Controls for 2-boiler setups

[mick]

How on earth did you guess? ;-) I've run the Crouzet demo software in a XP virtual machine under linux, but not for actual programming (the demo won't do it). It didn't want to run properly under wine (at the time).

Only if you happen to have them. I just happen t have a dozen or so assorted PICs and a programmer so I don't really feel the urge to invest in even more junk to clutter the place up.

Now that can be a problem ... However, it always is. It doesn't matter what form the logic takes, someone is going to have to maintain the system - even if it involves reverse-engineering your work. This is something that I used to get involved in as a control panel designer!

The most basic, and expensive, system is to use relays and timers. It costs a bomb but is quite easy to maintain. Design and assembly time is quite high. Pretty good on systems running off 110vDC tripping batteries though. Not much else can handle this. Also very good when all signals are at mains voltage.

PLCs can cost even more and aren't maintainable after about 10 years anyway (no bits made now mate!). Lithium battery usually supports clock/ calender and, sometimes, user-changeable values. That goes flat in about

10 years and the usr will *not* have changed it regularly you can bet! Quite often the logic is self-checking, with regular memory tests. PLCs get *very, very* expensive as soon as someone mentions "analogue i/o".

Intelligent relays have the programmable advantage of PLCs, but with a lot less cost. The internal lithium battery will run the clock/calender functions for about 10 years. I suspect that the logic isn't self- checking. Some can do analogue input & PWM output.

PIC & similar chips are a bit new yet. Very promising, but not self- checking logic so not good for replacing PLCs where it is critical. Can be *very* cost-effective though. You can always fit a socket and supply a few spares too, although if you've got the i/o designed right they'll probably never be needed. You could even do the whole jobby as a plug-in module now.

[mick]

It doesn't matter what we install in a customer's house - we are responsible for that installation whether we like it or not! :-) Also, there's really no reason why a well designed and built little PCB should be any more unreliable than any other piece of electronics - probably much more reliable than the boiler's PCB.

OK, reality: A 2-boiler system will almost always have the boilers connected in parallel for a start. If they aren't arranged like that then it's 2 separate systems.

YAPH is on the right track with his original post, except that the lag boiler is usually brought on by a stat in the common return main. Sometimes you may find that it is locked out by an external stat set at about 16C, so that the lag burner isn't brought in when the weather picks up. The 2 burners are usually switched to equalize the wear on them. This can be done automatically using a step relay or manually (much more common).

Quite often frost protection will run only the lead boiler, but if the return main doesn't reach a minimum temperature within a set time the lag boiler will also fire. If the system already has a frost protection system as part of the room stat (e.g. night set-back) or other temperature controller then you can probably ignore this. Don't depend on just the boiler's frost protection though - that's usually designed to protect the boiler, not the plumbing. A lot of systems used to use 2- stage frost frost protection. On air frost (outside stat) the heating pumps run. That keeps the water in the system moving and less likely to freeze, as residual heat from the warmer rooms protects the colder pipes. On low return main temperature the (lead) boiler fires until the low return main stat is again satisfied.

On bigger boilers you would need back-end valves to isolate them, but on something this size they aren't needed.

I'm not quite sure how the pump is handled in this application (the manual says that the pump must be connected to the boiler). It should probably be controlled by the lead boiler only, so 2 extra switching poles (on the lead boiler select switch) would be needed.

[geoff]

Nevertheless, in John's case, unless it's a favour job for a friend, I think that he should stick to off the shelf solutions

The reality (I like that word) is that if it goes wrong within, say, 12 months, and John, for some reason, illness or whatever, is not in a position to be able to rectify the fault, he would still be liable for someone else to come along and rectify the situation

Sometimes, you need to look a bit beyond the tip of your soldering iron

I should f'ing well hope so

[mick]

Hokay...

(Assuming that system has a timeswitch and room stat and that no further frost protection is needed)

Existing: Live to t/sw + room stat instead of feeding directly into SL on boiler feeds lead boiler signal.

Live from t/sw, before room stat, goes via new return main stat to lag boiler signal.

Now either specify boiler1 or boiler2 as lead and connect up or arrange switching (these boilers seem to be operated by a simple switched-live SL input). If not switched then pump is from whichever is lead boiler.

If switched, a 3-pole 2-way switch would be needed. 2 poles wired as a reversing switch from lead/lag to boiler1/boiler2 and the 3rd pole switching the pump live from one boiler to the other to keep it in step with the lead boiler.

Minimum cost, 1 return main stat. Optional 16C external hold-off stat in series with it. Optional 3-pole 2-way mains-rated switch for duty select (must be able to handle the pump load).

Each boiler should have it's own means of isolation, breaking both the continuous live and the SL connection.

That's a good reason to make the system easy to fix. As I've shown above, though, in this case you could make this work pretty well without adding more than a couple of stats and a switch. It's a pretty standard sort of circuit for dual-boiler heating systems and shouldn't flummox any heating engineer worth his salt too much. The important thing, above all else, is to document your work and leave a copy with the customer (who *will* lose it, but it makes them feel better) and another inside your control box/ panel/whatever. Even electronic solutions can be sorted out then.

yeah ...

[YAPH]

Actually 3 zones of CM907/927 progstats and 2-port motorised valves (plus HW zone) but yeah ...

Why? That would bring the lag boiler on when there's no call for heat.

Since the boiler each has its own pump only 2g2w switch required - standard-ish from elec wholesalers.

With external thermostat it would be simple enough: call-for heat (output of zone valves combined) goes direct to lead boiler's SL, and via external 'stat to lag boiler's SL (with the changeover switch swapping over the connections).

Yup, fan isolation switch to each.

Heh, yes - preferable glued in!

[YAPH]

Actually 3 zones of CM907/927 progstats and 2-port motorised valves (plus HW zone) but yeah ...

Why? That would bring the lag boiler on when there's no call for heat.

Since the boiler each has its own pump only 2g2w switch required - standard-ish from elec wholesalers.

With external thermostat it would be simple enough: call-for heat (output of zone valves combined) goes direct to lead boiler's SL, and via external 'stat to lag boiler's SL (with the changeover switch swapping over the connections).

Yup, fan isolation switch to each.

Heh, yes - preferable glued in!

[mick]

Oops! Yeah - sorry. After T/sw + room stat.

Personally, I wouldn't be happy with just an external stat for lag boiler control but YMMV. IMHO the vagaries of the English climate would leave you twiddling that stat until doomsday trying to get it "just right".

Wouldn't be the first time that's been done either! (by me ...)

The pumps still have me confused. The 15/60 can handle a greater head than the 15/50 so why were the external 15/50 pumps fitted? Is the head still too much for the internal 15/60s alone, and they were added to give them a bit of help I wonder? I could understand that if it was anticipated that only one boiler would run at once. I'm not a heating engineer, perhaps someone else can help? I suspect that the 15/50s will be set for full flow and the 15/60s throttled back to give about 1/2 the

15/50s flow each. Only a guess.

[mick]

Ah... yes... those pumps.

I've just been a-googling and come up with this:

"Centrifugal pump in series are used to overcome larger system head loss than one pump can handle alone. For two identical pumps in serie the head will be twice the head of a single pump at the same flow rate." "Series operation of single stage pumps is seldom encountered - more often multistage centrifugal pumps are used."

But:

"If you connect the pumps in series, the heads will add together, so the capacities must be the same or one of them will probably cavitate. You could also have a problem operating too far to the right of the best efficiency point with a possible motor "burn out"."

IMHO if the system is currently working ok with either one or both boilers running (without the noise of pump cavitation) then you should be ok as you are only changing the control method. However, if the system doesn't currently run like this (e.g. only one boiler runs at once) then I should be tempted to leave well alone for the moment. There may well be a way around this but I'm not qualified to answer.

[YAPH]

I haven't done the mass/flow calculations (and it would be hard to, given that I don't know where half the pipework goes) but it's basically a large-ish house with several parallel circuits, some of them single-pipe, which shouldn't be wildly demanding for a 15/60. If anything I guess it would want more flow than head - more like the 2 15/50s paralleled, except that the original installers instructions are to use only one of these at a time. (There's no reason you *couldn't* use both together, but the instructions suggest it wasn't designed to be used that way.) So, a mystery.

I'll probably take out the two paralleled pumps and see how it goes with the 15/60s fitted to the boilers, and have a re-think if that doesn't play nicely.

[geoff]

In message , YAPH writes

Don't forget that there is an 8-10min pump overrun from each pcb

[mick]

I was going to suggest that, but I thought you may not want to be messing with the plumbing! I can't see why they would be needed. You parallel pumps to get more flow and series them to get more head capability. Even putting a 15/50 in series with a 15/60 won't get you the full sum of their heads though, so there wouldn't be much to gain IMHO.

[YAPH]

Hmmm. There will be a temperature where it's still cold enough that the external stat is closed so both boilers get called, but the heat load is much less that can be supplied by both boilers. The system will heat up very quickly, and the boilers will cycle, and their anti-cycling mechanism may kick in. But I don't see that it's different to the situation in a smaller, conventional one-boiler system in warmer weather where the load is a lot less than the boiler output: you'll get cycling, and anti-cycling behaviour, in that situation too, n'est-ce-pas?

Well that's what the external stat is supposed to be - a summer/winter switch. But one that takes account of the actual temperature rather than what the calendar says!

[YAPH]

That should be OK if I use the pumps associated with each boiler. The 2 parallel external ones seemed to be wired to be on 24*7 (OK, switched manually on the wall) anyway! And hopefully I can do away with those.

[YAPH]

No, I don't know what the original installer was on but I really don't see how the existing arrangement could be necessary. With multiple, mostly single-pipe, circuits in parallel then, if you did need extra pumps you'd want to add them in parallel to cope with the flow not series to boost head.

[mick]

In theory the anti-cycling protection won't usually kick in much if the boiler is correctly rated for the load. It will do, but not often. In multiple boiler systems the effective boiler rating is, of course, changed by setting the number of boilers available to run. That is usually done "on the fly" by sensing the return main temperature as I've said before. That's done with either multiple return main stats (sometimes with delay timers) or a temperature sensor feeding a boiler sequencer box. Systems with 2, 3 or even 4 boilers are not uncommon.

In your case, using an outside stat, you *may* have problems. Not necessarily, because all heating systems vary. It will depend on your heating load, the response speed of the heating loop, hysteresis of the system etc. etc.

Your external stat position will be important. An external stat on a south-facing wall is always useless (artificially warm). The same stat in the centre of a north-facing wall (not above a window) will work perfectly. The same stat moved to an external corner of the same north- facing wall may work well for half the year, depending on the prevailing winds. The one thing that you do know for certain is that the return main temperature is always related to the heat loss of the building. If it's lower than expected then you need more heat input.

Try it. A single external stat is cheap enough and could still be used if you later decide to try a return main stat. If you want to try a "summer/ winter" select then just switch off the lag boiler.

And don't trust the weather ... ;-)

[YAPH]

---8 Try it. A single external stat is cheap enough and could still be used

Yup, that's my thinking. Start simple, if that's not good enough then's time for plan B. They're only down the road from me and nice enough folks, I think they'd understand if it takes a bit of iteration to sort out what they already know is a weird system.

Only snag I see with my plan A is I should probably arrange to only fire one boiler when there's only HW demand even when it's cold outside otherwise I'm pretty much guaranteed cycling. But that's probably not straightforward electrically. Dammit, now it's getting complicated again!

[Onetap]

I haven't read all this, so excuse me if I repeat an earlier post. What order of cost do you think would be reasonable for such a a 2 boiler control system? I'm not selling, just interested.

Some boiler manufacturers do controllers for multiple boilers, so that would be my first port of call.

The usual control system (in BMS systems) is a temperature sensor in the flow connected to a proportional (or proportional & integral) controller, with the lead & lag boilers set to operate at various percentage outputs.

The first boiler control system I came across was a proportional (output = gain x error) controller (Satchwell Monotronic) which put a

0-10V output into a valve actuator. There were microswitches operated by the actuator at various points and these operated the 3 boilers' hi/ lo fire burners. The gain and set point was set by pots on the controller. The disadvantage of proportional controllers was that there must always be an error from the setpoint to get an output.

When electronics got cheaper, this was replaced by BMS outstations which could not only control the burner but do all manner of clever stuff besides (automatic changeover of boilers, duty sharing, high/low temperature alarms, data logging, frost protection, etc., etc.,). BMS systems cost an arm and both legs and the programming is intentionally made inaccessible to non-trade users. There are some cheaper systems now (see adverts in BSEE magazine) so I'd try there too.

I have some 20 year old BMS outstations that could do this once programmed; it is tragic that such technology has been made unavailable to domestic users bt the UK building controls cartel.

[mick]

If I'm looking at the right beastie, those are simple boilers, not combis, so it shouldn't be too difficult.

I think you could use the HWS demand and heating demand signals to operate individual relays. You could then only connect the lead boiler for HWS under normal circumstances.

-------------------------------- | | | / heat / hws / stat | | | | | / heat +---------+ | | | lead boiler lag boiler

You can have fancy indicator lights for Heating & HWS too!

[YAPH]

If they were combis I wouldn't be fretting about my HWS switching!

Do keep up in the back there ;-)

Could do it with one relay. If the HW motorized valve had a changeover contact I wouldn't need even that:

---------------+---------------- | | | / combined call for heat | sat | from heating zone valves call o o------------+ \ | HW o | zone valve | / ext stat or relay | | | | lead boiler lag boiler

Yes, fancy lights are *good*!

I like the one in this

[mick]

Duhh ....

Nice. I used some a while ago. They were pretty expensive IIRC. We changed to some much cheaper 230vAC LED indicators. They are rather neat, with 2 strings of LEDs so you can still see them if a string has failed.

We used to use tranilamps ages ago, when they were still affordable and you could get bulbs!