View unanswered posts | View active topics It is currently Sat Nov 25, 2017 10:34 am



Reply to topic  [ 26 posts ]  Go to page 1, 2, 3  Next
Karri's "Wildcat" EMS for ModICE enclosure 
Author Message
LQFP112 - Up with the play
User avatar

Joined: Tue Jan 15, 2008 5:31 pm
Posts: 166
After a lengthy hiatus from most car-related activities (post-layoff blues and weariness and lack of funds with the engine build) I've decided to properly take on a FreeEMS HW implementation. To keep the feline themed naming, I though to use simple "Wildcat", googling "wildcat engine management system" only yielded hits on some snowmobile named Wildcat so I think it's fine.

The basis of my version is the rugged ModICE enclosure, protected up to IP69k rating; pressure washer proof. It would enable the use of the device in harsh environments like boats, motorcycles, snowmobiles and other such vehicles. Pin budget of 60 should be sufficient, but I've queried Cinch about the specific 112 pin ECU header faceplate for availability in hobbyist batch sizes; such a header would allow the board to have all the features modern OEM control units have and then some. The enclosure is sealed, so all connections must be made through the faceplate connector; also air pressure sensors must be external, including ambient absolute pressure.

Unfortunately, many of the goals are counterintuitive with the DIY spirit. Since the enclosure is tamper proof too, so once the unit has been assembled, it won't be easy to take apart for modifications. Because you can't really expand much the capabilities later on, flexibility needs to be thought of in the design. The PCB will be a 4-layer, as it is desirable in this case for their superior noise immunity due to the uninterrupted ground and power planes and better routing possibilities for top and bottom layer signal layers, which can also use thicker traces. The downside is 4 layer boards can't be made at home, but they have become much more affordable to obtain in small quantities recently. Preferred choice of components is surface mount, through hole are to be used only when there is no alternative; notable example being the connector itself and the TO-220 parts. Component sizes are selected so they can be soldered manually though.

Here's my initial list of features:

Output:
  • Twelve high power (IGBT) output drivers, used as follows:
    • Six precision drivers for injection and ignition, four example configurations:
      • 4 cylinders: 4 injectors, 2 ignition channels
      • 6 cylinders: 3 injector pairs, 3 ignition channels
      • 8 cylinders: Bank injection, 4 ignition channels
      • Bank/batch injection and distributor ignition for more cylinders (injection load spread over several pins)
    • 2 high current outputs for fuel pump and ignition/injection relay
    • 4 PWM outputs with 16 bit precision (boost controller, power steering speed data, PWM-driven BAC, EGR, etc)
  • General purpose moderate current outputs (for LED indicators or external relay drivers)
  • Malfunction Indicator / Check Engine light
  • Shift cue light
  • tach output
(Pin budget check: 15 defined pin functions, several undefined)

Input:
  • Pulse:
    • G and NE signals (opto and hall, no VR conditioning)
    • Vehicle Speed Sensor
  • Analog:
    • Intake air temp
    • Engine coolant temp / cylinder head temp
    • Manifold Absolute Pressure (external sensor)
    • Ambient Absolute Pressure (external sensor)
    • EGT (dual channel for the benefit of V engine builders)
    • EGO2 (narrow/wide band single channel)
    • Flex fuel sensor
    • TPS
    • Knock (if not pre-conditioned)
    • Any good cases for extra analog inputs?
  • Binary:
    • Map/Mode select; Street/Track power etc
    • Extra load request (higher power demand from alternator, eg, when heaters, blowers, A/C is on; increases idle speed)
    • User definable switches, internally debounced (eg launch control, datalog start/stop, datalog mark)
    • Spark event detection (some ignition modules support this feature, perhaps expandable to ionic sensing once that's figured out)

Pin budget update: 16 defined plus several undefined; (31 defined + several undefined; one header plug furnished.)

Data IO:
  • USB
  • CAN (hoping for OBD-II output code!)
  • ...Bluetooth? (there are OBD2-to-BT converters to be used with Android cockpit software.)

Pin budget update: 6 (37 defined, several undefined)

Electrical:
  • 12V continuous
  • Device ground (three pins)
  • High-current (MOSFET/IGBT) ground (six pins)
  • Return ground for sensors
  • +5V for 3-wire sensors

Pin budget: 12 (49 defined)

The final tally for undefined function would be 11, which could used for GP switch inputs, GP analog inputs and GP transistor outputs; number of each should be decided.


Thu Nov 28, 2013 7:58 pm
Profile
LQFP112 - Up with the play
User avatar

Joined: Tue Jan 15, 2008 5:31 pm
Posts: 166
Hardware layout

You don't design the board before doing the schematics, but the two do go somewhat hand in hand, and here's my take on the PCB layout. Top layer is for the really sensitive signals and most components, and is zoned for each function group; CPU is at the signal hub. The second layer is the ground plane which is split at the 12v area, the third layer is the power plane which is split into star-ish power distribution pattern. The bottom layer is for control signals, filter circuits and decoupling capacitors. Between the header and component zones is an area where signals from top and bottom layer are distributed over all four layers to help with routing the header.

I'll start the schematic design with the IGBT drives. I want to use one design for all high current outputs, and the IGBT needs to be pretty hefty. International Rectifier has a nice logic level IGBT capable of handling 125w and 20amps, I think it'll suffice. It's part number is IRGB14C40LPBF; here's the product page from digikey:

http://www.digikey.com/product-detail/e ... -ND/812232

So... better get to it.


Attachments:
File comment: Gnarly sketch of the PCB layout, Q&D department in action!
Wildcat-layout.jpg
Wildcat-layout.jpg [ 114.89 KiB | Viewed 2026 times ]
Thu Nov 28, 2013 9:43 pm
Profile
LQFP144 - On Top Of The Game
User avatar

Joined: Tue Mar 02, 2010 2:33 pm
Posts: 1201
Location: Australia
Interesting enclosure. . . . Reminds me of something. . .

What software package will you be using?

Also, you are correct, my experience with that enclosure is that the tabs do break if you open up the case often.

Nice ideas though. Code doesn't support your inj/ign configuration yet and I think you should include vr as the circuit for it doesn't consume that much PCB real estate.

You also need to address how map hoses etc are going to be fitted, unless you are going to support external map only.


Thu Nov 28, 2013 11:08 pm
Profile
Moderator
User avatar

Joined: Tue Jan 15, 2008 2:31 pm
Posts: 14880
Location: Home sweet home!
There's more wrong too, but I don't have time to reply. Hell, I didn't even read all of it.

_________________
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!


Thu Nov 28, 2013 11:31 pm
Profile WWW
LQFP144 - On Top Of The Game
User avatar

Joined: Thu Feb 25, 2010 3:57 am
Posts: 563
Location: Gosport, IN USA
I think a good place to start would be to define what the goals of your project are, ie: what type of injectors you are wanting to support (low impedance or high impedance), what type of ignition systems you want to support, etc instead of stating right from the start what I/O configurations (ie: IGBT for injectors and Ignition outputs).

I would advise you to look at the Ravage https://github.com/dvisser/Ravage and Jaguar https://github.com/DeuceEFI/Jaguar projects as both schematics have many of the circuits you propose already layed out and tested on running engines. If you look at the Jaguar repository make sure you are in the "dev" branch as it has the latest improvements based on feedback from many FreeEMS testers.

Take a look at what the Jaguar PCB currently supports http://forum.diyefi.org/viewtopic.php?f=67&t=2212. This thread also lists the current limitations of the freeems-vanilla firmware.

ST Microelectronics has a nice line of logic level driven MOSFET low side drivers that range from 6A to 20A, I used the VND7NV04 6A D-PAK version on the Jaguar, but you could use the 12A D-PAK version since it is the same package size. The larger rated 20A version is a larger D2-PAK package. These are in the VND series of SMD MOSFETs. I'm just curious why you want to have outputs rated for 20A?

Just let me know if you have any questions about what has been done on the Jaguar schematics.

_________________
Andy.
FreeEMS vehicle #11, 1932 Ford 5 Window Coupe with a 1996 GM 3.1L SFI V6 with DIS ignition
FreeEMS vehicle #16, 1996 Chevrolet S10 2.2L SFI I4 with DIS ignition
http://www.coolefi.com


Fri Nov 29, 2013 1:39 am
Profile WWW
Moderator
User avatar

Joined: Tue Jan 15, 2008 2:31 pm
Posts: 14880
Location: Home sweet home!
IGBTs are not appropriate for injectors, at all.

_________________
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!


Fri Nov 29, 2013 2:10 am
Profile WWW
QFP80 - Contributor

Joined: Tue Feb 12, 2013 5:43 am
Posts: 57
Location: Scarborough, ON
The Cinch ME enclosure is easily opened with hand tools (popsicle sticks ...), but has the same rugged construction. It is unfortunately much smaller and is limited to 48 pins. I am in the process of building a MicroSquirt Module-based prototype for a fellow, and it's a nice little box.

There IS a 'blank' header / faceplate available, to allow other connection options.

Remind me - Is there CAN communications for this design yet? You might consider running all the sensors (except cam and crank), plus some of the aux outputs to one box, with the primary box having the main processor with the cam and crank, spark and injection IO.

_________________
TTR Ignition Systems


Fri Nov 29, 2013 2:40 am
Profile WWW
Moderator
User avatar

Joined: Tue Jan 15, 2008 2:31 pm
Posts: 14880
Location: Home sweet home!
No CAN yet and that sounds like a horrible idea for most key sensors. OK for auxiliary stuff, though.

_________________
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!


Fri Nov 29, 2013 3:15 am
Profile WWW
QFP80 - Contributor

Joined: Tue Feb 12, 2013 5:43 am
Posts: 57
Location: Scarborough, ON
Which sensors change too fast for CAN? TPS, MAP, ...?

Another point to ponder with the ME is it has more useable headroom than the LE or SE, perhaps allowing stacked boards.

_________________
TTR Ignition Systems


Fri Nov 29, 2013 4:11 am
Profile WWW
LQFP112 - Up with the play
User avatar

Joined: Tue Jan 15, 2008 5:31 pm
Posts: 166
Ah, so there was to summary of the available features. And as I thought, I have a couple of feature ideas without firmware support (yet?).

The blank faceplate option was explored, turns out for all of the options headers (that I found) overlapped with the PCB holders on the faceplate, so that didn't turn out a solution to the pin problem.

I forgot the link for the enclosure from the start post (or it might've gotten lost during the edits, I edited it a lot.). Here's the manufacturer page:

http://www.cinch.com/products/modular-i ... -enclosure

The enclosure sets a number of limits. First, beside many other things, it's great at capturing heat too. I've selected a version which has heat sinks for twelve TO220 components. SMC transistors are fine where there's no significant current to be expected (like signal lights), but more high-current prone stuff will likely have use for the heat sinks. There are no provisions for tubing in the enclosure, so internal pressure sensors are not possible without compromising protection rating, which largely voids one of the key benefit.

Injectors will be high impedance. I wanted to do low-z earlier, but since Bosch introduced their EV series they've lost a lot of their older glory, and so I instead opted for a simpler circuit. Igniters are to be external too. Neither actually need the huge current rating, but it's there for ruggedness. About IGBT's not being appropriate for injectors, I'd like a bit more than an assertion on that. They're used in many solenoid-based devices, what makes them unsuitable to drive injectors?

The reasons I don't plan to include the VR sensors is because I also want to have a VSS, even though there's no support for it in the code (yet anyway). VSS sensors are normally either Hall or optic, since VR has problems tracking very low rotation speeds. Thus, I would need to include a separate buffer circuit for that sensor anyway; keeping the selection in just Hall or Optic, just one chip can handle all the channels, while adding a VR conditioning chip would increase individual part count (while I'm pretty sure this device will never be produced by pick and place machines, I'm developing it as if it was), create more tracks to route and use up two pins in the already crowded header. However an external unit would be quite possible.


Fri Nov 29, 2013 5:21 am
Profile
Display posts from previous:  Sort by  
Reply to topic   [ 26 posts ]  Go to page 1, 2, 3  Next

Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group
Designed by ST Software for PTF. ColorizeIt.