Source/Vendor/Part # of Maple "External" 18 pin 10.5 mm female stacking header?

LeafLabs Garden » Support » Maple Hardware Support

Source/Vendor/Part # of Maple "External" 18 pin 10.5 mm female stacking header?

(67 posts) (6 voices)

Started 5 years ago by leaflabsandy
Latest reply from gbulmer

« Previous 1 234 5 Next »

gbulmer
Moderator

My Quote: It would add 4 extra connections or I/O pins to the Maple.
Your reply: 3 not 4. VBAT is still VBAT, and is not an I/O pin.

My response: I was indirectly referring to the extra board connections
needed on the Maple.

VBAT (1), for crystal(2), & 1 connection for the osc. output. <----- This is needed for the calibration. Total of 4 connections.

VBAT - agree that would be needed on a header if the RTC is to be used.
Crystal(2) - both are already on the EXT header, though I agree it isn't likely to work well via a header, so I am suggesting them as candidates for removal from the 18pin EXT to make a 16pin EXT, which can be then put on a 0.1" grid.
Osc Output - this is TAMPER-RTC and is pin 1 of the EXT header.
So only VBAT is extra.

Posted 5 years ago #
leaflabsandy
Member

Quote: Have I messed up?
Reply: Nop - I was "assuming" the pinout docs produced by Leaflabs and the silkscreen were correct.

On Eaglecad Pin 18 is ground <----> Leaflabs docs D38 (18) is PC9 ???
Checking this pin with a meter and it is connected to ground???

Okie ... you got some explaining to do.

Posted 5 years ago #
leaflabsandy
Member

That is better yet ... just removing this ground and one I/O pin and moving the
connector north by 0.050 " should fix the external connector "off grid" mistake.

Posted 5 years ago #
leaflabsandy
Member

Someone mysteriously updated the Leaflab's internet Maple GPIO web page
and removed D38 / PC9 / Ground from the External header.
That's what I call magic.

Please note: I copied this same page in this topic link (above) just one day ago.

Posted 5 years ago #
bnewbold
Ex-Leaflabs

Hi Andy, the documentation mistake was mine, not okie's. D38 = PC9 was true for rev1 and I had not updated the pin mapping table for rev3. I have made these changes now, thanks for catching the mistake. I did previously change our "advertising" and made a note in the errata page, not sure why the pin-mapping wasn't updated as well.

Posted 5 years ago #
bnewbold
Ex-Leaflabs

It's not magic; you can view all of our documentation changes here:
http://github.com/leaflabs/maple-ide/commit/b611f8b7fd477429ab5407dcb4c531f1128e8a2b#diff-3

Posted 5 years ago #
leaflabsandy
Member

I can see by connecting digital I/O to D38/GND might cause some smoke!
Make sure you indicate that this external header pin D38 is ground!

BTW ... the magic was just a joke.

Posted 5 years ago #
gbulmer
Moderator

That is better yet ... just removing this ground and one I/O pin and moving the
connector north by 0.050 " should fix the external connector "off grid" mistake.

I am good with that.
Gets to an easy to source socket, on a 0.1" pitch, with minimal PCB changes.

I would add that I do like lots of grounds, it makes it easier to quickly wire things directly to headers.
I'd also still think the RTC Oscillator pins may be the least valuable pins on any header (I don't believe the crystal will work well on a daughterboard, so they are just low-current, non 5V pins)

I ditched them from headers on my board, but I intend to use the RTC on a version. (Yes, I know what you all think of the RTC, but I am a cheapskate :-)

Posted 5 years ago #
leaflabsandy
Member

Gbulmer: Check out the "off the shelf" ChronoDot by Macetech.com. <--- save your money

Posted 5 years ago #
gbulmer
Moderator

A mildly amusing side note ...

I went looking for suppliers of the DS3231 (Digikey were too costly).
I tried searching for "DS3231" at Future Electronics, and the search returned:

STMICROELECTRONICS - M41T Series 5.5 V SOX-18 Serial I2C bus Real Time Clock with battery switchover

Adding:

This part has been identified as a viable alternate to part you are searching for. In most cases it will be a drop in replacement but this is not guaranteed. Please check the technical specification on the datasheet to ensure compatibility before purchasing.

It's $2.93 one-off, much closer to my budget.
So I looked at the datasheet. It seems to be slightly better than the onboard STM32F RTC clock oscillator, but it also uses a similar technology and process to the STM32F RTC clock to quickly calibrate it :-(

It seems that I am not the only one who didn't realise the Maxim part was much more convenient; ST or Future don't either :-)

Posted 5 years ago #
gbulmer
Moderator

Andy, thanks for the link to Macetech.com, that's very helpful of you.

After you suggested the DS3231, I went googling for it, and came across that chronodot at Adafruit. I had assumed LadyAda had done it, but obviously not.

I'd like a low-cost RTC solution on my board.
My ideal solution would be a factory-calibrated, low-cost version of the DS3231 temperature compensated oscillator, in a small package, so I could track that into the STM32F RTC clock input. Predictably, all the ones I found started at a similar price to the DS3231.

If I can ever get around to it, I'll write a clock calibration program for the STM32F and my Mac. That might be an interesting rainy (long) weekend challenge :-)

Posted 5 years ago #
poslathian
LeafLabs

Hey guys,

1) docs issue, pin 38=gnd? - As bnewbold pointed out this was a docs error, the information was correctly reported in the schematics, reflected in the library, and printed on the back of maple! (you did notice the silk on the back of your rev3 didnt you!?). This docs error was our mistake, its fixed now! Thanks for pointing it out. Also, I love the github docs tracking, all magic secrets revealed!

2) 1/10" spacing issue - as suggested by llandy well discussed here we are removing pins 37 and 38 from the 18pin header, and moving the header north. The two pins will still be available on the inside of the 18pin header, without any plastic (just pads). No library changes will be needed. This change will means that Maple Rev4 (which we just got samples back) will be thrown out in favor of Maple Rev5 (which we are no ordering and includes the change). Starting in at some point in september, Rev5 will be what we are distributing. The layout is final - shield away. The arduino mounting holes are preserved. Hooray!

3) the rtc clock issue - Honestly I havnt read through yet. I would like to move this discussion on this thread. Perhaps ill migrate it at somepoint tomorrow. Good stuff though. Seems like RTC keeps coming up...

Posted 5 years ago #
poslathian
LeafLabs

oops! I mispoke there when i described the change. I said we were moving pins 37 and 38 off the header. This is false. What we actually did was move the RTC pins, 21 and 22, off the header. Those pins are now run out to the bottom right open field of the board and exposed as 1/10" pads. Theres enough space in that area to add onboard RTC circuitry one day, but A rtc 37KHZ clock should have no problem running the 1" trace. The important factor with the RTC is that the period is stable, which will not change with length. Since 37KHz is slow, there arnt any big parasitic issues there. Plus we were able to change some of the power planes and introduce some other movement to keep the noise nice and low (as it is). Additionally, the second VCC pin on the bottom left header has been replaced with VDDA, the analog voltage. This is 3.3V by default (same as VCC) however the analog and digital supplies are independent to reduce noise. This new feature allows you to use fancier, higher power supplies on daughterboards and jack them right into VDDA or VCC, for even lower noise operation.

We feel good about these changes, thanks for all the feedback and brainstorming!

Also a plethora of smart silk changes were made.

Posted 5 years ago #
poslathian
LeafLabs

Also to be clearer about the pin 38 issue. Pin 38 IS connected to pc9/the button. The issue is that this pin is not exposed on any header. You can still use pin 38 in your sketches to read from the button.

This is being changed on the silk to reflect that BUTTON=38

Posted 5 years ago #
gbulmer
Moderator

oops! I mispoke there when i described the change. I said we were moving pins 37 and 38 off the header. This is false. What we actually did was move the RTC pins, 21 and 22, off the header. Those pins are now run out to the bottom right open field of the board and exposed as 1/10" pads.

Yeah! Excellent.
Andy's suggestion was a 'stormer' (translation: v, v good). I am very pleased you've managed to implement it.

Sorry to be a bit 'bah humbug', but I'd like to ensure this is clear.
Do you really mean pins 21 and 22 (P$2 and P$3)?

IMHO, it would be better to track both RTC oscillator pins (22 and 23, i.e. P$3 and P$4, OSC32_IN, OSC32_OUT) down to an open area of the board. Connecting a crystal will be even more awkward with them split, one on the header, and one in the corner of the board. Also, the RTC-TAMPER on a header is a handy pin. I realise this might be harder, but I just want one last chance at getting "everything for Christmas, please" :-)

Well done for getting to the magic 0.1" pitch.

Posted 5 years ago #

RSS feed for this topic

« Previous 1 234 5 Next »

Reply »

You must log in to post.