RE: Strobit PCB
Stephen Eaton
widget discussion group, that way we can get further feedback.
> -----Original Message-----
> From: Konstantin [mailto:beowul...@gmail.com]
> Sent: Thursday, 21 May 2009 4:07 AM
> To: Stephen Eaton
> Subject: Re: Strobit PCB
>
> Done assembling the board. First impressions...
>
> It's working! I was able to program the controller and light the LEDs.
Excellent, yours is the first one in the wild that I know of :)
>
> It was a pain in the butt. 0805 components would have been
> much easier than 0603. I realize you're a die-hard
> electronics engineer and use oven to solder the boards, but
> us, simple types, do it with a plain old soldering iron :)
>
Lol not a diehard, but I just find it easier do all my smd either via a
hotplate/skillet or toaster oven. Actually now I find that SMD is so much
easier than PTH, all my soldering is done once. You don't really need any
special gear, a hotplate is very cheap and does the job, I hacked together a
cheap toaser oven and that more than surfices for what I want. I agree
hand soldering SMD does suck, no matter what the size lol
Changing/adding the 0805 will not be a biggy I may need to shuffle a couple
of things around.
> The SMD quartz is definitely overkill - I was unable to find
> it anywhere. Nor was I able to find the switch and the SMA
> connector, but I decided to not use them. But with the quartz
> I had to solder a regular through-hole quartz to SMD pads :(
I totally agree, it was the first thing I found and had put an issue in to
change it once I built my first board.
> I see no benefit in using those hard to find components.
> While almost anything can be ordered online, I really prefer
> to buy stuff in local shops. Making everything SMD would make
> sense for totally automatic assembly, but there are SIL
> connectors that would need to be hand-soldered anyway.
The switch is available via sparkfun, I've just listed the manufacturer
part#. And the quartz are available via mouser, I will put distributor
part# up on the wiki for some of the components, most of distributor part# I
have are for my australian distributor, X-ON (who is also the Mouse
distribtor in australia). I would like others around the world, like
yourself to let me know where you purchased your parts and their respective
supplier part# so we can put them online for others, I could possible look
at doing a kit of components if the need is there as I purchase most of my
components by the reel now and I can get the benefits of bulk purchases cost
which can be passed on.
I know that PTH components will make it easier for the DIY, however I will
be sticking with SMD for the time being wherever possible, it's a lot
cheaper and easier in the long run. Having said that, the schematics are
online and it would be good if someone could contribute a PTH design for
those wanting it.
Unfortunately I can't purchase anything locally, I live outside of all the
major cities so all of my ordering is online, so I'm used to waiting a week
or two for things to turn up, actually while I was waiting for my components
for this project to arrive I salvaged pretty well all of my SMD discretes
from an old motherboard.
>
>
> My suggestions for the next revision of the board would be:
>
> 1. 0805 components
It will be easy to change to the 0805 design, I may have both designs
running parallel, as 0603 will be better for manufacture on a larger scale.
But agree with your points.
> 2. Leaded quartz
I've been thinking of a couple of options for the quartz:
1) I had initially thought to use resonators, but for the cost and accuracy
I'll stick with quartz.
2) I am definitely looking at changing the quartz footprint, but may still
leave it as SMD, just standard HC-49 smd footprint, I don't need the
temperature stability of the ceramics, I originally designed them with
ceramics for the stability over a wide temperature range, as I also wanted
to uses these boards as a low cost WSN Node, however if I need the
termperature stability I can use the S2 version of the RFM12/B SMD package
which uses a ceramic quartz anyway and implement #3 below.
3) I'm also thinking of bringing the onboard RFM12/B external clock back to
the CPU and have it available via a solder jumper, that way I can get a
0-10Mhz clock without any extra components, anything more can be done as a
quartz?
> 3. Space for a 10K resistor to accomodate RFM12 and not only
> RFM12B
When you first emailed and told me that you're looking at using the RFM12 I
had made a mental note about adding in the resistor then, at least it will
give some aditional flexability.
>4. Space for voltage regulator in SOT-23 (or better
> TO-92) package in external power supply line (near RS-232 port).
Yes I blogged about the decision of putting in a LDO regulator here
http://blog.everythingrobotics.com/2009/05/06/pcb-arrived-and-hot-out-of-the
-oven/, I'm inclined to agree and considering it for the next board
revision.
> 5. Clear indication of part names on the board ("R1", "C1",
> etc) to make hand-soldering easier.
I've found that having an oversized placement drawing, like the one I have
on the wiki, I did not have any problems identifing what goes where, it
makes for a cleaner board. And really there are a not a lot of components
on the board, might be different if the board was faily dense. I will look
at your suggestion and have a play around when I'm making the next lot of
updates.
>
> Also, I think that vias under the RFM12 are not covered with
> mask. Or rather, some are and some are not. I am afraid these
> vias may short with the RF module. A friend of mine who works
> in an electronics manufacturing facility says that that the
> mask surface on your boards is not very regular which is a
> manufacturing fault. Maybe you can complain to the guys who
> produced the boards and they will make it better next time.
> To me it looks OK, though.
The PCB are a prototype deal through GoldPheonix so I expect that the
quality will not be the same as the end finished boards, once the design has
been finaliased then I'll be getting proper boards done.
>
>
> Keep up the good work! By the way, what are the other guys
> saying? Has anybody shared their impressions about the boards too?
>
You are the first to complete but I'm grateful for the feedback, it's what
I'm after. I really need the boards to be put through their paces, i.e. ADC
test, D/O tests.
If you or others find any problems please add them as issues or enhancements
on the project home page.
http://code.google.com/p/strobit/issues/list
Any other questions/suggestion please put them out on the stobit general
discussion as more people can at their 10c worth as well.
Regards,
Stephen
__________ Information from ESET Smart Security, version of virus signature
database 4092 (20090520) __________
The message was checked by ESET Smart Security.
Konstantin
> I hope you don't mind but I will crosspost this to the strobit wireless
> widget discussion group, that way we can get further feedback.
No, not at all :)
>> 2. Leaded quartz
> I've been thinking of a couple of options for the quartz:
> 1) I had initially thought to use resonators, but for the cost and accuracy
> I'll stick with quartz.
> 2) I am definitely looking at changing the quartz footprint, but may still
> leave it as SMD, just standard HC-49 smd footprint
HC-49/SM is definitely better that the one you have now, but I have
one more suggestion. Since the size of the HC-49/SM is the same as
HC-49/U (leaded) maybe you can make a special footprint that would
accomodate both? Like having pads big enough to be able to solder SMD
but also with holes for through-mount?
> 3) I'm also thinking of bringing the onboard RFM12/B external clock back to
> the CPU and have it available via a solder jumper, that way I can get a
> 0-10Mhz clock without any extra components, anything more can be done as a
> quartz?
Isn't that a big no-no? Like that HopeRF reported they had a major
problem with clock in their units?
Stephen Eaton
> >> 2. Leaded quartz
> > I've been thinking of a couple of options for the quartz:
> > 1) I had initially thought to use resonators, but for the cost and
> > accuracy I'll stick with quartz.
> > 2) I am definitely looking at changing the quartz
> footprint, but may
> > still leave it as SMD, just standard HC-49 smd footprint
>
> HC-49/SM is definitely better that the one you have now, but
> I have one more suggestion. Since the size of the HC-49/SM is
> the same as HC-49/U (leaded) maybe you can make a special
> footprint that would accomodate both? Like having pads big
> enough to be able to solder SMD but also with holes for through-mount?
Sounds like a good idea, best of both worlds. Will see what can be done.
>
> > 3) I'm also thinking of bringing the onboard RFM12/B external clock
> > back to the CPU and have it available via a solder jumper,
> that way I
> > can get a 0-10Mhz clock without any extra components, anything more
> > can be done as a quartz?
>
> Isn't that a big no-no? Like that HopeRF reported they had a
> major problem with clock in their units?
>
I vaguely remember something from integrations errata but will do some
further research, it would be good if it could be done, as it means for any
frequency quartz below 10MHZ you wouldn't need any additional components,
when I build my very first prototypes I used the this for my PICs clock
source.
Even if there is a problem with the current versions from Hope, future ones
may fix it and it can alwayys be turned off on the RFM12B if not required.
Stephen...
Konstantin
> I vaguely remember something from integrations errata but will do some
> further research, it would be good if it could be done, as it means for any
> frequency quartz below 10MHZ you wouldn't need any additional components,
> when I build my very first prototypes I used the this for my PICs clock
> source.
>
> Even if there is a problem with the current versions from Hope, future ones
> may fix it and it can alwayys be turned off on the RFM12B if not required.
How much does the quartz cost anyway? I can get it here for $0.50,
possibly cheaper if I buy 10. Also, there's always the option of using
Atmega's internal RC oscillator.
Stephen Eaton
> virus signature database 4092 (20090520) __________
>
> The message was checked by ESET Smart Security.
>
> http://www.eset.com
>
>
> __________ Information from ESET Smart Security, version of