Cost Reduction

From RepRap
Revision as of 07:48, 30 December 2014 by Glenn (talk | contribs) (+Category:Community)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

This is a page for documenting ideas for reducing the cost of RepRap builds. The first section is a short summary of the current cost structure and then each major category will be split up to be discussed separately. The estimates will be in dollars but the article is meant to be comparative so conclusions should be universal.

The main categories of reprap materials along with current cost ranges are as follows.

RP Parts $40-$100

Hardware $50-$100

Steppers $40-$100

Power supply $40

Electronics $100-$200

Hot End $40-$90

Heated Bed $30-$50

Total Cost: $300-$620

RP parts and other hardware

When the RepRap project was young, parts printed on a 3D printer were expensive or time-consuming or difficult to obtain.

When people first started pondering how Darwin could be improved, many people expected that as more 3d printers were built, so reprapped plastic (RP) parts became easier to obtain, future generations of RepRap printers would use far more plastic and far less of the other stuff ("vitamins").

Many people find the idea of a machine that, when fed the right instructions, can print out 100% of its own components, a fascinating idea. Many people made designs that got ever closer to this ideal: the threaded rail of the MultiRep, the printed case of the Tantillus, the Gada Prize requirements, etc.

However, other people feel that "making fasteners" out of RP plastic is a waste of time and money. Many people made RapRap designs that use *less* plastic than the Darwin, using it only in places in the 3D printer where it cannot easily be replaced with lower-cost off-the-shelf fasteners and frame material. These designs include the Test Tube Mendel, many of the category: LaserCut designs, etc.

RP Parts

The cost of RP parts have come down so far that alternate strategies like cast parts etc. may not be as attractive as they once were. Still if you are solely focused on cost savings then that may help a little. It seems as though the price of RP parts is at an equilibrium state. I doubt that a set of Prusa RP parts will come down much more.

The main path to cost savings in this category therefore is likely to come in new designs that minimize RP part count and print time. This would affect price in two ways. It would decrease the cost and difficulty to produce a set, which would have the added effect of increasing supply as more printers got in the game. Both factors would tend to decrease prices.

Another approach is to build a RepStrap, replacing the RP parts with wood or MDF, as in Adesina or replaced the RP parts with duct tape as in Duct Tape RepStrap, or replaced the RP parts with something else in other RepStraps.

TODO: An apples to apples comparison of print time for the main types of RepRap printers would be helpful here. Just need to skein each set with the same settings and put together a table of total print time as reported by skeinforge.

Hardware

This is already a minimal cost item. New designs with lower overall hardware requirements may bring this down a little but overall savings are limited since the price is already pretty low.

With people going from printed PLA bushings to linear ball bearings, going from threaded rods to aluminium extrusions for the frame, going from printed to bought pulleys I actually see the opposite direction here. People don't mind to make their RepRap more expensive. --Traumflug 07:48, 5 March 2012 (UTC)

Steppers

This will likely remain a constant. Small designs like the Huxley that use smaller steppers can reduce this cost somewhat but there may not be much we can do short or long term about this cost.

Moving away from steppers may save some money but seems likely to raise costs in other areas like electronics, and may reduce print quality.

Very cheap used steppers can be found on ebay for as little as $40 for a set of 5

Electronics

Now that the price of RP parts has come down so far, the cost of electronics has become the single largest line item in a RepRap build. This is likely to be a place where significant savings could be had.

Current electronics solutions are split up into four different solutions, organized here from most to least expensive.

1. Modular designs: These designs are not necessarily aiming for cost effectiveness but rather for easy repair and performance. Currently Gen3 is an old and deprecated set, while the Makerbot Gen4 is very expensive compared to other designs. This design paradigm makes it easy to experimentally change just one thing, so it is useful for initial experiments to see if some new (and perhaps significantly lower-cost) component will work at all. Alas, the connectors often cost more than the chips.

2. Arduino shield/daughter board designs: RAMPS is the most common example of this type of electronics; several others are listed at Pololu Electronics. These designs suffer from two different budget busters. First an Arduino must be purchased separately, and the Pololu driver boards are also separate purchases. This means that a RAMPS build will likely be in the $150 to $200 range for the final completed build with drivers and arduino board. The user can save some money by buying the RAMPS board as a kit and doing assembly at home.

A typical system includes (typical prices as of 2013):

  • €40 $40 (4*$10) Pololu A4988 Stepper Driver (typical price for 4 as of 2013)
  • €35 $50 RepRap Arduino Mega Pololu Shield (RAMPS) (typical price for 1 assembled as of 2013)
  • €40 $56 Arduino Mega or Taurino microcontroller board (typical price for 1 as of 2013)
  • ---------
  • €95 $125 complete RAMPS-based reprap electronic set (typical price for 1 as of 2013 -- [1]; [2]; etc. )

The cost of an Arduino-compatible microcontroller board that can allegedly drive a RepRap, as of 2014, has dropped below $10 in ones ( [3]; [4] ). The cost of stepper motor driver electronics also continues to fall.

3. All in one surface mount boards: Gen 6 is the most common example, though others like Melzi and printrboard are becoming more common. These designs are made for easy scalability of supply. Since they can be mass manufactured they are much easier to spec for large orders like with the recent IndyGoGo campaigns of Huxley and Printrbot. Unfortunately the fact that very few people are likely to be comfortable building their own surface mount board means that all such boards are generally sold assembled. The cost of assembly generally offsets the advantages of cheaper surface mount components, and integration of the driver chips on board. So most of these boards are likely to be in the $130-$200 range as well. With higher volume, this may be the future cost leader. A surface mount design in large quantity orders may even beat out the DIY boards below as the low price design, but not likely to be so in the short term. The current lowest price is the Printrboard from Printrbot for $129. Hopefully there will be kits for those comfortable soldering surface mount components that will bring the cost down even further.

4. All in one DIY boards: This is where cost savings gets serious. When you do a DIY board like Gen7 or Sanguinololu you are trading your time for cost savings. Unfortunately most DIY designs still suffer from Pololu lock in. You could probably build any of the current all in one boards like Gen7 and Sanguinololu for about $40 depending on how well stocked your junk box is, but you would still have to buy the Pololus which bring the total cost up to about $100.

The main area for cost savings in this category is in elimination of the pololu daughter boards and integration of the driver onto the board. This is the goal of the Sanguish electronics design along with the uncompleted Gen7T design and some others. There are not many choices of through hole(DIY) driver chips. The toshiba driver in Gen7T / Sanguish is one of the only commonly available drivers that is also available as a through hole component. Trading the four $13 pololu boards for four $4 Toshiba drivers saves about $36. So the cost of a complete electronics set could come down to $60-$70. In larger quantities the cost comes down even more though a DIY kit is likely to be inherently a lower quantity sales item than a complete manufactured electronics set.

While this was true in 2012 and early 2013, reduction in stepstick prices now makes this cost saving measure less attractive. In ebay or in other sites like banggood you can find stepsticks as cheap as $16 for a pack of 5 (ebay)[5] or less than $4 in banggood[6] . Similar deals can be found elsewhere. This reduction in prices make the choice of electronics with seperate driver more affordable and has the added advantage of a driver being easily replaceable in case of failure/burn. All in all current stepstick prices of $3-$4 per stepstick further drives down the cost of electronics


Recommendations: The sanguinololu and Gen7 designs are the current price leaders and are fairly well proven designs. For cost conscious buyers this is likely the way to go, but watch out for toshiba based DIY electronics to bring the price way down in the near future.

Electronics DIY Kit Cost Arduino Cost Driver Cost Total DIY Cost Total Retail Cost (Pre-Assembled)
RAMPS $25-$35 $30-$60 $16-$35 $71-$130 $120-$160
Sanguinololu $35-$55 NA $16-$35 $51-$90 $100-$120
Gen6 NA NA NA NA $135-$180
Printrboard NA NA NA NA $129
Sanguish (estimated) $50 NA NA $50 NA

In some cases, an Arduino clone such as a SainSmart Mega 2560 for Arduino can be used. But, be careful, and check reviews, because some of these will not work for our purposes.

Hot End

Commercial hot ends are all in the $60-$90 range except for the $17 hot end kit from Techzone. The Techzone hot end suffers from reliability issues due to the oak insulator which degrades over time from the heat and comes with a thermocouple rather than a thermistor which limits compatibility with existing electronics. This may be an area that arguably is not a place to play with cost cutting. A good hot end will more than pay for itself in avoided frustration. DIY options do exist and may work as well as the more expensive milled designs that are mostly sold through retailers. The main reason that hot ends have become a commercial item is the difficulty most people have in making their own. Often a lathe or drill press is required for some of the parts needed for the design which limits the number of people with the equipment and skills necessary to build the design.

The Wildseyed and North90 V2 hot ends are notable DIY hot ends requiring minimal tooling for the cost conscious, but might not work as well as a commercial CNC lathe built hot end. Hot ends built with a heatsink and no plastic insulator like this one allow all the parts to be found at a local hardware store (except for the tiny drill bits), but are very hard to build, specifically drilling the nozzle and threaded rod. It would make it much easier to buy a premade nozzle that can go onto your threaded rod, such as the J Head or Makergear, but this cost $25 and $7.50 respectively.

TODO: A nice table comparison of DIY designs with categories of materials cost, tools needed, and time required to build would be helpfull here.

Heated Bed

While it is true that you can print without one, most people likely will end up getting one at some time so this is an area for cost cutting. A DIY approach may lower the cost here somewhat. For instance using resistors or nichrome wire as a heating element and wiring it up yourself rather than buying a prusa heated bed circuit board. Another way to go is to get strip board and wire it up as in the ghetto HBP or etch your own PCB.


Overall

A RepRap design focusing on minimizing RP parts and hardware costs could shave $50 off the cost of a build, similarly the electronics could come down by about the same amount using a DIY integrated design. A lower cost or more accessible DIY hot end could save $30-$50. This would lower the floor to about $250 or so for a "cheapest" type RepRap build.

Further reading