DocScrutinizer05 has quit [Disconnected by services]
DocScrutinizer05 has joined #neo900
xman has quit [Quit: Leaving.]
qws-user-1228 has quit [Read error: Connection reset by peer]
qws-user-1228 has joined #neo900
jamm has joined #neo900
<jamm>
hello
<jamm>
quick question about the use of LMV221 on inbound
<jamm>
wouldn't it not be sensitive enough to detect the return signal? it only goes down to -45 dBm
<jamm>
or am I not seeing something about how the coupler is supposed to be working?
<jamm>
sorry, this is also regarding the neo900 proto v2 schematic
<Joerg-Neo900>
the concept is to detect direction of signal. If 'inbound' signal (actually outbound reflected on antenna) is so much weaker than the original outbound that the LMV221 doesn't see it anymore, we have no problem at all
<Joerg-Neo900>
if any inbound signal is so strong that its reflection on modem triggers the outbound detector, for sure the inbound detector will also see it
<Joerg-Neo900>
and will also rate it stronger than the reflected signal on outbound, so we avoid false positives
<Joerg-Neo900>
we don't want to trigger on arbitrary inbound signals, we only know one thing: they are not relevant for us
<Joerg-Neo900>
the purpose of the whole coupler is to reliably detect TX energy and discern it from inbound energy from nearby other transmitters
<Joerg-Neo900>
as an additional benefit you can use it to check other transmitters (phones) if they send RF, if you get close enough to them to couple sufficient energy from their antenna into Neo900's antenna
<Joerg-Neo900>
but that's not the main purpose
<Joerg-Neo900>
jamm: ^^^
chomwitt has quit [Ping timeout: 258 seconds]
<jamm>
hmm, i see, so is it for detecting a bad mismatch on the antenna? or for ambient rf levels (but wide band?)
<Joerg-Neo900>
it is for detecting outbound RF energy and discerning it from inbound RF energy that gets reflected on modem input and goes back the line through the coupler direction antenna from where it came
<Joerg-Neo900>
the reflected signal will be weaker than the original signal
<Joerg-Neo900>
so no matter how strong the inbound signal and how weak the outbound signal, we can tell apart the two
<Joerg-Neo900>
you wouldn't want false positives from your pub neighbors cell phone, right?
<Joerg-Neo900>
not even when he places his phone right next in touch to yours, on the bar
<jamm>
but i thought a coupler's purpose is to be directional, so you would only ever see your transmitted signal?
<jamm>
i'm not completely sure though... i have never thought about it being used where other signals are coupling strongly
<Joerg-Neo900>
a) the coupler never has an infinite separation b) we have reflections on modem input, the so called return-loss which is as feeble as 5dB when the modem is idle
<Joerg-Neo900>
so the reflected inbound signal in coupler is only a 5dB lower than the original inbound signal
<Joerg-Neo900>
even if the coupler was working perfect, we couldn't tell apart a true TX signal from a 5dB stronger RX aka inbound signal
<Joerg-Neo900>
unless we also probe for the inbound signal strength
<Joerg-Neo900>
which for inbound signals *always* is higher than the reflected signal
luke-jr has quit [Ping timeout: 265 seconds]
luke-jr has joined #neo900
<Joerg-Neo900>
in above diagram see how the light blue dotted signal also goes bouncing off the modem and going through coupler in TX direction
<jamm>
yes, I see, I guess I just figured this factor was much higher, and you didn't need to worry about enough that it interferes with your output power measurement
<jamm>
(very nice picture btw!)
<Joerg-Neo900>
thanks :-)
<jamm>
and maybe I just get to use nicer antennas with better return loss lol
<jamm>
oh yeah, 1 more question, about the NFC/RFID section with the TRF7970A, and the kinetis microcontroller
<jamm>
did you look at using an ic that has support in the linux nfc subsystem?
<jamm>
ah ok I think we needed better Android support where we used the nxp pn7120 instead of the TI part.
<jamm>
I think Android support is possible with the TRF7970A, but TI hasn't really done anything with it yet (not like NXP, anyways)
<Joerg-Neo900>
NXP is nasty when it comes to datasheets about their "smart" chips
<Joerg-Neo900>
almost all (Android) phones use the PN5dunnowhat with embedded CPU for the more complex protocols, and they don't share any info about the formware blobs to load in there, nor about the CPU
<Joerg-Neo900>
thus we used a 'dump' RF frontend chip and our own MCU to do the compelx protocols
<Joerg-Neo900>
'dumb' even
<Joerg-Neo900>
we could have done without MCU if we wouldn't want SWP for the SIM integration
<Joerg-Neo900>
but now we have a really nice NFC system that outperforms many RFID/NFC eval boards hands down
<Joerg-Neo900>
you can do pretty much everything possible with this design
<jamm>
yes, that is pretty awesome, especially with the sim integration!
<jamm>
yeah, I am still wondering how you are going to fit it all on the board haha
<Joerg-Neo900>
probably the RFID board will buy our devices for R&D and evaluating of new protocols ;-)
<Joerg-Neo900>
((fit on the board)) that's actually a huge concern and problem, yes
<Joerg-Neo900>
last we checked, we had a coverage of ~70%
<Joerg-Neo900>
so pretty much on the limit
<Joerg-Neo900>
we'll see :-)
<jamm>
it would be great to have all that functionality for sure, so best of luck...
<Joerg-Neo900>
thanks!
<Joerg-Neo900>
you wouldn't know of layouters who accept a job?
<gry>
didn't get that part
<Joerg-Neo900>
gry?
<gry>
do you want to hire a layouter? what does that do?
<Joerg-Neo900>
layout? :-)
<Joerg-Neo900>
we have Ahycka, but never place all eggs in one basket, you know
<gry>
web layout? electric chip layout? layout of a dinner table? what?
<Joerg-Neo900>
PCB layout, obvously. sorry for being fuzzy
<jamm>
not really unfortunately... we have another person at my company that does a lot of the layout (mostly Altium though,) but I don't know if we can do as complex a design yet
<Joerg-Neo900>
kicad pcbnew
<ds2>
KiCad for Neo900?
<gry>
ooh :-)
<Joerg-Neo900>
yeah, we would even accept altium if that's needed to get the job accomplished
<jamm>
i still have fears of doing DDR(2,3) layouts haha, even though TI gives you the whole design
<ds2>
oh my... condolences for the the poor layout guy
<Joerg-Neo900>
hehehe
<Joerg-Neo900>
jamm: DDR2 is harder
<Joerg-Neo900>
more timing contraints
<ds2>
did POP get dropped?
<Joerg-Neo900>
no
<Joerg-Neo900>
luckily not
<ds2>
then there is no DDR layout :D
<Joerg-Neo900>
exactly
<jamm>
yes...
<ds2>
just breaking out 300+ balls
<Joerg-Neo900>
:nod:
<ds2>
fun with VIP
<Joerg-Neo900>
VIP?
<jamm>
how is the mainline support for the DM3730?
<Joerg-Neo900>
linux? great
<ds2>
it is a 0.4mm BGA
<jamm>
we used it in one of our products a bit ago, but then TI got stuck on 2.6.37 and kinda dropped it
<Joerg-Neo900>
aaah
<ds2>
there really isn't room to do vias in between balls so you have to do Via In Pad
<ds2>
and probally have to do it blind
<Joerg-Neo900>
we will need in-pad vias, yes
<Joerg-Neo900>
blind capped in-pad vias
<ds2>
lots of pressure for the layout guy to not screw up
<Joerg-Neo900>
elaborate please
<ds2>
blind vias == $$$$$$ board
<Joerg-Neo900>
yes
<ds2>
layout person needs to be clever to make sure board is rescueable like appropriate place to wire around or cut traces
<Joerg-Neo900>
we don't hope for fixing stuff by cutting traces :-) If you knew which traces need to be cut-able you already could do stuff right instead ;-D
<Joerg-Neo900>
ds2: maybe you want to have a look at the schematics?