SN74LV8154: 64 bit counter for 10MHz clock

Using plain counters (74xx40x0) would require more chips.

Nowadays, this is solved with a microcontroller.

I'm using a Picaxe 40X2 at 64MHz in the picture to the right, what microcontroller can count as reliably as a logic counter at 10MHz with 64 bit precision while at the same time being able to send the value over a serial connection?

If I connect the Y bus of all counters in parallel to the same 8 pins of a microcontroller and have all register selection pins connected separately it would requite less pins on the microcontroller but would there be a risk of the Y outputs on the counters damaging each other or would it just not work well?

As shown in table 2, when all gate inputs are inactive, the Y output are Hi-Z. So you can connect them together as long as you access only one chip at the same time.

I'm looking at how I could replace the microcontroller with purely logic components which synthesize a serial output just out of scientific curiosity but I'm having trouble with component selection, these components appear identical for instance: CD74HC165 CD74HC166 how do I tell the difference between such similar items with a slightly different model number?

By comparing the datasheets. The '165 has an inverted output in addition to the normal output, while the '166 has a clear input. And the pinouts are different. [edited]

Surely because the triangles on the inputs of the 165 don't have a circle means they aren't inverting?

Hi David,

The inputs of the '165 are not inverted -- there is an inverted output available for the '165 (Q\_H), but it does not include a 'master reset' input.

165 FBD:

-

The '166 has a clear (or "master reset") input, MR\, but it does not include an inverted output.

166 FBD:

I find the function tables easier to compare (165 on left):

I misread Celemens reply.

Hey David,

I read it the same way when I first saw his response -- I figured just clarifying would help.

Thanks for the assistance, if I want the serial output to be inverted but idle low (binary 1) is there a shift register that would do that or would I need additional logic to do it?

I'm not sure what you mean by idle -- the output will be based on the data stored in the register at all times (except at startup, when it will be 'unknown').

When the parallel enable pin of the shift register goes high the serial output transmits the bits then goes low correct?

A shift register is "dumb" so to speak -- it just outputs whatever is in the registers. Each clock pulse advances the values in the registers.

If I want to turn a CD74HC4017 into an 8 bit ring counter can I connect the pin for bit 8 (pin 9) to the reset pin?

Since the MR input is asynchronous that should work, but you'll end up with an extra delay on the 8 -> 0 transition because of the time it takes for 9 to turn on and activate MR. For slower systems I wouldn't see it as a problem, but if you're trying to run at 10 MHz (100ns period), that might be a significant problem. Prop delay from CP to the output is about 50ns max, and for MR to the output it's another 50ns, so you'd lose up to a whole period.

Actually in this use case the serial clock would likely be 1MHz and this component one tenth of that for selecting the 8 bit registers with an added stop bit and 1 clock delay so 100KHz.

Hi David,

It seems that you don't really like any of our answers so we can't close out this thread -- and you keep asking unrelated questions to the original thread. Is there something I can do to resolve this so we can close it?

I'm sorry it got off track I had another idea but I do need to ask seemingly random questions and you are helping, would you rather I go somewhere else?

To be able to track the status of customers' problem, separate questions should be put into separate threads, and be marked as answered when they are answered.

Hey David,

While this forum is similar to others online, I believe it's the only one that's staffed entirely by engineers that are experts on our particular parts -- for example, I have been the world wide expert for TI on the logic portfolio for about 7 years now. Answering questions here is a big part of my job, however I handle a lot of others things internally as well for the logic product line.

My apologies if I implied that you shouldn't be getting help from us - that was not my intent. I'm happy to help - like Clemens said though, we like to keep one question (or set of questions pertaining to a single device) within one single thread, and then mark it resolved when it is completed. 

I should have mentioned this earlier in the thread, but your questions seemed somewhat related and I was ok with keeping the thread together - but it feels like this thread is getting very long with no resolution in sight. It's better if we can have each thread focus on one particular part / issue so that we can have a result for future engineers to find if they have the same question.

When you label something as "not resolved" it remains open in our system and is basically a negative mark against us. I wouldn't want you to mark something resolved unless you felt it was - but I want to do everything I can to get these questions answered. Sometimes the resolution is that we don't have a solution though -- not always something our customers will agree with, but there isn't much I can do about that.

As there is no clear answer to my question I selected the closest one.

Thanks David.

If you have a question that's related, please use the "+ Ask a related question" button at the top of this thread and it will link back automatically. If it's regarding a different device entirely, I'd say to just make a new thread using that part #.