Sum3D is Molasses

[ParkwayDental]

Sum3D is slower then molasses, it takes an hour to calculate a full puck. I mean really come on now!

I've heard it runs on a single core why? Why not up grade and get into 2013?

 

[Marcusthegladiator CDT]

Sum3D is slower then molasses, it takes an hour to calculate a full puck. I mean really come on now!

I've heard it runs on a single core why? Why not up grade and get into 2013?

I don't remember it ever taking that long, I would double check your process tree and make sure your strategies are the same and no one messed with it... Maybe someone dropped the roughing step and such you know?

 

[eyeloveteeth]

it also depends on how many different angles it has to calculate.

I've found and have been instructed by the wonderful blokes at CAP to bundle in small batches. - also great for productivity.

only time i run a full puck is overnight.

 

[ParkwayDental]

Nesting in small batches is crap, I need to load up a full puck and let it ride! I don't have time to wait an hour to mill a full puck or nest in a bunch of small batches.

In this case as far as calculation of time "Picasoft" blows Sum3D out of the water.

 

[eyeloveteeth]

just get a kid to nest and put units in sintering over for you. Then you'll have plenty of time. I used to do it all alone and it was so hard because i juggle too many things. Now that I have this kid, I don't even think twice - about 35-45% more productive.

Next step, he marks the margins for me and I just take the cases from there on.

 

[eyeloveteeth]

also, we've been nesting on 2 computers - 2 Rolands to each computer. pucks are saved in a centralized location.

 

[primus]

Probably could be fixed, to some degree?

Tyler, do you see your disk-holder loaded into the working scene around your material?
I suggest not loading the support equipment and other parts of the mill into the working scene.
This is done in your machine configuration, "codes for customization" LOADSUP=N, or something like that...and you could delete some un-necessary upper parts of the spindle and shank in the CAM tool area. This is of course your strategy is safe and all other things are worked out.

These extra parts add a bit of time for detection of collision during calculation. If in 4-axis, then don't worry about it. You'll be fine. Delete this data and run without collision detection. I am sure your CAM is limiting your maximum A and B axis to something. Maybe 20degrees or something? You'd see this also in your codes for customization. So, you still have another safty-net to prevent A or B axis from getting in bad angle and have collision anyways..

Also, do you have something like this in first milling operation's name....ROUGHING XXXXXX " <MATRES0.?> "
Try adding a wee bit to the value...If it is Material resolution is at 0.03, then try 0.05 or 0.06 even....

This takes a bit to explain, so I will save that for another day.....

 

[primus]

We nest up to Eight projects at a time (various thickness blocks of both HT and HS),and calculate all of them at once, while we set up more.....
Having multiple instance of SUM open really helps. Calculate half the disk on One monitor, and other half on another instance of SUM.

This would split-up the process to Two CPU now...and so on....

One of our CAM stations uses Four monitors. Some people laugh, but it really helps! Brad has up to Twenty programs/files open at once all day long.

 

[shane williams]

You know what Tyler I was in that same boat! P!ssed me off more than anything! But I did some changes to my setup. Adding different design stations, upgrading my pc's and so on. After I had CAP install SUM on my new pc the version they put on with my new pc kicks major a$$! Sometimes it still sucks, like when I send 10 or so to the mill, it brings up like 6 additional steps to each step! Upsets me to no end! I think sending bridges with crowns has something to do with that. I kinda agree with eyeloveteeth, someone here is usually designated scanner, I'm designing. The computer that I design on has SUM on it, so after I get a handful I send them off. I find it keeps the mills going all day rather than sitting around waiting for me to design 18-20, then try nest them and send at once. Plus I like putting crowns that can fit into a 10mm puck on 10mm puck. I hate mixing crowns with different sizes and just putting them on a larger puck. But that's just me. Who did you get your SUM through? Maybe they have a better version, or contact CAP and have them check it out. This thing seriously rocks man. I just nested 3 5-unit pmma bridges for stents, usually takes up to 20+ minutes but it calculated it under 5.

 

[NicelyMKV]

You can also allow it to start after the first pass is calculated

 

[ParkwayDental]

There is no lack of people in my Dept. I have 4 employee's one person scans all day, myself and two other techs design and run the mills.

It is not a people problem it is a software problem.

My Picsoft takes 15 minutes to calculate a full puck.

Sum3D is a powerful software there is no doubt about it but I cannot wait on the software when I have 100+ units a day.

Maybe I need to look at a new computer? I am still learning on Sum3D so I do not know all the ins and outs of how to modify things.

Even when I import the stl files in it sits at 93% for a minute before it finally brings in the file.

We aren't cheap asses and we run quality equipment. I am just bummed, I spend 15k on software and it doesn't perform to my expectations.

Even if it took 20 to 25 minutes to calculate a full puck I would be happy with that.

I nested a 12 unit screw retained CrCo bridge and it took over 5 hours I mean really?

 

[eyeloveteeth]

things that slow down nesting.

Different height crowns. if the rough has to cut through a lot of material first... i.e. 9.5mm crown in a 18mm puck...

bridges are in there, copings mixed in.

We do on average 180 cases a day . The lab outputs about 250-300 cases on low average. and we only have 3 machines juggling wax, zr, ultimate and pmma. Could I use 2 more machines? absolutely, but I'd say SUM3D is fine on our end.

I really do think you should try batching everything in as similar of a profile as possible. Especially since you have the staff available. Any "work" you do before the calculation is going to help it. But then you run into the issue of is it simply faster and more efficient to let it run an extra 15 minutes vs having some human try and do it.

 

[ParkwayDental]

^^^ This doesn't make sense though why should I have to nest in small batches? This doesn't make any sense. Sum3D is running on old architecture and Luigi needs to get off his ass and get in the game here and make the necessary changes!

Nesting in small batches is not the way it is suppose to work. The point of this software is to fill a full puck and let it ride. That is not a normal work flow and is ass backwards to work like that.

 

[zero_zero]

The main problem is that Sum3D is using quite a few licensed libraries (like all 5axis, simulation, import etc.),every call goes through by windows API's creating a bottleneck... haven't checked lately, but a few updates ago still had reminiscent DOS era code...

 

[Labwa]

Are you running Quadra cards on your CAM PC Scott?

 

[cadfan]

At the moment they have a problem the software runs under 64 bit but not supporting, the 64 bit is slower than the 32 bit version they try to fix it. sometimes the programm has problems with other programms the zz guys had a problem with acronis true image that was wrong and made it 10 times slower. If you have a lot of restmaterial calculation in the programm this takes real time but not 5 hours. Its not the easiest and fastest programm

 

[primus]

5-axis takes quite a while. Especially if you have loaded in the fixture CAD and all sorts of stuff for the software to consider during calc. for collision avoidance.

Anothet topic.....You can choose to make strategy that does NOT update the working part after every operation. (software will re-draw in CAD the new shape after an operation to be used by next operation)

There are Two ways to "attack" the milling strategy when using more than just One rough, and One finish. (Complicated milling, usually hard materials)

One way actually updates and re-calculates the exact new CAD shape after milling inbetween different tools/operations, and this can be set to a certian level of tolerance as well. Using the <MATRES.04>command and a value. This turns down the resolution of the newly re-worked model. (speeds up calc times!)
I think it stands for Material Resolution? ??

The next operation now looks at this new CAD shape as the material left to cut, and goes from there.

Another method is to manually type in what the previous tool's diameter and it's z-step in the milling parameters to determine what/where the next tool should be doing.
The CAD of the working model is NOT updated and used. Much faster because it skips this step, but easy to make situation that has too much material to be removed in some pockets when going down in tool diameter. You can "lie" about previous tool's geometry to provide a wee-bit of safty margin, so you don't overload a tool and fry it.

 

[zero_zero]

Yeah, wish we were supplied with a fat manual for that price... It takes forever to experiment with all that many options it has...

 

[primus]

yes, this is the problem.

Takes so long to experiment, indeed!

Three years and counting for me, and I am still finding functions and features hiding.

 

[Macron Dental Lab]

Sum3D is slower then molasses, it takes an hour to calculate a full puck. I mean really come on now!

I've heard it runs on a single core why? Why not up grade and get into 2013?

15-20 on my German cam soft -and 36g DDR3 , 3 sli nvidia cards and a monster processor a great combo.