I'm doing a project in my digital systems class where I need to divide a 10 khz clock pulse down to 1 hz. I thought I would use 4 cascaded decade counters for the job.

Then I found this one: SN74LS90N

But I can't figure out how to cascade 4 of these together, because I don't see any ripple carry or terminal count output. No enable input either for that matter.

Does that mean they can't be cascaded? Or is there some simple way of solving it?

Thanks

 

In my opinion this is a specialized counter for specific purposes. For clock division I usually use the 74191.

 

Great! So I bought a few of those, and it turns out I can complete the whole project using only them and NAND gates! The project is to design a control system for a traffic light btw.

I have a little follow up question though, just to make sure I understand the data sheet correctly.

It says that the load function is asynchronous. Does that mean that if I wanted to make one of these divide by 10, I would have to set the data inputs to binary 5 and connect the RCO to load. Whereas if the load function was synchronous I would get the same result by setting the inputs to binary 6?

 

In that case you have two options:

Either start counting from 0 and set the LOAD to trigger at 10. This way you have 0,1,2,3,4,5,6,7,8,9. The very moment the counter hits 10, the count will reset. If the LOAD whas synchronus you would wire the AND gate to trigger at 9, and the LOAD would be effective on the next cycle.

Or: you could try to count up from, say, 5, and use the ripple clock to activate the LOAD. However, I am not sure at wich point of the cycle it actually goes from HIGH to LOW, so you might lose a fraction of a cycle in every count period. A software simulation could resolve the matter.

 

In that case you have two options:

Either start counting from 0 and set the LOAD to trigger at 10. This way you have 0,1,2,3,4,5,6,7,8,9. The very moment the counter hits 10, the count will reset. If the LOAD whas synchronus you would wire the AND gate to trigger at 9, and the LOAD would be effective on the next cycle.

Or: you could try to count up from, say, 5, and use the ripple clock to activate the LOAD. However, I am not sure at wich point of the cycle it actually goes from HIGH to LOW, so you might lose a fraction of a cycle in every count period. A software simulation could resolve the matter.

But won't the clock make sure the cycle is complete? I mean, the moment the counter reaches 15, it will jump back to 5, but then it won't move on to 6 until the next positive edge of the clock pulse, right? So then each cycle should have exactly 10 states that are equal to a period of the clock pulse.

 

The problem isn't when the counter will resume counting. It is that you don't know at which point of the "15" cycle the ripple clock will go to low. It might be at the very end, it might be at the middle of the cycle. And when that happens, the number 5 will be immediately loaded.

 

But either way the 15/5 state will be just as long as the 9 other states, won't it?

 

No, the 5 state is completely independant from the 15 state and will stay for a complete clock cycle.