[OpenTRV-dev] OT: GSM remote control in car

Tue Mar 3 22:56:11 GMT 2015

Hi

just check the voltage input of MC34063A - it can tolerate up to 40V and
thus you shouldn't need a tranzorb protection device. If you are getting
40V spikes on your car battery then there is a problem.  If you still want
to add protection, you only need to make sure the tranzorb clamps to less
than 40V in worst case.  go for 26V or something like that - it will ensure
the transorb never conducts under normal crank or charging cases, but still
provides protection. I couldn't see where else +V goes to , so you need to
check the max operating voltage of whatever's connected there as well.

The MC34063A isn't a good choice for powering the system since (at 100kHz
switching frequency) it may not respond fast enough enough to handle the
high peak currents demanded by the GSM pulses (up to 2A bursts , I remember
times in usecs, but online says 1/8 timeslots * 4ms period).

VCC for the GSM modem is powered directly from the MC34063 and its output
is set for 4V which is probably right. Usually these GSM modules can be
powered directly from a Li-ion cell (max 4.2V / 4.35V). The modem should
continue to work down to around 3.5V so your battery voltage would really
need to drop for it to start failing.

You could try reducing C4 to speed up the regulators response to the fast
transients generated by the GSM modem as it transmits.  As Kevin suggested,
double C2 to help the regulator . Use ceramic caps since they have lower
ESR than large aluminum electrolytics. Datasheet recommends 1000uF on the
input (C2).

I see there is already 3x 470uF  (likely those 3x black ceramics in the
picture next to each other) +100uF  but they are on the main board and not
electrically close to the modem. If the tracks are not thick (really thick)
and the caps not close to the modem, they won't do much to help it. Indeed,
with b-2-b connectors in between there is too much inductance for them to
help for peaks.

>From the picture though , the modem board has its own local ceramic cap and
the schematics suggest 2 of them.  Really , that should be good enough.

Any chance you can measure the input voltage and / or GSM supply rail with
a fast meter as you turn on the car?

Hadyn

On 3 March 2015 at 20:11, Kevin Wood <EMAIL ADDRESS HIDDEN> wrote:

>
> I would go a little higher than 15V on the tranzorb. You want it to have a
> reasonable margin above the highest possible voltage you'll see, and a car
> system could get quite close to 15V. 18V would be reasonable, I'd say.
>
> Kevin
>
> > So far i lean towards adding a 22000uF 25V cap and a 15V transorber...
> > 15V is 1V below what the remote control board should be spec'ed for
> >
> >
> >
> > 2015-03-03 17:52 GMT+01:00 Kevin Wood <EMAIL ADDRESS HIDDEN>:
> >> OK. In that case, there probably isn't much else you can do but use a
> >> capacitor / battery to hold up the supply.
> >>
> >> You already have a diode there, so effectively increase C2 to give it a
> >> bit more resilience. I'm still slightly surprised that you have a
> >> problem,
> >> because I wouldn't expect the terminal voltage to drop below 8V or so.
> >>
> >> I would also put a tranzorb rated at maybe 18-24 volts across the 12V
> >> supply input after the fuse, just to catch any spikes from the car's
> >> electrical system. It's a very dirty 12V source and might spike high
> >> enough to kill something.
> >>
> >> Kevin
> >>
> >>> the remote control is direct to battery :-D
> >>>
> >>> or else... yes... it would loose power at engine start...
> >>>
> >>> it for a car preheater that has its own fuse box etc and not tied in
> >>> to the cars electrical other than at the battery terminals
> >>>
> >>> 2015-03-03 17:27 GMT+01:00 Kevin Wood <EMAIL ADDRESS HIDDEN>:
> >>>> I would check how you're deriving the supply voltage for the module. I
> >>>> think the SIM900 module itself only needs 3.3 / 3.8 V but the current
> >>>> is
> >>>> up to 2A peak when transmitting. Mine includes a regulator on the
> >>>> shield
> >>>> that drops this from a nominal 5V supply input.
> >>>>
> >>>> I'd be very surprised if your battery voltage is dropping low enough
> >>>> that
> >>>> an LDO regulator wouldn't be able to supply the module from 12V, but I
> >>>> suspect you might be regulating down to 5V, then down to 3.8, etc. and
> >>>> losing a bit of headroom with each step, which might cause you a
> >>>> problem.
> >>>>
> >>>> Dropping 12V to 3.8 at 2A might be a tall order for the regulator
> >>>> without
> >>>> better heatsinking, but I wouldn't mind betting the average current is
> >>>> quite low, given that the module won't work in every GSM timeslot, so
> >>>> you
> >>>> might well get away with it.
> >>>>
> >>>> Also, the accessory circuits on some cars are disconnected when the
> >>>> key
> >>>> is
> >>>> in the "III" (cranking) position, so it might be simply that you are
> >>>> losing the supply completely.
> >>>>
> >>>> BTW: Sorry for my absence on here of late. I'll update you with what
> >>>> I've
> >>>> been doing when I get a chance as it might be interesting.
> >>>>
> >>>> Kevin
> >>>>
> >>>>
> >>>>
> >>>>> its a sim900 based module and datasheet says 2A at max...
> >>>>>
> >>>>> how do i calc how big the cap needs to be for every X sec i want the
> >>>>> cap to power it? it not its quite extreme but better be safe than
> >>>>> sorry
> >>>>>
> >>>>> the module runs at ~4.8V, but need to check the regulator on the
> >>>>> mega2560 based board the module mounts in to
> >>>>>
> >>>>> 2015-03-03 17:10 GMT+01:00 Adrian Godwin <EMAIL ADDRESS HIDDEN>:
> >>>>>> Probably a cap and diode will fix it. Measure the current the GSM
> >>>>>> module
> >>>>>> takes (while transmitting) to be dure. If not, use a tricklecharged
> >>>>>> battery
> >>>>>> When you turn the starter, the battery voltage can fall very low and
> >>>>>> may
> >>>>>> well drop below what the module power supply needs.
> >>>>>>
> >>>>>>
> >>>>>> On Tue, Mar 3, 2015 at 3:56 PM, Bo Herrmannsen
> >>>>>> <EMAIL ADDRESS HIDDEN>
> >>>>>> wrote:
> >>>>>>>
> >>>>>>> Sorry for the Off Topic
> >>>>>>>
> >>>>>>>
> >>>>>>> But i have on the shelf a gsm remote control i build last year...
> >>>>>>> it
> >>>>>>> works fine but as soon i start the engine the GSM module goes
> >>>>>>> disconnects from the cell network
> >>>>>>>
> >>>>>>> i could just power that off and turn it on should it loose
> >>>>>>> connection
> >>>>>>>
> >>>>>>> but i wondered if a fix could also be a big cap and a diode ? or
> >>>>>>> maybe
> >>>>>>> a small lithium pack and a small charge circuit?
> >>>>>>>
> >>>>>>> /bo
> >>>>>>> _______________________________________________
> >>>>>>> OpenTRV-dev mailing list
> >>>>>>> EMAIL ADDRESS HIDDEN
> >>>>>>> http://lists.opentrv.org.uk/listinfo/opentrv-dev
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> _______________________________________________
> >>>>>> OpenTRV-dev mailing list
> >>>>>> EMAIL ADDRESS HIDDEN
> >>>>>> http://lists.opentrv.org.uk/listinfo/opentrv-dev
> >>>>>>
> >>>>> _______________________________________________
> >>>>> OpenTRV-dev mailing list
> >>>>> EMAIL ADDRESS HIDDEN
> >>>>> http://lists.opentrv.org.uk/listinfo/opentrv-dev
> >>>>>
> >>>>
> >>>>
> >>>> _______________________________________________
> >>>> OpenTRV-dev mailing list
> >>>> EMAIL ADDRESS HIDDEN
> >>>> http://lists.opentrv.org.uk/listinfo/opentrv-dev
> >>> _______________________________________________
> >>> OpenTRV-dev mailing list
> >>> EMAIL ADDRESS HIDDEN
> >>> http://lists.opentrv.org.uk/listinfo/opentrv-dev
> >>>
> >>
> >>
> >> _______________________________________________
> >> OpenTRV-dev mailing list
> >> EMAIL ADDRESS HIDDEN
> >> http://lists.opentrv.org.uk/listinfo/opentrv-dev
> > _______________________________________________
> > OpenTRV-dev mailing list
> > EMAIL ADDRESS HIDDEN
> > http://lists.opentrv.org.uk/listinfo/opentrv-dev
> >
>
>
> _______________________________________________
> OpenTRV-dev mailing list
> EMAIL ADDRESS HIDDEN
> http://lists.opentrv.org.uk/listinfo/opentrv-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opentrv.org.uk/pipermail/opentrv-dev/attachments/20150303/4f071315/attachment.html>