Hello! from the UK

[RadMx5]

Just want to say hi guys!

I have a Mazda MX5 1999, stock engine 1.8. It will be a drift car in the future, so ill need more BHP -> Turbo...

Already have visited the https://rusefi.com/wiki/index.php?title=Main_Page so I think I know how to start, lots of good info there!
I'm going to buy a second engine then the plan is to rebuild it and try to control everything with rusEfi. I hope rusEfi will help me achieve the goal.

Love this forum and the people here! This is such an amazing source of knowledge.

My English might by sometimes funny as it is my second language

[AndreyB]

Thank you for your https://github.com/rusefi/rusefi/issues/350 contribution - your picture is now on https://rusefi.com//wiki/index.php?title=Manual:Hardware_Frankenso_board#Power_Supply

Did you use Gimp to create it by any chance? Do you have the source file?

[RadMx5]

Not a problem quite enjoyed it!

Yes I have used Gimp. I should still have the source file on a different PC and I can't remember if I saved it without merging the layers...
Ill have a look tomorrow, will upload it here.

[kb1gtt]

Welcome along and pleasure to meet you. Thanks for the picture and thanks for contributing to our documentation.

No worries about the language barrier. My first language is American, my second language is English. However my primary form of communication is miscommunication

I typically comment about the wiki and offer a link, such that new comers can learn from this resource. It would appear you already know about the wiki. So I guess I'll instead ask if you found our Doxygen stuff? Just incase you have not, here's a link.
https://rusefi.com//docs/html/joystick_8cpp.html

[RadMx5]

As I said accidentally I have saved it with merged layers so I don't think you will be able to change it much...

I guess the Doxygen stuff is a description of the functions in the program which controls the STM32 microprocessor?
Also, have a question about how far are you guys with the Frankenso 0.5? Is it tested yet?

I found the Frankenso 0.5 PCB here but was struggling to find the components needed or schematics.
Have visited the: https://www.tindie.com/stores/russian/ and couldn't find anything about the 0.5 version.

I have the equipment needed to solder SMD and good experience with soldering, so all I would need is the PCB, components, and schematics.

Really appreciate your help Thanks !!

[kb1gtt]

The changes from R0.4 to R0.5 are fairly minimal. Mostly the power supply was given more surge protection. This extra protection comes at the expense of a less efficient voltage regulator, so the R0.5 regulator makes a bit more scrap heat, but could handle larger surge events. You can find a list of what has been changed on R0.5 at this link.
https://github.com/rusefi/rusefi/blob/master/hardware/frankenso/known_issues.txt

The big changes are 133) the surge protection, and adding a bunch of via's and some silk screen changes.

This list is basically anything that someone has noticed about the board, and feels is a potential design item. If I hear something, I go add it to the list for consideration.

I think you'll find a purchased assembled board is significantly lower cost, and much faster. There are many many components, and even if you can get each component installed in like 10 seconds each, you'll be soldering for many hours. An assembled board also includes some validations of known weak areas. For example, pin 3 or U1001 is on the ground plane. So it take extra heat to get a good solder joint. If this is not soldered properly it will dump 12V to the 5V rail. We protect the 5V by making sure it works and we install W23 after we have validated the 5V is functioning properly. There are things like that which can be a bugger for a DIY solder.

The R0.4 has allot more validation than the R0.5. I don't think anyone has installed a R0.5 yet, so I don't think it has any significant real world testing. However it's also not much different than the R0.4.

[AndreyB]

I have 0.5 boards for quite some time but assembly goes very slowly (insert any valid excuses here)

https://rusefi.com/forum/viewtopic.php?f=4&t=569&start=390#p26813

To be honest if would be great help if you decide to order a kit & make it run your MX5 - there is a good chance you will run your miata on 0.5 before I do. Just use the 0.4 option and add a comment "0.5 board"

[RadMx5]

Ok, so I'm waiting for parts now. Have ordered the r0.5 PCB from eBay, all the components from farnell.com and the 178764-1 64 - Pin Connector from RS.
All the parts should arrive this week.

Have spent all weekend on rusEFI wiki, rusEFI forum, Youtube, Google, Wikipedia trying to understand what do I need and why...
I have also tried to get the idea of the 1.8 1999 Miata engine schematic to see what needs to be controlled and what can be skipped. I want to make sure I understand everything before I start messing, so here are my maybe silly questions:

1. For the first run do I care about: EGR boost sensor, EGR solenoid, Purge valve solenoid, CDCV - Carnister Drain Cut Valve, TPCV(no idea what it stands for)?
or can I leave them for now, this also tells how many outputs of the frankenso r0.5 I need to use to begin?
2. To be sure to drive the ignition coils do I use four channels of the 8 channel high/low side driver? Frankenso schematic page 2.
3. This engine uses Hall sensors for camshaft and cranksaft, so can I use the op-amp solution to read these signals? Frankenso schematic page 3.
4. Do I need a wideband O2 sensor or can I use the stock one for now?

Before I start soldering, can I use this topic or should I create a new one once I get all the parts, so all the info and problems won't be in Hello World...

[kb1gtt]

1. You can probably ignore. I would be mostly concerned that they could allow air into your intake, but they probably do not. If you can get access to them, perhaps pull the hose and add a cap to plug the hole which would make sure you don't get an unexpected air leak.
2. Which schematic do you have? I don't know if your coil has an integrated igniter or not. Is this your schematic?
https://rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_1999
3. Yes you can use the op-amp trick. This is much easier than soldering the MAX chip. I seem to recall you'll need to use the op-amps with the jumper thing. This allows you to install a small jumper wire. I don't recall this being done on Frankenso recently, so I'm not sure if there are pictures of how to install this jumper wire. This jumper wire is much shorter than it used to be. When the time comes be sure to ask questions, and we would be grateful for a picture that we can put on the wiki.
4. O2 sensors are mostly emissions and long term adjustments and perhaps identification of lean operation caused by a fuel pump which cannot deliver enough fuel. You should be able to tune for max power via butt dyno, and that will typically cause you to run rich. However if your fuel pump is to small, then you may command more fuel and it simply would not make it into the engine. This is where an O2 can be useful, as you could see that you are running lean. In lieu of the butt dyno, the O2 sensor can be used to get your AFR close to peak horse power, or to allow Tuner Studio's auto-tuning to do it's thing. Tuner Studio auto tune will require a wide band. I would say don't worry about the O2 for now, but plan for it long term. The OEM narrow band is next to worthless. The NO2's are really just for emissions. Running rich is OK, and it's common that performance folks will run rich specifically using the fuel to help cool the engine. The two things you are looking for is unexpectedly running lean, and peak HP with out dumping extra fuel.

[RadMx5]

2. Which schematic do you have? I don't know if your coil has an integrated igniter or not. Is this your schematic?
https://rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_1999

Yest this is exactly schematic I was looking at.
The board has arrived, so as soon as I get rest of the parts I will start soldering

[kb1gtt]

Argh, the elusive great white. Be careful it doesn't get you by surprise and sink your ship.

The white looks nice.

[RadMx5]

Ok, so almost all parts have arrived as I made couple mistakes while ordering components... but there is some good progress with soldering.

Power supply: almost done waiting for the right size coils. I have ordered some 63uH coils from TME but they are the wrong size...too big, also ordered the wrong voltage regulator, I ordered the LM2596DSADJG which is an adjustable version, it should be LM2596S-5.0/NOPB, it was late night when I completed the order... I hope the PSU will work

High/Low side driver: Using only U601 and U602 for spark coils, the LEDs light up when doing a spark (Cyl1. Cyl2) bench test.

Op-Amps section: This is prepared for: MAP, IAT, CLT, TPS, VBAT. I will also try to get the: CAM and CRANK pulses from there. Have not tested this section yet.

12 Channel Solenoid/Injection driver: For now only soldered: Q409, Q408, Q411, Q412 for injectors, and Q406 for the fuel pump. I was going to use Q410 for fuel pump but I could not set the PD5 pin to drive this channel, as it looks like the PD5 pin is used by SPI?

All looks promising

Current work progress of the Frankenso r0.5

[kb1gtt]

Looks like Q1001 is cold soldered. It would likely have problems. This is connected to the thermal pads and such and can be a real problem to heat up with a small soldering iron. Any how, just a cautionary note.

Also what happened to Q1002? Looks much smaller than the expected one. If that one works, I'd be interested in learning it's part number. The larger one can dump allot of heat as it clamps the surge voltage. However we probably don't need to dump that much heat.

[RadMx5]

I agree the Q1001 looks like is cold soldered on the attached photo, but it really isn't. I have used a hot air gun to warm up the pads, then a soldering iron to actually solder the big thermal pad first. I have put a little bit too much solder on the pad, and it looks as it looks...

Q1002 I can't remember what I have put there, I'll check it when I get home. I know I was struggling to find the right component which is: DN2625K4-G Datasheet

P.S

Any suggestions for a replacement part just in case?

[kb1gtt]

If Q1001 flows correctly, I expect the excess solder to flow into the thermal via's.

I see several of Q1002 are available here. I tried a couple alternatives, but they did not perform as expected. I do not currently have an alternative for Q1002.
https://octopart.com/search?q=DN2625K4-G

[RadMx5]

Was busy and couldn't reply earlier...
A couple days ago I have received the coils
Q1002 the temporary replacement was a: DN3525N8-G which was not powerful enough as you said, it was only ok for a bench test witch no load. I have replaced it with: IRFR210PBF and it looks like it works fine see the photos. I have tested the frankenso r0.5 power supply with an adjustable current limit bench power supply, just to see it starts up. Nothing got warm so I increased the current limit, and I got nice bright blue LED The current drawn is around 0.08A with the Discovery board.

After quick bench test the outputs I gave it a go in the car last night and... Success well I have only tested spark, and injectors using rusEFI console and they worked.

Tonight I have made connections for the sensors : CLT, IAT, TPS with the 5V and ground references.

Questions:
1. Looks like the MAF sensor is supplied form 12V from this schematic: https://rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_1999 on the frankenso schematic the power to MAF is shown as 5V?
a little bit confused with this one...
2. Tachometer output - does the signal from STM32 must go through one of the 12 Channel driver to the tachometer pin? or directly from the STM32 pin?

Also the fuel pump in the 1.8 1999 EU miatas is connected to pin 3P and for the US to pin 3N. See attached photo.
I have noticed it when I had it connected to pin 3N and I tried to test the fuel pump.
The red floating wire on the photo is the MAF not soldered yet...

[kb1gtt]

I'm not sure about the MAF. If the MAF signal goes up around 12V, anything above 5V will simply read as full scale. If you have a 12V sensor, you may want to scale it down via a resistor divider. Most people are looking for performance, so they add a MAP sensor separately and do not use the MAF. MAP works better for performance, and MAF works better for OEM's who look for simpler tuning and lower cost compliance with regulatory emissions things.

Also I'm not sure about the tach question. Perhaps someone else knows these details. I don't know this car very much.

I don't think IRFR210PBF is proper. It appears it's enhanced mode, while you need a depletion mode.

Looks good and good to hear you have things like injectors working.

[RadMx5]

Will the MPX4250AP MAP sensor be ok? I know people use it with MegaSquirt.
I didn't know that MAP is used instead of MAF, I always thought MAP is used only in turbocharged engines as it only measures the pressure (not a mass of air).

In naturally aspirated engine (yet) there is no pressure in the intake manifold. Of course, the engine will be getting a turbocharger, but first I would like to run the car as it is before any further mods. WIll the MAP be still ok for this test?

I don't think IRFR210PBF is proper. It appears it's enhanced mode, while you need a depletion mode.

I will better find the right transistor then. Thanks!

[kb1gtt]

The MPX4250AP is a fine choice for MAP sensor. As well rusEFI can use either MAF or MAP.

The problem with MAF is that as you increase the velocity of the air flow through the sensor, turbulence will cause errors in the measured readings. As well MAF is very critical about how the plumbing is arranged. Things like tubing diameter, tubing bends, or ripples in an accordion style flex pipe will cause turbulence variations. When you vary these turbulent conditions, your MAF gets extra errors added to it's readings. The errors typically result in lean burns at heavy loads with high RPM's. Of course lean when heavy loaded commonly means less power, it also can result in engine failure. Lean burns commonly lead to detonation as the fuel is not there to pull the heat away. So not only does it result in less HP, it can also destroy your engine. However with MAP, your error condition under heavy loads at high RPM will typically cause you to dump extra fuel. Of course extra fuel is bad for emissions, but good for reliability. It is common that people and OEM's will use MAF for idle, and MAP for full load, with some kind of blending of the signals as you change from idle to full load.

[AndreyB]

Questions:
1. Looks like the MAF sensor is supplied form 12V from this schematic: https://rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_1999 on the frankenso schematic the power to MAF is shown as 5V?
a little bit confused with this one...
2. Tachometer output - does the signal from STM32 must go through one of the 12 Channel driver to the tachometer pin? or directly from the STM32 pin?

While technically MAF is supported, MAP is strongly recommended. On my green car I use the "GM 3bar" sensor from eBay - this case I have the sensor in the engine bay and only a wire goes to ECU.

Tach ouput - not directly via stm32, actually goes via one of the 8 channel hi/low drivers.

https://github.com/rusefi/rusefi/tree/master/firmware/config/engines would provide you with some ideas especially https://github.com/rusefi/rusefi/blob/master/firmware/config/engines/mazda_miata_nb.cpp https://github.com/rusefi/rusefi/blob/master/firmware/config/engines/mazda_miata_vvt.cpp

[RadMx5]

I'm assuming that the tachometer output must be 5V not 12V?

[RadMx5]

I have configured the tachometer high side driver to 12V and it works.

The current look of the frankenso r0.5:

IMG_1532.JPG

IMG_1531.JPG

There is another topic about this project: https://rusefi.com/forum/viewtopic.php?f=3&t=1282&p=27348#p27348

[AndreyB]

I am slowly catching up with my assembled 0.5 boards.

Can you please confirm that your self-assembled board works for 10+ minutes without a reboot while powered by 12 volts? We are investigating https://rusefi.com/forum/viewtopic.php?f=4&t=394&p=27976#p27974 nevermind, it was R1003 missing - problem solved.