I have a chip that can provide its own clock by connecting one of its outputs to one of its inputs. But the datasheet states the two pins must be momentarily grounded after powerup to initialize the chip and I want to learn how I can do this with an R/C network if possible.

After experimenting I found it's ok for the two pins to start out grounded at power up and then when un-grounded the chip runs. So I picked values for R and C and hooked them up, but it does not work since that way the pins end up at +5v and the chip isn't running.

I need a way to ground the pins for a moment and then leave them connected only to each other. My next thought was to use the R/C network to switch a transistor but I don't think one can switch from ground to 'nothing'. How about connecting the output via a diode to the input and seeing if my R/C network grounding just the input would work? But that deviates from the datasheet, and I think the +5v going into the input is my problem anyway.

Any thoughts would be appreciated!

 

Post a schematic of your existing circuit; include the part number of your IC and the values used for R and C. Describe the power supply that you are using.

 

The chip is a Nat'l Semi ADC0804 A/D converter:

http://www.ti.com/lit/ds/symlink/adc0804-n.pdf

Page 11 shows the 'Self-Clocking in Free-Running Mode' I'm aiming for.

Here's the schematic. My R and C choices yield 220 msec which I thought would be enough, but maybe it's too short?

The chip sends out a low pulse on _INTR_ after a conversion is finished; when connected to _WR_ that triggers another cycle. As is, the chip does one analog to digital conversion and then stops as the _INTR_ low 'go again' output pulse is being overruled by the constant high R/C output (I think).

Power is +5v coming from a regulated supply. Thanks!

 

Try using 50k (47k to 51k Ohms) and 1uF instead. Don't know how long your supply takes to go from 0v to 5v; if it has a slow turn-on time due to large filter caps in the output, that would effectively decrease the reset time available by the RC network.

You could also use a logic-level N-ch enhancement mode MOSFET like a 2N7000, with it's gate capacitively coupled to the +5v supply, and a ~100K resistor from the gate to GND; your leads to short to GND on the drain, and source connected to GND. That will momentarily turn the MOSFET on during power-up, shorting the leads to GND; the 100k resistor will discharge the gate after a period of time, turning the MOSFET off.

 

Thank you. I'm leaning towards the MOSFET since I think even with a longer delay, the logic high once the RC network is charged will interfere with the low pulse trying to get from _INTR_ in to _WR_.

So, how does this look: