Divide by N

[3v0]

You are arguing in favor of old technology. It may be fun and educational to use but is not practical.

 

[Mikebits]

It sure is practical, and still in use, of course CPLD's have replaced many of the discretes, but even then, there are times when a simple divider is all that is needed.

Consider a factory building a widget. One option is take circuit board using discretes, solder parts, and done. The other is take circuit board, solder parts, send board to another stage to program PIC parts. This adds a lot of cost, and then to maintain a configuration control on the code, what happens when the uP becomes obsolete, someone has to re-write the code perhaps, debug the code, then release it for production.

What do you do if you need a divide by N at say 300 MHz? How many PIC's operate at the frequency?

Bottom line, I think it is important to learn basic design, and no when to use a particular technology.

 

[mneary]

Consider a factory building a widget. One option is take circuit board using discretes, solder parts, and done.

Most devices need to be tested at least once after assembly. Testers can usually program, too.

 

[3v0]

Bottom line, I think it is important to learn basic design, and no when to use a particular technology.

I have not disputed this.

What do you do if you need a divide by N at say 300 MHz? How many PIC's operate at the frequency?

There are other programmable parts that do.

The obsolescence problem exists with all parts including logic chips No technology is immune. With the uC there is a very good chance that a pin compatible replacement uC will be provided.

A uC typically replace many logic chips. Even with the expenses you site there is a cost savings due to reduced PCB size, chip cost. It cost less to assemble a smaller board with few parts too.

But if a logic chip is all you need by all means use it.

 

[Mikebits]

Well, you make some good points, and I cannot deny the facts that even in the hardware world that coding is becoming the trend. Be that as it may, even with VHDL and other design languages, ASIC designers will still need a basic knowledge of digital design to properly implement their code.

I suppose the need to understand digi logic has been superseded by the power of the PC, and this is as it should be, and is a good thing, as we advance our tools and make life easier.

But I still contend, that for a casual circuit builder that wants to make a simple divider circuit that should take a hour to construct, why in bleeps sake would you have them order $50. worth of supplies, weeks of reading just to divide a signal. In this case, simple is best, not to mention they learn a little about the hardware they are using.

Once they can grasp concepts like And gates, Or,NOR, etc, and D-FF, then they will become a much better PIC designer.

I hope you realize my tone is not aggressive, and I harbor no hard feelings, nor do I wish any in return, just voicing my thoughts in a free exchange of ideas

 

[Nigel Goodwin]

But I still contend, that for a casual circuit builder that wants to make a simple divider circuit that should take a hour to construct, why in bleeps sake would you have them order $50. worth of supplies, weeks of reading just to divide a signal. In this case, simple is best, not to mention they learn a little about the hardware they are using.

Which is why I suggested a PICAXE, no programmer required, extremely low cost, and you could learn it in an evening.

 

[ArtemisGoldfish]

Circuit: Now with crunchy pieces of MSI and discrete logic.
U1 and U4's supporting logic act to reset them and increment the next significant counter. U6 acts as a Divide-by-5, and it's supporting logic drives a D-type flip-flop that acts as a Divide-by-2. Someone with experience tell me if that would work ;-)

 

[ArtemisGoldfish]

Okay, I'm stupid, I way over-complicated things. Here's a new schematic with a Divide-by-500 glued together, and a Divide-by-2 to latch it. Nevermind the part numbers obviously, you'd have to use matching TTL or CMOS parts to equal out the logic levels.

 

[ericgibbs]

Okay, I'm stupid, I way over-complicated things. Here's a new schematic with a Divide-by-500 glued together, and a Divide-by-2 to latch it. Nevermind the part numbers obviously, you'd have to use matching TTL or CMOS parts to equal out the logic levels.

hi AG,
If you havn't started the buy/build yet, look at the Intersil ICM7217I.
Its a 4 digit BCD counter with a preload compare register and an Equals output, etc.
If required it will also drive a 4 digit LED directly.

Datasheet from DatasheetArchive :: Free Datasheet Search Engine :: Electronic Components

The 74HC93/4020's in your present circuit are asynchronous counters, you could have problems with gate glitches.

 

[ArtemisGoldfish]

hi AG,
If you havn't started the buy/build yet, look at the Intersil ICM7217I.
Its a 4 digit BCD counter with a preload compare register and an Equals output, etc.
If required it will also drive a 4 digit LED directly.

Datasheet from DatasheetArchive :: Free Datasheet Search Engine :: Electronic Components

The 74HC93/4020's in your present circuit are asynchronous counters, you could have problems with gate glitches.

A: I'm not building anything, just trying to show the OP what logic you'd need.B: What do you think of the revised one with a single counter and only 4000-series logic?

 

[ericgibbs]

A: I'm not building anything, just trying to show the OP what logic you'd need.B: What do you think of the revised one with a single counter and only 4000-series logic?

Hi AG,

Sorry AG, its a case of my typing finger being quicker than my eye...

I think the 4020 ripple counter glitches will cause misoperation of the 4013 F/F.

You may have to add some simple R/C filters on the outputs of the 4020.

 

[chemelec]

Hello board ^^, just recently someone got me involved in a project where they are making hydrogen gas for their car with electrolysis. They got me building all the electronics for it, and they are doing the mechanical aspect. I have everything about 80% done, expect for one part:

One sub-circuit requires a divide by N where 1000 input pulses = 1 output pulse. The circuit is using a CD4059, the problem is that the schematic only shows a block diagram for that part o.o.

The chip itself is considered obsolete, I managed to order 5 of them, but I can't figure out how to get it to divide by 1000?

If anyone can help me, I will greatly appreciate it. Also, a modern or different way of doing it is welcome, but as I said I have 5 of the CD4059's.

Thanks

Here are the Connections for the 4059 in that Hydrostar Circuit.
But your Totally Wasting your Time and Money on this.
The Hydrostar DOESN'T WORK.

**broken link removed**

 

[3v0]

Well, you make some good points, and I cannot deny the facts that even in the hardware world that coding is becoming the trend. Be that as it may, even with VHDL and other design languages, ASIC designers will still need a basic knowledge of digital design to properly implement their code.

I agree.

But I still contend, that for a casual circuit builder that wants to make a simple divider circuit that should take a hour to construct, why in bleeps sake would you have them order $50. worth of supplies, weeks of reading just to divide a signal. In this case, simple is best, not to mention they learn a little about the hardware they are using.

Yes and No. I think that programing at some level has become (or soon will be) a required skill for anyone designing digital circuits. The programmer and parts should, maybe will, be on hand. It is realistic to expect these people to learn how to program rather then work around it.

But everyone has the right to do, what they want, their own way. As long as they are not working for a person who has final say.

I hope you realize my tone is not aggressive, and I harbor no hard feelings, nor do I wish any in return, just voicing my thoughts in a free exchange of ideas

No problem.

 

[Leftyretro]

Valid points made in this recurring argument of programmable chips Vs standard logic functions, by both 'sides'.

The main thing is that the tech has made it possible to have a single 8 pin pic chip (picaxe 08M in this example SparkFun Electronics ) available for $4 that can be made to replace a large number of simple logic functions. The software to program the chip (in a simple basic language) is free from the manufacture and the 'programmer' function is simply two resistors and a comm link to a PC. Programming is a one time effort and not needed after the function is placed into the chip. If the function required requires more input or output pins then there are picaxe chips available in several size and cost points all affordable.

The small pic chips have kind of become this centuries equivalent to the famous 555 timer chip as far a wide range of applications and low costs.

There may always be special needs for dedicated logic function chips because of super high speed or non standard voltage levels but most application can utilize a modern pic chip very efficiently.

PS: I guess one argument against using pic chips would be a single source supplier rule. Many companies use to not utilize a chip unless there were second sources for the component. At least that was a design rule used a lot in the 60s and 70s. Now a days that rule seems to not be the rule a all!

Lefty

 

[NEWBY08]

Im in the same boat with this problem, can anyone tell me the pin configuration to make the cd4059 divide by 1000, ive got the data sheet but im only new to this game!!!! please help:

 

[chemelec]

Im in the same boat with this problem, can anyone tell me the pin configuration to make the cd4059 divide by 1000, ive got the data sheet but im only new to this game!!!! please help:

WHY DON'T YOU READ MY PREVIOUS POST?
AND CLICK ON THE LINK!

It Shows the Pin-Out Connections of the CD4059 to Divide by 1000.

"Not that it will Help You, as the HYDROSTAR is CRAP"!

 

[JScaner]

CD4059A Pins connections

Hello Chemelec,

I saw you schematic drawing on the CD4059A pins hookup to 12 Volts and grounds to make N = 1000.

Now do you wire these pins connections perminantly ?

I heard people said it takes three clock cycles to program the chip, then removed the connections. Which way is the correct way ?

Thanks,

Johnny

 

[chemelec]

Hello Chemelec,

I saw you schematic drawing on the CD4059A pins hookup to 12 Volts and grounds to make N = 1000.

Now do you wire these pins connections perminantly ?

I heard people said it takes three clock cycles to program the chip, then removed the connections. Which way is the correct way ?

Thanks,

Johnny

The Pinout I show is Correct for the 4059 to divide by 1000, and Yes you wire it permantly.

But WHY would you bother with this??
The "HYDROSTAR" CIRCUIT IS CRAP And DOESN'T WORK AS STATED.
Its a RIP-OFF!

If you bought the plans for this, Try to get your money back!

 

[JScaner]

Cd4059a

Thank you for the information on permanentpin hook up.

I guest I gave you the wrong name, the circuit that I am building is call REStar PWM.

Someone has made over ten modifications to the circuit, now is called restar pwm circuit... this circuit has both fuel cell and torroid coil circuit build into one.

**broken link removed**

I have just completed building the circuit, fine tuning the frequency on the fuel cell to 20 Khz, but when i got to the coil circuit, it used CD4059A.

Measure the output frequency on the coil circuit and it read 20Khz, so the N = 1 right now.

I will hook these pins out, hopefully will get 20 hz out for the torroid coil.

Thanks for the information.

Johnny

 

[davidbball13]

i thought that a CD4018 was a divide by N

as shown here (you have to scroll down about half way)
**broken link removed**