PCI-6601 and noise spikes

[TML]

I'm trying to measure pulse period continuosly with PCI-6601. The input signal frequency varies from 10 to 100Hz and comes from slotted opto switch. The problem is that every single electrical machine around seems to generate noise spikes to signal. I'm using shielded cable and connector block, but they are not enough. I still get 10 to 2000 pulses when I switch on (or off) the ceiling lightning. PCI-6601 can use programable filtering, but it just reduces measuring accuracy and I still get unwanted pulses. This is very frustrating. Suggestions?

[bman]

TML

I would suggest using a Schmidt trigger if the filtering capibilities of the 6601 device is not suitable. What a Schmidt trigger will do is recognize the initial rising edge and ignore the rest.

You can find a lot of information about how to build a Schmidt trigger on the web. You can also use a pre-built Schmidt trigger such as the 74AC14.

Brian

[TML]

I tested MC74HC14AN schmitt trigger circuit and it reduced spikes, but did not remove them completely. I also made a test with empty cable. I connected pins 1 (+5V) and 7 (SOURCE_1) together in the end of 4m meter cable. Test panel program shows 2 to 5 pulses every time ceiling lighting is switched on. When cable is removed spikes disappear. It is strange that something can pull the voltage down from +5V so that counter recognizes pulses.

[bman]

What kind of cable are you using. Is it a shielded cable or a ribbon cable?

Brian

[TML]

All cables all shielded. The cable from PCI-6601 to SCB-68 connector block is SH68-68-D1. The cable from the SCB to the opto switch is shielded 5x0.5mm2. If SH68-68-D1 is replaced by ribbon cable there will be only a little more spikes.