I'm with Hero999 on this one: Use the CMOS (TLC) version.
If your circuit can't be powered from a wall wart, then batteries are a given. The best option (for the enviornment and your wallet) is rechargables, and that means the most bang for the buck: LiPos.
Problem is, LiPos are a pain to recharge, and to use one with a standard 555 means getting a two-cell version.
Unless you use the TLC555! It can be used with 3V cells - which is pretty much any LiPo. Since the advent of R/C cars and helicopters, small single-cell LiPos can be had for little money at most any hobby shop, and the chargers available are inexpensive and versatile (see Dave's video #397). If you already have these supplies, then just slap in a power connector to match your battery pack, and go!
Adafruit LiPo:
http://adafruit.com/products/258Adafruit Charger:
http://adafruit.com/products/280Now, you're not going to want to swap that battery out every day, or every other day, for that matter. So, low power draw is essential. Another CMOS advantage. I breadboarded the attached circuit, configured for a 33mA load (from your measurements), and played a bit with the values...
The pull-down resistor worked up to 10Mohm (the largest I had), so the trigger current isn't a problem, even at 3.3V (I was running from an 1117-33 regulator). Problem was, I was reading a draw of over 4mA. Until I realized I had hooked up my meter
before the Vreg. Once corrected, the quiescent current dropped down to an expected 90uA level, jumping to about 35mA during the timing (LEDs on).
Advertisement:
When playing around with CMOS timers, measuring current draw can be a pain with low-resolution meters! That's why there's uCurrent! Available at Aussie Platypus dens and http://adafruit.com/products/882 finer Chic Geek Boutiques!Now, we can also reduce the current through the R1/C1 tank by making R1 as large as possible, and making C1 lower as a result. The problem with that is that even the TLC555 can only use caps down to a certain value before it hits the frequency limit (~300KHz for NE555, ~3MHz for TLC555). The other problem is that if you have a leaky cap, say 1uA or so, and you've limited the current to charge it by using a high value for R1, say 0.8uA, then the cap will never charge enough to trip, and it'll be forever timing.
Having said that, the circuit I breadboarded, using the values shown, gave a working timer with a delay of just over 30 seconds. Using the previous 1300mAH battery above, and figuring that we'll be recharging it at about the 50% level...
.65AH / .00009AH = 7,222 (hours), or 300 days.
Of course, it'll be drawing 35-40mA every time the door is opened, and doing that for a minute every two openings (with my values), so...
4mAminutes x 3 (six openings per day) = 12mAminutes / 60 = .0002AH per day.
100 days = .02AH, 200 days = .04AH, and 300 days = .06AH, so:
@300 days, we're at 650mAH - 60mAH = 590mAH.
Wow. The self-discharge rate of the LiPo should be higher than that, so I'd just recharge every 2-3 months just to be sure.
Also, you may not want to use that BD201. All transistors have a leakage value, and the power types tend to be leakier than the small-signal or general purpose types. Try your (or my) circuit again using a common 2N3904 or 2N2222. Since you're only drawing a 35-40mA load, darn near any transistor can handle it - no need for one of those big boys. It's probably where that extra 150uA went. Note also that the TLC555 can only source/sink about 3-4mA @ 3V, so adjust the value of the base resistor accordingly to provide enough bias for saturation for your load. A 1mA input will allow 100mA or so through a 2N3904. Plenty.
nop
Although not designed with the TLC version in mind, this 555 calculator doesn't mind shoving high values into it. The calc output isn't exactly spot-on, but it's close enough to get you going...
http://freespace.virgin.net/matt.waite/resource/handy/pinouts/555/