No announcement yet.

hardwiring Labtop DC-DC converter directly to battery. ie bypass any switching relay

  • Filter
  • Time
  • Show
Clear All
new posts

  • hardwiring Labtop DC-DC converter directly to battery. ie bypass any switching relay

    Ok so here is my plan. I have a netbook kav60 acer that I will be adding to my build.

    The idea is to keep the computer in sleep mode when not in use which will be wired directly to the car battery drawing a minimal
    ~1watt at 19Vs or 50mA's.

    now a normal battery has around 70A/h, so putting a small 50mA drain on it would theoretically be an ineffectual drain on the battery. 10 days in sleep would still draw less then 20A hours when adding other components and taking P/S inefficiencies into account.

    SO, does anyone know of any reason that this idea would be a bad idea?

  • #2
    Originally posted by kdl View Post
    ... a small 50mA drain on it would theoretically be an ineffectual drain on the battery.
    Not so.
    Even Optima did an example of a 17mA drain on one of their ~70AH batteries. (IMO purely academic - and Optima would have failed way before that... LOL!)

    50mA is more than 2 LEDs, and even one LED can be considered a problem (viz 17mA example).

    Vehicles with ~50mA alarm drains have problems after a few days.
    Not to mention their shorter battery life.


    • #3
      How do tools like the CarNetix CNX-P1900 with built-in micro controller work then? assuming a micro along with all other circuitry must drain a few dozen mA's.

      This is kind of a big dissapointment. It changes my setup boottime from 3 seconds to ~45. Thats a huge change. From standby to playing music is 5 seconds, faster then my pioneer headunit!

      What about adding a 100-200ma solar trickle charger to the setup, tuck it away on the roof or something?


      • #4
        You'd have to ask others with Carnetix etc supplies. Maybe they don't leave their vehicle unstarted or uncharged for longer than one week? And most probably use a dual battery set up.

        I deal more with things that have ~10uA as a sleep/standby current, way less than a battery's self-discharge current.

        You can try the 50mA, your calcs are probably right.
        I'm just saying that your battery won't live as long. With 20 days of sulfation build up (probably being most significant after ~7 days)...
        Though a good run after should blow away the sulfates (unless hardened - eg, sulfates older than ~7 days), it's not a desirable situation. And one day your main cranker may simply collapse...

        Better with a timed converter that connects when the netbook's battery has discharged, or though to be near discharged. But they too have a limited number of cycles.
        Or a second battery for independence form the main, and cheaper replacement ($25 for 7AH etc), though then a battery isolator is required (which can simply be a $3 relay)

        Solar chargers... well...
        I reckon too expensive, and IMO they MUST have a regulator unless the battery's self discharge current or loads are guaranteed to exceed the solar output. (Do NOT want solar charging to exceed ~14.5V!!)


        • #5
          I like the idea of a timed circuit with a set number of mAh it will expend before simply letting the PC battery discharge. I know that I could setup the PC to auto hibernate when the battery reaches critical levels.

          but the circuits that I would need to monitor this setup will further increase the strain on the car battery. I do plan on buying a 9500mAh battery replacment for the netbook which will more then double what I have now, and It already last 1-2days in sleep mode.

          So what I was thinking was I could let the netbook run on its own battery down to ~%10 or so THEN hibernate. Figuring It will last in standby for a 2-4 days before It will turn off into hibernation. I use my car a lot more then that so it will likely never get down that low, seeing as the battery will change back up to %80+ within a reasonably short car ride/
          Last edited by kdl; 03-26-2012, 07:53 PM.


          • #6
            I use an eeepc and leave it in standby on the battery. It is a 2 year old battery that was used regularly when I used it as my daily PC around the house. Thus far, it has never gotten below 90%. I drive daily though and have a 30 minute commute. In the unlikely event I park it long enough to drop to 10%, it will hibernate. If I ever park it for a few days when I go away, I just hibernate or shut it down.

            I definitely wouldn't recommend leaving the charger running on the battery.

            Oh, weird thought. If you were up for the idea of the solar charger, you could use a relay with both normally open and closed. Then you could have input from battery when switched on and on solar when switched off. Spark might be right about price. I know that when bench testing, my charger was very specific about voltage. If I didn't have enough, it would cause the net book to convulse.


            • #7
              Originally posted by kdl View Post
              but the circuits that I would need to monitor this setup will further increase the strain on the car battery.
              No they won't.

              Why would anyone add a power saving feature that uses more power that it saves?

              I'm not saying that people don't (love those little mains AC energy savers eh?), or that people use standard 555 timers that use 10mA on standby (1,000 times the current of a PIC etc) or other stupid circuits etc.
              But a true power saver is a power saver.

              Chargers are ok left on batteries PROVIDED they have a cut-off or "drop to float" (voltage & current) upon the battery reaching full charge. (But how do they determine that..? IE - "intelligent" chargers.) [Hence the need for solar regulators - even for small panels.]

              I'd go for a 2nd battery unless a low voltage cutout was installed, though I might add that anyhow if the 2nd battery is expensive.
              The 2nd battery (thru an isolator) ensures cranking battery reserves. (Or rather, ensures the secondary battery and load does not discharge the main battery.)
              Many use a 2nd battery the same as the main, hence main battery [I]redundancy/I]. Unfortunately the 2nd often has to be an AGM type if internal etc (cabin, trunk/boot), and that's overkill (as is an AGM for cranking; in general), but a 7AH or larger will often do as an emergency cranker.

              Automotive relays typically use 250mA (~50R (Ohms)) but relays can be less , eg 80mA for 160R. But that is still high.
              MOSFETs are negligible (uA of not nA), and transistor can be negligible (uA - mA).


              • #8
                I was looking at a few solid state transistor relays and it is good to know that use such a small amount of current.

                I also stumbled on the watchdog feature in many micro controllers which allows the micro to run its code, which takes a few micro second, then completely power down for a set amount of time before resuming and running the code again. Now this time off can be set for any amount of time an uses only uA in the off mode.

                So say the micro will run for 8 hours "always on" with watchdog you can set it to be off for 2 seconds after it performs its duties. So say we set watchdog for 2 seconds off and the code takes a relatively long 10mS to run; 2S/10mS = 200 . Using the watchdog timer you could increase battery life from 8 hours to ~1600 hours or over 2 months!

                Now obviously this method is no good for any application that requires a very accurate and precise monitoring, but with something thats job is to monitor an ignition or control a set of relays a 2 second delay i very tolerable.

                I believe the watchdog also has a wake from sleep input monitoring mode, basically if a lines digital logic changes it will wake or sleep the micro. Its something to think about. This paired with a few low uA solid state relays would work perfectly in a carPC setup/
                Last edited by kdl; 03-27-2012, 10:58 AM.


                • #9
                  I sometimes forget I'm on a forum that isn't afraid of diodes, transistors & uPCs! I can skip preaching to the converted....

                  I've been looking at the PICAXE 08M2. Finally a PIC with enough program space etc, simple to use, and only 8 pins. (With SMD (surface mount( versions available). All for ~$3 plus the "overhead" to learn its language, though copy & paste & simple edits is usually enough for common applications. Oh - plus a $20-$30 USB converter if you don;t have a serial port.

                  But for an 8-pin chip plus its 5V regulated supply & 2 resistors & socket for programming (as with all PICAXEs) plus the "interface circuitry" that is common to any electronic solution, it's IMO the way to go.
                  Change delays or voltage thresholds or even the combinational logic that triggers an event or events - just reprogram - no trimpots or circuit changes or added chips.

                  You might find that a simple $2 MOSFET can be the relay. It's simple of it's a grounding relay (open collector) though +ve switching seems to be as simple. And with on resistances (Rds) as low as mili-Ohms, voltage drops aren't a problem.

                  In fact, there is your "proper" lighting regulator - a slid state switch that only passes current when the "average" voltage dips below 14.2-14.4V etc.
                  And there too is your headlight or other dimmer (eg - LEDs) if wanting to reduce power etc.
                  The latter applications would use the PICAXEs PWM output(s).

                  The sky is the limit. Even serial RAM can be added for datalogging or extra tables etc (eg, EFI or electronic ignition).


                  • #10
                    sounds very similar to the ATmega328 (arduino uno). Is the clock internal on the PICaxe.

                    The 8pin through-hole setup is really nice though. is it programmed in C? also is it loaded with kiel/? or does it have its own interface for programming

                    -- went looking at the AXE myself. Seems like a really nice little chip. cheap and has its own programming environment. unfortunately the cable to program this thing cost more then an entire UNO board
                    Last edited by kdl; 03-28-2012, 07:25 PM.


                    • #11
                      Nah - the Arduino is a full uPC...

                      PICAXE is all internal (including clock for up to 32MHz speed) which allows software protection (can't be read or sensed - no external address nor data pins).
                      2 resistors to enable serial programming.
                      The M2 series is the one to get - at least for the 8 pin (ie, PICAXE 08M2, the PICAXE 08 chip variants are generally designated PIC12Fxxx) - earlier had (IMO) too small a program space.

                      See sites like - eg, What-Is-PICAXE? & GettingStarted- PICAXE-Manuals.

                      (Bluddy PC freezes! Gotta love firefox's & this site's auto-save & recovery!!)

                      I'll eventually go SMD (in production, that's cheaper anyhow). EG:
                      The AXE230 module:
                      Click image for larger version

Name:	AXE23-PICAXE08ModulePCB-SMD-s50.jpg
Views:	1
Size:	10.0 KB
ID:	2283975

                      Or the RKP08c from via eBay...
                      Click image for larger version

Name:	RKP08c.jpg
Views:	1
Size:	48.8 KB
ID:	2283976

                      Though their RKP08c kit descriptions are somewhat ambiguous (whether the 2 versions are populated with the 08, and if the 08M2) and the rkeducation site is currently down (probably due to Americanisation).

                      Incidentally, I got 2 Arduino Duemilanoves a year (or 2, 3?) back instead of learning PICs and Fusion Brains etc (though I wanted stand-alone application). [Damn... Reboot #2... me thinks it's time for some PC maintenance...]
                      But the new 08M2 makes an attractive "single-core" (& PCB) for ANY typical application we find ourselves in - whether auto-up & -down dimming of puddle lights & LEDs, combinational switching (if lights but not engine unless temp exceeds xyz and mother isn't home...), even extra long timers (external time bases (crystals) can be used).

                      But I'll bail now before my PC does...
                      Last edited by OldSpark; 03-29-2012, 03:31 AM. Reason: Fixed images.


                      • #12
                        Ok, I've tightened the timing belt for my DRAM and rewound its elastic. Maybe I shouldn't have delete that desktop.ini under temp internet files the other day?

                        I've fixed the images above, but other PS stuff was lost.

                        But that's good - instead of writing how I was gonna get the 08M2 equivalent of a 2003 DS30487B for thePIC16F87/77 series (18-pin DIP etc), I found a close match for the 08M2, namely the DS41441B.
                        I liked (recently) reading the old 2003 doco because it had it all - chip variations, electrical characteristics, timing and waveforms, inter-dependencies, memory maps, module descriptions (PWM, timer/counters, ADC etc), commands & flags, etc.

                        Though ~212 pages (the newer DS41441B is 382pp), it's a good "selective" read. Question something, find its page and then follow whatever links seem relevant or of interest - a bit (or byte) at a time until interest wanes.

                        Did I say what a non-expert I am on the PICs, or did that get lost?

                        Speaking of lost, I've just lost my cursor keys. I can't wait for my new i5 or whatever it will be (it was supposed to be a "C2D exceeding 2GHz" when I intended to upgrade a few years back, but at least now I can skip XP all together (and Vista? - never touch a 3rd release from MS!)).

                        Off to watch the TV whilst the rubber band settles in and RAMps up...

                        Ah I miss the old static RAMs!


                        • #13
                          Damn! how old is that machine!

                          I am actually going to take a look at the RKP08c, for like a few dollars It gives you a nice board to program the chip in.

                          Does the picaxe utilize any watchdog like techniques? whats its power draw like?


                          • #14
                            From doc DS41441B:

                            Low-Power Features:
                            • Standby Current (PIC12LF1840): - 20 nA @ 1.8V, typical
                            • Operating Current (PIC12LF1840): - 34uA @ 1 MHz, 1.8V, typical
                            • Low-Power Watchdog Timer Current (PIC12LF1840): - 300 nA @ 1.8V, typical

                            As to my PC, it solved my y2k date issue when I got it circa 2002.
                            I figured I'd keep it till UNIX's "y2k bug" (2032??), but unlike commercial software and platforms, *NIX has already fixed that. (So no need to upgrade my Window 2000 PC eh?)