Page 6 of 13

Re: microRusEfi

Posted: Tue Sep 10, 2019 2:00 pm
by Ahmad
DonaldBecker wrote:
Sat Sep 07, 2019 3:42 am
kb1gtt wrote:
Sat Sep 07, 2019 1:43 am
See reverse polarity pmosfet noted at the below link. The pmosfet is typically better that a protection diode. https://rusefi.com/wiki/index.php?title=Manual:Hardware
In this case it's only negligibly better. MRE draws only a trivial current directly from IGN (BAT, BATPA, BATPB). Probably single digit mA when activating the main relay, and almost nothing after. The rest of the current goes through a series of linear regulators. Except during a severe voltage dip when cranking, every mW burnt by the protection diode is a mW that doesn't need to be burnt in the 6V pre-regulator transistor.

It would be different if we were using something like the TLE9471, which has a buck regulator for the 3.3V or 5V supply. There the efficiency of using a MOSFET instead of a diode is worthwhile, and the chip includes a charge pump to enable using a lower-cost N-channel device.
Next generation of All-in-one ICs is L9788.
it have Boost_Buck regulator that reduce power consumption and work well with low battery voltages.

Re: microRusEfi

Posted: Tue Sep 10, 2019 3:19 pm
by jbiplane
stefanst wrote:
Tue Sep 10, 2019 12:34 am
One semi-mystery is solved. The screws holding the housing bottom to the top do seem to be M3.5 x 0.6. Because M3s just won't cut it and M4s are way too big. I guess.
Have a look at drawings of ECU cases produced by https://kinkong-group.en.alibaba.com/?spm=a2700.details.cordpanyb.1.5a4c2693oGwJ0b
48 pin case, seems the same what used for this project?
CKKB48-1-A.pdf
(421.01 KiB) Downloaded 546 times
I guess holes should be calibrated by drill 3.3 (through hole) and tapped by M4 (M4x0.7)

For reference cases 32 and 24 pins
CKKB32-1-A.pdf
(413.51 KiB) Downloaded 529 times
CKK24-1-B.pdf
(26.87 KiB) Downloaded 545 times

Re: microRusEfi

Posted: Tue Sep 10, 2019 3:22 pm
by AndreyB
Just a friendly reminder that CKKB48-1-A.pdf is available in the top level folder of https://github.com/rusefi/hw_microRusEfi

Re: microRusEfi

Posted: Sun Sep 15, 2019 5:12 pm
by jbiplane
Encountered minor error in connector area of PCB.
The hole pitch = 2.5, not 2.54mm
PCB correction.jpg
PCB correction.jpg (86.95 KiB) Viewed 34147 times
PCB correction 1.jpg
PCB correction 1.jpg (63.5 KiB) Viewed 34147 times

Re: microRusEfi

Posted: Sun Sep 15, 2019 8:50 pm
by DonaldBecker
Where is that footprint from?

I'm seeing the same from
https://media.digikey.com/pdf/Data%20Sheets/Molex%20PDFs/500762-048x_Dwg.pdf

Your correction (2.5 mm pin spacing rather than 0.1 inch) is likely accurate. Could we get a confirmation from someone with a connector physically at hand? Ideally with a current version of the board where it's bad fit.

Edit: I updated the connector footprint on my branch of the board layout. This revision also consolidates options onto the bottom layer. The bottom now contains jumpers, option selection, optional components, and variable components e.g. pull-downs, VR biasing. The remaining bottom components are steering/protection diodes and a few redundant capacitors.

Re: microRusEfi

Posted: Sun Sep 15, 2019 11:07 pm
by AndreyB
we have fabricated at least one board of each known revision and if there is a footprint error, it's so minor that we were not affected in any way. header snaps into pcb and pcb goes into case.

Re: microRusEfi

Posted: Mon Sep 16, 2019 12:26 am
by stefanst
2.54mm vs 2.5mm over ten columns is 0.4mm. Hardly noticeable. So the footprint may be wrong and possibly should be fixed in future revisions, but it hardly matters in real life.

Re: microRusEfi

Posted: Tue Sep 17, 2019 11:46 am
by Ahmad
Is it possible to show functions of pins on schematic or an excel file or cubeMX file?
For example, PC6 may be PC6_TIM3_CH1.
Or a cubeMX file same as below:

Re: microRusEfi

Posted: Tue Sep 17, 2019 11:49 am
by AndreyB
normsl end users probably would not benefit, advanced users can figure it out

if you believe irs really useful please makd ghe list and i will publish

Re: microRusEfi

Posted: Tue Sep 17, 2019 11:52 am
by Ahmad
Sure, I will do.
but i don't know function of some pins.
for example: PA5, PC7, ...

Re: microRusEfi

Posted: Tue Sep 17, 2019 12:30 pm
by DonaldBecker
stefanst wrote:
Mon Sep 16, 2019 12:26 am
2.54mm vs 2.5mm over ten columns is 0.4mm. Hardly noticeable. So the footprint may be wrong and possibly should be fixed in future revisions, but it hardly matters in real life.
The error isn't split centrally. The existing footprint starts at the correct position on one side, and has the most error on the side with 8 wide contact pins. The tight fit is between the wide pins and the nearby alignment post.

The footprint mismatch looked pretty bad when I was changing my board layout, but I know that's misleading. Through hole footprints are sometimes deliberately misaligned to hold the part in place, and those footprints look like a wild zig-zag when you are zoomed to route traces.

Re: microRusEfi

Posted: Tue Sep 17, 2019 2:10 pm
by jbiplane
DonaldBecker wrote:
Tue Sep 17, 2019 12:30 pm
The error isn't split centrally.
I corrected PCB and case bottom part according chinese drawings. Fixing holes moved up to 0.45mm.
Cannot drop STEP file here. Can correct upper part of case as well (according drawings).
When I get real case I can scan and reconstruct it with precision of 0.02mm.
Corrected.JPG
Corrected.JPG (220.19 KiB) Viewed 34055 times

Re: microRusEfi

Posted: Tue Sep 17, 2019 4:31 pm
by kb1gtt
Scanned would be cool. The model we have now is just my calipers, and eyeball. I did not include the casting tapers. So it's not 100% accurate, but it's fairly close.

Re: microRusEfi

Posted: Tue Sep 17, 2019 10:00 pm
by AndreyB
jbiplane wrote:
Tue Sep 17, 2019 2:10 pm
Cannot drop STEP file here.
Please attach as .zip file

Re: microRusEfi

Posted: Wed Sep 18, 2019 2:37 pm
by jbiplane
russian wrote:
Tue Sep 17, 2019 10:00 pm
Please attach as .zip file
Attached. PCB connector and bottom part together. Corrected by factory drawings.

Re: microRusEfi

Posted: Thu Sep 19, 2019 2:18 am
by AndreyB

Re: microRusEfi

Posted: Thu Sep 19, 2019 6:08 am
by puff
a giant banana!

Re: microRusEfi

Posted: Thu Sep 19, 2019 10:07 am
by kb1gtt
I'm missing something here. Is the claim that the factory provided 3D model does not match the factory provided drawing? The model we have is from TE.

If the concern is that the PCB model is slightly wrong, that's driven by KICAD, and the KICAD lib should be updated, then that should drive the 3D file.

Re: microRusEfi

Posted: Fri Sep 20, 2019 11:41 pm
by kb1gtt
I miss spoke. It was not TE but Molex. I believe the 3D came from here.
https://www.molex.com/molex/products/datasheet.jsp?part=active/5007620481_PCB_HEADERS.xml

I understand the china component may have slight differences from the OEM Molex part.

Re: microRusEfi

Posted: Wed Oct 02, 2019 8:43 pm
by Andreas_28023
I am about to start a project on an Opel CIH 2.2 injector engine. I am trying to decide wether to wait for µRusEfi or go for the Frankenso kit. I don't really have a timeschedule on this but eventually I aim to put the engine in a Manta A. Do you have any suggestions?

//Andreas

Re: microRusEfi

Posted: Wed Oct 02, 2019 10:51 pm
by AndreyB
if you do not care for size, looks or ETB - go with Frankenso

Re: microRusEfi

Posted: Fri Oct 04, 2019 9:28 am
by Andreas_28023
Well... I do care about size and looks. How far from a release are you?

Re: microRusEfi

Posted: Fri Oct 04, 2019 12:16 pm
by AndreyB
Andreas_28023 wrote:
Fri Oct 04, 2019 9:28 am
Well... I do care about size and looks. How far from a release are you?
There is no budget and no planning in this project. Nothing was ever "released" in rusEfi. The only things you can rely on are the things you either see exist already or the things you will make yourself.

Re: microRusEfi

Posted: Sat Oct 05, 2019 2:07 am
by mck1117
Three microRusEfi v0.3 boards are done! All started and ran my car great, and seem to behave identical to the v0.1 board I've been running.

Image

Re: microRusEfi

Posted: Thu Oct 10, 2019 6:40 pm
by AndreyB
Very special orange pre-production unit available for sale at https://www.ebay.com/itm/333356685951

Re: microRusEfi

Posted: Wed Oct 16, 2019 8:29 am
by AndreyB
@DonaldBecker is trying JCBPCB pre-assembly feature with his 2-layer improved MRE

Only available in green :(

Re: microRusEfi

Posted: Wed Oct 16, 2019 1:36 pm
by DonaldBecker
I had a long description composed that was just deleted by a new kitten. Doh!

The layout includes every listed errata and suggestion except for the SPI connection to the TLE9402 e-throttle. That would involve a change in the STM32 pin function assignment.

Notably reverse polarity protection is included. Most of the board is protected by diodes (D36 and D37, plus D1), while the high-current TLE9402 is protected by the main relay. To use that protection you need separate 12V-IGN (key signal) and 12V-relay (12V output of the main relay) power inputs. Previously those two pins were internally connected.

Note that the MRE only activates the main relay coil when the battery polarity is correct, and reverse voltage protection requires that injectors and most other engine devices are powered by through the main relay. This is the typical configuration, but it's worth repeating.

The pictures above are the boards as assembled by JLCPCB, with only the USB connector and programming header added by me.

They still need the TLE8888 (main 100 pin chip), TLE9402 (e-throttle driver) and pre-regulator MOSFET added. These are specialty chips not available through JLC. As of October 2019 the TLE8888 is only available from Mouser or directly from Infineon.

I redesigned several times to make it more suitable to the part availability and assembly limitations. The goal was to get back a board that was ready to partially test, and required relatively little tedious work to complete. My estimate is that assembly takes care of about 85% of the solder joints, and well over 95% of the component placement effort.

I'll cover the other changes (standby power, wake-on-CAN, clock support, headers) in a future post if there is interest.

Re: microRusEfi

Posted: Sat Oct 19, 2019 9:29 pm
by atntpt
Are you willing to share the files of the modified board , SMT placement and bom ? I wanted to tryout myself :)

Re: microRusEfi

Posted: Sun Oct 20, 2019 12:41 am
by kneelo
Lots of really great progress... Wish I had some time to play along at home!

Re: microRusEfi

Posted: Sun Oct 20, 2019 1:16 am
by DonaldBecker
atntpt wrote:
Sat Oct 19, 2019 9:29 pm
Are you willing to share the files of the modified board , SMT placement and bom ? I wanted to tryout myself :)
I think that most of the files have been integrated into the master branch.

I thought that I had added the fabrication ZIP file to my repository at https://github.com/DonaldBecker/hw_microRusEfi/ but I'm not seeing them there. I'll check on that.

If you are considering having a batch fabricated, let me know first. I have a small set of minor tweaks. Most are silkscreen changes, but I also want to change the USB connector to a different type that is easier to hand solder and inspect. And perhaps change the pushbuttons to a type that has legs that splay out rather than tuck under, again to make them easier to hand solder.

Boards that are to be hand soldered need to be designed with that in mind. Aligned side-by-side components with easy access to the components ends is far easier than unaligned parts and 'T' positioning. That's easy to think through with some parts, but I didn't consider it with the USB connector. The specific part used has the leads tucked under the metal shell and I made it even more of a challenge with the nearby protection diode. It's not especially close, but it still mean using a narrow soldering iron tip that had trouble transferring enough heat.

I think that this type might be easier to use (and less expensive and available from more sources)
https://www.aliexpress.com/item/32633184095.html?spm=a2g0o.productlist.0.0.1c6621bbO7Vbup&algo_pvid=208a3adc-0f89-433f-9df7-cbc973fa2944&algo_expid=208a3adc-0f89-433f-9df7-cbc973fa2944-3&btsid=3a00b98e-885f-4761-9949-6f562d7bab29&ws_ab_test=searchweb0_0,searchweb201602_7,searchweb201603_52

The pushbuttons aren't especially bad, and I had a few in my supply, but this type might be easier to solder:

https://www.aliexpress.com/item/32696590759.html?spm=a2g0o.productlist.0.0.12835a82H4qLLE&algo_pvid=030dc37d-0998-4ce4-aa2f-855da014de75&algo_expid=030dc37d-0998-4ce4-aa2f-855da014de75-7&btsid=f5eb8abf-a375-48db-a8b0-c8d1b640a151&ws_ab_test=searchweb0_0,searchweb201602_7,searchweb201603_52

Apologies to future readers when these links are no longer valid.