capacitor charging timer

[lhlblue]

i want to realize a timer using capacitor charging method, not by a osillator. that is, i use a current I1 to charge a capacitor C1, according the relationship: I1*t=C1*V1, then the time t is proportional to V1; however, for a long time timing, for ex, 10ms, then, I1 will be very small, or C1 will be very large(cannot be integrated in the chip); for I1=100nA, t=10ms, V1=1, C1=1nF, 1nf is too large to integrate in the chip inside. then, is there any way to solve this problem?
i know a method called capacitor boosting, but if i boost a capacitor by 1000 times (1pF --> 1nF), then amplifier gain will be 1000, for such a big gain, the power consumption will be large, and the resisitor area will be large too(10K vs 10M), and any other methods?
thanks all.

 

[Electro nS]

i know from filter design there is a problem of ingrating coils capacitors and big resistors so :
there is something called switched capacitors , that is using transistors to make integrated capacitors

and thers is transconductance or something , that uses transistors inverse gain to generate virtual resistors ..
check these out

 

[ElectroxX]

Use an RC series circuit:
https://en.wikipedia.org/wiki/RC_circuit

1. Pick a rather large R as that will reduce the circuits energy consumption. It could probably be as high 0.5-1.0 Mohm, if you can find a suffiently small capacitor to match it.

2. If you apply a voltage U it will take approx 5*R*C seconds until the voltage across the capacitor is U, provided the voltage across C is 0 at start. That means 5*R*C is your desired delay T.

3. Solve for C. C = T/(5*R).

4. In case you cand find a suffiently small C, pick a smaller R and try again.

5. You will have to reset the delay somehow. Use a digital I/O and a series resistance if youre using a microcontroller.

Note that parasitic board capacitance will generate a larger error when C is smaller.

 

[erikl]

... I1 will be very small, or C1 will be very large(cannot be integrated in the chip); for I1=100nA, t=10ms, V1=1, C1=1nF, 1nf is too large to integrate in the chip inside. then, is there any way to solve this problem?

You could scale down the current to 1nA; even 100pA working currents are used nowadays.

On the other side: why not use an external cap? It's so cheap and costs just one more pin. If you want to hide it, consider a hybrid solution in the same package.

BTW: there are silicon caps available with much higher cap values, made with special 2D or 3D processes: ipdia.com.

 

[Electro nS]

.............
............

.

nice name , very Original LOL

 

[erikl]

nice name , very Original LOL

Like yours, Electro nS . Probably abbreviates nanoSiemens, isn't it? Or nanoSievert? ;-)

 

[Electro nS]

Like yours, Electro nS . Probably abbreviates nanoSiemens, isn't it? Or nanoSievert? ;-)

hehe i was being sarcastic , what are the odds that a person answer the post after you with a similar name (liked the "xx" , sexy electronics ?? probably )
, i really wanted to name Electron (the small particles in atoms that move) and Nanosecond (high speed) so i combined both !!

 

[erikl]

i really wanted to name Electron (the small particles in atoms that move) and Nanosecond (high speed) so i combined both !!

Right so, fast electron! But shouldn't you have used a lower case "s" then? ;-)