Low Pass Filter for Pulse Counter (Update)

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Greg

Well-known member
Joined
Jun 7, 2004
Messages
1,784
Location
New Orleans, LA
hi:

Please refer to this diagram

So I've got a little problem. I'm working with this input module that interfaces with a PLC. My input is across pins 1 and 2, and what I'm trying to do is count pulses (0-24V, max 500Hz). Apparently this input signal has some kind of carrier around 2kHz, which can be changed to a higher frequency if necessary. This carrier is messing up the pulse counter, so I was thinking about adding a low pass at the input to only allow the pulse to get through.

I was thinking about using a basic RC filter, but I doubt that would get rid of all the 2k hash. But it might get rid of enough. The other concern I have is the loading an RC network will have on the previous output stage, which I don't really know much about. Ideally, I'd want to keep the RC filter with a fairly high input impedence.

Also, the input requires 7mA for the device won't operate properly.

Any suggestions? Good idea? Bad idea?

I have feeling PRR is going to lay down the law on this one... :green: .
 
RogerFoote: I'll check into that.

moamps: After looking at datasheet, would that device be used to read or create the pulse? It seems that would be used to create the pulse ?!? If so, I have no control over the sensor being used, that's outside the scope of my system. I just have to read what's already there.

PRR: I can't remove Q2 and Q3... I can't modify the high speed module. Also, Q2 and Q3 are part of the protection circuit for the opto... why did you suggest I remove them?

What about an emitter-follower as a "buffer" (since I'm single ended here) with a simple low pass. I just want to strip off the damn carrier.
 
[quote author="Greg"]After looking at datasheet, would that device be used to read or create the pulse?[/quote]
It create the pulse.
If the lenght of the pulse is set to be longer than carrier's lenght and the monostable multivibrator is retriggable type, resulting output signal will be "1" for the carrier present and "0" for carrier absence.

Regards,
Milan
 
Thanks for the replies, but try to understand that I'm limited in what I actually have access to. Take a quick look at this guy and tell what you think:

Buffered Low Pass Filter - 1st Order

The main piece of information I'm waiting on is what my + rail will be, and the amplitude of my input pulse. As soon as I get this, I can calculation some values. So R1, R2, and R3 for setting bias (if necessaary), and R4 and C are my LPF.

Also, I chose a BC550 because it's really the only NPN I have lying around. As I stated before, my main concern to make sure I have a high input impedence.

Thanks.
 
hi Roger:

Thanks for your help. I solved for my cutoff without a problem, and that worked great. I perfboarded a design prior to your 100K and 10K. In my original design, I used R1 = 33K, R2 = 39.2K, and R3 = 3.6K. The circuit worked well, but it had below unity gain... perhaps I have a bias problem.

I'm planning to rebuild using the circuit with the 100K and 10K resistors, but what it the ideal voltage at the base... or before the 100K. I looked at the datasheet for the BC550, but didn't see bias info. I'm not very experienced with transistors... I'm much more familiar with opamp designs.

Here's a link to BC550 if you need anything from there:
http://www.fairchildsemi.com/ds/BC/BC550.pdf

I would appreciate some guidance.
 
[quote author="Greg"] The circuit worked well, but it had below unity gain... perhaps I have a bias problem.[/quote]

The EF has gain below 1.
 
So I breadboarded the following circuit and it seems work very well. I am a little concerned about the loading of the transistor.

Buffered High Pass #2

With no filter network, the EF had nearly unity gain; however, changing the 10K from the emitter to GND had an effect on this. Also, with the filter network attached, changing the mentioned 10K had a drastic effect on the output. I also tried 8K and 22K and the output became very non-symmetric. This resistor is setting the current from collector to emitter... correct?

I think this design works well, and I think I learned alot in the process. Any suggestions for improvement are also welcome.

I also learned from searching around the internet from the Geoff Tanner site than a similar EF was used in the 31109 (?) as a buffer. Neat-o!

EDIT: I also have a DC blocking cap on the input, but it's not shown. 47uF/50V.
 
We're on the same page so to speak.

I was trying that very thing when you posted, but I can't seem to get it to work just yet. Here's another updated schematic.

Buffered High Pass with Switch

My input from the sensor pulses from 0-10V @ 0-500Hz. At the filter output, I have around 7.9V. I'm having trouble setting R1 to get the damn thing to switch properly. If you can offer anything, please do so. I'm going to step away from this for a little while to clear the cobwebs.
 
Thanks... but the switch still isn't working properly. I'll see if I can search around the net and see what's my problem. :?
 
Back
Top