Clean Sheet Setup.

[Cadmonkeychris]

Hello All,

I am about to join a fledgling startup where I will be wearing a few hats, including engineer and designer. One of my early tasks is to choose and implement a CAD system. A clean sheet of paper!

I can’t divulge details of what we are working on but we will require a system that can handle all of the usual mechanical stuff along with piping, wiring harness design, thermal (flow) and stress analysis, good surfacing capability etc. Ideally there would be a native or closely tied PLM system that could start with a small and simple(ish) setup and scale as the business grows.

Local manufacturing tbd but I imagine built in CAM would be a good thing. Good ECAD integration is also necessary, though perhaps not immediately as the (very experienced) sparky is set in his design tool ways (he’s knocking on a bit).

Several of us have a 20yr+ background in SolidWorks, including the founder, so we are well aware of the pros and cons of that system. We have been watching DS’ behaviour with dismay so our conclusion is that we go anywhere but SW, 3Dwhatever and DS in general.

My (our) immediate thoughts turn to NX or possibly Creo but with the ongoing transition to the cloud, I wonder if we would be missing a trick by selecting older tech. Not that we are particularly keen on cloud based stuff…

Assume that money isn’t a particular issue, though we wouldn’t be averse to enrolling on one of the CAD for start-up schemes available.

TLDR: if you had a clean sheet of paper and had to select a bells and whistles, future-proof CAD setup tomorrow, what would you do? Anything we should ask or be aware of?

Thanks,

Chris.

[zwei]

My 2 cents when it come to selecting CAD...

1. Take "CLOUD" or whatever it is called with a grain of salt... the last thing you want is someone holding your data/design hostage because of a "CLOUD" or server issue...

2. Learning Curve matters... Don't expect people to switch CAD and be able to work right-away if they do not have experience with the CAD system in the past... No CAD system is created equal, some is harder to learn than other

If money isn’t a particular issue, NX or CREO seem like a good fit. You might also want to look at SOLIDEDGE.

One thing to note is that CREO will have a steep learning curve and take some time to get used to, I have seen people, including myself, struggle when switching from SOLIDWORKS to CREO (I had lost count of how many times i get mad at "NOT PART OF CREO FUNCTIONALITY" answer when asking for the most simple stuff like note numbering...)

I dont have much experience with NX so i cant comment on that.

[Cadmonkeychris]

Thanks for replying.

we're not keen on keeping everything on "the cloud" though it has obvious advantages. I have freelanced for years and find this aspect of F360 very useful, if just as a means of allowing clients to spin around and modify SW designed files (interactive mark-up).

The learning curve is an issue, as is the availability of designers that have relevant experience. SW's ubiquity means that pretty much everyone who would end up working on this project will have had some exposure to it. A couple of us have significant individual investments over the years in SW (I have 16 yrs of subs behind me, another 20+) so this would be a big commitment for us. We have some idea of the fun and games that will arise with everyone picking up a new product at the same time, but now is the time to do that, I think.

I used Wildfire a while back and I am not surprised to hear that Creo still has a learning curve. This does worry me as Wildfire required a more rigorous approach to get stuff done correctly whereas SW allows for all sorts of sins...

I have worked with and for people coming from an NX background over the years and get fed up with hearing complaints about SW's lack of functionality. They all end up on SW due to initial cost and the overwhelming need to supply clients with SW native CAD.

I am looking through SE stuff at the moment.

[SPerman]

I have experience with Ideas (RIP), NX and SW. Of those 3, I would chose NX every time, if cost were not a factor. I doubt there is a challenge you will encounter that NX can't solve.

[AlexLachance]

My 2 cents when it come to selecting CAD...

1. Take "CLOUD" or whatever it is called with a grain of salt... the last thing you want is someone holding your data/design hostage because of a "CLOUD" or server issue...

2. Learning Curve matters... Don't expect people to switch CAD and be able to work right-away if they do not have experience with the CAD system in the past... No CAD system is created equal, some is harder to learn than other

If money isn’t a particular issue, NX or CREO seem like a good fit. You might also want to look at SOLIDEDGE.

One thing to note is that CREO will have a steep learning curve and take some time to get used to, I have seen people, including myself, struggle when switching from SOLIDWORKS to CREO (I had lost count of how many times i get mad at "NOT PART OF CREO FUNCTIONALITY" answer when asking for the most simple stuff like note numbering...)

I dont have much experience with NX so i cant comment on that.

I wish I could sit with you and have a talk about your knowledge, you seem pretty balanced both in terms of softwares and sharing opinions.


Back to the subject at hand, it all depends on what you are drawing too. If it's complex, then of course you won't draw it on a sheet of paper with a pen, but if what you require to draw is not complex, maybe you can keep as simple as DraftSight

[Cadmonkeychris]

I have experience with Ideas (RIP), NX and SW. Of those 3, I would chose NX every time, if cost were not a factor. I doubt there is a challenge you will encounter that NX can't solve.

I keep hearing this. Currently advised that cost isn't an issue...

[Cadmonkeychris]

I wish I could sit with you and have a talk about your knowledge, you seem pretty balanced both in terms of softwares and sharing opinions.


Back to the subject at hand, it all depends on what you are drawing too. If it's complex, then of course you won't draw it on a sheet of paper with a pen, but if what you require to draw is not complex, maybe you can keep as simple as DraftSight

Site installation drawings might well end up being on DraftSight or similar - keeping that as a separate thing for later as we're not aware that any of the packages deal with architectural type stuff well. Feel free to correct me if you know otherwise.

The system / products themselves will need more than 2D. There will be ID work as well as complex multi-disciplinary hardware design going on. SW could probably handle most of the ID work but I have had enough of SWs flakiness on complex geometry and the kludges often required to get to a desired form.

[HerrTick]

I doubt there is a challenge you will encounter that NX can't solve.

One challenge NX can not solve is the availability of skilled NX users. I just spent 4 years doing NX. New users were not adapting as well as they ought. Recruiting NX talent is not easy.

While you dislike SW for obvious reasons, you have a lot of compelling reasons to continue with it. Most notably, you already have a nice pool of SW knowledge going in. You won't become indentured to someone else's expertise. You can interview (more readily available) designers with confidence. You can assess problems more readily.

OnShape seems to be maturing nicely. Worth a look. If your product isn't too complex in the surfacing department, probably a good bet.

I would avoid CATIA.

[KennyG]

I think all of them are working to insure they will be cloud capable when there is actual demand for it, so I would not be too afraid of going to PTC or Siemens because of that. I think you really need to decide if a mainstream app like Solid Edge or SolidWorks will suffice or do you need advanced features provided by NX, Catia and possibly CREO. It is really a decision of simplicity and economy vs. complexity and expensive. If you don't need the complexity, then avoid the expense and stick to the mainstream apps which really are SolidWorks, Solid Edge if you also want the other stuff you mentioned including robust PLM as an option.

Now as far as Solid Edge is concerned, you mentioned you were looking at it so here are a few things to consider...
They do offer a program for startups with free software for a year (Premium licenses, not sure about other stuff in the portfolio)
Solid Edge CAM Pro 2-1/2 axis included with Classic or above licenses under maintenance (NX CAM). Obviously more advanced capabilities available if purchased.
Siemens owns Mentor Graphics, so SE Electrical and Wire Harness are those tools. Look into the live linking between them and the SE mechanical model... it's cool stuff.
Teamcenter is the PLM option unless you just want workgroup level SE data management which is built in (SEDM). SE development team writes the integration with TC and it is very thorough and tight. TC offers on prem or cloud hosted. RapidStart template for on prem establishes PDM capaboilities quickly and economically but can expand to full TC suite if needed.
Solid Edge Simulation based on FEMAP and NX Nastran does steady state and transient thermal. CFD available too.

[Lapuo]

One challenge NX can not solve is the availability of skilled NX users. I just spent 4 years doing NX. New users were not adapting as well as they ought. Recruiting NX talent is not easy.

While you dislike SW for obvious reasons, you have a lot of compelling reasons to continue with it. Most notably, you already have a nice pool of SW knowledge going in. You won't become indentured to someone else's expertise. You can interview (more readily available) designers with confidence. You can assess problems more readily.

OnShape seems to be maturing nicely. Worth a look. If your product isn't too complex in the surfacing department, probably a good bet.

I would avoid CATIA.

Why would you avoid Catia?
I am curious.

[HerrTick]

CATIA does not play nicely with other systems. Translating CATIA to other systems, even by neutral formats (e.g. STEP, IGES) does not work well.

Internal tolerances of CATIA's modeling kernel often lead to problems when sending data to outside systems. The problem is rooted in the CATIA kernel itsel. What are often seen as translation problems are the result of bad source data translated accurately.

Good luck recruiting CATIA talent. Worse market than NX.

The only clear case I see for a small enterprise to use CATIA is if one is doing business in an existing CATIA ecosystem, such as major automotive or aerospace suppliers.

[zwei]

CATIA does not play nicely with other systems. Translating CATIA to other systems, even by neutral formats (e.g. STEP, IGES) does not work well.

Internal tolerances of CATIA's modeling kernel often lead to problems when sending data to outside systems. The problem is rooted in the CATIA kernel itsel. What are often seen as translation problems are the result of bad source data translated accurately.

Good luck recruiting CATIA talent. Worse market than NX.

The only clear case I see for a small enterprise to use CATIA is if one is doing business in an existing CATIA ecosystem, such as major automotive or aerospace suppliers.

I agree with this... Never had good experience importing CATIA data. even when they are converted to STEP.

[zxys001]

SolidEdge.

[XHawkeye]

[snip]

While you dislike SW for obvious reasons, you have a lot of compelling reasons to continue with it. Most notably, you already have a nice pool of SW knowledge going in. You won't become indentured to someone else's expertise. You can interview (more readily available) designers with confidence. You can assess problems more readily.

[snip]

This is the best reason to stay with SW, you'll be up and running in no time.

======

PDM: If you do stay midrange check out DDM. They are from your neck of the woods and it works with SE, SW, IronCAD, Creo Parametric, Acad and Inventor. At the current job DDM was used with Creo & Acad but when they switched to SW it was an easy connect to DDM.

======

Be sure to let us know how it goes.

[Frank_Oostendorp]

Whatever you choose, preferably NX, you need to invest in proper training. Find a guy with 10+ years of experience in real use of the chosen CAD software in a similar business. Let him train your users good. You can not expect someone to be trained in a new software within a week. It will take like 4 hours a week for a whole year to get good educated. Send a few people to extensive external training to become an expert. If money isn't a problem, you could easy convince an expert to leave his present job.

[jcapriotti]

There isn't really enough to go on here in regards to the product you're developing. IMO, the software you needs is really dependent on what you're designing/manufacturing. It also depends on how big you expect your company to grow and how quickly.

Most here would say NX if money is no object but money isn't important until it is. NX is powerful if you need it's capabilities but you will pay for it, and its steeper learning curve. Ironically I found it's method of modeling to be somewhat similar to SolidWorks and picked it up easier than I thought.

[SPerman]

Whatever you choose, preferably NX, you need to invest in proper training.

The last place I worked decided to send the CAD admin to training, and then let him train the masses. 2 problems with this. 1) He had been an admin for so long he didn't really have much experience as a designer. 2) He was a horrible instructor. So the company saved tens of thousands of dollars in training, and in exchange got a workforce that didn't know much more about the software than they did before wasting a week in training.

[Cadmonkeychris]

Thank you to all that have replied. I can't tell you precisely what we will be working on due to commercial sensitivity and I have been wracking my brain to find something similar.

I appreciate the comments about retraining existing users, finding talent (and paying for it...) I meet few designers that will attempt to keep a clean SW feature tree and not leave a tree of pus and blood festering in their models.

The potential for the product / system is such that I think that we should be looking at higher-end package. I would expect our data to have to play nicely with other manufacturers' systems as our product is intended to be stand alone, adaptable or retrofittable. My quick delve into donor systems suggests that the majority of manufacturers use NX and Creo.

I agree that staying with SW has its attractions. I wonder if other systems are a bit better at cracking down on / preventing user misbehaviour (modelling, external geom. etc.) I spend a fair amount of my time correcting others' modelling so would like some way of nipping this behaviour in the bud!

I have downloaded the community version of SE. So far so good but as usual, whenever I try to get to grips with a package, I get swamped by short term deadlines suddenly popping up!

[HerrTick]

One big plus for NX vs Creo is the direct modeling tools. If you need to manipulate unparameterized solids, NX has great tools (and has had them for over twenty years).

NX also has a good suite of mesh manipulation tools and a good mesh-to-surface capability. I find NX does a better job of importing and sorting meshes, and also does great at creating best-fit surfaces from meshes, including best-fit planes and revolves. But SW is still better at making whole solids from meshes. NX and SW seem to complement one another well on such projects.

[tsmith]

The import tools in SE are impressive, especially when pulling in SW models.

If I could change us from SW to SE I would. I think mainly because the more rigorous nature of SE helps to ensure that users are making things in a *closer* to good way.

The other side of the coin is that SW lets you get away with things, which as long as you understand the ramifications of what that will stop you from doing later, it can be managed.