i worked a lot with KUKA and adding external axes is pretty trivial. Use KUKA servo motor and do whatever you like - make custom turn table (or linear rail or whatever) with whatever gear ratio. Some have used parts of old robots (like A1).
One can also integrate non-KUKA servo actuators or motors even stepper) but this will require some DIY and this will be no "real" axis (not coordinated motion with robot). But for many applications this is not needed - it is often enough to be able to move the external unit to some position and then have robot do its magic. When real coordinated motion is needed (high dynamic), stick with KUKA drive/motor and you will be fine.
KUKA configuration software is only needed for newer systems that run on KRC4 or KRC5 architecture. but that software is free - WorkVisual is included with each system (check D:) or it is a free download from KUKA website. on older generations of hardware (KRC1 and KRC2) no software is needed. configuring external axes on those system is simply editing text files and you can use included editor or notepad.
this is not to be confused with KSS - KSS is the system software required to run robot. but this is included with each robot. only if you wish to change it a different major version it will be expensive. changes within same major version are free.
software versions are expressed as four numbers such as 8.5.3.155
first two numbers represent major version (so software changes like 8.5.17.x to 8.5.42.x are free as long as you are willing to do the change yourself).
but changing major version (like going from 8.2.x.x to 8.3.x.x or 8.5.x.x) means new KSS license is needed. also compatibility check for hardware and any options is needed (it is pretty common that changing major version means any installed options will likely need to get new license too).
if you are buying NEW system and for some reason certain functionality is not working, KUKA normally grants version upgrade/downgrade free of charge.
I know lots of people say used robots aren't worth the trouble, but for some it's the only entry. I think if I get one that's shown to be working, from a reputable dealer, I should be ok.
A used robot is a risk. Take KUKAs -- great machines, but there are a lot of KRC2 on the 2nd-hand market, for whom parts have not been manufactured in over ten years at this point. KUKA no longer supports them, and good 3rd-party support is getting harder to find as the experienced users age out, or just forget. Buying a discontinued model of used robot is asking for a lot of extra pain. This applies to every brand.
Especially if you want to start making modifications. KRC2s were great for adding extra axes to, the base software would handle up to 12 axes... but still required add-on hardware from KUKA. Not just the actual KUKA servo motor, but each servo had to have its own servo amplifier and cables. And mucking about with the old hardware raises the odds of destroying a component that could be very hard to obtain a replacement for.
Then there's the "inexperienced buyer" problem. Most old KRC2s are "stock," or pretty close to it, but there were multiple generations whose hardware is only somewhat cross-compatible for swapping parts, and that's before we get into the "special editions" made with unique customizations for Volkswagen, GM, BMW, and other customers for whom "regular" wasn't good enough.
So, if you're going to buy a used robot, my advice is always: try to buy a model that isn't discontinued yet (KRC4s, in KUKA's case). If you have to buy a discontinued model, try to buy at least two of the same generation and "edition", so one can serve as a spare parts source for the other without falling into version-match hell. And if you want extra options (extra axes, special communications, etc), try to buy robots that come with those options, because finding the right bits for a discontinued version could be difficult, expensive, or both.
Also, be aware that every brand and generation has its limitations. For example, KRC stock KSS can accept multiple extra axes, but some brands require a paid option to be added to the operating system, and past a certain age, the manufacturer doesn't sell it anymore. And getting that software installed is going to be much harder than finding spare parts on the 2nd-hand market. OTOH, KRCs don't support simple TCP/IP communications without a paid software option, but some brands include it in the base OS.
The mechanical side doesn't bother me, as like you said, I could just buy a second for spares. It's more the control cabinet parts that worry me.
As long as the arm hasn't been abused, many of the better brands stay mechanically usable for a very long time. However, wear&tear does add up, and even brand-new, robot arms are far better at repeatability/precision than accuracy. I see a lot of people coming from the CNC world misunderstand this and get bitten.
In reality, do the controller parts need replacing often, are there common issues?
Most items in the controller don't wear the way the mechanical bits do, but certain items tend to fail, and can take out other components or software when they do. KUKA KRC2s, for example, ran on standard computer hard drives, which could fail unpredictable after decades of solid service. And when they do, you lose the core software and the special version of Windows that runs the user interface. So the first move on any used KRC2 that boots up should be to clone the hard drive.
Also, capacitors tend to age out over time, and when they go, you've lost the entire board they were part of, unless you can find an electronics expert willing to try fine-point repair. Cables tend to get crushed, kinked, etc, until they fail, and possibly kill some electronics.
It wouldn't be a big deal if I couldn't figure out a turntable, after all I'd have 6dof 😜. I could just do multiple setups.
This hinges greatly on what kind of motion you want. If you just want your turntable to be an indexer (robot and TT never move together), then doing that with a simple 3rd-part motor and some I/O is easy. This would be like a regular 3-axis CNC with a relay-controlled "flip jig" to work the opposite side of the stock.
If, OTOH, you want the TT to be part of the robot, to get something like CNC simultaneous 5-axis milling, that's a much bigger hump to get over. It essentially requires complete integration of the external axes into the robot, which can usually only be done using all the correct brand-specific hardware and software.
