# Measure a two megabit signal!!!

I want to build a simple construction for measure of a two-megabit clock pulse, the clock pulse is also differential. I want to measure that the signal toggles and maybe also check the frequency. I also want to check or measure that the amplitude is correct.  Can I do something with a PIC controller or with a PLD, a A/D-converter and so on??? I also have a 8kHz signal which I can use as a reference.

Best regards

Yes I can use an instrument to measure the signal, but the problem is that I want to measure this frequency on a several circuit boards and in several hours. It will be too expensive to buy hundreds of instruments. Yes it is important to know if the signal rises to a proper level. Example: If the signal rises to one (1) the level should be above for example 1,5 volt and a zero (0) should be under 0,5 volt. The frequency should also be around 2 MHz, not under 1,9 MHz and not over 2,1 MHz. To indicate that something is wrong it should be OK to light an LED.

Best regards.

I have looked at your proposal, but I have still the problem to measure the level of the signal. Do you have any solutions for that?

Best regards

###### Who is Participating?

x

Commented:
If the level is not correct then the monostables wont fire.
But to check the level directly, you could histogram the signal in a pc, or, since simple hardware is good,
you could use a thing called a 'window comparator'.

A window comparator uses opamps to compare the input with two reference voltages.

Since the voltage must go through 1V on each transition, the output has to be either filtered or latched to avoid false positives. If the latched comparator output goes high you catch it with a 7474 and light an LED.

the book
The Art of Electronics
by Horowitz and Hill

is good for cooking up this kind of circuit and is a very good read.

as
0

Commented:
You've left out VITAL information here.

HOW PRECICELY and ACCURATELY do you need this measurement to be made?
0

Commented:
A 2 MHz clock pulse? Some kind of oscilloscope, provided it can handle differential input might work. There are a few thingys for attaching to a PC that also might do the job - look for Pico Technology (www.picotech.com possibly).
Regards
/RID
0

Commented:
use a ttl counter to divide the signal down to a few hz and count the pulses in a know time with a pc. Or use a longer counter and read the msb of the counter bits on the parallel port, make sure to avoid glitches, you might want to read twice, or latch the data before reading

National Instruments have cards that do this.
0

Author Commented:
Yes I can use an instrument to measure the signal, but the problem is that I want to measure this frequency on a several circuit boards and in several hours. It will be too expensive to buy hundreds of instruments. Yes it is important to know if the signal rises to a proper level. Example: If the signal rises to one (1) the level should be above for example 1,5 volt and a zero (0) should be under 0,5 volt. The frequency should also be around 2 MHz, not under 1,9 MHz and not over 2,1 MHz. To indicate that something is wrong it should be OK to light an LED.

Best regards.
0

Author Commented:
Yes I can use an instrument to measure the signal, but the problem is that I want to measure this frequency on a several circuit boards and in several hours. It will be too expensive to buy hundreds of instruments. Yes it is important to know if the signal rises to a proper level. Example: If the signal rises to one (1) the level should be above for example 1,5 volt and a zero (0) should be under 0,5 volt. The frequency should also be around 2 MHz, not under 1,9 MHz and not over 2,1 MHz. To indicate that something is wrong it should be OK to light an LED.

Best regards.
0

Commented:
A pc is a good tool for automating testing.

You can get cards for a pc with multiple counters on a board. The simplest of these uses the 8254 counter timer chip, (the same chip that divides the 1.19 Mhz clock in a pc for interrupt zero). The 8254 has 3 counters on it, and costs about \$4.

look at
http://www.boondog.com/%5Ctutorials%5C8254%5C8254.htm

to see how to build an isa card yourself.

You could add a multiplexer or relay box to select between mutiple inputs.

The cost per counter channel of the most expensive boards is about \$100.

Since the accuracy required is 5% there is a very cheap solution, A retrigerable monostable set for 5% longer than the clock period will never return high, if it goes high you can catch that edge with a 7474 and light your led. That will tell you if the frequency ever drops below target. A high frequency can be detected by a 7474 latching the output of a second retriggerable monostable (set for a %5 short period) on the clock edge.

There are all sorts of ways to implement the monostables, digital and analog. A digital crystal controlled monostable could be very accurate. Also check out the line of factory trimmed monostables from Dallas Semiconductor, easy to use, I recomend them. You may have to cascade them to get the timings you want. The 74123 and the 555 are very popular analog monostable circuit blocks.

The cost per channel here is obviously less, but you might still want to use a pc as a 'smart led'

as
0

Commented:
the 75176 is a good differential ttl line reciever.
as
0

Author Commented:
0

Author Commented:
I have looked at your proposal, but I have still the problem to measure the level of the signal. Do you have any solutions for that?

Best regards
0

Author Commented:
Thank you!!
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.