Good afternoon everyone,
We are going to get the DMT! It feels like Christmas in July! You all have been telling us to get this tool for years, and we finally are moving on it.
In anticipation, I would like to get your advice again on how I might use DMT to perform a specific action.
First, a little background. For whatever reason, all of our part operations have been set to a default operation. So every operation references a resource group instead of a dedicated operation.
Moving forward we are setting up operations to replace (almost) every resource group. The operation will call out a capability which will call out a resource by priority.
My job is to update all of our old part/rev masters by exporting the existing operations, their resource group, and their operation comments. Then using the DMT, update all the part masters to use dedicated operations instead of default operations, and pull the old op comment into the new operation.
Once I get the DMT setup in Pilot I will do some testing. What do you all think? Have you updated lots of part masters using the DMT?
Thanks for your time!
Nate
I have used the DMT to reload our e10 data from e9 so it was interesting. Some of the things that caught us was the order of things for sub assemblies and things like that. once we go that figure out were off to the races.
We had to have revision created before we update the BOO (Bill of Operations) and you need to have a ECO group for maintenance like an import or update group. Because the DMT will check out and check in the BOO for you and approve the Revision.
Since DMT templates use business logic, in general they need to jive with the manual entry steps you’d use in the engineering workbench. So simpler (for me) to use multiple templates - JobHed, JobAsmbl, JobOper, JobOpDtl… instead of the combined template. oops… should have specified part revs… e.g. PartOpr, PartOpDtl
Other than that, I think it will be pretty straightforward after you get thru the basics.
I’ve done mass MOM updates via DMT. Good news is that you’re going to be doing the work in the ECO tables, so if you screw up, you just delete the group and re-copy from the part tables rather than re-pull the live database.
Create new ECO Group in Engineering Workbench solely for the mass update.
Export to Excel the most recent active MOM (incrementing the Rev#) and import via DMT to the ECO Group in step 1. This will probably be a few steps to this. PartRev ->EcoRev will need to be done before you can do PartOpr → EcoOpr, for example. As a general rule, I skip PartOpDtl (or use OpMasDtl), since the op master should be set up to pull in the correct details by itself.
I then make the changes to the Eco tables via DMT. Unless it’s extremely elaborate and not conducive to scripting, in which case I send it back to SME’s to handle manually.
Then I check everything in via Workbench.
Final step is to mass-deactivate the previous rev.
Ooohhh I know our engineers are gonna hate this! They keep insisting that our rev can’t be changed. Is there any way to update an existing rev without incrementing?
Sure… but the entire point of having multiple revisions is so you can keep your history. If that’s not important or necessary, then you can revise your revisions indefinitely.
Ooohhh I know our engineers are gonna hate this! They keep insisting that our rev can’t be changed. Is there any way to update an existing rev without incrementing?
My honest, professional opinion is to tell them to go to hell.
I had an entire engineer staff with the same mentality. What did I do? I disabled the ability to check in previously approved revs, forcing them to increment. Granted, this was initiated by a customer audit finding, but I took the strictest possible interpretation, implemented it, and stood firm against attempts to roll it back.
Two years later, I still get an occasional thank you from the same engineers that resisted the most. Because the benefits of a clear history outweighed the minor inconveniences. Accidentally checked a bad MOM in? Re-pull the good rev and start over.
It also dramatically helped scheduling, because they could now see which jobs (which might have been cut months prior) need to have details re-pulled.
And… I think I saw a post by you a couple days ago, about adding in inspection op with sequence number 900… i.e out of sequence?
Not sure if DMT still does this or not but…
it used to ignore out of sequence op numbers (added ops) in templates and instead auto-increment by 10.
Not a big deal to work around if you know ahead of time
but of course… the first time it happened to me after my first pass importing a few thousand older methods… structures were 12 levels deep, lots of subassemblies and out of sequence op numbers. Rework was not great fun… so I check now, make sure Op Seq are coming out right up front.
I am (pretty) sure it was just a problem for added ops, and no problem changing existing ops… i.e maybe two templates in your case…
first one to add the Op, and a second one to update the OpSeq(s) to 900?
I really like this approach and I will work hard with our team to make sure we do the right thing.
I’m familiar with the mentality of “the rev only changes when the customer rev changes”. I feel like it’s an old school aerospace thing, though the big players are very much against that kind of mentality these days. The added twist is that ERP doesn’t really separate between external and internal revisions either.
Feel free to message me for details. I’m not going to pretend I came up with the best solution (and I’ll have to eventually rework some of it for Kinetic) but it’s survived a couple point-release updates and numerous patches without a hiccup.
Yes, You can make it look like it never happened. We do keep the rev matching our customer as we are a contract supplier and changing processes they don’t care about, only material. I set the approver and approval date back when I am finished so I don’t get a hundred calls about why I am touching boos and boms.
Regardless of going this route you still need to do a query or baq of all approved and all unapproved revs since at the end of this process they ALL will be approved since DMT is doing an approval and check in.
Also, your failures will be left in the DMT ECO Group you setup.
Part Combined is usually my go to for DMT mfg parts. Templates for different types helps Usually I setup a BAQ with the DMT-able template and column names test and validate once it looks good setup a session on Terminal server and let it run. DMT is not fast for lots of records as it hits multiple tables at once, but you can split the file into 200-1k record chunks and run multiple sets at the same time (have not tried CLI setup to check speed difference though).
This is interesting, Greg. Why would DMT change the approval status for revs? Can it be configured to keep the same rev approval status? It seems pretty simple. I did this with a BPM inside a UBAQ. But that was when I was trying to build workarounds for not having the DMT.
Reminds me… in some cases I used separate templates for “regular” ops and “subcontract” ops.
Also, Excel can reformat your data, depending on the template type used ( .csv, xls, etc…) AND the steps used to view/process data.
Not all the time but the more “fiddling” you do, the chances of reformatting will increase.
e.g. a vendor number with a leading zero 012345 might turn into 12345
or if job number use dashes they might turns into dates, etc…
Intermittent and VERY annoying when it happens, which is why I’ve gotten in the habit of keeping before/after copies of all my populated templates.
My standard practice when pasting in data to a sheet is to paste, look for columns that might be affected by Excel’s reformatting, format those columns to text (or date), and re-paste the data using values only. When I forget to do that, I almost invariably regret it.
SPECIFICALLY… Excel LOVES to drop the leading zero on Northeast US Zip Codes… and if your phone numbers are just numbers with no alpha characters (like 8881234567) Excel will show them in scientific notation. There are certainly others as well.
And this one really makes work so much fun ~ our alpha then 3 number suffix material code at end of part numbers (e.g., E110) looks like scientific notation on paste…
It’s a lovely find after working the data over and getting it all ready for upload!
Usually I setup a BAQ with the DMT-able template and column names test and validate once it looks good setup a session on Terminal server and let it run. DMT is not fast for lots of records as it hits multiple tables at once, but you can split the file into 200-1k record chunks and run multiple sets at the same time (have not tried CLI setup to check speed difference though).
