After spending quite some time trying to work out what was going on he concluded that it was the condensate pump that was the culprit.
What he showed me:
- The boiler is fed from the main (240v) by a flex.
- Every time the pump is called into action the 240v to the boiler goes down to 0v.
- The controller relay, which gets its power from the boiler, then stops working as its (the boiler's) power is cut off (see my original message).
- Once there is no demand any more for the pump (after a few minutes) the power to the boiler is restored and it goes through the "boot" cycle and starts again until the next time the pump is called.
He thinks that the pump causes some sort of a short on the power feed to the boiler.
Makes sense to me... What do the experts think?
PS: I have already replaced the pump once since it was fitted 5 years ago. Unfortunately it is located under the house with quite tricky access, and since I just had a minor operation I will have to get someone in to replace.
2mA would require 120k?, which at 50Hz is 0.0266?F. 0.022?F is the NPV. Modern LEDs don't need more current for indicator use.
Then you size your R for the peak current under any circumstances. If we allow 0.33A at a peak of 330v we need 1k?.
Next: capacitor ratings. A mains capacitor requires a 250v ac X2 rating. 400v caps were formerly used but fail too often on mains, and lack safety features.
Now resistor ratings. 1k? at 2mA rms drops 2v rms & dissipates 4mW, so power rating is no issue. A 400v rating is a minimum, and typical through hole parts are rated to 200v, so 2x 470 ohm in series will do the job.
Finally the diode that goes back to back across the LED. It must cope with 0.33A peak and a few volts in reverse - not a demanding spec, but germaniums & some small signal diodes are unsuitable.
Sounds like the flex feeding the boiler isn't direct from the mains, but from some other relay, thermostat, multi-port valve, whatever that he hasn't understood ...
I just had a closer look inside the spur fuse/switch box feeding the boiler.
4 cables going in/out of it:
- A feed cable coming out of the wall behind and going into it (feed from the ring main I believe)
- 1 x thick flex going to the boiler (240v from the box to the boiler I believe)
- 2 x thin flexes going from the box to the cond. pump (feed and return?)
I am not 100% sure, but I am 99% certain that the wiring is as in this diagram:
formatting link
Based on that it looks as if his diagnosis is probably correct.
A fault in the pump causes it to cut off, resulting in breaking the live feed to the boiler. After a while it "recovers" and the feed is resumed.
Not the behaviour I would expect from the pump (does it not have a fuse?), but possible...
The penny dropped for him when, with no power to the boiler, he disconnected the discharge pipe from the boiler to the pump, blew into it (removing debris perhaps?), causing the pump to start working and the boiler came back to life.
I saw the guy checking the fuse on the 240v feed inside the boiler, which was fine.
See the 3 flexes in the bottom right of the photo. Also, the thick live (to the boiler) connected to the thin cable coming from the pump using the connector block.
Well I guess its possible that the condensate pump has a switch in it that discoinnects the boiler if it trips. Could be overtemeparytuer or overcurrent.
Or this (wiki)
"Condensate pumps usually run intermittently and have a tank in which condensate can accumulate. Eventually, the accumulating liquid raises a float switch energizing the pump. The pump then runs until the level of liquid in the tank is substantially lowered. *Some pumps contain a two-stage switch*. As liquid rises to the trigger point of the first stage, the pump is activated. If the liquid continues to rise (perhaps because the pump has failed or its discharge is blocked), the *second stage will be triggered*. *This stage may switch off the HVAC equipment (preventing the production of further condensate), trigger an alarm, or both*. "
Maybe time to say what the part number of the pump is and do some reasearch on it.
What it looks like is that you have a second stage switch and it thinks that the condensate is not getting punped out. And perhps it isn';t.
Personally life being increasingly too short, I'd get a whole new assembly and replace the whole thing. Probably less than £100 and not hard to do.
I agree 100%, life is too short, and am happy to replace.
Unfortunately I am unable to get to the pump at the moment as it requires crawling under the house for 10m, and I am after a minor op... I'll get a man in.
I did however manage to find the receipt from when I bought and replaced it myself a couple of years ago (the float mechanism died after 3 years). All it says is Jet HighFlow (£60 at the time). I do remember that it looked very much like these (most probably this is the one):
formatting link
No much info online though (it looks as if it had been discontinued), apart from a single sheet on the Wolseley website above.
PS: The WB guy suspected that it the pump that was the problem as soon as I described the symptoms to him. However, after he went under the house with a bottle of water he returned to say that it seemed to be working fine.
Only when the boiler started to turn itself on and off for no apparent reason whilst he was looking at it did he start investigating the power feed and concluded that it looked as if it was the pump that was faulty after all.
That's not credible. It could only cut off the boiler by a short if the fuse blew/MCB tripped. However, the condensate pump may have an additional NC level switch which is higher than the design maximum level in the tank, so that if the pump fails the "overflow" level switch opens and cuts off the boiler live feed which is wired in series with it. The actual pump motor is wired in parallel with the boiler always live supply (and in series with the normal level switch) and the issue will then be failing to pump, either due to motor failure, level switch failure or (my recent case) blockage of the pump with bacterial slime. Blockage of the pump outflow would be another possible cause.
I don't know if wiring to condensate pumps varies, but this one shows that the live is fed to the pump, and then via a "safety wire" to the boiler, so the boiler depends on the pump being "happy"
If the pump is working well, it is beginning to sound like a loose connection either at the pump or boiler end of the wires to the pump. Other possibilities are an intermiittenly failing pump motor or float switch, neither being likely. Or a blockage in the condensate pipe to the pump. The latter is excluded if ithe technician demonstrated that the live supply to the boiler from the pump was failing. Might be worth checking the mains supply to the pump, which is probably not from the boiler but might be, and the cable to the pump overflow switch. You can check the terminals at the boiler end even if the other end is currently inaccessible. Competent replacement of the pump will sort the connections at that end, even if it wasn't the pump that failed!
A purpose made LED analogue of a neon indicator lamp could be fabricated using a pair of same colour LEDs mounted on the same header wired anti- parallel[1] with a 10K resistor and a 2nF Y class cap in series. That would draw 300 microampres or 0.3mA. Using a 10nF cap would draw 1.5mA with the 10K 500v rated surge limiting resistor wasting all of 22mW with the LEDs consuming a mere 4mW or so.
The 10K surge limiting resistor is required to limit the worst case switch on transient (cap charged to -340v when switched back on at a positive peak of 340v giving a total of 680v across the resistor) current spike of 68mA.
Silicon rectifier diodes are usually specced to handle 1 microsecond current surges around about a hundred times the average continuous maximum current rating rating but this is with a Vf of only a quarter of that of a LED so I'd expect a LED to only have such transient peak ratings around 20 to 30 times their maximum average rating. For a small indicator LED, which these days may only have a 25mA rating, this would correspond to a circa 750mA 1 microsecond non-repetitive peak rating so a
1K surge limiting resistor could be utilised to reduce the waste energy to a mere 2.2mW.
Alternatively, with the ultra high efficiency LEDs we have today, a half milliamp of drive current would likely suffice and the 1 or 2 nF cap could be dispensed with and a single 300v rated 470K resistor used in place of the cap and surge limiting resistor.
The total dissipation would rise to 122.5mW (1/8 th of a watt), mostly in the dropper resistor. Though a magnitude larger energy consumption compared to the capacitor dropper circuit, it's still less than the 200mW or more consumed by the classic neon with 220K/180K ballast resistor mains voltage driven indicator lamp.
Fun Fact: The neon lamp in a recently purchased three way switched mains socket extension adapter draws some 200 to 250mW which is noticeably more than the 100mW or less standby consumption of a "Pound"land 2.1A USB charging wallwart I recently purchased for 2 quid.
Such USB wallwart chargers with sub 100mW standby consumption have been available for a decade or so now, rendering the advice to switch off or unplug such chargers when not actually in use, in order to reduce the UK's energy demand, rather *more* pointless than it was at the turn of the century.
[1] If such LED based indicator lamps are ever put into production (for all I know, seeing as how this is such an obvious application, they may well be available right now), the 'no-brainer' improvement to efficiency would be to integrate a string of 4 or 5 such LED pairs to convert a larger fraction of the mains voltage into light, allowing even lower lamp current requirements such that a 5M ohm ballast resistor will both simplify and reduce the manufacturing (and running) costs of such modern day replacements for the humble neon indicator lamp.
Using a 220 ohm surge limiting resistor, it's no wonder you managed to blow your test LED! :-)
That's a worst case scenario of 2.49 amps spike current through an LED probably rated for a maximum average current of 25mA with a possible non- repetitive 1 microsecond surge rating as high as 1 amp. You should have chosen a 1 to 10 K ohm surge limiting resistor to protect against such extreme switch on surge transients.
When I wanted to modify a couple of two gang and a single gang 13A mains sockets with LED indicator lamps for use on a dedicated 2KVA UPS protected supply nine years ago, I did give the surge current limiting resistor value some thought. I think I chose 10K (possibly just 1K - it was almost a decade ago!) with 47nF caps and 1N4000 series diodes in anti- parallel with 3mm red LEDs recovered from 70s vintage IT kit (no such thing as "super bright" LEDs back then).
They're still going strong after nine years of continuous service. To be fair, they haven't had to withstand many transient switching events in all that time compared to other mains powered kit that gets switched on and off on a daily basis. However, they must have gone through scores of switching events just the same in all of that time so ample opportunity to succumb to any miscalculation on my part wrt my choice of surge limiting resistor value.
If I'd been prepared to accept blue LEDs for this job, I'd have probably used just a 470K resistor sans capacitor (with a 1N4000 series diode wired anti-parallel across the blue LED).
That's not under *any* circumstances otherwise you'd have used a peak (worst possible case value) of 680v (2 times root 2 of the nominal rms mains voltage) so, for a non-repetitive peak of 330mA, you should have chosen a 2K2 resistor, ideally with a 600v rating but a pair of 1K1 200v rated resistors in series would have sufficed in practice.
The 1N4000 series of any PIV rating (50 through to 1000 volts iirc) are
*the* go to choice of diode for this job. Its rated average forward current is an amp with a peak around 2 orders of magnitude greater for 1 microsecond non-repetitive transients (typical of all silicon rectifier diodes).
An LED with four times the Vf is unlikely to have more than 1 1/2 orders of magnitude peak to average ratio, plus in this case, we're not discussing the sort of LEDs used in GLS typically rated for 350mA average current. Indicator LED lamps are typically rated between 20 and 50 mA average forward currents so you might see peak transient values somewhere from 330 to 1000 mA - 330mA is a safe conservative choice but, imo, you underestimated the *worst case* scenario by a factor of two. :-)
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.