4 ms·

In most manufacturing, you have a prototype hall or floor where you have somewhat more general purpose/flexible machines with a lot of manual intervention to ma

In most manufacturing, you have a prototype hall or floor where you have somewhat more general purpose/flexible machines with a lot of manual intervention to make small quantities of prototypes, and factories with automated machines set up to make the exact parts you need for your full production. This was always my experience in electronics - you, as the development engineer, set up small-scale production (going from prototypes to EV or engineering validation articles, which are basically the final product produced by these manual/slow methods), and then manufacturing engineers work with you to transfer your processes to an automated assembly line making who knows how many units a minute. It's not unusual to produce a special prototype in the small/experimental floor; I mean, Apple probably goes through hundreds (or whatever) of prototypes and test articles before releasing their next phone, and they are certainly not stopping their million-per-day manufacturing plants to make them. However, it does show impressive commitment to the customer on the part of Fujitsu. Also, electronics is more cyclical than cars (3-year-old model may be completely different and useless compared to the current one), so it is common to produce everything in batches rather than "the Toyota way" with just-in-time. In other words, if a laptop model is sold for 2 years, the whole run may have been made in 6 months, and then you have that manufacturing equipment sitting idle while the next prototype is worked on, so you may have equipment and engineers to spare. Only a few of the highest-volume factories (e.g. Apple again, I assume) run at full capacity all the time.

notoothpaste 10y ago
Thank you for your sincere and informed response. I understand that EEs create countless prototypes over the course of the development of a product, and that development PCBs need not (especially for debugging purposes) bear any physical resemblance to the finished product. What is impressive to me is that Toshiba could: Redesign and test a PCB Set up a manufacturing process for pick and place etc Create an injection-molded case Put it all together Deliver all of this, on a whim, to a customer in the single digit quantities And apparently do so to similar standards as their mass-produced machines Did they just eat the cost of the most expensive PCB masks and injection molds, or is there something more interesting going on behind the scenes? That's what I want to know. EDIT: I tripped the (whatever the fuck random sequence of characters you need to bypass the filters) filter and now I can't post any more for a few hours, otherwise I'd happily call all of you out as the ignorant yuppie douchebags you are... no, worse, the 19 year old CS majors pretending to be ignorant adult yuppie douchebags. tl;dr I just wanted the chance to hear from experienced engineers. Please fuck off if you don't know what you're talking about, for the sake of anyone that wants to ask a remotely relevant question on this braindead hellhole.
- kw71 10y ago
  Some of the external ports in my lifebook are simply mounted on the housing and connected to the boards by cable assemblies, so: They may have stuck the needed connector on a wire and used an NC mill to make an appropriate cutout from an existing housing part. It is literally less than 1 hour to design and tape-out a inch-square pc board to accomodate a connector and cable termination. I am a little guy and have to wait a week and pay five bucks to have this manufactured. Fujitsu probably had this done the morning after a solution to the request was devised and approved.
  - Spooky23 10y ago
    I think you are right, but the idea that they were willing and able to actually execute was incredible to me. Alot of competence and strong relationships were in place in the sales and manufacturing organization for that to happen. As a customer, I think we bought ~1,000 units in the end. I've rarely seen that type of capability in my career, and it's too bad that the realities of the market have likely nixed it.
- yaakov34 10y ago
  You seem to want answers at a level that only the engineers at Fujitsu (not Toshiba, according to the OP) can provide. I can comment on some of the specific questions, but obviously I don't know what went on at Fujistsu. 1) The final development PCBs (the engineering validation articles) are actually identical to the production PCBs. Your final EV revision becomes revision 0 in production; only the manufacturing method changes. So the prototype floor definitely has the ability to make production-quality PCBs. 2) The design change may or may not be complicated. Adding a port typically refers to a port already supported by the chipset; so you add a ribbon connector for it and change the layout only. 3) You generally wouldn't have full pick-and-place automation for making prototypes. There are manual tools (sort of semi-automated) for making a few PCBs. A large company is making prototype PCBs all the time. They know how to do it, and they don't stop manufacturing for it. 4) For the case, the easiest would be to rework an existing case (add a slot). Creating custom injection molds is kind of a big deal. Then again, if Fujitsu does that in house, maybe not that big. Don't know. 5) The expense would not involve a cash outlay, but it would definitely involve diverting engineering, manufacturing and sales resources to even communicate and deliver this whole thing, let alone make it. The engineers may have been pulled off something, or may have been waiting for work; business hasn't been that great for most manufacturers, if you haven't noticed. No, I don't know how Fujitsu accounts for this or why they made the decision to go ahead. I am not a manager at Fujitsu. I can tell you that most companies are very loathe to let prototypes and one-offs to leave their hands at all. So it's definitely impressive on Fujitsu's part that they would go to that length.

notoothpaste 10y ago

Thank you for your sincere and informed response. I understand that EEs create countless prototypes over the course of the development of a product, and that development PCBs need not (especially for debugging purposes) bear any physical resemblance to the finished product. What is impressive to me is that Toshiba could: Redesign and test a PCB Set up a manufacturing process for pick and place etc Create an injection-molded case Put it all together Deliver all of this, on a whim, to a customer in the single digit quantities And apparently do so to similar standards as their mass-produced machines Did they just eat the cost of the most expensive PCB masks and injection molds, or is there something more interesting going on behind the scenes? That's what I want to know. EDIT: I tripped the (whatever the fuck random sequence of characters you need to bypass the filters) filter and now I can't post any more for a few hours, otherwise I'd happily call all of you out as the ignorant yuppie douchebags you are... no, worse, the 19 year old CS majors pretending to be ignorant adult yuppie douchebags. tl;dr I just wanted the chance to hear from experienced engineers. Please fuck off if you don't know what you're talking about, for the sake of anyone that wants to ask a remotely relevant question on this braindead hellhole.

kw71 10y ago
Some of the external ports in my lifebook are simply mounted on the housing and connected to the boards by cable assemblies, so: They may have stuck the needed connector on a wire and used an NC mill to make an appropriate cutout from an existing housing part. It is literally less than 1 hour to design and tape-out a inch-square pc board to accomodate a connector and cable termination. I am a little guy and have to wait a week and pay five bucks to have this manufactured. Fujitsu probably had this done the morning after a solution to the request was devised and approved.
- Spooky23 10y ago
  I think you are right, but the idea that they were willing and able to actually execute was incredible to me. Alot of competence and strong relationships were in place in the sales and manufacturing organization for that to happen. As a customer, I think we bought ~1,000 units in the end. I've rarely seen that type of capability in my career, and it's too bad that the realities of the market have likely nixed it.
yaakov34 10y ago
You seem to want answers at a level that only the engineers at Fujitsu (not Toshiba, according to the OP) can provide. I can comment on some of the specific questions, but obviously I don't know what went on at Fujistsu. 1) The final development PCBs (the engineering validation articles) are actually identical to the production PCBs. Your final EV revision becomes revision 0 in production; only the manufacturing method changes. So the prototype floor definitely has the ability to make production-quality PCBs. 2) The design change may or may not be complicated. Adding a port typically refers to a port already supported by the chipset; so you add a ribbon connector for it and change the layout only. 3) You generally wouldn't have full pick-and-place automation for making prototypes. There are manual tools (sort of semi-automated) for making a few PCBs. A large company is making prototype PCBs all the time. They know how to do it, and they don't stop manufacturing for it. 4) For the case, the easiest would be to rework an existing case (add a slot). Creating custom injection molds is kind of a big deal. Then again, if Fujitsu does that in house, maybe not that big. Don't know. 5) The expense would not involve a cash outlay, but it would definitely involve diverting engineering, manufacturing and sales resources to even communicate and deliver this whole thing, let alone make it. The engineers may have been pulled off something, or may have been waiting for work; business hasn't been that great for most manufacturers, if you haven't noticed. No, I don't know how Fujitsu accounts for this or why they made the decision to go ahead. I am not a manager at Fujitsu. I can tell you that most companies are very loathe to let prototypes and one-offs to leave their hands at all. So it's definitely impressive on Fujitsu's part that they would go to that length.