Why is it secret? I hope it's not another HHO device as these sort of shenanigans often turn out to be.

 

Because there's always the chance that no one's ever done it and I want to be able to have the choice to patent it if I so desire. In fact i'm fairly sure no one's done it, at least not the way I have.

 

But if you really wanted to know, the simple explanation is that the device is a lock and key. I want the key to be separated from the lock with no feedback as well as amplify enough to put a zener diode in it so that changes in voltage do not change the lock's sensitivity.

 

Good luck, I'd avoid a patent as well... Google "the case against patents"

Also what's the 32kHz for?

PS optocouplers are for signal isolation. You can't send any practical voltage through them.

 

The 32khz forces a similar oscillator to oscillate at that frequency. Practical voltage quite depends on how much is needed to do a task. About 100mV is enough to do what I need. It's the frequency that matters more than the amplitude. So, any ideas?

 

Is this an existing electronic lock? If not is there something similar you can show us an example of?
Here's a popular electronic lock, it doesn't use frequencies but instead uses a 1-wire crypto-lock iButton. Is there anything new your design brings to the table?

 

If you "key" simply depends on coinciding crystal frequencies, it is hardly very secure.

 

If you "key" simply depends on coinciding crystal frequencies, it is hardly very secure.

Very true, modern locks are more along these lines.
http://en.wikipedia.org/wiki/KeeLoq

 

Well, it has a 2Hz bandwidth per crystal where it will open. However, the final design will be 4 sets of matching crystals that could be essentially any value from high Hz all the way to about 500kHz or so. And you'd have to get all 4 right with no feedback at the same time. Considering rise times and other natural delays in the lock, there's about a 1/4 second delay between input and response, which limits frequency sweep. Another positive is that it cannot be hacked in the traditional encoding sense because it's not encoded at all. The whole idea was to create an analog electrical lock that is more difficult/time consuming to hack than whatever it is locking.

 

Can repliers please post useful responses to my question? This is not about the viability of my project, but more making it work. It matters for graduation afterall. The idea behind why it's hard to break is simply this: if you have a 10 digit decimal code, then you have 10^10 possible combinations. With my lock, if you have 4 crystals, you have a range possible per crystal of about 500khz. With a 2hz band, that's about 250,000 possible bands (it's analog, but for simplicity, assume it's discrete). Then it's 250,000^4 combination at about 1 choice every 1/4 second. That is a long hack time.

 

Those iButton crypto key locks are extremely difficult to crack, they are extremely popular and millions have been sold worldwide. The iButtons require no power (parasitic) and are waterproof and almost indestructible and cost only a few dollars per key & fob.
I don't see your system bringing anything new (seems to be a step backward) to the market and marketing it would be a nightmare.

Ahh it's a school project.

 

In other words, does it matter to you what happens with it after I finish? Probably not. The question as posed is: amplify a 32,768khz signal from 1.25VPP to 3VPP with the right DC offset to power an LED to send the frequency signal across. I'm still fairly sure it hasn't been done. Sure, I probably could have done it with a micro-controller (the lock) and been simpler. But I'm poor and don't have any funding to buy the development board. Therefore I came up with a way around it that works. All I need to finish my project is this amplifier.

 

I'm using as many free parts as I can that the school provides for me. I still think the idea behind the lock is cool. Being an analog device inherently makes it more difficult to hack than a digital lock. Just by decreasing the frequency range slightly, I can make it magnitudes more difficult to break. I have neither the care or patience to make it more accurate than 2hz. And all my prototype will have is 4 sets of 32.768khz crystals because they are cheap and the oscillators were made for them. It will prove the concept however.

 

Finally, blue, I am an engineer, not a businessman. The most I hope from this project is that it will help land me a job in 8 months. I want a working prototype to be able to use in interviews and just because. There was no need to post someone else's lock because I'm not competing with me. If I decided to get a patent, it would be either to prove I made something or to sell to someone else to develop further.

 

Sets of crystals matched well enough to make that work consistently would probably come expensive, compared to the relatively cheap components used in a digital lock.

 

Expensive as in about $2.50 a crystal or about $20 in the lock. Again, not my problem right now. It all works with gain. If crystal A has a certain frequency and crystal B has a close one, there will be some gain associated with it. However, with a simple potentiometer, I can reduce to gain bandwidth as desired so that only frequencies very close have enough gain to start the internal oscillator. It's magic.

But in all seriousness, no one has any actual comments to help with the problem I have. I need a way to amplify a 1.25V square wave that is between 0 and 1.25V so that it is about 3VPP and still can turn on an LED.

 

Hard to help without you posting your schematics.

 

I posted all I have. An optocouple and the oscillator with a blank in the middle for some kind of amplifier. There's nothing I didn't post in my first message that is needed to visualize the circuit. Once the amplifier works, I think I can get the zener diode part to work fairly quickly.

 

Well good luck with your project. My crystal ball just isn't working today.

 

Then you shouldn't have replied. I asked for suggestions and help on what amplifiers may work that I haven;t tried or things to look at for why the ones i've tried may not work so I can look into it. You've giver nothing and I've explained and shown all the diagrams i could find.