Loss in weight and S7 sample time method, OB or timer?

Hi all,

I am building a loss in weight system for one of our 3 layer extruders. I am using a S7 CPU. I have a question about the sampling time accuracy. I was thinking of 4 samples a second. The unit of measure to the operator is kg/h. This machine will eat about 250kg/h of material.
Would a 250ms timer be accurate enough or would I be better served calling a cyclic OB? (Module diagnostics say my current scan time is 2 - 4ms)
N.B this system won't be controlling the extruder speed just yet...

Hi I would probably use a timed interrupt block for accuracy, but having said that, how accurate do you actually need to be.

Your feedrate is 250Kg/hr which is (250/60/60) 0.069 kg/second so your 250mS would be a resolution of 0.017Kg which sounds pretty good. Would a slight variance for scan time etc be a major problem?

Whats wrong with every scan?

Peter
I read the question as the op was questioning the accuracy of using a standard recycling timer vs timed interrupt. However as before I would suggest the highest accuracy isnt required but only the op would know that.

Cheers

Hi
I have done LWS with 100 msec. I startet up with 10 msec but found out that 1oo msec was ok. I had to ad at lowpass filter and a awerage blow to convert to kg/h with a smooth resault.
If you are filling when still dosing change the PID to man. mode and ad a timer to let the weighing system stabilize.

Banker

Thankyou for the replies.
Some jobs will run for 8 days 24/7 so using the yard stick of 250kg/h i'm looking at 48 tonnes of material which needs to be reconciled against the deliveries to the silos.
Thanks, you guys have helped answer my questions.
Yes, I do need to be accurate. Cyclic OB it will be.
Resolution, i'll start off at 250ms. The sampling will be muted whilst re filling the metering hopper (18kg hopper). I'll code the running average (30 or so sample period) to carry over while re filling. (Refill time approx 3 seconds)

There's an "art" to what your trying to do.
After you struggle for a few weeks -
Look at the Siwarex FTx

cjd1965 said:

Peter
I read the question as the op was questioning the accuracy of using a standard recycling timer vs timed interrupt. However as before I would suggest the highest accuracy isnt required but only the op would know that.

Cheers

Click to expand...

I asked the question hoping to get more info from the OP. I've worked on batching systems where loss of weight method was used by just adding the difference every scan, for just requiring a target weight of product this was OK for coarse fine type of delivery.

A timed method would only be of benefit if you are trying to blend something at a ratio, i.e rate of delivery is important, I assume this is the case but the OP didn't fill in any detail on why he wants a rate or what he's trying to do.

Peter, my please accept my apologies for being vague.
Please allow to me to attempt to descripe the system.
The material arrives pre blended from an upstream station. The metering hopper that I am building sits on top of the extruder (isolated with load cell) and discharges directly into the extruder throat. The extruder speed governs the discharge rate, i.e I am not metering the material. The extruder speeds are currently set by the operator. Rate delivery is critical as there are three extruders feeding material (each with the same LIW metering hopper), which have a set ratio. For example; Extruder 1 (inner layer) 25% (100kg/h) Extruder 2 (core layer) 50% (200kg/h) Extruder 3 (outer layer) 25% (100kg/h).

At the moment the system will be "dumb" just reporting back material usage.

P.s (edit) Once I am happy with the system I will be posting back asking for assistance with setting up a control loop to automate the discharge rate (via the extruder speed). That will be my first go at a look up table with PID trim.
Thankyou all.