Race to the bottom contest

Sorry about the dual post, but I'm worried the new competition forum won't be as widely read.

A new contest for the electronics designers out there. Please feel free to talk this up anywhere and everywhere you like. Good publicity is very important to a contest like this. For example, during the sign up period of the Vertical X Axis contest, after a month of near daily posting everywhere I could think of, while enduring the grumbles of the very active reprapers about spamming, and "We've all heard this before, just shut up already." I still got "Contest, what contest?" reactions from people in the community. I even had a designer who would have entered but was too late because he hadn't heard about the contest in time.

Rules and prizes are described here

Race to the bottom

There's been a significant change to the contest. SMD designs are now allowed. I was convinced that SMC designs could be just as easy to assemble at home as through hole designs. But to be clear this is still a contest for DIY designs, so any SMD designs should be significantly different from the current crop in terms of ease of assembly.

At the risk of annoying the easily annoyed among you. I'm going to post here one more time to announce a significant upgrade to the first prize award. It will now include steppers and a nice power supply in addition to the Sanguinololu electronics and Prusa heated bed. I have dug deep into my personal stash in an effort to push the contest over the top and get a few more entries before the deadline. Hope this will convince those out there that were on the fence to enter.

Also I will post one more request here for a mention in your blog, etc. Publicity is very important to a contest such as this.

I don't this is annoying at all! Nice to see someone putting effort in motivating developers.

Keep up the good work!

Thanks

Great idea for a contest, it would be good to see some innovation in electronics either on price or performance, the community seems to have converged on an "Arduino clone + allegro drivers" pattern.

I am struggling to see how I could build a stepper driver cheaper than the $10 cost for pololu/stepstick/clone. Chinese CNC stepper drivers using TB6560 come in around $14 an axis, I figure the Chinese assembled price is cheaper than I could buy the parts for.

I toyed with the idea of using L298N and a dedicated Cortex M0 to drive several of them, but that probably would not give PWM current control and only has half-stepping. I am not sure that microstepping is really that necessary, but have not done any testing to prove that. The general consensus is that microstepping is useful, so I'll go with that....

After $40 for stepper drivers, that leaves $30 for all the rest to get to the top end of the cost range, quite a tough target.

You're right that separate driver boards are not likely to decrease in price. Once you factor in the cost of the circuit board and assembly, a $12 Pololu board is a bargain. The savings come when you integrate the components onto the main board, and use components that are DIY friendly. A TB6560 is $3.36 in Qty 10 at Mouser. So there is quite a bit of room to bring price down if you're not paying for a separate board and/or manufacturing. You basically trade your labor for a lower price. The problem is that with current DIY designs the cost of the manufactured Pololu boards eats away most of the DIY savings, making the DIY payoff relatively small. The other problem which is quite nice really, is that manufactured boards like printrboard are pretty darn cheap these days. This also eats into the DIY cost savings. So the point is to have some designs that a DIY enthusiast could build at home and save some serious money over just buying a prebuilt board.

My point is that there is not a serious amount of money that can be saved with TB6560 just because it is through-hole. The A4988 costs $3.28 at digikey, and this is for single quantity. The cheapest course is to merely clone an existing pololu board, rather than design something new.

There is a point where DIY cannot compete with mass production techniques, since the labour cost is a minor component, and through hole components are more expensive than surface mount, plus the DIY does not have benefit of bulk purchase etc. If cost is the goal, I think it would be cheaper to build a system from off the shelf commerical modules than to build a DIY single board.

Anyway, *I* can't see where the significant cost savings can be made with DIY, and perhaps the lack of entrants reflects a wider view. Unfortunately the way the points are assigned in the competition rewards firmware compatibility more than cost, so it would be easy to win the competiton with a minor variant of an existing board which has no improvement on cost.

Quote

There is a point where DIY cannot compete with mass production techniques

This is and always was the case since RepRap started. Still, a Prusa-like design is apparently a lot more popular than a 3D-Systems-Cube or UP! design. Some people started to mold Prusa parts (cheaper, more sturdy parts possible), it didn't pick up.

There's a reason why people print parts instead of using cheap, traditional techniques. Likely, the same reason applies to electronics, too, just not as strongly.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

bobc: I take your point about DIY vs mass manufacturing, but I don't think we've reached that point yet. Just as one example, The BOM for printrboard lists the total price of components at $26 or so. Add in $10 for a PCB and you have a $36 board. So why is the board selling for $129? Looks to me like there's signifficant room still for the DIY enthusiast to save some money. But I think most hobby level builders would be intimidated by the tiny SMC parts sourced for the printrboard. It just wasn't designed with the DIY builder in mind.

Of course, after posting the contest, I was convinced that SMD could be designed to be easy for DIY so it is allowed, but I don't think printrboard, and the like are the type of designs that would encourage large numbers to build on their own. It would need to be a ground up rethink of the design process, and some good education of the DIY community.

That's not to say I'm against mass produced boards. I hope volume goes through the roof and the printrboard, melzi etc. can be sold for $50 at a profit. That would be a great place for reprap to be. I also hope that we have DIY designs that can be made as cheaply so that those who want printed parts and homemade hotends have a DIY design that is still practical to make at home. I think it would be sad to get to a place where the DIY solutions were actually more expensive to build than the manufactured boards are to buy.

bryanandaimee Wrote:
-------------------------------------------------------
> bobc: I take your point about DIY vs mass
> manufacturing, but I don't think we've reached
> that point yet. Just as one example, The BOM for
> printrboard lists the total price of components at
> $26 or so. Add in $10 for a PCB and you have a $36
> board. So why is the board selling for $129? Looks
> to me like there's signifficant room still for the
> DIY enthusiast to save some money. But I think
> most hobby level builders would be intimidated by
> the tiny SMC parts sourced for the printrboard. It
> just wasn't designed with the DIY builder in mind.

That is an interesting example, I think it illustrates one of your points and two of mine. That BOM is for bulk Q's (1000's), for Q10 probably would be at least $50?. Plus, as you observe, it is SMD. Clearly, even at $60 there is a reasonable mark up on final cost, as is expected for a commercial product. But the other points are mine, to reduce costs, buy in bulk and use SMD. Bulk buying is impractical for DIY's, although group purchase is possible. The other avenue to cost reduction is SMD, which there is significant resistance to. But 0805 is easily doable with DIY. SMD components are significantly cheaper than TH, and use less board area.

> Of course, after posting the contest, I was
> convinced that SMD could be designed to be easy
> for DIY so it is allowed, but I don't think
> printrboard, and the like are the type of designs
> that would encourage large numbers to build on
> their own. It would need to be a ground up rethink
> of the design process, and some good education of
> the DIY community.

I would say the problem is willingness to learn new things, not board design. Cost is only a motivation up to a point. Otherwise people would be eagerly learning SMD techniques to take advantage of their low cost. Right now you can buy a bare SMD board, order parts from Mouser, and save money.

> I think it would be sad to get to a place where the
> DIY solutions were actually more expensive to
> build than the manufactured boards are to buy.

Sad, but true for everything else on sale in the shops. The quality and price of mass produced goods long ago put out of business skilled craftsmen in most industries. There is a general idea that DIY saves money, but in practice the advantage of DIY is not cost but getting a customised solution.

Quote

The BOM for printrboard lists the total price of components at $26 or so. Add in $10 for a PCB and you have a $36 board. So why is the board selling for $129?

I think you miss a few points here. While this might be an extreme example - Gen7 parts add up to a similar price, yet DIY-Kits can be bought for $100.- - - also don't forget the Pololus - it isn't valid to simply add up cheapest prices from a number of distributors and consider that to be all. For example, about every distributor has shipping costs. 3 distributors -> $15 shipping.

Then, making kits from a pile of parts takes time, too. So far I've seen nobody who invests a couple of hundred dollars to buy this pile of parts, just to sort and put it in bags for free.

Third, selling/distributing it's self is quite some work. You have to put these bags in envelopes, stick an address on, talk to the purchaser, answer his questions, run an internet shop, deal with international money transfers ... and so on.

To get a reasonable price expectation, I'd substract the Pololus from a current electronics and add the new stepper drivers in. So, ( - $8 + $4) * 4 = -16. A price reduction of $16, perhaps $25 is what I think you can expect.

If you really want to make electronics as cheap as you propose, buy a couple of hundred ATmegas, four time that many Pololus, along with the passive parts, solder all that together and sell that at the price you paid yourself for the raw parts. Can be done today.

Actually, there are people who source parts them selfs and make an electronics entirely without kit. I know this, because they sometimes email questions about Gen7's layout files. But these people are apparently rare, 99 of hundred want at least a kit, 90 of hundred buy "preassembled" (RAMPS, Sanguinololu).

With all that said, I consider this contest still to be a great effort. Even if it saves a $15 per board, only. It also gets rid of some industry dependence.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

I've used these $14 (or cheaper) TB6560 ready-made boards. They work pretty well with everything I've thrown at them and they work with 12V supplies and can do 1/16 microstepping. They come with a generous heat dissipation and opto isolated inputs that work great with 5 volts inputs from an Arduino board. Definitely stronger and more durable than Pololus (also bulkier). I am planning to use them for a ShapeOko CNC machine I am building.

Yes, I totally agree. In fact that was the point of my statement. Why are printrboards $129? Because of labor, overhead and profit. Only self sourced DIY builders can get out from under that cost as Traumflug has said. The point of that statement was that labor, overhead and profit for manufactured boards is not a negligable or minor part of the final cost as bobc argued but actually a large part of the cost. So just designing a made for manufacture board out of SMD parts will not be the most effective way to bring the total cost down as low as possible. It is unfortunate that so few have the confidence to self source and build their own electronics, but at least we can encourage that as much as possible.

Get one of these "shields" [www.ebay.com] and you can trade one PCB for some wiring if you use drivers like this: [www.ebay.com]

Thermistor and mosfets would need to be dealt with though if building a RepRap (or not if you are building a CNC).

However I do understand that someone who takes the effort of designing a new board, sometimes with changes in the firmware required and taking the risk of manufacturing the PCB may want to get to recover their investment making a profit out of sales.

Misan: I'm not sure if you understand the object of the contest. Replacing pololu daughter boards with TB6560 daughter boards at the same price doesn't really solve the problem I'm trying to solve. The point is to create electronics solutions that are

1 DIY

and

2 Cheaper

The most likely candidate for such a solution would be an all-in-one board with integrated through hole drivers or smd drivers that are easy to solder yourself.

PS. Our third contestant has entered, so the contest is officially on! You can still enter till the 13th. That's less than a week away!

Brianandaimee: My experience is that making one unit of a PCB is only cheap when you already have the required setup and you draw it by hand. Anything else (toner transfer, photo-etching, pcb shop) and you are easily beaten by the low cost of a pololu or similar.

I do not think a TB6560 is going to be any cheaper than an allegro-based board, as that chip is larger which may lead you to a larger board too. That is why I was suggesting you not to have any PCB at all and to buy modules that can be wired with screw terminals.

But you are right, my answer was trying to be 1 DIY and 2 cheaper, but it was not creating a new PCB and it was not based on discrete components but on pre-made modules.

Anyway, I was not trying to enter the contest but just sharing some info of what I have learned. Maybe I should have stated this first. I'll stay tuned.

I totally get it. Thanks for joining in on the discussion and I hope you enjoy following the contest. Now that we have the required three contestants I'm buckling in for a wild ride. I'll have to start working on Sanquish in earnest now I guess. (Fortunately my life has calmed down a bit now. )

Bryan

PS. If there is anyone still on the fence the entry deadline is this Thursday the 13th of September. All Aboard!

bryanandaimee Wrote:
-------------------------------------------------------
> PS. If there is anyone still on the fence the
> entry deadline is this Thursday the 13th of
> September. All Aboard!

I have to concede that you are entirely right about the cost overhead of a commercial board/kit vs a self-sourced project. There clearly is a big gap.

Since I was interested in building a board anyway, I have been busy working out just how to create a design which is both DIY and low cost. I have an idea how to reduce the cost of stepper drivers, but initially I will be using chip carriers. There are A4983 drivers on ebay now for £4.73, I can't buy the components cheaper than that. But getting hold of Allegro chips at all is difficult. My first design has a BOM around £35/$55 including drivers.

For want of a better name, I would like to enter "Sango-bc" to the competition.

Hi Brian

I have added the Thames Valley RepRap User Group electronics to the competition, just recently (today!) named Phoenix. Our main focus was not cheapest cost, it was through-hole build, home-maintainability/fixability, safety and extendability - so we have support for 6 axes (actually, no one has tried this), as well as heated bed, hot end, SD card etc. It uses Atmel644 Arduino (can be upgraded to 1280) and the Toshiba TB6560 drivers.

I hope that we have already developed it, and that there are already 40+ boards in the wild, doesn't preclude us from the competition. We haven't advertised it previously as we have wanted to test it thoroughly. Though I think I gave you a few hints from our development when you were doing your TB6560 board!

Regards, Ian.

A hearty welcome to our two new contestants! To be honest I didn't think so many would sign up

Bobc: I like the idea of a cheap carrier board. It seems like a good middle ground between difficult SMD and expensive prebuilt daughter boards.


Droftarts: I remember the top secret info, thanks again. Glad to have you aboard. I think having a head start shouldn't be a big deal as everyone should have enough time to get through the scored tasks if they are determined. If we need to there is provision for extending the finish line by a month.

So the contest is on! We had one last minute entry. He was a bit late but he asked so sweetly that I figured why not. Plus the tibot entry brings the total projects including the Hare to 7. And as we all know 7 is a very auspicious number. So the entry phase is now closed and the contest phase begins. A hearty good luck to all the contestants. I can't wait to see what you all come up with.

And a quick note to the contestants. Please PM me an email address I can use during the contest to contact you with announcements, questions etc. I should have posted that as a requirement to enter. It just makes my job a lot easier if I have that info early on.

Bryan

Edited 1 time(s). Last edit at 09/17/2012 01:17PM by bryanandaimee.