Well, it looks like I've stumbled into here again. Most of my previous projects have ended up failing, but this one I think has a chance of success.

I have a small alarm clock that I can mount comfortably within my computer, then I can take off a small drive cover and access it easily, which is perfect. After a while of poking around inside the alarm clock, I've found that there's a constant voltage of ~1.5v (from the AAA battery) going through the alarm speaker, and then when the alarm goes off that changes to a (presumably DC) waveform. Knowing this, I hooked it up to a transformer which I believe has equal number of windings on both sides. Then, when the alarm turn on, the other side of the transformer has current, but not when the alarm is off.

The problem is, I need to use this to drive two C1815 (NPN) transistors.
The first transistor was linked into the switch that turns the alarm off, so that my computer doesn't turn on then off again over and over again. The second was hooked up to the two power button pins on my computer mobo.

The power button pins on my computer work when the +5v pin is connected to any ground. The 0v ground pin of the power button is pretty much irrelevant. Since then, I have switched them around to where the second transistor turns off the alarm, and the first and second are both wired to the alarm pins. When the alarm goes off (it turns on for probably about a fourth of a second, then turns off) my multimeter read 4v going through the switch in the alarm. This leads me to believe that the transistors are switching the current correctly, however the computer does not turn on, nor will it turn off if it's already on.
I don't really understand how that's possible, but since the C1815's (and I believe most transistors) take a base voltage of 5v, not 1.5, that it's probably not enough power to fully switch the transistor.

Any tips on why it might not be working? If I was correct and there's not enough voltage, any idea how I could fix that?

I don't know if they make transistors that will take a base current of 1.5v, or if they make step up transformers that will step 1.5v dc up to 5v dc.

I really want to get this working =P

There are programs to start your computer at a given time. Many newer computers even have this built into their BIOS.

RDP

So what you're saying is you want me to start up my computer, wait for it all to boot, load all my startup programs, just so I can set it to turn on, then turn it back off?

That sounds like a smart idea :rolleyes:

And I don't feel like bothering with tons of bios menus each time.

I'd rather just set an alarm and leave, not to mention it's much cooler.

That transistor idea (no current when voltage is constant) is really clever. Is there a way you could modify the source voltage (from the alarm clock) to be 5V instead of 1.5?

I don't know if this is a wall-powered or battery-powered alarm clock, but if it's a battery powered one (and its batteries total to more than 3.5V) this'll be easy. If it's mains-powered and the primary transformer is more than 3.5V, this will also be easy. But you'll run into trouble if it's less (i.e. if the whole damn thing is powered by a 1.5V power source). I think augmenting the power from the alarm side is the easiest way to go. Trying to boost the power AFTER the first transformer seems difficult.

Otherwise, you could make a rube-goldberg-style alarm clock that physically presses the power button on your computer! That would be really fun!

First, it is easier to make pot out of baking soda (with or without Raid-soaked lima beans) than it is to get 1.5 volts of current in any circuit. Voltage is measured in volts; current in measured in amperes!

Let try to give you a few tips:

Whatever signal you are getting out of the transformer is AC, not DC. It is nice to know that the transistor is turning the alarm off. This is a good indicator that you are on the right track.

Small signal silicon transistors will start to turn on at a Vbe of around 0.6 volts. I can only grasp a vauge idea of what your circuit looks like based on the description you present. Are you simply paralleling the base and emitters of the two, or have you incorporated seperate base resistors for each one? Simple paralleling might work, but is not a very sound design.

More directly to your 'problem' goes back to the AC/DC (no, not the band) issue. It is likely that your alarm has some debounce circuitry that allows it to see the pulsed output of the one transistor, but the PC does not.

I would suggest connecting the transformer output (for the PC transistor) through a diode (like a 1N914 or 1N4148) anode to coil, cathode to a capacitor (~1uF) then through a resistor to the base of the transistor. This should give you a steady turn-on of the PC control.

The only remaining problem I can anticipate is a potential time-race... where the clock might turn off before the PC has time to recognize the start signal. This can be fixed easily enough, but I don't want to get too far ahead of you at the moment.

So what you're saying is you want me to start up my computer, wait for it all to boot, load all my startup programs, just so I can set it to turn on, then turn it back off?

That sounds like a smart idea :rolleyes:

And I don't feel like bothering with tons of bios menus each time.

I'd rather just set an alarm and leave, not to mention it's much cooler.

First, the BIOS settings are a one time thing and take around a minute to do, second the programs are all set to automatically re-occur just as the BIOS would. Sorry for suggesting something that is guaranteed to work!

RDP

But you'll run into trouble if it's less (i.e. if the whole damn thing is powered by a 1.5V power source).


It's running off a single AAA battery. No luck there =(


I would suggest connecting the transformer output (for the PC transistor) through a diode (like a 1N914 or 1N4148) anode to coil, cathode to a capacitor (~1uF) then through a resistor to the base of the transistor. This should give you a steady turn-on of the PC control.

Ok, I'll try this some time soon.

I can provide a picture of my circuit if you'd like, though it's kinda messy.

First, the BIOS settings are a one time thing and take around a minute to do, second the programs are all set to automatically re-occur just as the BIOS would. Sorry for suggesting something that is guaranteed to work!

RDP

I don't need it to turn on at the same time every day, it varies. Sorry for being an ass in my reply post, I was in an ass mood that day.

bet you could even find a program that turns your computer on at a certain time and you would have more space for something other than a stupid alarmc lock.

The diode idea didn't work. One side of the transformer is connected to the ground of the alarm clock battery, and the other side is connected to the first transistor. I soldered the diode facing into the transistor from the transformer. =P

Yeah, the diode alone won't do squat without the capacitor.

How about you post a sketch of what you are doing? Even if it is a bit 'messy'.
That'd be a big help.

Maybe a relay.

5v to close the relay on one end making a connection on the other where the 12v supply is. You could also possibly have a couple wires running off of it to close the connection with the button that stops the alarm. Capacitor usage probably but RDP would be the best person to work that out.

Or am I talking out of my arse? Hope not.

The relay idea is perfect, and I thought of it too, except that that the alarm is only 1.5 volts, which is not enough to switch the relay. I could wire it with a transistor to use to computer's +5v on the pin, but I tried that and it's not enough to switch the relay I have.

I can get a capacitor and try that later, and I'll also post a sketch of the circuit.

http://www.improvementscatalog.com/home/improvements/792928169-digital-light-timers.html

They make digital timer like the one above for christmas tree lights and plants.

Google search terms:
"westinghouse holiday"
"westinghouse digital timer"
"digital bar timer"

I imagine you like your current alarm clock and only want to set 1 timer a day and just want your computer to warm up before your alarm clock goes off. That's not gonna be easy. I would probably try and find an audio triggered relay like a "clapper" and fuck with the microphone until the alarm clock sets it off.

If you only have a couple different times you wake up: then multiple timers on a power strip could be turned on or off individually based on what day it was.