Because it matches EXACTLY both the boiler and fault scenario for a job I looked at today, I've revived this old thread.
In today's case, the boiler was installed in late March 2010, to replace a cast-iron floor-mounted type, of about the same output (28 - 30kW). It's apparently never worked properly and every man and his dog, including Vaillant, have looked at it, tinkered in various ways but failed to solve the problem. The system included mostly old rads and pipework and WAS open-vented. It was Power-flushed in an effort to fix the S53 / no heat problem and the owner says she thought the flush actually made it worse! The system water looks pretty clean now but NO evidence of any inhibitor chemical in it (can power-flush companies be SO cynical that they leave a system vulnerable to MORE corrosion in the hope of more work in future - you bet you a*se they can!). The system has also been converted to the sealed type, with a correctly-sized expansion vessel.
Apart from some electrical problems (eg. boiler responds to the room stat irrespective of state of the programmer!) the main problem is that is can never get the rads up to full temperature! The boiler fires, the Flow temperature goes whizzing up, reaches the 16 degree Flow / Return delta-T shut-down point and then overshoots up to around 20 degrees delta-T and the boiler goes to S.53. The problem is that it stays in that state for several minutes (just a timeout, unaffected by falling Flow temperature) and then fires again, and the same cycle repeats ad nauseam. Due to the long S.53 shutdown periods, the Return temperature only creeps up and then levels out at about 38C degrees.
I tried REDUCING the system output (D0 setting) and found that at 16 kW (down from 28kW), the S53 went away and it went into a more normal 16 degree delta-T shutdown - pump overrun - wait awhile - restart cycle. This in turn meant that the boiler was actually running for more of the time and the system started to stagger into life, with some of the rads getting at least a bit hot. The reason for this was that the gentler Flow temperature increase did not cause the overshoot after the delta-T got to 15 - 16 degrees.
My conclusions so far:
1.It's very unfortunate that Vaillant apparently did NOT include any 'ramp-up' logic in the Ecotec software. Other makes (Atag, Viessmann?) limit the burner output so that Flow temperature only rises at 2 or 3 degrees per minute. This would ensure that on a 'big and slow' system, with lots of rads and a big volume of cold water sitting waiting to come down the Return, that the boiler would not go crashing into the S.53 state all the time.
2. This system (21 rads on three floors, plus DHW cylinder, plus underfloor heating in a new extension) is too big for the boiler. Although the available heat output is OK, the boiler has too much internal resistance. As a result, there can NEVER be enough flow through it to keep the Flow to Return delta-T within limits: too much heat will be lost from the rads even when the system is still quite cool, so the Return temperature will probably never exceed 45 degrees! This will stop the boiler whenever the Flow gets to 61 or so, and the boiler will 'hunt' at around those temperatures indefinitely.
3. Far as I can see, the cheapest / best solution is a low-loss header and a second (modulating) pump on the downstream side of it. The existing 15/60 type would be fine as a shunt pump to drive the loop between boiler and header. Once set up that way, the boiler can run up to its proper setpoint quickly without hitting S.53 or high-delta-T conditions.
4. Based on a previous, similar problem with another system (large but not with a Vaillant boiler) I don't think that just fitting a larger pump will fix it. Previously, I found that even with a large pump (Grundfos Magna, driving 5 zones and 25+ rads) the boiler's resistance put the pump curve into the nearly vertical area, so that increasing the pressure could NOT increase the flow. A header DID work for this other system, although there were other, unrelated design issues. Also, if you put a big pump on a high-resistance system, it gets noisy.
Anyone have more / different information to invalidate these conclusions or possibly offer a better solution?
