Help With Spec Sheet: MM74C926

This is the place for any magazine-related discussions that don't fit in any of the column discussion boards below.
EPA III
Posts: 39
Joined: Tue Aug 17, 2004 1:01 am
Location: Beaumont, TX
Contact:

Help With Spec Sheet: MM74C926

Post by EPA III »

I am looking at using a Fairchild MM74C926 to implement a counter circuit for use as a tach. for a stepper motor. The Fairchild spec sheet shows a number of waveform times for operating the IC, but the thing I am confused about is that all the minimum values are larger than the typical values. The maximum values column is blank. For example, the Reset pulse width is given as: 100 ns typical, but 250 ns minimum and no maximum is shown. How should this be interpeted or is this just a mistake/misprint? What would be a good operating value for this? Since maximun values are not given, perhaps I should just make it a lot longer, like 500 to 1000 ns?
Paul A.
Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Re: Help With Spec Sheet: MM74C926

Post by Bigglez »

EPA III wrote: I am looking at using a Fairchild MM74C926
Wow, is that part still around? I used it when Reagan was in the Whitehouse.
EPA III wrote:I am confused about is that all the minimum values are larger than the typical values. The maximum values column is blank. For example, the Reset pulse width is given as: 100 ns typical, but 250 ns minimum and no maximum is shown. How should this be interpeted or is this just a mistake/misprint?
You must provide a 250ns wide reset pulse to guarantee operation.
Production devices typically work with 100ns, so you can speed up
your system by reducing the pulse width, but Fairchild can't be
blamed if it doesn't work all the time. Silly CYA policy...
EPA III wrote:perhaps I should just make it a lot longer, like 500 to 1000 ns?
Good idea. If your system can support slowing everything down to
get longer times you'll have peace of mind (if you go into volume
production).
EPA III
Posts: 39
Joined: Tue Aug 17, 2004 1:01 am
Location: Beaumont, TX
Contact:

Re: Help With Spec Sheet: MM74C926

Post by EPA III »

Thanks! Yes, it is a bit old, but still available. So is the 741 which is a lot older: at least 40 years and that is a lot before Reagan was in the White House.

If you have a better way to make a quick and dirty counter, I am all ears.
Paul A.
Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Re: Help With Spec Sheet: MM74C926

Post by Bigglez »

EPA III wrote:If you have a better way to make a quick and dirty counter, I am all ears.
The simple answer is to add a microcontroller (uC) to your project.
Costs about one eighth of the MM74C926, based on Digikey pricing.
EPA III
Posts: 39
Joined: Tue Aug 17, 2004 1:01 am
Location: Beaumont, TX
Contact:

Re: Help With Spec Sheet: MM74C926

Post by EPA III »

Bigglez wrote:
EPA III wrote:If you have a better way to make a quick and dirty counter, I am all ears.
The simple answer is to add a microcontroller (uC) to your project.
Costs about one eighth of the MM74C926, based on Digikey pricing.

I have been experimenting with some uCs and I did think of that. But then you have to interface the uC to the display. And displays are not cheap. Microprocessor friendly, serial displays are even worse. I have some seven segment LED displays in my junk box. Oh, and you have to program it too. And debug. And I want switching for several different units of measure. I think a hardware solution makes sense here and the 74C926 connects directly to the 7 segments with four transistors and seven resistors (also in my present stocks). I plan to just switch different timing resistors in a 555 timer circuit (got those too) for different ranges. Two chips, four transistors, a rotary switch, and a handfull of Rs, Cs, and Ds and it is done. I suspect it will be half the price of a uC circuit.
Paul A.
Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Re: Help With Spec Sheet: MM74C926

Post by Bigglez »

EPA III wrote:I have been experimenting with some uCs and I did think of that. But then you have to interface the uC to the display. And displays are not cheap. Microprocessor friendly, serial displays are even worse. I have some seven segment LED displays in my junk box. Oh, and you have to program it too. And debug. And I want switching for several different units of measure.
Let me play devil's advocate. We can discuss your project using
hand written notes, and the US mail. We'd not have to learn
typing and computer skills, own a computer, connect to the
internet, and pay the higher electricity bills.

But the Internet thing is so much better overall, that anyone
reading this post has already overcome the objections and
hurdles noted above.

Same deal with the uC project. Once you have programmed
one chip you are set up to do more, do them different, and
do them well for an unlimited variety of new projects (within
reason).

I have driven twin seven segment displays directly from the
small uC chips, and adding ony two transistors and seven
resistors (which you likely have with combinational logic
drivers).

I have driven multiple alphanumeric displays from a uC with
a few dedicated driver ICs. The power of changing fonts,
special characters, and other manipulations is shifted to
firmware coding.

I've used industry standard LCD modules directly on the
uC ports, with both new and surplus (cheap) LCD modules.
I haven't used the serial LCD modules yet, there seems
little advantage over the parallel variety.
EPA III wrote:I think a hardware solution makes sense here and the 74C926 connects directly to the 7 segments with four transistors and seven resistors (also in my present stocks). I plan to just switch different timing resistors in a 555 timer circuit (got those too) for different ranges. Two chips, four transistors, a rotary switch, and a handfull of Rs, Cs, and Ds and it is done. I suspect it will be half the price of a uC circuit.
A medium size 28 pin DIP AVR uC costs $3.66...

Once you have a processor in your project all that other
glue logic and timers etc. goes away. Complex and
costly rotary switches are replaced by simple switches
or better yet a rotary encoder (digital 'volume control').

The transition is not easy. Some companies that made
analog only test equipment took a long time to drink the
all-digital hardware and software Koolaid.

Should be interestnig to see which path you take. They'e
both valid.
EPA III
Posts: 39
Joined: Tue Aug 17, 2004 1:01 am
Location: Beaumont, TX
Contact:

Re: Help With Spec Sheet: MM74C926

Post by EPA III »

My overall project is a manual controller for two or perhaps three stepper motors that will automate a small milling machine. It will have provision for future computer control, but for now, I want to get the manual control working well first. The tach circuit will be to display the speed of the movements in one of several units (like inches per second, mm per second, also per minute, etc.).

So, a couple of things here. First, this is a one-off project for my personal use. I really don't know if I will ever make another one. I guess that depends on how well the first one works out. Definitely no more than two or three, ever. Next, there are some user friendly things I really want. These include controls that work the way I want them to, like dedicated, single switches for each function instead of combining operations via levels of operation or multiple button presses. I want to be able to operate all the controls without looking at them as my attention will be on the cutting action on the part being milled. I plan to spend perhaps $100 on switches alone to achieve this ease of operation.

I want a single, multiposition rotaty switch to select the range of the tach, not some successive series of button presses. It seems to me that doing this with a uC would require using several input pins just for the tach control.

I also want a big, bright display that only needs four numeric digits and a selectable decimal point position. Doing this in a straight forward manner from a uC would tie up between 12 and 14 output pins. Adding an interface chip(s) would add complexity. Does anyone make a single chip: serial in and 7 segment X 4 out?

I am willing to consider alternate methods. I just think implementing the tach with a uC would wind up being more expensive and more difficult to operate. I would love to find a more modern IC that would implement this function, perhaps one that combines the timing necessary as well as the counter and display interface. ??
Paul A.
Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Re: Help With Spec Sheet: MM74C926

Post by Bigglez »

EPA III wrote:I also want a big, bright display that only needs four numeric digits and a selectable decimal point position. Doing this in a straight forward manner from a uC would tie up between 12 and 14 output pins. Adding an interface chip(s) would add complexity. Does anyone make a single chip: serial in and 7 segment X 4 out?
Reviewing your requests one at a time, the large bright display
is easy. Here's a thread with a couple of PIX of LED displays.
Note that in each case the uC uses only three IO pins to
operate the displays. 64 LEDs in the small project, 208 in the
large one.

SMT LEDs cost eleven cents each, less for 100 qty. Dot matrix
modules cost under five dollars each (5x7 LEDs). Each example
had much less than $100 BOM cost.
Engineer1138
Posts: 458
Joined: Thu Feb 05, 2004 1:01 am
Location: Minneapolis, MN
Contact:

Re: Help With Spec Sheet: MM74C926

Post by Engineer1138 »

If you are building a controller for the stepper motors then you would already know how fast they are running. It seems that you wouldn't also need to measure their speed, just display it. That should be a much simpler requirement to meet. What are you using to set the motor speed?
rshayes
Posts: 1286
Joined: Tue Mar 04, 2003 1:01 am
Contact:

Re: Help With Spec Sheet: MM74C926

Post by rshayes »

You might consider using a CPLD instead of a microcontroller for this type of application. These consist of a number of cells, each of which has a cluster of logic gates and a flip-flop. They are programmed by software using a schmatic diagram of the logic desired as the input. Counters and timing chains are much easier to implement than using a microcontroller, especially if several operations are being done at the same time.

The XC9500 series from Xilinx costs about the same as the microcontroller. The programming software is available from Xilinx as a download from their web site. The web site also has the schematic for a programming interface, which uses less than $10 in parts. The 44 pin PLCC package is available with either 36, 72 or 108 cells, which should be enough for this application.

One precaution is to allow some ventilation for the package. I have had one overheat when it was mounted parallel to and close to an aluminum plate. The heat radiated by the chip was reflected back into the package instead of being radiated to the outside world. Drilling a grid of holes in the aluminum plate alleviated the problem.

Some effort will also be needed to learn how to use the programming software. It generates a large number of files, most of which you won't need unless a program detects an error.
Dean Huster
Posts: 1263
Joined: Wed Dec 05, 2001 1:01 am
Location: Harviell, MO (Poplar Bluff area)
Contact:

Re: Help With Spec Sheet: MM74C926

Post by Dean Huster »

You must provide a 250ns wide reset pulse to guarantee operation.
Production devices typically work with 100ns, so you can speed up
your system by reducing the pulse width, but Fairchild can't be
blamed if it doesn't work all the time. Silly CYA policy...
I think that in all fairness, the manufacturer is not implementing a CYA policy. They're offering an ironclad Sears & Roebuck Co. guarantee: This part will always operate as described with this pulse width (or wider) within the specified operating temperature, altitude, mounting position, power supply voltages, input logic levels, etc. -- or your money back.

We had a student once, who was doing computer installations (before networking outside of Xerox' word processing twinax Ethernet) using RS-232 connections. He was proudly boasting that he was able to exceed the specified cable lengths by over 100% and the systems still worked. We told him to not do that. The last thing he needed was for the customer to move something around funny, change a data type somewhere, change out a printer at the end of the line, change the computer, run the room cooler -- whatever -- and have the system start glitching up. The maximum cable lengths for the data rates were there not as a challenge to go farther, but as a guarantee that the system would work reliably under the worst of conditions, making for a happy and satisfied customer. Turns out that he was sending data to printers six times faster than they could print it out, so it was silly to force the speed issue anyway. Customer satisfaction is the name of the game.

Dean
Dean Huster, Electronics Curmudgeon
Contributing Editor emeritus, "Q & A", of the former "Poptronics" magazine (formerly "Popular Electronics" and "Electronics Now" magazines).

R.I.P.
EPA III
Posts: 39
Joined: Tue Aug 17, 2004 1:01 am
Location: Beaumont, TX
Contact:

Re: Help With Spec Sheet: MM74C926

Post by EPA III »

Thanks guys for starting the thought process. I made a parts list for my original tach circuit and priced it today and was surprised to find that my cheap solution was not that cheap. It came to about $40. I was hoping for $20 or less.

So I started to sketch a generic circuit with a uC and was again surprised. It needs more work, but it looks a lot simpler. A BCD rotary switch should provide the range selection I need with only two or at most three INs. I am looking for a driver IC to convert binary or BCD to seven segment. I think the four digits could be selected via individual OUTs or a simple chip would decode two OUTs into one of four lines. I think I will still need current limiting Rs and possibly transistors on the digit select lines. Oh, and of course the decimal point. Two or three chips, one switch, and a few Rs and transistors for the tach.

Anyway, it looks encouraging.

As for the data source, no I am not getting feedback from the mortors. Instead, I intend to just count the driving pulses for a given period of time. They are generated by an IC oscilator/divider chip (4060 and a few passive components). It provides about 10 outputs at sucessively slower rates. These are picked off by switches which can be patched to any of the 10 outs. A pot in the oscillator section will provide about 10:1 speed adjustment and by switching capacitors I hope to extend this range to 1000:1 or even greater. Additional switches will provide single step operation and perhaps a fixed number of pulses. I may look at a uC for this also, but it does not seem as easy. The external parts count would not be reduces by much as I want the switches and pot for easy control.

The tach readout's range and units can be changed by simply changing the period used for the count. Originally I intended to do this with switched Rs in a 555 timer circuit, but it would be very easy to implement in code.

Thanks for the comments.
Paul A.
Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Re: Help With Spec Sheet: MM74C926

Post by Bigglez »

EPA III wrote:A BCD rotary switch should provide the range selection I need with only two or at most three INs.
BCD? Why? Wouldn't a rotary digital pot and human feedback
from the existing readout LCD/LED suffice? Turn pot up or down,
fix value by hitting 'enter' button adjacent to the pot. Or, use a
digital encoder pot with a switch in the shaft - three functions
from one component!
EPA III wrote:I am looking for a driver IC to convert binary or BCD to seven segment.
Why? The uC can do any code conversion and also use LUTs
to generate segment patterns for alphanumeric and symbol
characters.

Only three uC IO lines are required for any length display,
as serial data and two clocks control external shift registers.
EPA III wrote:I think I will still need current limiting Rs and possibly transistors on the digit select lines. Oh, and of course the decimal point.
Yes. However, octal power logic SRs are available as low-side
drivers for common cathode LEDs, leaving a minor requirement
for digit select high-side drivers. Do you need a diagram?
EPA III wrote:As for the data source, no I am not getting feedback from the mortors. Instead, I intend to just count the driving pulses for a given period of time. They are generated by an IC oscilator/divider chip (4060 and a few passive components). It provides about 10 outputs at sucessively slower rates. These are picked off by switches which can be patched to any of the 10 outs. A pot in the oscillator section will provide about 10:1 speed adjustment and by switching capacitors I hope to extend this range to 1000:1 or even greater. Additional switches will provide single step operation and perhaps a fixed number of pulses.
I'd need a schematic or at least a block diagram to get
my head around that idea.
User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Re: Help With Spec Sheet: MM74C926

Post by MrAl »

Hi,


Wow, 1000 piece pricing is up around 12 dollars US each.
LEDs vs Bulbs, LEDs are winning.
EPA III
Posts: 39
Joined: Tue Aug 17, 2004 1:01 am
Location: Beaumont, TX
Contact:

Re: Help With Spec Sheet: MM74C926

Post by EPA III »

Bigglez wrote:
EPA III wrote:A BCD rotary switch should provide the range selection I need with only two or at most three INs.
BCD? Why? Wouldn't a rotary digital pot and human feedback
from the existing readout LCD/LED suffice? Turn pot up or down,
fix value by hitting 'enter' button adjacent to the pot. Or, use a
digital encoder pot with a switch in the shaft - three functions
from one component!

I said "rotary". I meant thumwheel switch. About $5 and it is the only component you need. Seems to make sense to me.
EPA III wrote:I am looking for a driver IC to convert binary or BCD to seven segment.
Why? The uC can do any code conversion and also use LUTs
to generate segment patterns for alphanumeric and symbol
characters.

Only three uC IO lines are required for any length display,
as serial data and two clocks control external shift registers.
EPA III wrote:I think I will still need current limiting Rs and possibly transistors on the digit select lines. Oh, and of course the decimal point.
Yes. However, octal power logic SRs are available as low-side
drivers for common cathode LEDs, leaving a minor requirement
for digit select high-side drivers. Do you need a diagram?

OK, you come out serial and then what? Do you know of some real cheap LED display (four digit at least and nice BIG numbers like 1/2 inch) that incorporated serial input? I don't and would love to have a part number and/or spec sheet. Am I missing something? If a $3-4 uC will be dedicated to this task, what's the harm in dedicating six outs to doing the display? Do you have any part numbers for "octal power logic SRs"? I am presently looking at a 4511.
EPA III wrote:As for the data source, no I am not getting feedback from the mortors. Instead, I intend to just count the driving pulses for a given period of time. They are generated by an IC oscilator/divider chip (4060 and a few passive components). It provides about 10 outputs at sucessively slower rates. These are picked off by switches which can be patched to any of the 10 outs. A pot in the oscillator section will provide about 10:1 speed adjustment and by switching capacitors I hope to extend this range to 1000:1 or even greater. Additional switches will provide single step operation and perhaps a fixed number of pulses.
I'd need a schematic or at least a block diagram to get
my head around that idea.
Perhaps I confused you talking about the source of the pulses. It does not matter where they come from. Simple idea really. Steppers are driven by pulses. One pulse equals one step and in my case one pulse equals 1/200 of a revolution. Speed equals RPS or revolutions per second. Count pulses that are sent to the motor control chip for a given length of time (1/200 of a second) and the count equals speed in RPS. Different periods of time will allow translating that RPS figure into revolutions per minute or, taking the lead of the lead screw which the stepper is driving into account, you can calibrate for inches per second or per minute or whatever units you want. Hence the switch to select the units of the display.

I am looking at components. So far, it seems like the tach circuit can be done for under $20 including a thumbwheel switch. I will post more later.

I do appreciate all the comments. They have been a great help.
Paul A.
Post Reply

Who is online

Users browsing this forum: No registered users and 22 guests