Post by thread:[EE]: Battery Powered IR Xmitter, Comments wanted.

I've been thinking about putting togther a door sensor with an IR transmitter. This particular door doesn't have access to the basement, so I can't easily hardwire the door sensor like I've done for the others in my house. Obviously a wireless sensor will require battery power. So power management is the critial issue. I wanted to get some opinions on the subject. Basically I'd like for the sensor to indicate 3 states: heartbeat while the door is closed, door opening, and a heartbeat when the door remains open. I'm figuring that the closed door heartbeat will do fine with a 1 to 2 minute interval. The open door is more critical so should be faster, say every 10 seconds. But how would I start to determine the running time of such a setup? Let's throw out some numbers that I pulled off the RatShack specifications page. The nominal Ah rating of a AA alkaline is 1.7 Ah. I'm thinking of using a Linear Technology LT1173 switching regulator due to it's extremely low 110 uA supply current. My circuit will be driven by a 16F84 (simply because I have them on hand) with the WDT enabled. The nominal current is 7 uA in this case. So I envision the basic operation to be: 1) Sleep until the WDT goes off or get an RB0 interrupt indicating that the door has been opened. 2) If the door heartbeat timer goes off, send an IR blip indicating a heartbeat and reset the timer, go back to sleep. 3) If an interrupt, send a door open event, go back to sleep. My preliminary numbers indicate that 1700 mAH/.117 ma -> 14529 hours which is about 605 days. That's absolutely fine with me. However I have no real clue how to determine the power for the IR Led. My first crack at it is to try to compute the 100 percent duty cycle time and the real duty cycle and divide. So say we drove the IR led at the max 25ma that a PIC I/O pin can handle: 1700 mAH/25 ma -> 68 hours. Now each day we'll get 1440 IR blips for closed door heartbeats at 1 per minute. Let's say we're transmitting 10000 BPS serial with 3 bits (start bit, heartbeat/ interrupt, door status). Finally since the closed door heartbeat is the most common case, we conserve power by just sending the start bit as active. So in essence we'll pulse the IR LED for .1 mS. So the total active time for the LED for a day would be 144 mS. So the effective current draw would be 25 ma * (144 mS/1440000 mS) -> 2.5 uA. So then the effective time is 1700000 uAH/2.5 uA -> 68000 hours which is practically forever. I'm just wondering if I missed something. It seems I should easily be able to run such a setup for a year on a single battery. And with the heartbeat I should be able to get a fairly fast indication of the battery being drained. Lastly John Payson posted an article describing what he called a 3-6-3 encoding for IR pulses that's self timing. Honestly I didn't really catch it. Here's the relavent discussion found in the article located here: http://www.infosite.com/~jkeyzer/piclist/1999/Jan/1745.html --------------------------------------------------------- Every six on/off cycles encodes a byte (or if you want to push things, one of 400 states). Three of the on pulses are longer than the other three pulses, and three of the gaps are longer than the other three gaps. To receive the data, store the lengths of the received pulses and gaps, and look for the longest three of each. Although the interleaved 3-of-6 approach requires sending more pulses than direct async communication, its reduced requirements for precise pulse lengths would allow the pulse rate to be increased without aff- ecting reliability. It would remain to be seen, of course, just how far the technique could be pushed, but it should be reasonably effect- ive. ----------------------------------------------------------- Maybe a visual example can clarify it for me. Any ideas on this one? Looking forward to hear your thoughts. BAJ -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> I've been thinking about putting togther a door sensor with an IR > transmitter. > This particular door doesn't have access to the basement, so I > can't easily > hardwire the door sensor like I've done for the others in my house. > > Obviously a wireless sensor will require battery power. So power > management > is the critial issue. I wanted to get some opinions on the subject. Unfortunately I can't help much with your question (it does sound about right though, I'm by far NO expert but I didn't notice any glaring errors) however, I was wondering whether you would be willing to post the details of this project when it is completed, sounds VERY interesting!! :) Thanks, TTYL -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> Obviously a wireless sensor will require battery power. So power management > is the critial issue. I wanted to get some opinions on the subject. > ... > But how would I start to determine the running time of such a setup? Let's > throw out some numbers that I pulled off the RatShack specifications page. > > The nominal Ah rating of a AA alkaline is 1.7 Ah. I'm thinking of using a > Linear Technology LT1173 switching regulator due to it's extremely low 110 uA > supply current. My circuit will be driven by a 16F84 (simply because I have > them on hand) with the WDT enabled. The nominal current is 7 uA in this case. > > So I envision the basic operation to be: > > 1) Sleep until the WDT goes off or get an RB0 interrupt indicating that the > door has been opened. > 2) If the door heartbeat timer goes off, send an IR blip indicating a heartbeat > and reset the timer, go back to sleep. > 3) If an interrupt, send a door open event, go back to sleep. > > My preliminary numbers indicate that 1700 mAH/.117 ma -> 14529 hours which > is about 605 days. That's absolutely fine with me. Hmmm... this doesn't sound right. You say that the LT1173 draws only 110uA. However, that is when the LT1173 is in shutdown. If this buck regulator is in shutdown, how can you keep your PIC alive? The output voltage when the LT1173 is in shutdown goes down to about 1.4 volts (1.7 volts minus a good diode's drop) since there is a DC path through the inductor, and diode, to other devices in your circuit. I don't know what this switcher draws, as a minimum, when your load is only the few microamps that a PIC draws when sleeping. Often, these boost regulators don't operate efficiently at *all* in the low end of their current range, typically only providing 10-20% efficiency. I don't see that in your computations. Unfortunately, I don't see a system efficiency curve for this device in a quick perusal of the data sheet, so I would suggest that you talk to an FAE, or perhaps do some prototyping to qualify this. Also, as a "one of" project, you should be aware of the cretinous low ESR surface mount tantalum capacitor supply crunch that exists right now. We are having a tough time getting unit quantities of these capacitors for *our* buck regulator designs. This regulator is extremely sensitive to the quality of the capacitors you bypass your battery and output with. *Do* check for capacitor sources before committing artwork to make sure you can stuff your circuit when you are done. > BAJ > On the other hand, I am using an LT1304 to power my handheld device. I have been pleased with its performance thusfar. My 0x02 cents. -d -- http://www.piclist.com hint: To leave the PICList spam_OUTpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu>

{Quote hidden}> > > Obviously a wireless sensor will require battery power. So power > management > > is the critial issue. I wanted to get some opinions on the subject. > > > ... > > But how would I start to determine the running time of such a setup? Let's > > throw out some numbers that I pulled off the RatShack specifications page. > > > > The nominal Ah rating of a AA alkaline is 1.7 Ah. I'm thinking of using a > > Linear Technology LT1173 switching regulator due to it's extremely low 110 > uA > > supply current. My circuit will be driven by a 16F84 (simply because I > have > > them on hand) with the WDT enabled. The nominal current is 7 uA in this > case. > > > > So I envision the basic operation to be: > > > > 1) Sleep until the WDT goes off or get an RB0 interrupt indicating that > the > > door has been opened. > > 2) If the door heartbeat timer goes off, send an IR blip indicating a > heartbeat > > and reset the timer, go back to sleep. > > 3) If an interrupt, send a door open event, go back to sleep. > > > > My preliminary numbers indicate that 1700 mAH/.117 ma -> 14529 hours which > > is about 605 days. That's absolutely fine with me. > > Hmmm... this doesn't sound right. You say that the LT1173 draws only 110uA. > However, that is when the LT1173 is in shutdown. If this buck regulator is > in shutdown, how can you keep your PIC alive? The output voltage when the > LT1173 is in shutdown goes down to about 1.4 volts (1.7 volts minus a good > diode's drop) since there is a DC path through the inductor, and diode, to > other devices in your circuit. According to the data sheet that's not how the part works. It has a comparator that detects the voltage dropping. When it sees the voltage dropping it then kicks the inductor/switch/oscillator in to boost the voltage. Then it turns off again until the voltage drops below the threashold again. It short it refills the glass when it runs low. There's 50 mV of hysteresis so that the comparator doesn't go nuts trying to run the engine all the time. It said at low current draws that it'll operate in quiescent mode for several seconds, then kick in for a few cycles to refill, then go back to sleep. It seems perfect for the application I have. It's lazy and power conserving when its load is power conserving by can deliver up to 80 mA instantly on demand running from 2 AA batteries. In their design note 63 they show an example giving 26 hours run time on a 30 mA load. My design goal is a load between 150 and 200 uA. This projects to between 160 to 220 days runtime. I certainly can live with changing batteries once every 6 months. > Also, as a "one of" project, you should be aware of the cretinous low ESR > surface mount tantalum capacitor supply crunch that exists right now. We > are having a tough time getting unit quantities of these capacitors for > *our* buck regulator designs. This regulator is extremely sensitive to the > quality of the capacitors you bypass your battery and output with. *Do* > check for capacitor sources before committing artwork to make sure you can > stuff your circuit when you are done. Surface mount? Artwork? As you point out it's a one off. Most likely I'll wirewrap the unit. I may stick it on a 1/2 RatShack preetched board simply because they have a board/case combo that seems to fit the bill. Planning on using a standard radial tantalum. This is definitely hobby work. No professionalism required, or allowed ;-) BAJ -- http://www.piclist.com hint: To leave the PICList piclist-unsubscribe-requestKILLspammitvma.mit.edu>

Byron A Jeff wrote: > > You can't probably use the full Ah-capacity of the batteries. A normal dry > > cell (ZnC) drops the voltage more or less continously during discharge, > > while a NiCd have an almost flat discharge curve. According to my > > experience there are always some "empty" Ah:s in a dry cell. > > All I really need right now is off the cuff estimates, more order of > magnitude than precision. It's looking like 6 months for a pair of AA > batteries is a easily reachable goal. > > BAJ (Maybe I'm weird, I'd be tempted to think "Lead-Acid Gel Cell recharged by a small solar panel" here probably <G> There're tiny 2V Lead cells - something like 500mAh or more? - used for some flashlights, just keep the battery in "Float" charge and you're set. Assuming there's enough light there <G>) Mark -- Detest spam? Take the Boulder Pledge, boycott SPAMmers. http://www.zdnet.com/yil/content/mag/9612/ebert9612.html -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]: PIC only [EE]: engineering [OT]: off topic [AD]: advertisements

What's the deal with these Energizer E-squared cells, the ones with the titanium core ? I had a look at Energizer's site (http://www.energizer-e2.com) a few days ago but couldn't find any specs. Plenty of pat-on-back bumpf but no figures that I could see, even in the FAQ section. My FAQ - "Are they any good ?" > > > You can't probably use the full Ah-capacity of the batteries. A normal dry > > > cell (ZnC) drops the voltage more or less continously during discharge, > > > while a NiCd have an almost flat discharge curve. According to my > > > experience there are always some "empty" Ah:s in a dry cell. > > > > BAJ > > (Maybe I'm weird, I'd be tempted to think "Lead-Acid Gel Cell recharged > by a small solar panel" here probably <G> There're tiny 2V Lead cells - > Mark -- http://www.piclist.com hint: PICList Posts must start with ONE topic: "[PIC]:" PIC only "[EE]:" engineering "[OT]:" off topic "[AD]:" ad's