How long to design new products? 5

[hygear]

I'm trying to get an idea of how long it should take to design new products from other engineers. The reason for this is that I feel like it is taking my company an incredibly long time to go from design to mass production even on simple projects. My goal here is to get input from others and try to figure out where we can improve.

To give a little background, I am part of a small design team that designs accessories and add-ons for heavy equipment. This could be anything from designing a hydraulically actuated implement to creating a special guard to protect the machine during off-road use. We typically spend anywhere from 1-3 months creating a design(including FEA) but our real issue seems to be outside of engineering. It takes around 1-3 weeks for our purchasing department to process a prototype order,1-4 weeks for quotes, anywhere from 2-6 months to get prototypes, another 1-4 weeks to get through quality inspections, and another 1-2 weeks to get released to our testing department.

After we finally receive parts, our testing department will run testing on the piece of equipment until we reach the full design life of the parts (while also collecting strain, acceleration, and pressure data). That means if it is rated for 500 hours, we will run for the full 500 hours which can take around 16 weeks to finish.

The two main issues here is that it is taking us around 31 weeks to go from concept to prototype and another 16 weeks to finish testing. When you add in design review meetings, documenting design issues, and getting manager approvals for each step of our PLM process, we are usually running over 1 year to finish even a simple project.

Here are some of the questions I have for others mechanical engineers that design welded assemblies and mechanisms:

1. How long does it usually take you to get fabricated parts? My team is good at designing for low volume so we utilize a lot of tab and slot to reduce tooling, minimize bends, and minimize welding. In my mind it shouldn't take 6 months to get most of the parts we order since I could build them in my garage in a week.
2. Does your company have a policy that mass production parts must be produced by the same supplier that creates the prototype parts? In my opinion, this shouldn't be necessary both suppliers can meet the drawing and the design is correct.
3. Does your company require full life testing or do you do some kind of accelerated testing? We do fairly detailed FEA of all our different loading conditions and the loading conditions are based on previous testing. My feeling is that we design way too close to our target life which is the reason we need to do full testing. I also believe that increasing material and safety factors would be much cheaper than running all the tests (the managers see otherwise).
4. Does your company do 100% inspection of prototype parts or do they just check key features and release them for testing? Currently our QC department does 100% inspections including all of the child parts. When you have a fabrication that is welded of several parts, this can be extremely time consuming even with a Faro arm.
5. Does your company increase complexity of a part just to reduce material? My company has the mentality that reducing material reduces cost but I know this isn't true. I would rather increase material or use a higher strength steel than add lots of gussets or extra bending steps.
6. Does your company avoid the use of loctite? One of the biggest design issues we have is loosening bolts due to painted surfaces. Over time the paint wears away and the clamp forces will reduce slightly. When this is coupled with the fact that we insist on using spring lock washers, everything gets looser faster and we require more bolts or more testing. At my previous job we used loctite on pretty much all the fasteners and never had issue of loosening bolts even after thousands of hours.

Thanks for any feedback you can provide!

 

[TheTick]

Locknuts

 

[TheTick]

Sounds like the company as a whole doesn't quite grasp what a prototype is.

#2 is just madness. That's half your problem right there.

Most places I've been don't let purchasing touch proto quoting except under direct supervision of engineering. Prototyping is not production. When you buy a prototype, you are buying knowledge. Bean-counters who get bent out of shape about per-piece costs far exceeding production costs should be kept far away from the process.

Saving material saves weight. That's usually good. But if you are not casting/molding and not reducing blank size, reducing part weight doesn't help.

Using higher strength steel does not make anything stiffer than plain carbon steel. It takes geometry changes to change stiffness. I'll bet this has come up.

 

[3DDave]

You need faster suppliers. I expect the bids are going to the cheapest, not the fastest. This makes procurement look great for saving money and engineering look bad for taking so long.

1) It should take a week or two, but that may depend on an on-going relationship with a supplier and a willingness to spring for head-of-the-line costs.

2) It wasn't a policy, but we did really low volume. The danger is that a prototype house will use more care and be closer to nominal, while a production house or even production from the original maker will be sloppier. I've seen a number of designs that worked great in prototype only to require some unplanned for debugging of production parts.

3) It was Military contract work, so they specified the testing, usually accelerated in sometimes unrealistic ways. For example, wear tests run so fast that heat buildup and lubrication loss became limiting factors. Many companies turn the prototypes over to an early user group who will beat the daylights out of it in ways not easily specified. Your results will vary.

4) 100% checking is a good idea. As I mentioned in #2, the prototype parts may vary less from specified than production parts and it is useful to know what the differences are. I've seen items where problems came from areas that weren't 'critical' to function - surprise! But if you mean accounting for each square millimeter, then maybe not.

5) Material reduction is only useful when the material costs are a dominant part of the overall cost and fabrication costs are not. For example, injection molding can produce complicated shapes for the price of the heat and the power to run the injection process. For small variations in complexity the production cost doesn't change, but if one can cut half the material then that's half the material cost saved. On the other hand, if one takes $10 worth of time on a CNC and doubles that to save 50% of the material and the material saved is prorated at $3, then it's a bad choice. And if the savings means moving to a higher degree CNC (3-axis to 6-axis) so it's now $20 per unit, then it's even worse.

This isn't to say that other costs should be ignored. Lighter parts are less costly to ship, but a complete analysis of all cost factors is required, including the additional time to document the weight saving features, any inspection, and any additional difficulty in processing such as painting or plating.

6) Locking compounds can help or not. I've seen instances where loosening fasteners got locking compound only to have the fastener end up failing from fatigue, leaving field service trying to remove the glued in stub fractured, broken off even with the surrounding surface. Sometimes loosening is a sign of a misunderstanding of what the fastener is resisting. I would look to higher strength fasteners and hardened washers with higher installation preloads. Masking would also be a help a would ensuring enough friction is available between the parts or reconfiguring the design so the fasteners resist less shear.

 

[IRstuff]

Note that normal lock washers DON'T WORK, period.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[CWB1]

1. Depends on the commodity type and what youre willing to spend. Standard lead time on machined parts and weldments is usually 4-6 weeks, castings and forgings usually are 8-12 depending on size and complexity. Using personal relationships developed during supplier visits, sales pitches, free lunches, etc to beg favors and paying 0-30% markup can get me down to <2 weeks on any commodity in a pinch however.
2. No, using a production supplier for prototyping is simply absurd. Inexperienced managers always try to justify it as potential savings on tooling costs but in reality the earlier the production supplier gets involved the more costs get padded, literally they step over a dollar to pick up a penny. Lead times are also usually significantly longer and the pool of suppliers willing to bid gets smaller, again raising costs significantly.
3. Full life-cycle testing? Do they also test a large quantity of prototypes to provide for possible anomalies? I've worked quite a few highly regulated projects between defense, mass-transit, nuclear power, and the oil-patch and never done so. A combination of simulation and engineering judgement based upon history is carefully applied to design and implement accelerated testing.
4. No. Proto inspection is done only at the behest of product engineering, manufacturing engineering isnt involved until the part is released for production. Inspection varies by part, non-critical parts may not require inspection at all while critical parts may need full batch or additional inspection not normally done in production (ie 3d scanning). Most parts however do not require anything near the partial-batch, partial-print inspections done during production.
5. Depends on the part. This is another time that great relationship you developed touring suppliers and listening to their sales pitches is essential to get open, honest feedback. The goal isnt part complexity or material, its cost reduction, and this is where the supplier's expertise with THEIR processes and THEIR industry is vital.
6. There's no reason to rule out a possible design tool, especially on accessories for equipment held together with fasteners loctited by said tool. The biggest issue I've seen with loctite is simply designers choosing the wrong one, often not considering temperature and chemicals the joint might see.

 

[EdStainless]

Engineering should have relations with houses that can build protos.
Design, evaluate, and then get some built. No purchasing involvement.
From start to having the protos should be 2-4 months.
It should get thorough your inspection in a couple of days.
And the test plan should already exist, full testing should not take full life, you can tell if something is failing.
If the bonus money for purchasing, testing, engineering and such all depended on new product development you would not have issues. The problem is that Sr mgt does not care if you do this or not. Ask you boss if it is OK with his boss that this takes so long.

Ban the use of lockwashers, use loctite or locknuts.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[TheTick]

Like Ed said: build relationships with proto vendors. There are things that can be done to expedite proto parts:
[ul]
[li]Preferred drawing and file format for each vendor[/li]
[li]Specific contact person[/li]
[li]Willingness to adapt to use readily available material (i.e. what's on the shelf)[/li]
[li]Common fasteners[/li]
[li]Get creative with laser cut sheet metal. You can make quickie weld fixtures and assembly aides[/li]
[/ul]

One more thing: NEVER ask for free stuff. NEVER make changes in mid-order without asking how much it costs. NEVER try to expedite without paying for the speed.

Do these things, and eventually your vendors will start doing the things you never ask.

 

[MikeHalloran]

6. If indeed 'we insist on using spring lock washers', 'we' are doomed.
3. If full life cycle testing is considered valid on a sample of one unit, your managers do not understand statistics, and again 'we' are doomed.

Get your resume in shape and get out.
You can't fix stupid.


Mike Halloran
Pembroke Pines, FL, USA

 

[GregLocock]

I'll answer from two perspectives, first is big automotive OEM, second is a small successful industrial product

1 4 month/3 weeks
2 the eventual supplier must arrange the prototyping whether in house or externally/ No
3 accelerated/accelerated
4 No/No
5 Yes to some extent /yes at significant expense
6 Yes/ Not applicable but we would have used it where necessary


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[btrueblood]

1. How long does it usually take you to get fabricated parts?

4 weeks. Castings and forgings take longer, as do other parts requiring complex tooling...but we can and have accelerated prototype castings using a 3D printing service to make patterns.

2. Does your company have a policy that mass production parts must be produced by the same supplier that creates the prototype parts?

No, not usually. We do have a local shop we work with, and I have tried to ensure that certain parts go to him (he helped me figure out how to make the part at best economy, and he is usually within a few percent of the low bid, so it makes sense...having a regular cash flow helps him stay in business, and makes him more willing to bail us out of jams when we need him).

3. Does your company require full life testing or do you do some kind of accelerated testing?

We do accelerated testing, typically at worst case "corner conditions" (hottest, harshest, highest pressure) of the spec envelope.

4. Does your company do 100% inspection of prototype parts or do they just check key features and release them for testing?

We check first articles 100% on production orders, then do samples after that. Prototypes...vary. Typically it's just the critical features.

5. Does your company increase complexity of a part just to reduce material?

We simplify parts (combine two or three parts into one cast part for instance) to reduce parts count and labor...sometimes it reduces material...rarely do we add complexity. "Simplificate and add lightness" should be every design engineer's motto.

6. Does your company avoid the use of loctite?

Opposite since I started. And nobody I know trusts or uses lock washers.

 

[drawoh]

btrueblood,

My new employer is determined to use helical spring lock washers. Also, they are determined to use calibrated torque wrenches to install screws. They (we?) do vibration testing in-house, so they have documented results. Maybe the torque wrenches make up for the lock washers!

--
JHG

 

[Tmoose]

" Maybe the torque wrenches make up for the lock washers! "

I think it's the rabbit's foot dangling from the rear view mirror in the PM's car. Or maybe the fuzzy dice.

 

[CWB1]

drawoh, given that one of the most common causes for issues with lock washers is over-torqueing then yes, torque wrenches should help in that regard. Another common mistake is poor joint design that is overly elastic. From experience I can attest the machines the OP's attachments will mount to have many split lock washers and they are a good step between applied torque alone and locking compounds but like anything else they need to be used properly. Personally I am always amazed by other engineers making blanket statements for or against something, given the challenges we often find ourselves with I would think they would want as much design flexibility as possible. OTOH, I have a reputation for developing products and technology others said couldn't/wouldn't work much less make production, so part of me also enjoys these statements.

 

[tbuelna]

In most of my experience, testing of prototypes that do not conform to a production configuration is usually avoided. Primarily because the results from testing a non-conforming prototype article are not valid for the qualification/certification of a production design configuration. Testing is expensive and time consuming. Ideally, it is most efficient to get your production design squared away before building any test hardware. Testing should primarily be used to validate your design and analysis work. All test articles should be thoroughly inspected, documented and controlled, if you expect to end up with acceptable test results.

 

[rb1957]

There's some sense to having the prototype made by the guy who's going to do the production manufacturing.

A separate prototype manufacturer will quite naturally focus on solving the prototype issues, without much emphasis on the production manufacturing implications.

A manufacturer doing the production run is in a better place to "recover" costs of the prototype. Production tooling could be a massive expense; would someone want to use someone else's tooling ? not "plug and pray" I'm sure.

another day in paradise, or is paradise one day closer ?