Sometimes I’m just incredibly lucky. Tabletop RA4 roller transport processors are unobtanium these days and if you find one, it is likely to be insanely expensive. I happened across a Durst RCP20 (which in fact is a Thermaphot machine, but Durst sold them under their own brand). These machines have a few drawbacks though, which boil down to them being darn old pieces of equipment. First and foremost, to be able to use mine, I had to convert it to run at the right speed for the current RA4 color paper process. Here’s how I did this.
First of all: don’t tell me how lucky I am to have found one of these machines. I wasn’t really looking for one, and before I know it, I had…two of them! One I had to buy, the other was a gift (!!!) And no, don’t ask me to sell or donate you one. I already did, so I’ve only got one left and that one I intend to put to good use.
Now, let me start by whining and complaining a bit, because that’s what I like to do the most. These machines are really old – the ones I got were both around 40 years old. This means that the rubber rollers on the racks are generally toast, to a greater or a lesser extent. I suspect someone already did some refurbishing work on the one I kept, because some of the rollers look like they’re coated with a different material than the original black rubber. The problem is that these rollers start to bloat, which causes friction in the rack, and ultimately this will result in paper jams and the motor burning out if the racks get too heavy to turn. Essentially, these machines are ticking time bombs, in a way. I got a quote from a company to refurbish the rollers, but this came out at around €55 per roller. There are something like 16 rollers in a machine like this, so errm…well, let’s just…not. I just hope that mine will remain operational for some time until I figure out how to refurbish these rollers myself.
Then, there are numerous other things that may go wrong, get lost in time etc. One of the machines I got lacked the little couplers that link the drive shafts of the individual racks to the machine as well as the white bar that keeps the shafts in place. This is the bit I’m talking about:
They’re simple enough components, so a replacement could be fashioned from junkbox parts or (partly) 3D printed if they are missing.
Another issue is the thermostat. One of the machines I got had a faulty thermostat. In fact, it did work, but the temperature setting was way off. I had to set it to around 20C for the developer bath to remain at 35C. Still, it kept the temperature stable, so it wasn’t much of a problem. But I’m not sure if a replacement thermostat is easy to find if it really goes pushing up daisies. Interestingly, the other day I replaced the thermostat on our refrigerator (my sweet girlfriend had almost ordered a €300 fridge as I was just finishing the order process on a €15 thermostat for the otherwise perfectly fine machine…) and the thermostat I exchanged looked pretty much exactly like the one in the Durst RCP20. Now, the temperature range is obviously different, but it makes me hopeful that perhaps a replacement could in fact be found in the marketplace, still. I haven’t looked for it, yet.
But now for the really major issue. These machines are originally geared for the EP2 color process, which was far slower than today’s RA4 process. In addition to the slow speed being unnecessarily annoying, your RA4 paper will in fact hopelessly overdevelop (and throw nasty color casts, fog etc.) when developed for as long as EP2 paper took. So forget about just dealing with the low throughput, pretend it’s still 1978 and run your RA4 paper through an unmodified RCP20. Something needs to be done.
Now, fortunately, again I simply have to steal the good idea that has been published long before I ever thought about any of this. In fact, there’s still a very nice web page online by a German fellow who published the essential details on modifying an RCP20 to run RA4 process: https://www.vogelstimmen-wehr.de/rcp20.htm Not only is the page still online, the exact parts he used are also still available from the same online store! Here’s that parts list:
- One aluminium cogwheel, diameter 54mm, 30 teeth, 8mm bore hole.
- Another aluminium cogwheel, diameter 24mm, 10 teeth, 6mm bore hole.
- One drivebelt for the wheels above, 330mm length, 65 teeth.
Alright, prices are not the same anymore, but still, for less than €50 at the time of writing you can get a (nearly) complete Durst RCP20 modification kit. Nice! So that’s what I did. Now…where do those things go, exactly? Let’s have a look.
If you open up the RCP 20, which works best if you turn it upside down and take off the bottom, you find this area around the motor assembly where the original gears and drivebelt are:
The motor is at the top of the picture; note the inspection date of 9 Nov. 1976 – did I mention these things are old? The motor drives a transmission consisting of some metal gears, and that in turn drives a small plastic cogwheel. This is connected through a drivebelt with a bigger cogwheel (bottom of the photo), which is attached to the actual driveshaft. The driveshaft extends outside the bottom right of the photo to three worm wheels that drive the racks. Note also the plastic fan mounted directly onto the rotor of the motor in the top of the picture. Our new parts will fit right in place of the old parts, like so:
Note that the bigger wheel will go at the motor/gearbox side, and the smaller one goes onto the driveshaft. This creates a 30:10 transmission and that in itself is something we should take into consideration. You can see that the cogwheel order is pretty much the opposite of what it was in the original situation. In the original machine, the gearbox drives a small cogwheel that in turn drives (through the belt) a larger wheel. This means the gearbox and therefore the motor experiences a low torque situation. If we then flip things around in the new situation, the motor actually ‘sees’ a higher torque – which stands to reason, because the actual racks run much faster in the new setup, so it’s logical that more force is needed from the motor.
The consequence of this is that we need to consider cooling. The motor already tends to run warm in the original setup, and having it work harder won’t make things much better. The original fan is nice, but more for decoration purposes. Its blades are at a straight angle on the axis, which means they flap about a bit without creating a well-directed airflow across the motor at all. Probably this has to do with the more rudimentary injection molding possibilities in the mid-1970s. Now, we’re going to lose that fan anyway, because it’s in the way of our new cogwheel. Out it goes! The fan can simply be pulled (with some force) from the shaft as it’s a simple friction fit. But make a mental note that we fix the cooling issue later on. I know that the German bloke chose to just make the fan smaller by cutting the blades, but I don’t think this is a good idea at all. The fan isn’t there just to look pretty (they would have made it prettier if it were).
Sadly, it’s not a plug & play affair, this one. The 8mm bore hole on the bigger cogwheel is a little too small for the shaft that extends from the metal gearbox, so it needs to be machined to a larger diameter. I forgot, but I think it needs to be 9mm instead of 8. Just measure the shaft and machine accordingly. It’s worth mentioning I got help from the great friend who gifted me one of these machines – not only is he a really swell guy (and he really is!!), he is also mighty experienced with machining stuff. I’m more of an electronics person, so I’m over the moon if someone not only has a machine tool, but also knows exactly how to tweak its knobs!
Then there’s the drive shaft of the racks, the long one with the worm wheels on it. This one needs to be machined down a little to accept the small cogwheel, which has a smaller bore. Again, if you know how to work a machine tool, this is a simple affair.
Once the machining and drilling are done, it’s time to mount everything together. Now it gets a little tricky, because you want to tension the belt as well. Especially in this new, higher torque situation, you don’t want that belt to be all loose and flappy. Besides, the new belt is longer than the old one, so something had to be done anyway. The motor assembly is bolted to the chassis, and it turned out to be the easiest solution to just lengthen the mounting slots a little so we could slide the motor along the same axis as the drivebelt. This way we could easily fit the belt, and then slide the motor backwards until the belt was tensioned, and then fasten the motor firmly.
So now for the cooling stuff. Well, I got this 3D printer not too long ago, so I did it the fancy way and started by replacing the original fan with a brand new one. I started up Autodesk Fusion 360, googled a bit on how to design a fan and engineered one to the exact measurements I needed. Yes, it’s a friction fit, just like the old one, so it slid right onto the motor’s shaft. Moreover, it turns out to run really well and make a lot more airflow than the other one. Only problem is I designed it the wrong way around, so it sucks air from the motor towards the developer bath, while I intended for it to be the other way around. Ah. Well. Doesn’t matter too much, because it still works this way. Besides, I also added a separate fan that blows across the motor’s coil, which is the part that actually heats up. The original fan, even if I had done the replacement job right, would never really reach that area very well anyway. I added a small 12V power supply and fuse for the Intel CPU fan (I think it was the stock fan from a Pentium 7?) Here’s what the final result looks like:
Note the new cogwheels on the motor shaft and the rack drive shaft (bottom of the photo), and the 3D printed fan in transparent PET-G. Yes, there’s ample clearance between the new fan, the developer bath and the drive mechanism! In the center of the image is the little power supply for the CPU fan right above it; there was a convenient empty space it could be strapped into with a couple of tie-wraps. For the power supply I used a Chinese off-the-shelf low-power SMPS comparable to what you’d find in a typical phone charger, but 12V instead of 5V of course.
And that’s basically it! The rest was just a matter of remounting the base plate, flipping the machine over, fill’er up and go! I haven’t done all that much RA4 printing lately, only a couple of sessions, but the machine has worked like a breeze so far. I used to do RA4 in trays at room temperature and while that worked quite alright, this is much faster. Yes, I’m limited to 20cm print width, but for me, that’s OK. I generally print small anyway.