What's new
What's new

Machine Data Collection

GiroDyno

Cast Iron
Joined
Apr 19, 2021
Location
PNW
Anybody using it? We want to be able to benchmark our spindle utilization so we can analyzes changes we make on the shop floor in ways other than sales numbers. Especially as we lean into automation, would be nice to have concrete numbers before and after.
Been browsing the literature, seen some demos for Cimco MDC, Machine Metrics and Scytec, lot of numbers thrown around and pretty graphics.
Curious how and why are you using it? What are the benefits? Drawbacks?
 
Have you used MM?

Yeah, my company demoed it after seeing it at IMTS2018. Nobody in IT liked it. It was constantly disconnecting, and a pain to get working with our machines. I think our demo lasted a few weeks and we told them to take it back.

Not to bad mouth the product, maybe it works wonderfully in other implementations, and we just had bad luck. I'm pretty sure the demo is free, you could reach out to them.
 
We have a couple Yasda machines that have software to monitor utilization. I believe it is something that Yasda itself offered when we bought the machines.

My take is this: machine monitoring is great if you are ready for it. We hit the ground running with pulling and analyzing data, and quickly realized that while the data was correct and accurate, it wasn't needed. Spindle running at 27%. Didn't need software for that. I know it didn't run much because I walked the floor and saw it sitting all the time.

Unless you are already running at an 80%+ utilization rate and trying to get more, I dont know that machine monitoring is going to be very useful. You have to get all the low hanging fruit first.
 
Having an operator be able to specify causes of downtime seems like it would be valuable in finding the low hanging fruit. Obviously we could go around at the end of the day and compare spindle hours to find utilization, but knowing when/why there was downtime might get everyone on the same page.
Operators know how long it takes to change clumsy fixturing, or that first articles take xxx% extra time, if we can assign hard data points to that it should be easier to get managements attention when things need to change.
 
It was constantly disconnecting, and a pain to get working with our machines.

What kind of machines, and how many? Half or equipment is Haas with the "classic" control so it will need some kind of serial adapter, everything else is Fanuc and should be ethernet plug-n-play.
There's 4 different brands we're looking at, they all seem to collect the same data, have similar installation requirements, similar prices, seems like it might come down to service/reliability.
 
I will say this: be careful. There can be a few good things about monitoring spindle time, but it can also be a distraction (i.e. there might be better/more important things to measure). One of the most important things is that the operators, if they know this is being measured, might be able to "fake" higher utilization (yes, depends on their ability). The main thing is to make sure that you're collecting useful data and, ideally, have something else to correlate it to (e.g. intended "up-time" from work orders to actual spindle time).

If your software/operating system can't generate spindle time, look into IoT devices from something like Monnit, cheap little sensors (huge variety) that can send data over WiFi into a spreadsheet (you should be able to run a sensor off the electrical power to the spindle motor or something similar).

Good luck!
The Dude
 
What kind of machines, and how many? Half or equipment is Haas with the "classic" control so it will need some kind of serial adapter, everything else is Fanuc and should be ethernet plug-n-play.
There's 4 different brands we're looking at, they all seem to collect the same data, have similar installation requirements, similar prices, seems like it might come down to service/reliability.


Multiple brands. All 5 years old or less. Some Mitsubishi Meldas controllers, some Heidenhain, some FANUC. We only installed on a couple machines to test, installing on every machine would be a waste of time and effort for a demo.
 








 
Back
Top