Simple controller for backup battery ?

[jondee86]

I am in the habit of disconnecting my car battery to stop the alarm
running it flat when the car is not used for a week or more. However,
this means I also lose all the pre-set values in the head unit. So I am
looking at putting in a simple gel cell backup battery, just to keep
voltage on the HU memory (yellow/constant live) wire. I was thinking
of something like this...



This circuit is said to trickle charge at 120mA, and I was looking at
something like a 6Ah gel cell. HU will always be off when the main
battery is disconnected, so the current drain on the backup battery
should be just a few mA.

Anyone tried this or care to comment ?

Cheers... jondee86

[Grrrrrrr!]

The 100ohm resistor is way too high, 10-15ohms would give you about 120ma charging current on a flat battery, and about 0.4A on a dead flat battery. Personally I'd just run the one diode and hook the gel battery straight to it. So long as you are driving it at least every couple of week for a decent amount of time that gel battery will never get anywhere near flat.

[jondee86]

So, what you are saying is that since the gel cell should always be
close to fully charged, I can forget about the resistor and D2, and
just put the cell in parallel with the load ?

I'm thinking that if the car sits for a week with the battery connected,
the backup voltage will drop in line with the main battery voltage.
When it starts up, the charging voltage could be high enough to do
some damage without the resistor ?

Also wondering what happens when the engine is cranking ? With the
gel cell in parallel to the main battery, will the cell be taking a big hit ?

Yeah... I think someone made a typo on that diagram, as I see the
resistor listed as 10 Ohm 10 Watt elsewhere.

Cheers... jondee86

[Grrrrrrr!]

No, the one diode in the circuit will stop the gel battery from being drained by the main battery and rest of car. The main battery will be dead in a week or so, but the gel will still be at 11+ volts for at least a month unless the headunit drains a lot more power than it should.

[sergei]

You have to watch out the two things:

1) Current rating of the diodes vs what the device behind them consumes (make it at least double).
2) Each diode has a voltage drop ~1V-1.5V.

Say your gel cell battery is at 11V (still usable for low power device) after diode it might be 9.5V, which is often at the cut-off for many 12V rated devices.

[jondee86]

There does not seem to be any way to find out the draw on the
yellow "keep memory alive" wire. Most opinions seem to be that
once the HU has been switched off, 20 - 40 mA should be about
all that is required. So I figure that 1 Amp is the most that is ever
likely to be flowing anywhere in the backup battery circuit, and
that I would use Schottky IN5822T 3A diodes.

But 20-40 mA out of a 6 Ahr battery does not seem like a great
hold-up time to me, so I think I will measure the parasitic load
just so I know

Cheers... jondee86

[Grrrrrrr!]

20-40mA? oh hell no!

if it draws more than 2mA to keep the memory alive the designer should be put against a wall and shot. It should in the microamps.

[jondee86]

Well, now I can contribute something to the knowledge pool

Seems that this head unit (Sony CDX-GT650UI) must run a self-diagnostic
or something similar when you put power on the yellow wire. What I found
was that when I put the meter on the battery negative, I had a initial drain
of around 203mA, but this dropped to 32mA after about ten seconds. Then
I tested the radio fuse and got 172mA dropping to 1mA.

So the often quoted "normal" parasitic battery load of 20-40mA is in the
ballpark for the whole car, and in my case, most of that is down to the old
alarm system. And grandmaster Grrrrrrr! is right on the money with 1mA
to keep the radio memory alive.

A 6Ahr gel cell would keep that alive for months... I might even look at a
smaller battery to make packaging easier.

Cheers... jondee86

[whynot]

If you are using an after market radio make sure that D1 is rated for 10 amps as the radio draws the bulk of the current through that yellow memory wire. It's also possible the radio might not perform as well with a diode in that wire as there is typically a 0.6v drop across a silicon power diode.

[Bling]

Could you rewire the alarm power source so you can flick a kill switch to cut it's brain power? That way you can leave the battery connected and the head unit will keep it's memory. Just another idea anyway.

[jondee86]

Could you rewire the alarm power source so you can flick a kill switch
to cut it's brain power? That way you can leave the battery connected and the head
unit will keep it's memory. Just another idea anyway.

Yes... that is about the point I have reached in my thinking. If the alarm
stays on line, the backup battery will be carrying the alarm drain. So I need
to take the alarm out of circuit. If I do that, the only thing other than the HU
memory that would be drawing power would be the clock, and that can't be
much of a draw.

Then I can forget the whole backup battery idea, and just leave the main
battery in place with the alarm powered down. Any idea how much current
a modern alarm system draws while armed ??

Cheers... jondee86

[Grrrrrrr!]

Yes... that is about the point I have reached in my thinking. If the alarm
stays on line, the backup battery will be carrying the alarm drain. So I need
to take the alarm out of circuit.

No, the diode stops the backup battery from carrying anything except the headunit. Diode = one way valve. Power can flow into the battery and onwards to the headunit, but no power goes the other way.

There is a tiny bit of leakage, but thats in the uA range.

[jondee86]

Ahhh.... I should have posted this first... it shows the setup I was
referring to in my last post above...



I wanted to piggyback off the existing rather than intercepting it totally.
But without a diode in the yellow wire, it basically turns the gel cell into
a backup for the HU and the alarm system. So this lead to having a kill
switch for the alarm, which lead to not needing the gel cell

Cheers... jondee86

[sergei]

Ahhh.... I should have posted this first... it shows the setup I was
referring to in my last post above...



I wanted to piggyback off the existing rather than intercepting it totally.
But without a diode in the yellow wire, it basically turns the gel cell into
a backup for the HU and the alarm system. So this lead to having a kill
switch for the alarm, which lead to not needing the gel cell

Cheers... jondee86

The alarm will happily discharge the gel cell via D2 in your diagram, unless you plan on disconnecting yellow wire when not in use.
You can connect yellow wire to red (after switch), so yellow wire is disconnected unless accessories are on.

[jondee86]

Yes... to make it work I would either need to put a diode in the yellow
wire, or a self-latching relay that would open circuit when the battery
was removed and the yellow wire powered down.

I might try a kill switch on the alarm for now, although the whole idea
of having an alarm and not being able to leave it set seems pointless.
Maybe it's time to move up to a new alarm. Judging by how quickly this
one pulls the battery down, it must suck a fair bit more when it is set...
I'll do a bit more checking.

Cheers... jondee86