I: About the De Agostini/Fanhome R2-D2 Kit

This is a highly detailed review of “Build your own R2-D2,” a “partworks” kit designed and sold by De Agostini/Modelspace/Fanhome.

Introduction: those childhood memories

Indulge me for a moment with a slight detour. You know when you were a kid, you saw some awesome product advertised on TV or wherever? It was for some awesome thing from a movie! Or your favourite show. Wow!

So you begged your parents, or saved up your pocket money, or whatever. And you were so excited when you got the box home or the thing arrived in the mail. Gosh!

And then you opened it up and discovered that the toy was actually kind of lousy. The proportions were wrong. Colours were out. Details were missing or oversized. Big joins and seams all over. Stuff missing. Etcetera. Then the cheap plastic thing broke. What a letdown.

But maybe an important life lesson in how advertising always lies!

But as time has passed, and fans age, SF/fantasy fandom products, as opposed to children's toys, are generally getting better and better. Products are often fairly accurate in appearance, and of decent quality. Many models, replicas, and toys are built from original blueprints or scans of original props. The quality, on average, is definitely going up. They're certainly getting more expensive.

So. Is the De Agostini/Fanhome R2-D2 costly kit any good? Is it a toy-like disappointment, or a modern-style accurate and high quality product? Read on!

What does this R2 kit get you?

A nearly half-sized droid of your very own - make those childhood dreams come true! The idea of having a little robot rolling around the house is appealing to many of us nerds, and since it's a kit you get the fun of building it yourself.

Aside from the obvious facts that a full-sized R2 is expensive, difficult to make since it's not available in kit form, and takes up a surprising amount of room, this small model claims to offer a high degree of automation, accuracy and ease of construction. And it succeeds to varying degrees.

The De Agostini product is a model engineering kit and not a plastic model. It isn't a static shell with some detail parts glued on – it’s actually got a metal subframe with a plastic skin and electronic and mechanical subsystems for lights and motors. It even has a one-piece seamless motorized metal dome. That makes it more interesting, and arguably more educational, than just a hollow plastic kit. De Agostini also trumpet that the product is the "ultimate authentic replica", which is a mighty big claim.

Incidentally, it doesn't come with a restraining bolt; something of which I can only approve. Be free, little robot! Be free!

Three legs good. Two legs bad.

The most unfortunate design choice is that the model is always in three-legged mode for stability. The movie R2 could switch from two-legged (Kenny Baker style) standing mode, to three-legged (radio control style) mode. In real life this was performed by some awkward trickery and film edits, since the mechanical props never quite worked right.

It’s quite complicated to engineer what fans call a 2-3-2 mode: a robot capable of switching leg modes back and forth. This is because the middle leg must move up and down, and the side legs must change in angle as well. It’s understandable that the rather old-fashioned De Agostini didn’t want to engineer a fully automated product like that, as it'd be expensive to do. (unlike the clearly more sophisticated and deeper-pocketed engineers at Sphero, who brilliantly pulled off such a stunt for their tiny remote-controlled R2 toy)

But instead De Ago designed the feet and ankles as plastic shells with no pivot capabilities whatsoever. In fact, they made the foot and ankle joint one plastic piece on each side, so they don't even look separate. Whole recessed areas are missing. Argh! The shoulder pivots are also fixed.

Enough internal space could have been reserved to accommodate a foot that could be manually moved up. And it’d be easy to have a simple cutoff switch that would disable the foot motors if the middle foot is retracted. It’d then be possible to have a nice standing R2. Oh well. Perhaps the overdone dome engineering, with all of the metal framework and metal dome, means it would be too top-heavy. Still. It feels like a lost opportunity.

Strangely it appears that the initial design might have featured a moving ankle. Here's a part that was sent to a customer by mistake - an early production test piece that clearly shows a proper symmetrical ankle with a pivot point. I believe it was also the part supplied during an initial test production run. But at some point De Ago decided to cheap out and make the fixed ankles. Dammit.

Could you modify the droid yourself to be permanently two-legged, with no descending middle foot? Maybe. It would not be easy. Unfortunately, as noted above, the feet and ankles are basically one piece parts, so you'll have to cut them apart and then build your own bits to fill the gaps. You'd then have to cut, reconstruct, and re-engineer the angled joins where the feet meet the ankles. Finally, you'd have to cut and alter the shoulder joint cylinders to the new angle. In short, it'd be a lot of work, and of course you'd lose the motorized wheels.

Automation and communication.

The kit is also advertised as an autonomous (actually "AUTONOMAOUS" according to the De Agostini UK website) robot, and it has an on-board CPU so it can drive around. Very noisily - its motors and gears are far from silent. It has under-foot sensors to keep it from falling down stairs, wall sensors on the shoulders and body front, and can operate in a “patrol” mode. This latter is just a glorified random mode - it rolls around in various directions, halting and then changing trajectory if it encounters an obstacle.

Frankly it's not as impressive as its marketing makes it sound. It certainly has no actual AI: another De Agostini advertising lie. It can't learn and adapt to its surroundings or anything. It doesn't even have a built-in mapping program the way a robot vacuum cleaner does. Oh well.

In addition to wandering the house theoretically unattended and freaking out your cat, it has the ability to interface with smart phones (Apple iPhone 5S or later, equivalent iPads and iPod touches, running iOS 11 or higher, or Android smartphones or tablets running OS 4.4 KitKat or higher) using WiFi so you can drive it around. Initially it was supposed to support Bluetooth as well, but that capability was dropped. Note that only 2.4 GHz WiFi is supported – 5 GHz sadly is not. This could be a problem if you want your R2 to connect to your 5 GHz-only WiFi base station, but I suspect most people will have the R2 connect to their phone's WiFi directly.

There were some billing and support concerns about the product announcement, with De Ago initially claiming that the app would be time-limited with a paid subscription required to run it thereafter, but after a public outcry this unreasonable limitation was apparently dropped, which is terrific.

Of course, it's highly unlikely that De Agostini will keep the phone apps updated to future releases of phone operating systems past the "three years from date of release" that they have pledged. It's probably wise to keep an old phone around solely for R2 remote control purposes, so you can keep running the droid and not worry about the app becoming obsolete. The app came out in 2018, so at some point in 2021 it'll probably be discontinued. It did get updated in 2020 for new phone OS releases, though I don't think feature limitations or bugs were fixed. At this time it isn’t known if the 2021 re-release of the product under the “Fanhome” name will extend the supported life of the app or not.

The robot has a small low-quality video camera in its head, so you can watch where it’s going and take spy photos or whatever. It supposedly can recognize voice commands, though predictably this is not an entirely reliable feature. Same with facial recognition, which is an advertised feature as well. One thing that makes it particularly aggravating is that it repeatedly makes the same confirmation beep noise when it receives a remote command or when you push a button. Over and over and over. This wears thin really fast. A classic example of a product never having actually been used by its developers.

Sadly the device is not programmable in any way, which is really a lost opportunity for education. The upper CPU uses an inexpensive Orange Pi board, for those interested. Unknown at this time what OS it's using and how hackable it might be. The “coin slots” on the front of R2's body are cleverly used as pushbuttons. The buttons control "patrol mode," front arms in/out, light sabre up/down, pairing mode for WiFi connections, switch connection mode, and overall power on/off.

Finally, it has no published debug mode. A lot of people have been having problems with erratic or random behaviour of their droid, often because a sensor lead isn't plugged in properly or whatever. If De Ago had had the foresight to add some form of debugging tool available to the public it'd save a lot of grief. For example, if the droid were to beep a specific pattern on power-up if it fails to detect an ultrasonic sensor, say, then you'd be able to focus on that specific subcomponent and make sure it's plugged in. As it is, everybody is having to unplug and reseat all the dozens of cables by trial and error. It's a serious shame this simple feature is missing, since it would help customers and also save De Ago a lot of money in support costs!

Complexity and ease of assembly.

It's easy to put together until the final assembly stages. In fact, it will probably annoy experienced or advanced builders, who'll find it much too simple to assemble. It's designed for the casual amateur; someone with little or no experience in miniature engineering. Basically each edition of the magazine comes with a plastic bag containing one to several small parts that you put together with a small Phillips screwdriver. Occasionally superglue (not supplied) or double-sided tape (supplied) is required. There is no soldering, welding, sawing, etc, required. Painting is also optional, since most parts are pre-painted (see below).

The final stages are a bit more complicated, as you have to connect up a whole bunch of wiring and assemble the final bits inside a fairly crowded dome and body. Allow some time for this stage, and don't rush things! It's easy to, for example, plug a connector in upside-down, which can permanently damage internal components!

The subscription is supplied with a few other useful things, such as a cheap screwdriver, a special stand for holding the dome upright while you work on it, a plastic storage box for parts, and a small battery pack so you can test various electrical components before all the parts that make up the whole system have arrived. Note that you should keep the styrofoam shell used to protect the dome in shipment - it doubles as a support stand while you're assembling the thing.

To summarize: if you want something simple you can work on with your family, this is the kit for you. There's a lot of stuff in the kit - it's just not difficult, step by step. If you want something complex, sophisticated, and challenging to assemble, this ain't it. That's not a criticism at all – just a statement regarding its intended market!

Mechanical quality of the build.

The components are reasonably well made and designed. It's quite a heavy kit, clearly engineered with sturdiness in mind, and should withstand ordinary use. Internal frames are plastic and metal (mostly Zamak or a similar zinc-based "pot metal" type alloy, and some steel), and the legs, feet, and exterior body panels are white ABS plastic. Most components are screwed in to keep things together. The iconic dome is a single piece of turned and seamless metal (some type of steel – it's definitely not aluminium as the originals were), though it has an unnecessarily heavy metal and plastic underframe that makes the finished product top-heavy.

Note that its design is nothing like the original movie prop droids. Most of the Star Wars robots were made from sheet aluminium, with two-layer aluminium domes. The Empire droids were mainly fibreglass, and most had aluminium domes. The De Ago approach of a metal frame with plastic panels on the outside was not used.

The mechanics for the parts seem good. The motors for each outer foot appear to be solid and substantial. The unit uses slightly knobbly hard rubber wheels, and not treads or anything like that.

Mediocre plastic moulding.

The injection molded parts are okay, but are not created using the most advanced processes. Let's just say it's not exactly Bandai quality here.

Here’s an example. The body panels that make up the main body barrel are shipped as a number of narrow plastic pieces, since two large halves would be more costly to mail. You then have to join the panel pieces up, resulting in some gap and alignment problems. Check this closeup of one of the joins.

Yes, that’s right. The recesses in the panel don’t even line up! One side is wider than the other.

Or take the feet, which are about the worst examples of plastic moulding in the kit. The De Agostini droid feet are made of two halves split vertically, but since the tops of each part aren’t perpendicular to the sides - they angle upwards slightly to a narrow seam - they look frankly toylike. (there’s a noticeable draft to the injection moulding, in other words)

The photo on the left shows the two actual parts that make up the De Ago R2 foot. Those seams are not befitting a £900 product, and the angles are wrong. The rectangular silver plate is also rounded on one side and flat on the other. And what's with that big curved piece on the top right half of the foot?

The middle photo has been Photoshopped. I altered the angles to show what the foot should actually look like, if it didn't have draft issues.

The photo on the right is one of the original ANH R2 feet. This one was for Kenny Baker to operate, which is why it has the extra cylindrical rollers on the bottom. Note how there's a big notch at the top of the actual foot for the ankle pivot to fit into. De Agostini didn't even bother to replicate this! There’s also a large curved protrusion to conceal the motor drive assembly.

The feet screw together from the sides, and come with plastic plugs to conceal the holes, rather than having the screws fit into concealed locations. For example, the feet could have been engineered such that the screws reach up from the underside, hiding them. Instead, this conventional and lazy product design makes the feet look like ordinary Earthbound commercial products rather than something from a galaxy far, far away.

Metal parts versus plastic.

Some detail components that are metal on the original design are also metal in the kit, especially those which ship earlier in the build, in order to impress buyers. But other parts are moulded plastic. All parts are sprayed a uniform silver colour to conceal the differences between metal and plastic. There are a number of consequences to this.

First, some of the solid metal parts look really great. Especially if you strip off the paint and polish the metal. The result is a very convincing and realistic component. I used ordinary gel-type household paint stripper for oil-based paints. I then used medium to fine grades of sandpaper and emery cloth to remove the cast texture, and sand off the seam lines. Finally I used a superfine foam sanding block to give the finished brushed appearance, always moving the part in the same direction across the block.

Some of these pieces, such as the shoulder pistons, also look extra good when trimmed back on the underside (carefully, using a Dremel-type rotary tool) to reflect their actual prototype appearance. Parts like this really feel worth the money! Unfortunately most of the larger pieces are plastic rather than diecast metal.

Second, although the genuine metal parts look great when sanded and polished, they’re not all cast from the same alloys. They all turn dull grey as they oxidize, but in different shades and darknesses. That may be okay if you’re going for a weathered old R2 look, but if you want an R2 fresh from the shop this is a problem. I don't know if this oxidizing would be halted or at least slowed by spraying on clear coats.

Third, if you treat the metal by stripping off the paint you’re going to have a continuity problem with the plastic pieces, which will still look like bits of painted silver plastic. You may have to spend some time painting them with more realistic metallic paints and concealing the differences with weathering.

How visually accurate is this R2?

If there's one word to sum up the problems with the De Agostini R2, it's “mixed bag". Okay, so that's two words. In terms of general visual appearance compared to the films, it's fairly good. But if you look closely many details are totally wrong.

And now we get into really fiddly minutiae. Most people will not, to be fair, notice any of the problems I've listed below. And even people familiar with Star Wars probably won't notice these issues unless they directly compare the kit to a photo of R2-D2.  But I list them on another page, mainly because I find these inaccuracies really frustrating, given how expensive the model is.

De Agostini advertise the product as "authentically detailed", and a "movie accurate... replica". It is neither.

Some bits are okay.

Some parts of the kit are pretty decent - sturdy, fairly screen-accurate at first glance, well finished. Other parts are merely adequate, or deviate randomly and unnecessarily from the appearance of the original movie props. This lack of consistency dogged their Millennium Falcon kit as well. It’s almost like some parts and sub-components are designed by their main team, whereas others are designed by their “B” team.

On the whole the R2 is much more screen-accurate than the De Ago Falcon. Their Falcon kit was riddled with tons of really minor errors and deviations that, to be honest, only really obsessed fans would notice. The R2, being far simpler and more geometric in design, is mostly correct.

Some bits are garbage.

Some changes are really disappointing. For example, the advertising photos (below left) show a prototype kit with pretty screen-accurate front vents. These are one of those things - they're not something that most people could draw accurately, but when they're wrong it's immediately obvious that something isn't right. But the kit shipped with toylike crap instead (below right). No joke.

I suppose there are two reasons for the change. First, the production vents are designed for cheap one-piece injection moulding. And second, they probably had to enlarge the holes in the back of the vent to get the (highly visible) ultrasonic sensors to work properly. Either way, it's bait and switch from the advertised product. To this day De Agostini/Fanhome are advertising the kit using photos of the screen-accurate vents rather than the ones which actually ship.

Other R2 errors are also super aggravating, since they simply aren't necessary. Some visual mistakes are obviously done the way they are to save money, or to make the parts cheaper to build at the factory. But others don't make manufacturing any less expensive, and are simple sloppiness.

Here's a typical example. The rear silver-coloured "coin return" panels should line up with the upper compartments. In other words, their sides should align as the red line indicates. But they don't. Not only that, but they put the seam line at a goofy angle because of it. Double your random inaccuracies! And look at those awful screwholes and gaping gaps on the legs and battery box. Even when filled with plastic plugs, and screwed tightly shut, the pieces look like toys. And why is there random stuff like a rectangular recess in the leg at the top interior?

So. If you're familiar with R2 these sort of errors look cumulatively wrong, even if you don't know exactly why. Obviously there was a lot of variation in detail across the movie props, so it's not possible to get 100% screen-accurate. That's fine - but De Ago also just made random stuff up a lot of the time, which never looks great. They also cut corners and altered details which are consistent across all the droids of the Original Trilogy.

Here's another super-obscure error. Hardly anyone is going to know that the De Ago part on the left has an extra step or groove in it for no reason whatsoever. It's supposed to look like my modified part on the right. Why did I screw around fixing it? God knows.

It's the sort of thing that bothers some people and not others. But my observation is that whenever De Ago deviated from the appearance of the on-screen props it never looks better. It always looks cheap. And if they'd marketed this as a toy, like the Sphero, that'd be fine. But they market it as a replica and model.

Which movie did De Agostini base the kit on?

De Agostini doesn’t explicitly say, but their kit has a mixture of features from across the Original Trilogy - Star Wars: a New Hope (ANH), the Empire Strikes Back (ESB), and Return of the Jedi (ROTJ). For example, some of the holo projectors have black or blue rings and are therefore ANH style, and the white lower body panels also follow the ANH design. But the octagon ports are closer to the ESB ones, and the half-moon details on the feet have some ESB-style details. The sabre ejector was only seen in ROTJ. The middle leg is too long, like an R5 droid rather than an R2 droid.

This is all fair enough I suppose, since there is no one true R2-D2. So all in all, the kit is mostly in keeping with the general design of the Original Trilogy R2, with a number of discrepancies and errors.

Paint.

Most of the parts are already pre-painted. And aside from noticeably inconsistent white and blue shades from one body panel to another (a typical partworks problem – the bits aren't all manufactured and painted at the same time), the paint job is okay. I guess it's supposed to be a factory-new droid and not a weathered old unit that has been trundling determinedly across the Jundland wastes after an Imperial assault. It’s an industrial spray finish, and fairly even.

However, it’s not an amazing job either. The blue parts, in particular, are quite dull compared to the movie R2 units. They're dark and slightly purplish in tone, and not metallic at all. The original R2s had a metallic purple-blue tone which depended on the light. There are also noticeable differences in appearance between the metal plates painted blue and the plastic pieces painted blue.

Left: one of the panels. It’s cast in some inexpensive pot metal/zinc alloy. It’s then painted so opaquely that you can’t tell it’s metal and it might as well be painted plastic. Note how the outer area is sprayed with a slightly yellowish metallic paint, to approximate the colour of the steel dome. The inner area is sprayed a satin purple-blue over top. Middle: a panel piece that’s been dunked in chemical paint stripper for an hour. All the paint peeled off in minutes. Right: polished up to resemble a lightly brushed metal finish. Note that it looks this shiny immediately after polishing. However, within minutes it dulls up as the metal oxidizes.

The problem is that if you want to repaint your droid it’s often awkward to do it as you go along, as you may want to paint all objects at one go for consistency. You also can’t assemble all the sub-components until the exterior surfaces are all painted.

Observations on the parts.

Because the kit is designed to be built across 100 issues, and because some electronic components are quite expensive to manufacture, many issues don’t really include much at all, just so that per-issue costs average out. Or they just include one or two parts that fit together in seconds. Unlike, say, a model ship which consists of a ton of small detail parts with a couple of larger hull pieces, the R2 kit includes some very expensive components such as motors, the steel dome, a CPU, and a video projector.

There are times when an issue arrives and it's got all kinds of interesting stuff and it's quite fun to assemble. And then there are times when something like this shows up. And, after you've spent your 7 seconds putting that solitary component in place without any tools you think to yourself, why the hell didn't De Agostini make this a 50 issue subscription?

Also, many parts of the model which would normally be produced as one or two large parts for a conventional product - the dome subframe, for example - are included in the subscription as tons of smallish pieces you need to bolt together. This is all fairly pointless make-work kind of stuff that also adds a lot of weight to the model. The finished R2 is actually quite heavy, in large part because of the additional complexity of its internal frame. Look at all those unnecessary metal plates! This is inside a freakin' sturdy steel dome - there's no need for all that additional heavy crap up top.