Double signal on an Apollo Flow Batcher??

[beattsmjk]

Hi,

I work in the bakery industry (as a process engineer) and we have one of these Apollo flow batcher units for dispensing quantities of liquid (water, milk etc) into various mixes.

http://www.apolloflow.co.uk/instruments/BC7batcher.htm

The system operates as follows.

1. A signal is sent from a mixing unit to dispense liquid
2. The Apollo unit dispenses the liquid (pre defined quantity - via a dosing valve)
3. A signal is sent back to the Apollo unit to say batch complete.

and so on, in general the batcher dispenses about six times an hour (depending on what is being mixed etc)

Lately though we have been getting double signals which results in two quantities of liquid being dispenses which invariably results in the mix being ruined.

I can't think why this is happening. I would say it does it about twice a week, but the implications can take us an hour or so to rectify.

Could noise or interference be causing these double dose signals. They generally happen within a few seconds of each other and I cant find any pattern.

There is quite a bit of other electrical equipment around the cabinet housing the Apollo such as a 3 phase motor, a cabinet housing numerous small invertors, a large cooling fan (bakeries get warm).

Is there any way I could shield the entire cabinet from erroneous signals perhaps?? All cable within the Apollo cabinet is shielded.

All ideas gratefully welcomed

thanks

 

[zappedagain]

You could be picking up noise, but at that timing I'm very suspicious of a software fault. I suggest monitoring the control signal with an oscilloscope. If you are getting nice clean pulses when you are not supposed to then my software suspicion is supported. If you see tiny little pulses that are triggering the mixer then noise pickup is more likely.

If you can swap the Apollo controller that might help to isolate the problem. Does it stop with the new controller (old controller is bad) or does it continue (old controller is good)?

I see from the web-site that the 'Scaled pulse' output is the open collector type. That means it must be connected to a resistor somewhere, with the other side of that resistor to a power supply. If you are getting intermittent connections on that resistor you could get extra pulses. Or if that output is driving a relay directly you may have a relay going bad.

Z

 

[itsmoked]

What do you mean "They generally happen within a few seconds of each other"? Doesn't a 'dose' take more than a few seconds anyway? Or do you mean a dose is triggered and when it's finally complete a few seconds later it triggers again?

Seems odd that if noise is triggering it it doesn't spontaneously occur throughout the day.

You note a "pattern". Could it be that after a dose something starts up and that is your mis-trigger generator.

You really should put an oscilloscope on the trigger line. You will see immediately if noise is a concern.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[beattsmjk]

Thanks for the advice chaps,

In relation to your advice 'Z' we already swapped out the apollo unit for a replacement and the problem still exists.

I agree with you on setting up an oscilliscope, however I will have to aquire one that logs over a period of a few days to get a good result as the problem has never occured when I have been watching it.

I might also like to add we have this exact same system in at least six other bakeries with no problems whatsoever, which makes me think that it could be noise.

This bakery is in need of a lot of repair, as I mentioned previously there are a lot of other electrical components near by that could theoretically generate a false signal perhaps and their electrical housekeeping is poor to say the least.

I dont really understand what you mean regarding the scaled pulse, excuse my ignorance.

in relation to your advice Keith,

The dosing takes on average around thirty seconds, and then the mixer operates for ten minutes. The second dose normally occurs upto a minute after the first dose has completed.

the only patterns we have to work with is, it generally happens on night shift 90% of the time, and also it only happens when the bakery is using wholemeal flour!!

If the bakery is using white flour, it is dispensed into the mixer and no problems occur

If wholemeal is used, then the mix is usually always 80% wholemeal and then 20% white flour (normal make up for wholemeal bread). Therefore the flour is dosed twice on wholemeal.

 

[itsmoked]

Never did like whole wheat..


What exactly starts the dose?
What exactly terminates the dose?

You could always brute force it. Put a time delay relay between the batch controller and the valve that doesn't allow a second dose in less than x-minutes.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[zappedagain]

"Scaled pulse" is the name of the output in the datasheet that you linked to.

If the stray pulse is "within a few seconds" then as long as you trigger on the first pulse you don't need to collect very much data.

John (Z)

 

[beattsmjk]

thanks again,

Keith,

thats exactly what we have done. We fitted a timer so that any second pulse generated would be ignored for a time period after the first dose had started (we made it something like 7 mins). However sometimes the doses need to be carried out manually, when making samples etc, and in this case the operator would need to wait seven minutes between doses.

The dose is started via a 24v signal from a San cassiano Mixer (PLC output) The output signal is a long pulse which we have shortened to a 2sec pulse.

The dose is terminated when the quantity is reached (this info is fed into the Apollo with a Siemens flow meter)

 

[CharlieGill]

Some of the notes above don't seem to agree with the BC7 data sheet and manual.
The 'scaled pulse" is an output from the BC7, not the input to start a batch. You get one pulse per unit of flow.

The 'remote start' input should be a dry contact, not 24V, according to the manual.

"It is possible to duplicate the functions of the front panel buttons using remote switches. This allows the instrument to be operated by PLCs or from a control room.
Run Button: Connect a momentary switch across terminals 12 and 13."

If you are just applying a 24V signal to pin 13, the operation will probably be erratic.

Charlie


Charlie
--

http://www.calibrator.com--

 

[beattsmjk]

Sorry Charlie,

I meant that the signal to the output is a momentry switch across 12 and 13. This is derived from a relay output on the mixer.

 

[CharlieGill]

Well there went all my great ideas
Unless the relay is in the PLC and you have a few hundred feet of cable to the BC7, I doubt noise is the culprit. If it is wired that way, I'd move the relay to the BC7 end as noise can rarely energize a coil.

Relays are vibration sensitive. Can anything strike the cabinet where the relay is located?

It still could be a software issue even though you have other similar systems without a problem. A coding error can sometimes give very pattern sensitive problems (I know, I've done it).

You should also consider mischief, particularly when it occurs only at night or on a particular shift. We had a helper once who thought it funny to just push a stop button at random when he was walking through the plant at night. Drove us crazy until someone saw him do it.

Charlie

Charlie
--

http://www.calibrator.com--

 

[beattsmjk]

do you think I could use a latching relay to overcome this problem?

 

[MacGyverS2000]

We had a helper once who thought it funny to just push a stop button at random when he was walking through the plant at night. Drove us crazy until someone saw him do it.

Do these people ever get fired? From what I've seen, they get a slap on the wrist (at best) and continue along their merry way...

Dan - Owner

http://www.Hi-TecDesigns.com

 

[CharlieGill]

He did get fired. He was the son of a regular employee (who was mortified by the experience) but no official report was made.

I don't see how a latching relay will help. Either the PLC is telling the batch controller to run or the BC is starting by itself. If your PLC has data logging capability, turn it on for the output bit and see if the extra outputs are logged. This will tell you if the PLC is the culprit. If you don't have data logging, try adding a counter to the output logic then compare the counter to the production logs each day to see if the count matches the intended number of batches. If the problem is the PLC, look at the code harder. It's easier to spot mistakes when you are sure that there is one.

If it's not the PLC, are you sure you are getting extra batches and not just an out-of-spec batch. Could the flow valve be sticking open? Could the flow meter be dropping out? The start button on the BC can be disabled, try that, or is that used for manual batches?

Always look for loose wires and miswiring (missing neutral or ground), 90% of all electrical troubles are mechanical in nature.

A general purpose data logger would certainly help your troubleshooting. You can rent them and in the past I have seen free or shareware software to turn a PC parallel port into a data logger (with suitable signal conditioning).

Good luck.

Charlie

Charlie Gill
--

http://www.calibrator.com--

 

[itsmoked]

Great points Charlie!

Keith Cress
kcress -

http://www.flaminsystems.com

 

[ScottyUK]

90% of all electrical troubles are mechanical in nature.

Oh, I want that framing and putting above my desk.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[zappedagain]

And the other 30% are software in nature! (90+30=100)

 

[MacGyverS2000]

And the other 30% are software in nature! (90+30=100)

Only if you're using an old Pentium...

Dan - Owner

http://www.Hi-TecDesigns.com