AB analog input card response time too slow compactlogic

First post.

We recently upgraded our machine from an SLC to compactlogix. In the SLC 5/05 we used analog 1746-ni8 cards for pressure and position sensors. When we moved over to the compactlogix 1769 l36erm with 1769-if16c cards we seen issues where the analog values had very little definition accuracy. This can be resolved by changing the filters but then the response time of the card is to slow and we miss data.

All of the relevant code from these sensors are in their own task and ive tried periodic, event and continous tasks but none of this helps.

The strange thing is with the SLC we never seen these issues.

Any ideas how i can improve the reponse time when the filters are set low?

That's weird. Something worth trying is delete your module and any code associate with it then recreate the module and accept the default IO parameters and look at the raw values coming in.

The first step is to put an Oscope on the analogs, and see if you have any noise problems.
Second is to configure them for differential mode (if possible, I didn't look up the module), and see if some common mode noise is present.

From your post at Mr. PLC (http://forums.mrplc.com/index.php?/...put-card-response-time-too-slow-compactlogic/) it looks like you are setting different sections of the input card to different frequencies.

From what I can tell in the manual (http://literature.rockwellautomation.com/idc/groups/literature/documents/um/1769-um018_-en-p.pdf) it looks like there are really only 2 Analog to Digital converters which are multiplexed on the card. So, if I'm right, the card reads channel 0&1, then reads 2&3 then 4&5 etc. up to 14&15.

You don't say the breakup between pressure inputs and position inputs but I'm assuming there are 8 and 8.

So the 8 position inputs you say are at 50Hz gives you 4 pairs. From the manual, each pair at 50Hz takes 70ms to update so the 4 position channels take 4*70 or 280msec. The 8 pressure inputs you say are at 16Hz would be another 4 pairs at 200msec each for 800 msec for the pressure inputs.

Total time to get around the card with all inputs would be 280+800=1080msec.

What I don't understand is that you mention that when you set the filter to 1365Hz that everything updates fast enough. Why don't you leave it at 1365Hz?

Correct on the pairing of the channels.

The breakup of position to pressure sensors are not equal on the card in question. A second card has 6 postion and 6 pressure this card suffers more thatn the first card.

The issue with the 1365hz filter is that is does not give the correct definition/ resolution when i take a snap shot of the pressure and position.

The position sensor monitors how long the cylinder is advanced for and usually it should be 680-720ms. However if i change the filters it states that the cylinder is only advanced for ~100ms, which is not the case.

Hi

I am battling some of the same problems with analog input, I am using
1769-IF4/B cards in a Compactlogix L3

I am running an application with some analog acceleration censors and needs a fast response.

If I only enable channel 0 on 500 Hz the signal is perfect, with a average response time around 2 milliseconds, but if I enable more than one i gets to slow:
1 Channel = 2 ms average
2 Channels = 23 ms average
3 Channels = 40 ms average
4 Channels = 200 ms average

So for me, each sensor needs it's own input card, well aware that there is a limit to how many modules I can add before the reaction time on the back-plane drops.

Regards
Holm

Holm, those measured times exceed the documentation for the 1769-IF4 by a little.

The 1769-IF4 non-isolated 4-channel Analog Input modules for CompactLogix have just one A/D converter chip. The main microcontroller has to command that chip over a serial bus to configure, sample, filter, and report each enabled channel.

When the configurations are the same for each channel, the re-configuration data doesn't need to be sent, so that helps.

The CompactLogix Analog Input Module User Manual says that with the 500 Hz filter enabled and all channels configured the same, the sampling takes 4 ms and the channel-switching takes 10 ms, so you should be getting (4 + 10) x 4 = 160 ms updates for the whole module.

If there's a real-time-sample setting for the module that is greater than the total sample update time, the module will report that data to the CompactLogix at the RTS rate.

There are a couple other options for this platform; the 1769-IF8, the Spectrum Controls universal analog module, and the isolated 1769-IF4I.

I think the Isolated module is your best choice. It has four independent A/D converters, so there's no channel-to-channel scanning delay.

Page 3-18 of the User Manual for the 1769-IF4I suggests that with the filter set to maximum (360 Hz), you'll get an update from all four channels every 10 ms.

Hi Ken

I was wondering a bit that my measurement was a bit off, i just have a timer where i copy the .acc value and reset the timer every time i see a change in the analog input.

When I made the test I only had 2 modules in the system, my module update time is set to 0,5 ms, maybe I am seeing something different from the module, but it is the raw input from the module i am reacting on, see attached

I will do the same test with the 1769-IF4I, not that I expect any difference, but as I understand this module have a separate converter for each channel.

/Holm