search

travisgoodspeed / goodwatch

Replacement board for Casio Calculator Watches using the CC430F6147
509 stars 55 forks source link

Added u.fl footprint for coax antenna connection, plus a bunch of updates. #104

Closed esden closed 6 years ago

esden commented 6 years ago

This adds a u.fl footprint to allow for connecting a coax cable to the board. Decided to mount it away from the SMD antenna footprint as otherwise it would introduce GND pads under the SMD antenna degrading it's performance. Also this position adds the shortest possible stub to the antenna trace.

edit: This PR also contains a bunch of updates that were discussed in #100

travisgoodspeed commented 6 years ago

Sweet! I'll verify this and merge before manufacturing.

esden commented 6 years ago

The patches as they stand now contain all the changes I have mentioned before.

One more thing I would like to take a stab at is to replace as many of the 0201 parts with 0402 parts as possible. Everything that is on the top side (where the main IC is) can be 0402 because we have the space vertically for the parts. I want to also see if I can move any more parts from the back of the PCB to the front so that we can use the 0402 parts for them.

I don't think we can get rid of all 0201 but decreasing their count might be worth the effort.

esden commented 6 years ago

Ok, I think I have made all the changes that I could do on my own.

I am missing the value for the L1 inductor. Is it only meant to be a ferrite bead? It was the only 0402 part on the back under the battery.

Otherwise I am looking forward to feedback. Let me know if I missed something, or you want me to change anything.

travisgoodspeed commented 6 years ago

L1 is a 1kOhm 0402 ferrite bead, https://www.digikey.com/product-detail/en/murata-electronics-north-america/BLM15HG102SN1D/490-3999-1-ND/1016259

travisgoodspeed commented 6 years ago

BOM.txt will need to be updated for the new component sizes before this can be merged.

UlfWetzker commented 6 years ago

Two weeks ago I read one of the tweets from Travis about the goodwatch and took a look into the kicad files. The idea of adding an u.fl connector float around in my head since I saw this round component on the back of the pcb (is this a second speaker?). u fl_connector Based on some rough measurement within the screenshot it should be possible to fit the connector insight the housing of the watch without modifying the plastic frame. To avoid a stub caused by the u.fl footprint the block cap after the lowpass filter could be used as a jumper. The down side of placing the connector at this position is that the !RST and TST pads have to move to the top side only and therefore giving up the possibility to connect via test clips.

I still have to wait some days for my watch to arrive since I need to take some more accurate measurements.

esden commented 6 years ago

Yes I looked at that position. It would be great to put the u.FL/UMCC connector there but it is quite far away from where the RF output is unfortunately. I do not think this is doable. I would rather find someone with good RF design background to help us design a good PCB trace antenna we could have on the board.

That said, it might be worth the effort to make a test version that only has the UFL connector (get rid of the SMD antenna that can't be used with the watch case anyways) and it is routed to the proposed position, and then see what happens regarding the RF output. I don't think we will know what the impact is until we test it. :)

esden commented 6 years ago

@travisgoodspeed I have updated the BOM.txt. I have also added the information about the non trivial parts to the schematic symbol fields including datasheet links.

I ran across one issue. The 26MHz oscillator is a 12pf part. The load capacitors in the schematic are 19pf. They probably should be at most 12pf and if we want to be super exact even less to take parasitics into account. I usually just go with the value from the part datasheet and this was working fine. The 19pf capacitors are also a rather rare and expensive species so it might be a good idea to use something else that is easier to get. But I might be missing something. :)

travisgoodspeed commented 6 years ago

Probably an oversight on my part. Feel free to lower the cap and choose a more easily available part.

esden commented 6 years ago

@travisgoodspeed thank you! I made the adjustment. Let me know if you have any questions. :)

UlfWetzker commented 6 years ago

By eyeballing the size and position of the hole I did a quick POC (esden/goodwatch#1) to place the u.fl connector insight the watch. I think using C10 as a jumper might improve the rf design by avoiding the stub introduced by the u.fl footprint.

In the past I did some miniaturized antenna array designs (a very long time ago) but I am not an expert in micro strip antenna designs. I think it is hard to design an antenna for the 70cm band with an expectable SWR by using double-sided 0.6mm FR4 and being restricted to less than 200mm². Building a proper wideband (300MHz-1Ghz) antenna sounds impossible to me :/

edit: Maybe an external flex PCB based antenna that sits beneath the wristband?

travisgoodspeed commented 6 years ago

@esden Merging this now. Thanks!

@UlfWetzker A flex PCB against skin is going to get very uncomfortable, very fast. Maybe we should find someone who makes watchbands, then work out how to toss the right shape of metal into the design?

UlfWetzker commented 6 years ago

In my pull request esden/goodwatch#1 @esden suggested that I should place the pull request here. In order to do so I need to perform some measurements to find the right position of the u.fl connector. Could one of you send me some photos of the top and bottom side for both plastic frames (using a desktop scanner and a ruler next to it would be perfect)? Based on this images I could create a drawing and finalize my pull request.