Counting very short TTL level pulses

R

Thread Starter

RSW

We are using an analyzer that has a raw output signal consisting of a variable rate (0 to 120 pulses per second) TTL level pulses. Pulse width is 1.5 microseconds. We want to convert this into a 4-20 mA signal proportional to the rate.

So far, all the frequency to current converters and rate counters that I have seen have a minimum pulse width requirement of at least 9 microseconds. Currently, we use a high speed counter module in a PLC, but would like to switch to something simpler, preferably DIN rail mounted, or a panel mounted unit with a digital readout. Does anyone know of anything that would work in this application?
 
R

Robert Scott

Since the highest rate is 120 pulses per second, there is plenty of room to stretch the 1.5 microsecond pulses to way over 9 microseconds using a typical one-shot. You can mount it on a proto board in an holder by Phoenix Contact to make it fit on a DIN rail. Then you could use one of the off-the-shelf rate converters to follow it.

If, however, you really want to have complete control over the conversion process, we can modify one of our DIN-rail smart boards to do what need. One of the characteristics that we can address is the sampling time. If you count for one second, you will get 1 in 120 resolution. If you count for longer, you will get better resolution, but worse update time. The ideal situation is not to count pulse frequency, but to measure the period between the 1.5 microsecond pulses and infer frequency from that. This gives the best resolution for a given update rate. We can customize this tradeoff in whatever way you need by averaging over however many pulses there are in a given period of time that you select. If this interests you, contact me privately through the moderator.

Robert Scott
Real-Time Specialties
Embedded Systems Development
Ypsilanti, Michigan
 
C

Curt Wuollet

The easy way would be to use a TTL monostable multivibrator, (one shot) on the output to stretch the pulses to a width that your signal conditioner can deal with. One with an open collector output would be most convenient especially if it could handle 24 volts. A 555 could be used in this configuration if the pulses can be inverted easily. Otherwise something in CMOS that can run at a supply voltage that will give you output amplitude sufficient to trigger standard 24 V modules might be your best bet. You might also google pulse stretcher as there may be someone who makes a unit for that.

Regards
cww
 
C

Curt Wuollet

Oh, and that's a very long pulse in TTL terms. No problem triggering anything except in the extremely slow automation arena.

Regards
cww
 
Top