[OpenTRV-dev] ready to eyeball

[Mike Stirling]

----------------original message-----------------
From: "Damon Hart-Davis" EMAIL ADDRESS HIDDEN 
To: "Mike Stirling" EMAIL ADDRESS HIDDEN 
CC: "Damon Hart-Davis" EMAIL ADDRESS HIDDEN , "Bo Herrmannsen" EMAIL ADDRESS HIDDEN , "Closed list for developer discussions" EMAIL ADDRESS HIDDEN 
Date: Tue, 2 Jul 2013 14:18:50 +0100
-------------------------------------------------
 
 
> Hi,
> 
> Thanks very much for the commentary!
> 
> On 2 Jul 2013, at 14:03, Mike Stirling wrote:
> 
>> Schematic:
>> 
>> - What's with the AVR symbol? It looks kind of non-standard - is it some kind of 
>> weird Arduino version?
>> - AREF should be decoupled
>> - Since you have loads of spare IO, why not connect the RFM22B IRQ pin?
> 
> This is the circuit I have operating, so for good or ill that's exactly what I want 
> on the first PCB!
> 
> For future reference what decoupling would you put on AREF? Note that I can't tie 
> it to any fixed voltage as I switch it between different sources.

100 nF to ground would be fine. The ref is present on that pin even if you are using an internal source, so decoupling is appropriate even if you aren't connecting to an external supply.

> 
> As to the IRQ, there's really no need as polling works fine, but I could make it an 
> option for rev 1; thanks.
> 
>> Board:
>> 
>> - Why not ground flood both sides?
> 
> Is there any harm in not doing so for now? (I was wondering the same myself, 
> actually, but forgot.)

No not really, as long as you via grounds straight through to the other side (I didn't really check that).

> 
>> - A lot of the tracks could be thicker. If there is no need to make them thin then 
>> don't - you will get better board yield. I acknowledge that they have to be thin 
>> for connection to the TMP112, and where you pass between pads on the RFM22B 
>> (which I would have avoided).
> 
> For rev 1 I want to use the net class minimum width trick (though the TMP112 causes 
> annoying warnings), and yes, we could bump everything up that isn't (say) 
> capacitance sensitive I guess. Is there anything likely to care on this circuit. 

Nope.

> I'm going to 
> 
>> - The jumper is underneath IC1 on the same side of the board, so will be 
>> inaccessible.
> 
> Agreed: it will have to be done first. Would like to do without it entirely if Eagle 
> could be convinced not to logically merge the unrelated nets.

They're electrically connected and therefore it isn't logically merging them - they are the same net. As I mentioned in the other email to Bo, the solder blob will prevent the chip from seating properly. You may get around this by using a chip socket, but I'd just move it (or get rid).

> 
>> - You have components on both sides. Not really a problem but it will 
>> complicate assembly if you want to do a run by pick and place (because one side 
>> needs to be glued down). You could argue that the through-hole stuff could go on 
>> the top and the SMT on the bottom, but in that case the RFM22 is still on the wrong 
>> side.
> 
> Yes, will have to think more about that for a larger run, eg rev 1.
> 
> So, aside from all those annoyances/etc, nothing should stop me sending this rev 
> 0 board off then?

No, should be fine. I didn't really spend any time checking the orientation of the USB socket, but Bo has spent so much time staring at that I assume it's right!

> 
> (CCed your comments to the list for posterity and will collect suggestions into a 
> REV1-TODO place...)
> 
> Rgds
> 
> Damon
> 
> 
>> Mike
>> 
>> 
>> ----------------original message-----------------
>> From: "Damon Hart-Davis" EMAIL ADDRESS HIDDEN 
>> To: "Mike Stirling" EMAIL ADDRESS HIDDEN 
>> CC: "Damon Hart-Davis" EMAIL ADDRESS HIDDEN , "Bo Herrmannsen" 
>> EMAIL ADDRESS HIDDEN 
>> Date: Tue, 2 Jul 2013 13:31:58 +0100
>> -------------------------------------------------
>> 
>> 
>>> 
>>> On 2 Jul 2013, at 13:04, Bo Herrmannsen wrote:
>>> 
>>>> Hi,
>>>> 
>>>> Me and damon have just finished the board, but we would like you to
>>>> eyeball it before we send the gerbers off to china :-D
>>>> 
>>>> if anything should have special attention it would be the usb
>>>> connector. fab house have agreed to mill the slots
>>>> 
>>>> damon says to check it all :-D
>>> 
>>> Well whatever you have time for! And "no" is an acceptable answer.
>>> 
>>> Either way I'll try to send it off later today.
>>> 
>>> (I checked in the Gerbers in the "gerbers" sub-dir, FWIW.)
>>> 
>>> Rgds
>>> 
>>> Damon
>>> 
>> 
>> --
>> 
>> 
>> 
> 
> 

--