The 74xx4060 Counter Chip

[German_EE]

At first glance this chip appears to be really useful, a 14 stage binary counter and an RC/ oscillator in the same package, but some weird decisions have been made:

1) The Q0, Q1 and Q2 counter outputs are missing, you therefore can't divide by anything below 8 and you can't divide by odd numbers.

2) The Q10 output is missing so you can't divide by 1000 using a single chip (this would have been really useful)

Does anybody know the history of this chip and have any idea why these specific outputs are missing? I've tried contacting NXP support and they have no idea.

[MrSlack]

This is basically the same as an old CD4060. The Q1,2 outputs (/2, /4) are missing because it keeps package size down I suspect. Annoyingly I find that it's hopeless as a timebase as your average watch xtal is 32.768KHz so the max divider output is 2Hz from that. So you want 1Hz? Have to use half a 4013 as well. Another package on the board.

There's a 4040 as well which has more divider outputs but is less flexible with oscillator decisions as it only has a clock input so you need at least one external inverter to throw a clock source at it.

I think this sort of shit is why everyone uses non 4000 series 74hc these days.

[nali]

Yep they predate NXP by quite a long way. Not sure when it was released but here's a link to a 1983 RCA data book (page 611)
http://www.introni.it/pdf/RCA%20-%20CMOS%20Ic%20Databook%20SSD%20250C%201973.pdf

I suppose they just put as many outputs as a "standard" 16-pin package would allow

[MrSlack]

The first 4000 series appeared in about 1968!

[danadak]

Cypress PSOC, just design with GUI or Verilog or both, make it anything you
want, Up, Down, Bidirectional counter, and assign (again GUI) routing to pins.
OpAmps to create your own oscillator, or extensive clk creation off of master
precision clock, like external xtal, or internal +/- 4% clock. Then using mux
and com onchip components shift counter value out serially or parallel or
USB.... So many options....component list of internal functions available,
attached.

Share counter pins with other I/O needs, add another mux, declare pin as
bidirectional....

Dev SW tools free, low cost boards.....

Regards, Dana.

[dannyf]

How fast would the cypress go?

[danadak]

The spec sheet says for 16 bits 59 Mhz, 32 bits 41 Mhz.

Counters with more features, Up, Down, parallel load, capture, run around
33, 10 Mhz for 16 / 32 bits respectively.

Regards, Dana.

[dannyf]

The spec sheet says for 16 bits 59 Mhz, 32 bits 41 Mhz.

Counters with more features, Up, Down, parallel load, capture, run around
33, 10 Mhz for 16 / 32 bits respectively.

How fast would a typical 4060/4040/4020/... count?

[danadak]

Google the datasheet for answers.


Regards, Dana.

[Gyro]

They're really useful for making simple, low current, long duration timeout switches. Sleep timer or charge timeout type applications.

Edit: Actually vanilla CD4060s are even better due to their extended supply voltage range.

[dannyf]

"Google the datasheet for answers."

It is not about me knowing the answers - I already do.

It is about getting YOU to see why sometimes a 4060 is more desirable than a psoc.

Hope it helps.

[MrSlack]

They also don't require programming and cost bugger all!

[danadak]

"They also don't require programming and cost bugger all!"

Yes, in the case of simple logic functionality like I showed, if no further
general functionality is used other than counter takes 3 lines of code to
start and run.

Yes, has to be programmed, once to work.

Price, if using PSOC 4 < $ 1 typical, depends on how much other circuitry
you want to pull inside this fantastic series of processors with analog and
digital and DSP and com and display and..........

In general, if all you have in a circuit, on the board, is a 4060 go with the 4060.


Regards, Dana.

[danadak]

"It is not about me knowing the answers - I already do.

It is about getting YOU to see why sometimes a 4060 is more desirable than a psoc."


CD4060BC max clk rate 5V     =   4 Mhz
MC74HC4060A max clk rate ~= 25 Mhz

Not sure why you wanted me to see these inferior clk rates.


Regards, Dana.

[German_EE]

Thanks for the replies so far but I would like to point out that the 74HC4060 has the same pinout as the old CMOS part. As for using the configurable Cypress part I have a straight choice:

1) Buy a few blank chips, learn how to program them then burn the new logic into them using an as yet unknown process.

2) Walk 2Km down the road and buy a 74HC4060 for 61 cents, 50 cents each if I buy more than 5








It looks like the reason for the missing output ports is now lost in the mists of time.

[edavid]

It looks like the reason for the missing output ports is now lost in the mists of time.

The reason some outputs are not connected is that when RCA designed the CD4060, they didn't have any package larger than 16 pins.

The reason they didn't connect the lowest outputs is that it's not so likely you would need both the lowest and highest outputs (especially since it's a ripple counter).

The really mysterious one is Q10 (which is the divide by 2048 output, not 1024).

[danadak]

"2) Walk 2Km down the road and buy a 74HC4060 for 61 cents, 50 cents each if I buy more than 5"

Totally agree, if all you have in design is a 4060 go for it. And in my case your method would give
me some good exercise, which at my age is most valuable.

The reason for working with tools / programmability these days is driven by market demand
and integration as we all know. One only has to look at the decline in discrete logic sales, the
obsolescence in many logic families, to realize their days are numbered. But then I have to remind
myself that the 8051 UP, still in production in various flavors, since 1981. But its days too are
numbered, for sure, in light of vastly better cores and silicon utilization.


Regards, Dana.

[Kleinstein]

In many applications you don't need  the low divider outputs, there are other chips that offer less decision. The choice is somewhat odd, but this what you get for a few cents and very little power.

In some applications a small µC might be an alternative to day, but this is usually replacing more than just the counter. Though hard to beat a 20 cent chip if the 74HC4060 does the job.

[Zero999]

CD4060BC max clk rate 5V     =   4 Mhz
MC74HC4060A max clk rate ~= 25 Mhz

Not sure why you wanted me to see these inferior clk rates.

All right, I'll give you a clue:

CD4060 maximum clock rate at 15V =  10MHz
MC74HC4060A maximum clock rate at 15V = 0Hz + lots of smoke.

[SeanB]

Simple reason why they did not have all the outputs was to fit it in a DIP16 package. When it was designed the next step up from DIP16 plastic parts was a DIP24 which was both mostly available in a CERDIP package ( where the package alone cost $5, without the lid and the wire bonding to fit the chip in it)  and also would need a separate packaging line to do this single part. Thus the reduced pinout to reduce the number of outputs.

Remember that TTL was also constrained by package size, the more complex logic was only available in DIP24 and DIP28, or things like microprocessors which came in a DIP40 package, and which were a very large step up in price over the small 14/16 pin devices, both because of the larger package and the large die inside there, which needeed the large package. Package size constraints meant you got either 4 2 input devices, 3 3 input devices or 2 6 input devices, or a single function ( plus an inverted output) in there. Only recently do you get a single logic gate in a package, or a microprocessor in a DIP16 package, the size of the die dropped so much that they were able to fit in the constraints of the package.

[Zero999]

Simple reason why they did not have all the outputs was to fit it in a DIP16 package. When it was designed the next step up from DIP16 plastic parts was a DIP24 which was both mostly available in a CERDIP package ( where the package alone cost $5, without the lid and the wire bonding to fit the chip in it)  and also would need a separate packaging line to do this single part. Thus the reduced pinout to reduce the number of outputs.

Remember that TTL was also constrained by package size, the more complex logic was only available in DIP24 and DIP28, or things like microprocessors which came in a DIP40 package, and which were a very large step up in price over the small 14/16 pin devices, both because of the larger package and the large die inside there, which needeed the large package. Package size constraints meant you got either 4 2 input devices, 3 3 input devices or 2 6 input devices, or a single function ( plus an inverted output) in there. Only recently do you get a single logic gate in a package, or a microprocessor in a DIP16 package, the size of the die dropped so much that they were able to fit in the constraints of the package.

That doesn't make any sense. The CD4060 is CMOS, not TTL and there are plenty of other parts in the family which have 16 pin packages: CD4046, CD4063, CD4017 and interestingly the CD4022, which could be easily in a 14 pin package but it's 16 pin with two not connected.

http://www.ti.com/lit/ds/symlink/cd4046b.pdf
https://www.intersil.com/content/dam/Intersil/documents/cd40/cd4063bms.pdf
http://www.ti.com.cn/cn/lit/ds/symlink/cd4017b.pdf

[T3sl4co1l]

That doesn't make any sense. The CD4060 is CMOS, not TTL and there are plenty of other parts in the family which have 16 pin packages: CD4046, CD4063, CD4017 and interestingly the CD4022, which could be easily in a 14 pin package but it's 16 pin with two not connected.

Ah, but 4022 is very similar to the 4017, possibly a variant produced by burning fuses.  Such could've been done as late as final packaging, or even after, if an undocumented pin function exists (like driving pin X five volts over VDD).

Tim

[MrSlack]

Just to add, a lot of the devices were made available over the space of several years starting in 1968 and not in order of part number. Many of the larger devices didn't appear until very late in the production.

[SeanB]

4022 and 4017 are a similar die, just the final metallisation layers are different, giving the changed function.

Remember plastic DIP is almost fully automated manufacture, and the main constraints are you need an entire line per size, and with the majority of parts being 8, 14 and 16 pins, having 3 lines next to each other was workable cost wise, as you could run multiple separate types with only the 3 lines. As well the leadframes were supplied only by a few ( very few, a small market) manufacturers, and they had mass supply of the narrow leadframe, which came on a reel and were cut to size only after final encapsulation. That is why you see on some older DIP plastic packages there are 2 metal stubs top and bottom of the pack, they are there for support during packaging. They get cut off in the final separation from the carrier, after which the IC goes for final QC functional testing.

Having a 4th line for a few parts only was expensive, you tended to run that line only to make parts that were costlier to get the ROI on the specialist line. Thus it was used for microprocessor support, where the higher selling price was worth the line, and where volumes were lower.

The 4060 would not have been a good seller at 4 times the price, along with not fitting nicely in a line with the other 0.3 dip packages, it would take up a lot of board space for a part that either had 4 NC pins on it, or which had 4 extra divider stages that most uses would not use. The gap in outputs was probably the best way they could do it, while still having the headline count of stages.

Here is an odd one, the 4045. Supplied in a DIP16, even though you only have 8 pins used. Done because it would not fit in the DIP8 die paddle, as it has 4 large mosfets to give it a 47mA drive capability on the 2 outputs. If you had a car clock in the 1970's or 1980's with an analogue display then this was the chip doing the driving of the mechanism. 5 resistors, 3 capacitors, a crystal and a zener diode for both reverse polarity and overvoltage and you had a clock.

http://www.ti.com/lit/ds/symlink/cd4045b.pdf

[MrSlack]

Interesting trivia - thanks for posting