How much in Machine Inspection are you doing?

[rgrin]

30 minutes ago, Leon82 said:

Back in the day Moore Tool had a jig-bore model for inspection if I recall 

Moore M48 CMM. NIST has one buried underground.  Machine Thinking made a youtube video earlier this year about it. Crazy precision.

Link: 

 

[MIL-TFP-41]

A large percentage of our parts require our CMM's to duplicate what the customer CMM's read. This presents a huge challenge to on machine inspection, Sure, it might be possible, but the cost would be enormous.  It would require multiple probe heads with multiple configurations of styli. I have yet to see a machine tool that can change styli...so figure on some of our parts 6 or more different probe heads. Then say you have a couple of different operations on the part you are working on, it would be impossible to measure features that workholding would restrict access to. So then you would have to come up with a whole separate operation just for inspection, fixture the part like you would on a CMM so you can access all the features from both operations.

All the time you are doing that you are losing the main thing a spindle is supposed to be doing, making chips. The cost of that downtime would pay for another CMM, nevermind the hardware cost.

Now for doing a quick check, something like measuring a bore or position on a first article part, yes, we will do on machine.

[crazy^millman]

On 12/22/2023 at 1:39 PM, JParis said:

Yup... still trying to teach this aspect to an old dog...

Why I will have a job doing what I do until the day I don't walk the face of this earth.

[cncappsjames]

@MIL-TFP-41 I think you're limited to 4 probes in a machine. The limitation is the number of variables required/probe software.  IIRC. 

[Bob W.]

I know this post is a little old but I'm bored today and can't help but add my $.02.  We inspect on the machines all of the time and our machine results (Makino) match the CMMs to .0002" typically.  We do this for process control because we might be running 5-10 identical parts overnight and if the second part is bad we will end up with 4-9 bad parts in the morning.  If a part checks out of spec it will alarm the machine and text our production manager along with the rest of the brass.  The raw material on these can be from $200-$1000 per part so it gets costly quickly.  We also have the measured results stored in a macro variable that gets exported to a database via MT Connect so we can see trends pretty easily and monitor the process remotely.

I know the goal is to have spindles produce chips but we want them to also be producing chips while making GOOD PARTS, not scrap.  Having the machine self-check also reduces the operator skill level because they don't have to know how to use a bore mic or other tools that require some skill to use well.  We typically inspect features requiring tolerances tighter than +/-0.001", because we can see that much drift over night if we are making several of the same part, depending on material.  We also have automatic tool comping in certain instances.

Our ultimate goal is to reduce the span and WIP between machining and inspection but this isn't always practical and we might have 8-10 parts that have been machined but not inspected.  Surprises can really suck when this happens and in-machine inspection can reduce this dramatically.  Like was said earlier in this thread, the efficiency of a shop is a result of the process as a whole, not just the capability of one department.  Machining 10 parts and ending up with 10 good parts is worth a lot, even if it is a little slower.

[Kyle F]

@Bob W. Great post, thank you. At my job shop we are about to embark on our "lights out" journey. As with most stuff, everyone around me is saying "now that we bought the machines, we can run lights out" without *really* understanding all the variables. 100% of our multiaxis mill parts have been set up and ran by me, and now I'll probably be programming full time trying to keep these two pallet mills fed with parts. (but I'm sure if a tricky 5axis part is in the pipeline I will want to personally set it up)

I have operators at my shop who load drills into dirty er collets with rust on the taper, and don't even bother doing a quick S200 runout EYEBALL check haha. These are the guys going to be setting up my stuff?! oh lawd

I also have setup sheets with specific holders and specific stickouts and it baffles me that sometimes they just straight up ignore it. Yeah, go ahead and put that 1/32 endmill stuck .500" out in a side lock holder, sure. Wonder why the finish sucks?

My *manager* was running my machine while I was programming and I gave him a torque wrench and told him the spec I was using,... I walked out 30 min later to see him using a regular wrench on the vise. makes me want to pull my hair out. Yeah, it's an OP10 and the part would be fine, but it's the principle... Sooner than later he's going to do that and it won't be fine, or the parts will vary in flatness/thickness/whatever. I could go on for hours lol.

Needless to say it's got me excited, and equally nervous. So I'm watching this thread like a hawk looking for tips and tricks 

[Aaron Eberhard]

2 hours ago, Kyle F said:

@Bob W. Great post, thank you. At my job shop we are about to embark on our "lights out" journey. As with most stuff, everyone around me is saying "now that we bought the machines, we can run lights out" without *really* understanding all the variables. 100% of our multiaxis mill parts have been set up and ran by me, and now I'll probably be programming full time trying to keep these two pallet mills fed with parts. (but I'm sure if a tricky 5axis part is in the pipeline I will want to personally set it up)

I have operators at my shop who load drills into dirty er collets with rust on the taper, and don't even bother doing a quick S200 runout EYEBALL check haha. These are the guys going to be setting up my stuff?! oh lawd

I also have setup sheets with specific holders and specific stickouts and it baffles me that sometimes they just straight up ignore it. Yeah, go ahead and put that 1/32 endmill stuck .500" out in a side lock holder, sure. Wonder why the finish sucks?

My *manager* was running my machine while I was programming and I gave him a torque wrench and told him the spec I was using,... I walked out 30 min later to see him using a regular wrench on the vise. makes me want to pull my hair out. Yeah, it's an OP10 and the part would be fine, but it's the principle... Sooner than later he's going to do that and it won't be fine, or the parts will vary in flatness/thickness/whatever. I could go on for hours lol.

Needless to say it's got me excited, and equally nervous. So I'm watching this thread like a hawk looking for tips and tricks 

I'll steal Bob's thunder, by simply quoting previous Bob wisdom.    Here's a thread where he replied to me a few years ago:

On 12/22/2021 at 4:02 PM, Bob W. said:

As for managing the tool crib and other shop related stuff.  We have detailed SOPs (standard operating procedures) and employees are expected to follow them.  If they don't follow them they get fired pretty quickly, end of story.  We have ZERO patience for guys on their own program.  If they don't like how things are we discuss how to make them better but they are always expected to follow the processes in place.  We are an aerospace shop and we run a really tight ship and all process violations are documented and resolved.  Three strikes and off to become another shops problem.  The result of this was quick turnover initially but we now have a very stable group of employees that operate at another level.  I wouldn't trade one guy (or gal) in my shop right now for anyone.

How I would interpret that is if you can't get your manager onboard with processes and procedures, you basically have no shot at true stress-free success.    All you can do is document everything, have meetings, make sure everyone from the owners down (say they're) on the same page, then pray

[Jake L]

1 hour ago, Aaron Eberhard said:

Here's a thread where he replied to me a few years ago:

Do you have a link to that thread? Looks like you quoted a snippet from a post with other good information, would love to read the whole post and thread. 

Edit, found it:

https://www.emastercam.com/forums/topic/103732-whats-everyone-using-for-tooling-management-as-in-inventory-usage-etc/page/2/#comment-1327457

[Kyle F]

8 hours ago, Aaron Eberhard said:

I'll steal Bob's thunder, by simply quoting previous Bob wisdom.    Here's a thread where he replied to me a few years ago:

How I would interpret that is if you can't get your manager onboard with processes and procedures, you basically have no shot at true stress-free success.    All you can do is document everything, have meetings, make sure everyone from the owners down (say they're) on the same page, then pray

That sounds about right!

Thank you aaron (and bob of course) That makes perfect sense. I may have to show my boss some of these posts lol

and I'll be reading through that thread as well @Jake L I also saw your bump on the old g54.4 thread and there is a ton of great info in there 

[Zoffen]

Just ran some unattended parts with some tight tolerance 5 axis features and used probing to make it all possible. The parts had a +/-.0005 Dia bore that had to be machined from both sides. Good ole haas love to heat up and move around so relying on just COR for this wasn't gonna work.

I machined half of the bore from one side, spun it 180 and did a wash down on the bore with that tool.

Grabbed the probe and then from this side reached through to the the bore I just did on the opposite side, set a Work offset for just this feature, checked the size and position(more for mixxxx/chips hitting the probe. Then came and machined the other side of the bore. At the end of the program I had the probe check the whole bore from one side for size and position. This worked great. Both sides were within .0002ish from side to side and only drifted .0001-.0002 in size throughout the 30 part run (verified with a cmm)

In process probing IMHO is way underutilized compared to the overall payback. I'm probably going to check more features as an in-process check when running unattended now.

 

 

 

[Leon82]

11 hours ago, Kyle F said:

@Bob W. Great post, thank you. At my job shop we are about to embark on our "lights out" journey. As with most stuff, everyone around me is saying "now that we bought the machines, we can run lights out" without *really* understanding all the variables. 100% of our multiaxis mill parts have been set up and ran by me, and now I'll probably be programming full time trying to keep these two pallet mills fed with parts. (but I'm sure if a tricky 5axis part is in the pipeline I will want to personally set it up)

I have operators at my shop who load drills into dirty er collets with rust on the taper, and don't even bother doing a quick S200 runout EYEBALL check haha. These are the guys going to be setting up my stuff?! oh lawd

I also have setup sheets with specific holders and specific stickouts and it baffles me that sometimes they just straight up ignore it. Yeah, go ahead and put that 1/32 endmill stuck .500" out in a side lock holder, sure. Wonder why the finish sucks?

My *manager* was running my machine while I was programming and I gave him a torque wrench and told him the spec I was using,... I walked out 30 min later to see him using a regular wrench on the vise. makes me want to pull my hair out. Yeah, it's an OP10 and the part would be fine, but it's the principle... Sooner than later he's going to do that and it won't be fine, or the parts will vary in flatness/thickness/whatever. I could go on for hours lol.

Needless to say it's got me excited, and equally nervous. So I'm watching this thread like a hawk looking for tips and tricks 

We have parts where if the dovetail isn't torqued it affects the flatness of datums

[Jake L]

12 hours ago, Kyle F said:

I'll be reading through that thread as well @Jake L I also saw your bump on the old g54.4 thread and there is a ton of great info in there

That thread I linked was an awesome (albeit lengthy) read, tons of great info in there related to ERP's and tool management software's. 

And as far as I've found that g54.4 thread is currently one of the best resources on the web, also well worth a read thru.

[Aaron Eberhard]

2 hours ago, Jake L said:

That thread I linked was an awesome (albeit lengthy) read, tons of great info in there related to ERP's and tool management software's. 

And as far as I've found that g54.4 thread is currently one of the best resources on the web, also well worth a read thru.

Yeah, sorry about not linking to the original thread there.  By the time I saw you asked for the link, you had already found it

[Newbeeee™]

On 12/27/2023 at 4:44 PM, Jake L said:

Gonna hijack the thread real quick. Is it true for ISO that each company writes their own rules? If so, why not be super vague to allow the most flexibility?

Ref ISO9001, there were 7xmandates when we got approval (2x man shop) back in 2007.

It (9001) was updated in 2015 and the mandates were changed, but at the time I remember the assessor saying he knew of 1x UK company that was a 1x man band who had got approval and another company whose manual was 2 pages....the manual then consisted of flow downs to other documents which specified/controlled the said mandates. My manual was total of 35 pages, which was very padded out as the 1st half was design requirements and the tail end were copies of the things like CofC, Job Traveller, Invoice etc - I reckon I could have consolidated it to 15 pages if I'd removed the padding, as the design stuff was only for "sales" as we weren't "ISO approved" for that.

But yes to quote Margaret Thatcher...."sometimes it is best to be specifically vague"....ie if you state ail job cards need sign off in ink, just state ink. If you state black pen, someone will use blue and you'll unnecessarily fail the audit (simplification but you get the idea).

AS9100 (aero) was the next step with the only real main difference (at the time) being stock control - you had to control every inch of material, every rivet and screw and washer etc - ISO9001 you could just state (for the same material batch number) job 12345 had 10", job 12346 had 20", job 12347 had 10", of material batch number AXXXX and that was okay.

AS9100 took that further by stock control monitoring so you had to detail incoming delivery of batch number AXXXX was 50", and you used 10, 20, and 10" on the 3x jobs above, so you now have 10" still left in stock (unsure of exactly how you get around tolerance of cutting and width of saw blades....for billets, washers, screws (ie "items") it's easy).

Initially...."getting approval" can seem daunting, but if you break it all down to bite size chunks, the dauntingness rapidly diminishes. Certain practices you should really be doing anyway - such as material batch traceability gauge control and calibration, and office things like "contract review" which catches things like repeat purchase orders which have a revision/change, so you don't make at previous (old) issue....

Overall, it helped focus my business and got a 2x man band approved to be Tier 1x supplier for some major OEM's. Which then allowed us to grow but with control and focus.