 540cc injectors, clean them and install them.

Cheers

Gwendal
Crazy Striker
kit assembler
Posts: 145
Joined: Wed Aug 08, 2018 6:05 pm
Location: Brest, Britanny, France

Re: 2004 Mazda MX5 1.8 VVT #50

Post by Crazy Striker »

Hi all !

Little update on my build.

I recently decided to go Turbo instead of Supercharger, and bought a IS20 turbo from a Golf 7 GTI. It has electronic waste-gate (ETB style) and electronic bypass (solenoid).

I'm also running the car using Hellen72 (Rev. C) PnP board.
I decided to use the OEM "boost sensor" located in the engine bay and which is used on stock cars for EGR system. It's a 1 bar (approx) absolute pressure sensor and is thus suited for running speed-density on a naturally aspirated car. File used for calibration is available as attachment to this post.

I had to modify the H72 board and my harness slightly to make it fully functional.
First i cut the 4T wire on the OEM ECU plug and routed it to a 12V source (in my case CAM sensor supply), so that ECU gets a reading of battery voltage, even when alternator isn't running. H72 rev D should not need this mod because Vbatt sensing will be taken from 4AF ECU supply.
After that car was running ok-ish but tachometer and alternator charge control weren't working properly, as you can see in following video :

https://youtu.be/gXZ1wQmEwUI

Then I had to remove the 12V PROT circuit protection by removing Q313 transistor, R313 resistor and D304 diode, circled in red in the following image :
117037781-db773900-acd4-11eb-9e23-73f0ad681494.png
117037781-db773900-acd4-11eb-9e23-73f0ad681494.png (535.83 KiB) Viewed 11290 times
This mod made alternator and tacho working again, allowing me to drive the car to work to make some logs and assess how it was running.

https://youtu.be/Gs9rVYSrhH8

I felt the car less powerful in mid range, and had issues with it struggling to maintain idle when cold. Turned out VVT and IAC weren't properly working. This was due to the absence of fly back diodes on VVT and IAC. After bodging something i'm not too ashamed of (see following picture), I can now report the car as 100% functional. Pulls great, runs fine and is definitely daily driveable.
I had to add two 1N4007 diodes, between 2P and 2Q for IAC (stripe facing 2P), and between 4D and 4R (stripe facing 4D) for VVT.
117421190-99d9ce80-af1e-11eb-9716-d07e980c73a6.jpg
117421190-99d9ce80-af1e-11eb-9716-d07e980c73a6.jpg (4.43 MiB) Viewed 11290 times
117421169-95151a80-af1e-11eb-893a-d9ea0109c26f.jpg
117421169-95151a80-af1e-11eb-893a-d9ea0109c26f.jpg (4.2 MiB) Viewed 11290 times

Overall I can say H72 is a very potent board for miata NB2, allowing a fast and easy install. The minor issues should be corrected in future version.

You can see tunes and logs of the car running H72 here:

https://rusefi.com/online/index.php?vehicleName=H72_CS_NB2&user_id=853&sorts[uploadedMsq]=-1&sorts[uploadedLog]=-1
Attachments
NB2 OEM MAP curve.xlsx
(15.82 KiB) Downloaded 119 times
Post Reply