encoderaudio:
That is right Chris, that is the target.7 ms...you know, I just sad something. 10 ms is also OK.
Modern Buchlas using 4 ms triggers, in Eurorack modular the average is (I would say) 2-10 ms.
Basically, 7 was just a number, not critical, but it should be constant.
Ok, and that's all you need to do is read a pot, and output a pulse stream, right? If that is the case, then I would go ahead and use delay(), and burn this onto an ATTiny 85 chip (a smaller version of the Arduino's brain chip.) I would also use an external, temperature insulated ceramic resonator instead of the internal oscillator. The code is almost trivial. Spend another day looking at threads about blink-without-delay/millis-based-timing on this site. You are already so close.
I realize that using delay - ever - is controversial. But in this case, where the output need never go faster than the natural frequency for this timing, AND the input doesn't need a faster rate either * it's fine. You've got a lot of syntactical errors in your code. Fix 'em.
Now, if you want to run anything else on this device - temperature compensation, additional user input, any other internal analysis or control of the output, then you will probably need to vanquish the delay! So write your code with the seven second delay with your output in the high state hard coded with delay. But work toward keeping even this under software control. This is the essence of what AWOL is advising.
ps, this sounds like a great project! I have a littleBits/Korg modular synth, with over 200 pieces, including 2 Leonardo Arduinos and 3 'cloudBit' Pis. ... Mostly way to complex for me! Your simple pulse device seems downright comfortable.