1) One of the pistons was sticky. This alone would make the brake not self-adjust correctly and/or drag on one side.
2) The pushrod locking grub screw was not tight, meaning that the correct pushrod adjustment could not be maintained.
3) The pushrod adjustment was not correct. It was wrong by about 1/2 turn, which meant that the piston(s) would (almost) self adjust on the way out, but it was difficult to push the pistons back, (i.e. the lip of the MC piston was acting almost like a one-way valve). If the 'tight' piston was pushed back, the other one would be pushed out, rather than brake fluid be pushed back into the reservoir.
4) There were signs of oil loss from the caliper, eg there was oil around the MC area
5) once the pushrod was correctly adjusted, the brake still didn't work correctly, in such a way as suggested that there wasn't enough fluid in the system
6) once the reservoir was opened up, there indeed wasn't enough brake fluid (mineral oil) inside
7) the brake fluid was dark brown and murky (opaque in fact) as if it were badly contaminated with rust
8 ) the brake fluid was also very clearly contaminated with water. {NB oil floats on water, so when there is water in the MC reservoir, it suggests that there is a load more in lower parts of the system].
Hydraulic pistons can become sticky in any event; this is not uncommon with this model of brake IME; they commonly need fettling when new pads are installed. However once the piston(s) is/are sticky, the result can be a negative pressure in the hydraulic system when the brake is released, such that any water sitting on the back of the MC pushrod may get sucked into the system. Obviously there is very likely to be water pooled there with this model of brake if the caliper is mounted on the chainstay; water can simply run down the pushrod past the rubber boot and then it will just sit there. A similar problem afflicts the brake cable in this installation also.
Once there is water in the system, there will be corrosion in various places and this greatly increases the chances that seals will be damaged and oil will leak out, making more room for water and the whole thing descends into an evil spiral of doom. I'm actually amazed that this caliper offered any resemblance to a brake at all, the condition it was in.
The takeaways from this are severalfold;
1) a chainstay mount is to be avoided with this model of brake, unless you envisage use in dry conditions only
2) it seems that very small issues (such as the pushrod being 1/2 turn out of adjustment, or one piston being a bit sticky) can precipitate a very evil chain of events with this brake system
3) In common with a lot of other hydraulic systems, when the system is failing, there are few external signs of what (if anything) is going wrong.
The last of these alone is sufficient to cause me to make the recommendation that, unless you are prepared to invest time and effort in detailed inspection and maintenance (proactive rather than reactive) of systems of this type then they are best avoided. By contrast if you maintain a cable operated caliper on a more reactive basis, there are only a few things that will bite you badly in the bum (such as cable problems or the possibility of a seized FPA in a BB series caliper, for example).
The other thing is that I'd say is that both in conception and execution (esp placement on the frame) this shows all the hallmarks of a 'typical modern bike design' in that it is really meant for folk that ride on the weekend and don't often get the bike wet, rather than all weather riding, leave alone a claggy daily commute. That is what happens if you leave you design work to a bunch of Californians....
cheers
