1762-IF4 conversion speed

I am working on a machine with 2 analog inputs

1) 4-20mA pressure transducer
2) 0.5-4.5 Volt LVDT

The machine captures a reading from the LVDT at a certain pressure, delays 0.75s, then captures another reading to calculate the distance travelled.

The problem I am having is conversion rate. The controller is a ML1400 with a 1762-IF4 expansion card. According to the manual my fastest cinversion is 130ms and I have all channels set to maximum speed, but it is still way to slow. I really need something that converts in about 20ms with at least a 15 bit resolution.

I can't find the conversion rate for the 1762-IF2OF2 module. I considered the mV module and use a voltage divider, but that will most likely introduce a bit of noise.

My real problem is that I do not have enough space to put anything much bigger than the ML1400.

What are my options?

Archie said:

I am working on a machine with 2 analog inputs

1) 4-20mA pressure transducer
2) 0.5-4.5 Volt LVDT

The machine captures a reading from the LVDT at a certain pressure, delays 0.75s, then captures another reading to calculate the distance travelled.

The problem I am having is conversion rate. The controller is a ML1400 with a 1762-IF4 expansion card. According to the manual my fastest cinversion is 130ms and I have all channels set to maximum speed, but it is still way to slow. I really need something that converts in about 20ms with at least a 15 bit resolution.

I can't find the conversion rate for the 1762-IF2OF2 module. I considered the mV module and use a voltage divider, but that will most likely introduce a bit of noise.

My real problem is that I do not have enough space to put anything much bigger than the ML1400.

What are my options?

Click to expand...

You won't get 20ms, but you can start by disabling the channels you are not using.... These won't be in the scan, and will result in quicker updates.

-Brian

IIRC, those built in analog inputs are only 8-bit resolution and very slow. You may want to consider a different manufacturer for a small PLC with better analog performance. I am at home on a slow connection, so it would be futile for me to attempt to find the info, but hopefully another member can chime in with better details.

Brijm said:

You won't get 20ms, but you can start by disabling the channels you are not using.... These won't be in the scan, and will result in quicker updates.

Click to expand...

I looked for a way to disable channels on the IF4, but couldn't find a way to do it.

OkiePC said:

IIRC, those built in analog inputs are only 8-bit resolution and very slow. You may want to consider a different manufacturer for a small PLC with better analog performance. I am at home on a slow connection, so it would be futile for me to attempt to find the info, but hopefully another member can chime in with better details.

Click to expand...

I did try the built-in analog channels on the ML1100, but they were definitely too low resolution to do the job.

If I can't disable channels on the IF4 card, one thing I am considering is to use separate 1762-IF4 cards for each analog input. The idea is to wire the same input into all four channels and with a little programming I could in theory get a scan rate down to about 33ms (130ms/4). Since both analog inputs are increasing in value throughout the test cycle, I should be able to check all 4 channels and use the channel with the highest value.

I am getting a bad feeling I may have to change brands of PLC. I did this same machine last year with an Omron brick PLC and it worked fine. This current customer prefers AB PLCs, so I chose the 1100 and didn't realize how slow the A/D was.

Plan B has failed

I got an IT4 (Thermocouple/mV) card thinking I could use some resistors as a voltage divider to drop the voltage down to the 100mV range, then get my 15 bit resolution along with 7ms update time.

The theory sounded good, but in practice the noise level is horrible below 53ms update. I measured the voltage with a meter at the module terminals and it is rock solid to 0.01mV. My raw counts in the PLC fluctuate by at least 150, which calculates to about 0.4mV.

Any ideas why I am seeing these fluctuations? I am using metal film resistors to reduce noise. One resistor in 82K and the other 2.3k.

Tomorrow I am going to Plan C and trying an IF2OF2 card. It is only 12 bit, but has a update time of 2.5ms. The 12 bit is fine on the pressure transducer, but not enough for he LVDT resolution I need. If I amplify my signal by a factor of 2, it will suffice. Anyone know of a good DIN rail amplifier to take my 4.5V to 10V?

Take a look at the Spectrum IF4OF4, not sure about the resolution, but I think it's 16 bit. I would download the manual, it has a lot of info.

http://www.spectrumcontrols.com/ab-1762.htm

From the data sheet:

Update Times*
With four channels enabled
0.495 sec @ 4 Hz
0.135 sec @ 17 Hz
0.045 sec @ 62 Hz
0.020 sec @ 470 Hz

Ken Moore said:

Take a look at the Spectrum IF4OF4, not sure about the resolution, but I think it's 16 bit. I would download the manual, it has a lot of info.

http://www.spectrumcontrols.com/ab-1762.htm

From the data sheet:

Update Times*
With four channels enabled
0.495 sec @ 4 Hz
0.135 sec @ 17 Hz
0.045 sec @ 62 Hz
0.020 sec @ 470 Hz

Click to expand...

I'm not sure I understand their update times. On the data sheet it shows what you have stated, but in the user manual it states:

Module Update Time=Conversion Time * (1+Number of Enabled Channels)

So with a conversion Time of 20ms, 2 channels enabled, my update would be 40ms?

I finally found hardware that worked. I was unable to make any combination of IO work with the Micrologix, so I had to replace it with a Beckhoff CX1010.

I greatly mis-estimated the scan time and conversion speed I really needed. The final program had to scan at 500 microseconds (1/2ms). I used EtherCAT analog terminals that could do a conversion in as little as 50 microseconds.

That became quite a learning experience in high speed analog with a PLC.

Thanks to everyone for the recommendations.