[info] 72 pin 1123038-2 connector board

[AndreyB]

This 72 pin is used on 2001-2005 Mazda Miata oem ECUs

http://www.te.com/usa-en/product-353830-5.html
http://www.te.com/catalog/pn/en/1123038-2





https://svn.code.sf.net/p/rusefi/code/trunk/hardware/1123038-2-connector/

https://www.tinkercad.com/things/j9J40VMtGdJ-rusefi-72-pin-connector-break-box-r01-31
https://www.tinkercad.com/things/aiqxh6EDABo-rusefi-72-pin-connector-break-box-r01-32
https://www.tinkercad.com/things/5c0VGAJTA5I-rusefi-72-pin-connector-break-box-r01-33

[AndreyB]

(2 years 9 days later) Bump. Now we really need this board. We have one 2001+ Miata on the forum & within an hour away from me already, I am seriously shopping for a depressed 2001+ Miata.

From experience one set of holes for each trace would be enough. From experience we need holes for 20 AWG wires - bigger than holes for 0.1" connector.

See also http://rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_2001 for an example how these pins are used.

[stefanst]

I also think these stock ECUs (and with that their cases) are cheaper than the ones we need for the current edition of Frankenso. Added bonus

[AndreyB]

@ has posted a draft, paper test passed.

[stefanst]

If I understand this right, this is an adapter-board for the 72 pin connector and will not fit into our standard ECU housing- correct?
Anyway, it looks much better than what I was planning on doing- just flying wires to the connector. Do you think it's ready to order? I want to get soldering on the Frankenso for the '03 Miata.
I assume the files here https://svn.code.sf.net/p/rusefi/code/trunk/hardware/1123038-2-connector/ are the ones I need to order from Oshpark? Or has there been any further development?

[AndreyB]

I've tried ordering but oshpark has not accepted current version, @ is planning to fix today-tomorrow. I will order right away & post an upadte here. Each oshpark is 3 boards so we can share.

[kb1gtt]

I understand that I selected the wrong layers for the MFG files. Ooops, should be easily fixed tonight when I get home, but I need to remember to do it.

[stefanst]

@kb1gtt: Don't forget now

[AndreyB]

Just ordered https://oshpark.com/shared_projects/22YPoMy2

By the way oshpark has a new 2oz/thin board option at no extra charge now (I've ordered regular 1oz just to get my boards sooner)

[kb1gtt]

Cool about the 2oz. For future orders I would suggest that, as I had to thin up the traces on this. They should be OK, but still the extra copper would decrease the temperature of the traces if you ran 1A continuous through them. At 1oZ I'm expecting a 25C rise in temperature. I generally design with a lower temperature rise.

[AndreyB]

Shipped on October 17th 2016

[stefanst]

Yaaay!
I'll pick mine up in two weeks

[AndreyB]

Looks good

[kb1gtt]

Sweet, and keep me posted if you need anything changed.

[AndreyB]

More silkscreen and I will order the 2oz version? See http://rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_2003 for ideal pinout naming.

[AndreyB]

Got 0.2 boards made in 2oz trim - that's about 20 days with oshpark, comparing with about 10 days for normal 1oz.

[AndreyB]

Miata pad naming makes little sense, but I am pretty sure

https://svn.code.sf.net/p/rusefi/code/trunk/hardware/1123038-2-connector/known_issues.txt

1) what we have as 4Y is actually 4X (MAF in case of Miata)
2) what we have as 4X is actually 4Z (TCM in case of Miata)

The picture on top of http://rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_2003 seems to have it right, validated with real harness wires.

[AndreyB]

I have a problem with random pads being connected to the fill area after I solder harness into connector board. Either it's my soldering technique or lacquer or clearance or else? Need help

Two boards affected - 1oz rev 0.1 and 2oz rev 0.2
Other boards from same batch seem fine, I did not check boards completely before soldering wires but I really suspect the issue is not present with the boards prior to me adding wires. Wires are cut at the time of test, you see a scratch in top-left corner of both pictures which is my test pad.

[AndreyB]

I was able to remove enough solder to get rid of unexpected continuity and here are some pictures of blank boards which probably demonstrate the problem. Is that lacquer offset? Is that something I should complaint to oshpark about? Do we need to increase gaps?

[kb1gtt]

I would say this is a OSH quality issue. It was built to OSH parks specifications, and they missed it by allot. They have an offset on the mask which does not need to be there.

I generally design with a 12:12 trace width and gap goal. I do that because occasionally an 8:8 PCB fab has that tolerance on the X and Y, so if you have a 45 degree, the tolerances could be more. However on this board I pushed that to the MFG claimed 6:6 and I see they missed that by allot on the X and Y. This is not an issue with the 45 degree traces. I could get back to my normal 12:12. However that would force this to be a 4 layer design. Or we can ask OSH to build them close to the spec. I'm pretty sure OSH park would consider this a quality issue on their end.

How hard would it be to add a little mask / nail polish to cover those traces? I seem to recall they make some paint pens that might provide enough of a solder mask to cover these areas. However I do not know the part number.

[kb1gtt]

Oh, that was easy to find. See below CircuitWorks CW3300B - Conformal Coating Overcoat Pen Dispenser, Blue, 4.9 Grams

https://www.google.com/shopping/product/10857752901624037855?q=solder+mask+touch+up+pen&sa=X&ved=0ahUKEwim-bb5jorRAhWGOiYKHez8A3QQkjAIBigB

[AndreyB]

Just emailed oshpart & ordered https://www.amazon.com/gp/product/B0111O7AJM MG Chemicals 419D-P-GR Green Overcoat Pen, 0.17 oz. (4.8g)

[kb1gtt]

Per OSHPark

Are you using KiCad, by any chance? We usually recommend a mask clearance of 3 mils (0.003 inches), but Kicad's default is something like 7 mils, which is ridiculous.

Turn out this board was 0.010 which is way crazy. I have updated the MFG files to a 0.003 in mask per OSHParks. See below before and after pictures. Also note this was PCB layout issue and OSHPark manufactured this properly.

This is the original MFG files. Notice how the blue covers the trace, this shows the mask overlapping the trace which we do not want.

Mask_0.010in.png

This is the R0.2.1 MFG files, which no longer overlaps.

Mask_0.003in.png

Sorry about that, my bad, and WTF KICAD, why don't you have something like 0.003 by default.

[AndreyB]

Can you please post a screen shot of the evil KICAD mask clearance setting dialog so that I can check on all boards next time we mfg?

[kb1gtt]

Should look like this.

[AndreyB]

4) wiring silkscreen is needed on both sides, not only top

[AndreyB]

4) wiring silkscreen is needed on both sides, not only top

Picture shows how silkscreen is on the other side. Half of board on this side is covered by connector but it would be nice to have silkscreen even under connector - that would be useful while assembling.

The main reason why I am looking to reorder - something is happening with these little traces between connector holes and wire holes. I am sure I had all injectors working until yesterday - yesterday I've suddenly lost connectivity in the 2A trace - on the picture you see my ugly shortcut wire. I believe this board is the 2oz version of oshpark, I believe I had similar issues with damaged traces with previous 1os versions. Are these too thing? What else could be the problem? That could not be explained by lacquer mask issue right?

[kb1gtt]

Yes the solder mask issue could be a contributing factor to this trace issue.

[AndreyB]

How?

[Rhinoman]

Has the trace fractured at the pad? it is a thin trace and it is not thickened before it joins to reduce stress.