Honeywell ABV setting

[decomplexity]

I have an open-vent S-Plan CH system (22 radiators, 2 DHW cylinders and long pipe-runs) driven from a new 35KW Viessmann 100W WB1B boiler. When most TRVs were open, the Grundfos UPS2 15-50/60 pump was struggling to keep an adequate flow through the boiler (the latter modulated and cycled excessively) so I changed it for its big brother the UPS2 25-80 which I currently run at speed ‘1’ in constant speed mode.

I also added a Honeywell DU144 ABV (the previous boiler didn’t use pump overrun but the new one does).

To determine the ABV’s setting, I needed to estimate a minimum flow when (e.g.) the room thermostat stops demanding heat or the programmer switches ‘Off’ (i.e. when all the flow would be directed through the ABV) . The boiler’s recommended minimum flow is 0.4 cu.m/hr. Using this and the pump’s speed ‘1’ head of 6m (which is constant for 0 - 2.3 cu.m/hr flows), the ABV’s setting chart gives a setting of 0.5, which is what I am now using.


If this isn't total rubbish, it leaves me wondering how to set the ABV if I want to use a faster pump speed (say speed ‘3’, which gives an 8m head). The ABV setting chart says that for an 8m head and with the largest available setting on the ABV (0.6), the flow rate is 0.8 cu.m/hr - nearly 60% of the recommended maximum for the boiler!

I just think I am misunderstanding something very basic!

 

[John.g]

ABV's work on the principle that as pump flow decreases then the head increases and you can then select a operating point on the pump curve where you want the ABV to give you the desired minimum flow, then as the TRVs/zone valves open the flow rate increases and the pump head decreases so the ABV closes.
Grundfos (among others) combine "fixed speed" with constant pressure to give large flow rates. (A true fixed speed pump will only give its maximum head at zero flow rate).
Because of this constant pressure operation you now will always have this minimum flow rate of 0.4 m3/hr irrespective of the heating requirements flow rates and 0.8 m3/hr at a 8M pump head.
The 0.4 m3/hr will have a impact on the boiler return temperature and the condensing effect but you will have to accept this with the present set up or 0.8 ms/hr @ 8M head, you may be able to source a ABV with higher settings or you could remove it (if allowed) and replace it with a combined manual bypass valve and flow meter and set this up depending on the pump head selected.
You could also dedicate one (or it may require 2) rad(s) (without a TRV) to give this minimum flow rate at all times, you will then get its heating benefit without affecting the condensing benefits.

 

[decomplexity]

Excellent answer John.g - thank you.

 

[doitmyself]

To determine the ABV’s setting, I needed to estimate a minimum flow when (e.g.) the room thermostat stops demanding heat or the programmer switches ‘Off’ (i.e. when all the flow would be directed through the ABV) . The boiler’s recommended minimum flow is 0.4 cu.m/hr. Using this and the pump’s speed ‘1’ head of 6m (which is constant for 0 - 2.3 cu.m/hr flows), the ABV’s setting chart gives a setting of 0.5, which is what I am now using.

An ABV will not work with your pump as the pressure remains constant over the working range of the boiler (0.2m³/hr to 1.5m³/hr - where did you get 0.4m³/hr from?). An ABV will only work if the pressure across it varies, which will not happen with your pump. At a setting of 0.5 there will either be a permanent flow of 400 litres/hr or no flow at all, depending on accuracy of the pump and ABV. Either way is not what is required.

 

[John.g]

The ABV will never close with a constant pressure, if 0.4m3/hr is the
boilers recommended minimum flow then a index setting of 0.5 with a pump head of 6M will (theoretically) give a constant flow rate of 0.4m3/hr, if 0.2m3/hr is the minimum boiler flow rate then the index setting will be ~ 0.55 with a pump head of 6M. The problem above is that if the required system pump head is increased to 8M then even with the index at its maximum setting of 0.6 the flow rate will be 0.8m3/hr.
A simple screwdriver slotted ball valve might be a better option but would need a fair bit of setting up, there are also "taco setter" type flow meters.
The best option might be a towel rad if the flow/return are upstream/downstream of any zone valves, the required minimum flow rate can then be set up by using the lock shield valve?

 

[decomplexity]

Thanks for these most useful comments.

Following Honeywell’s instructions for the ABV using the pump at its lowest speed setting at least gave an answer, even when the implication was that the ABV would be open all the time. Perhaps it would have been better to use a low-loss header and two smaller UPS 15-50/60s – one on the boiler side and the other on the system side. The boiler’s pump circuitry would then need to power at least the boiler-side pump, and possibly also the system-side pump if the header were too close to the boiler to achieve worthwhile heat loss on overrun (there would then presumably still be an need for an ABV on the system side).


The 0.4 m3/hr lower flow rate for the boiler was the best I could find from Viessmann’s technical literature (it is given as the flow rate lower than which a low-loss header is strongly recommended – as is at the upper end 1.4 m3/hr). I obviously didn’t look hard enough!


The towel-rail workaround is one I am currently using as – fortuitously – one was originally designed in on the boiler side of the CH motorised valves and gives what I hope is acceptable heat loss on overrun (there is also a further 7m of 28mm pipe run to the boiler…) . The rest of the system has grown haphazardly over 35 years.


Quite how the Grundfos UPS2 25-80 180 achieves a constant head from zero flow to around half maximum flow rate while maintaining a constant speed is intriguing (the head of the smaller UPS2 15-50/60 starts tailing off much earlier). Given the number of TRVs and DHW cylinders, I did look at exploiting the pump’s proportional pressure mode instead, but doing so and also catering for pump overrun would seem to need some sort of ‘inverse’ ABV that opened rather than closed when the system flow shut off!.

 

[chris watkins]

Close Coupled Tees = a lot cheaper that a LLH.
If you go with this option most likely you could have got away with the UPS2 15-50/60, once the resistance through the boiler was dealt with.
Heat loss on overrun will not be an issue just as long as the water is moving through the boiler.
The system pump (ups2 25-80) can then be set to work in auto which would deal with all the system pressure / flow changes leaving the boiler to run it's own pump & modulation to achieve the required delta T.
Strongly recommend that you close the system, there is normally no good reason not to.
Maybe should have purchased a system boiler & added additional expansion.

 

[John.g]

Re boiler minimum flow, if you assume a modulation ratio of 5:1 and a boiler deltaT of 20C then a minimum output of 7kw would require 0.3 m3/hr, so 0.4m3/hr would result in a deltaT of 15C. so close enough I'd say.

I think your towel rail set up is very good especially with that long run of 28 mm piping and even if you changed to PP mode then you should still get adequate minimum flow, you might try PP2 mode @ 7.5M. The attached file may be of some help in determining which one to try, PP3 would seem far too powerful.

Re "constant head/constant speed", no centrifugal pump can do this, and from the horses mouth:
“The UPS2 XX-80 is a 140Watt design and the curve is limited to 8meter until the 140watt is reached by the product QXH and then it follow the normal constant speed curve down.”

 

[decomplexity]

Although my CH system is larger than the typical domestic one, the same issue presumably applies to anyone with a boiler that needs pump overrun (most of them?) plus a pump that has a constant pressure regime at the ‘low flow’ end. The ubiquitous Grundfos UPS2 15-50/60 for example has a head virtually flat from zero flow to 0.6 m3/h on the highest of its three constant-speed settings (or 0.35 m3/h on its slowest setting). The roughly comparable Wilo Stratus Pico range has a similar characteristic, as does the Grundfos Alpha3 (the ‘clever’ version of the UPS2).

So I am a bit perplexed how installers handle flow bypass in this common combination of pump and the boiler’s need for pump overrun.

 

[John.g]

Yes, the trend now seems to be partially constant pressure, I have a Wilo Yonos Pico 1-6 which also flows 0.6 m3/hr @ 6 M head, I have a 20 kw oil boiler (no overrun needed) running 10 rads + HW on a PP setting of 4M (2-4). You can still get pumps like the DAB Evosta 4-7 which has true "fixed speed " settings.
I did a few calcs sometime ago based on a pump running at a PP setting of 6M (3-6) and if one set the ABV to 0.1mbar then combined with a suitably sized orifice/valve you would get a flow rate of 0.4m3/hr @ 3M and 0.63 m3/hr at 6M with the same orifice and ABV setting, however if a orifice (with no ABV) was sized to give 0.4m3/hr @ 3m then this would give 0.56 m3/hr at 6M, so I dont see any advantage in fitting a ABV combined with the throttling device.
It would appear that constant pressure with a fixed throttling device or towel rail is the only way to achieve a minimum by pass at all times short of getting the boiler controls to open a zone vale or something of the sort and I'm surprised that this isn't becoming more popular now.

In your first post you said "When most TRVs were open, the Grundfos UPS2 15-50/60 pump was struggling to keep an adequate flow through the boiler (the latter modulated and cycled excessively)" why was this occuring with the UPS 2 presumably at speed 3 which, even if the head required was 6m would still give 0.6 m3/hr?. If the boiler had modulated down to ~ 7 kw then the deltaT would only be 10C and one wouldn't expect excessive cycling especially with all TRVs open.

How is the system running now on the 8M pump and what setting.?.

 

[decomplexity]

Thanks again John; I learn a little more every day!

Re “… short of getting the boiler controls to open a zone valve or something of the sort…”, a solution would appear to be to install a motorised valve where an ABV would normally be and a logic ‘gate’ in the wiring centre that said “if power to the pump AND no demand to the boiler THEN open valve ELSE release valve”, and the pump could have an additional Line input from the switch on the valve rather than being driven directly from the boiler.

If as you implied boilers had an extra output circuit powered only when pump overrun was on, this could open the valve directly. Come on Viessmann, W-B et al, are you listening?


Re “why was this [modulation and cycling] occurring with the UPS 2 presumably at speed 3 which, even if the head required was 6m would still give 0.6 m3/hr?. If the boiler had modulated down to ~ 7 kw then the deltaT would only be 10C and one wouldn't expect excessive cycling especially with all TRVs open” We found when the smaller UPS2 was uninstalled that the impeller was significantly gummed up with black deposit (even though the system has around 2 x 4L of Fernox MB-1) which may have been part of the problem. When the larger UPS2 was installed and set to constant speed 1 (virtually the same low-flow head as speed 3 on the smaller pump), the difference was obvious and the boiler works steadily with mild modulation only. I have left it pro tem on speed 1 and next want to rebalance the system before changing things further.

The main design difference is of course that even on the lowest constant-pressure speed, the larger pump maintains a constant 6m head to around 2.2m3/h flow which is well beyond what I would want to expose the boiler to!

 

[John.g]

It would be very interesting if you could measure the pump power (watts), I don't think that particular model has a power display but might be measurable with a power meter.
The main thing if doing so, is to ensure that the pump isn,t running out of steam (power) on whatever curve you are using, ie < 80W @ 6M, < 110W @ 7M or < 140W @ 8M., that way you will be certain that its running at its rated head, if on speed 1 for example and say drawing 70W then you can read off the exact flow rate (@6M) from the pump manual, you could then enter those numbers in cells D14 & E14 (in attachment) and you can see the exact performance for each PP setting, it would make for very interesting reading for all "PP fans".

 

[decomplexity]

Thanks John

When I had a look at the .xlsx you kindly attached to a prior post, I understood how the calculation worked but was unclear exactly what it was trying to achieve (and how). Perhaps you could kindly elaborate.

The input variables ‘Assumed or Known Head’ and ‘Assumed or Known Flow’ give rise to an ‘Actual Flow’ CURVE and the comment that “The intersection [POINT] of the two curves is the actual head and flow ”. Also what please is the background behind the term SQRT(PP head/actual or assumed head) that pro-rates the ‘Assumed or known flow’ to give the Actual flow?

 

[John.g]

A bit long winded........

What I am trying to achieve is trying to see what would be the most appropriate PP setting at any given “known” or “assumed” head & flow. PP mode gives nice savings especially when only one or two zones are open. It is important that the PP setting (1,2 or 3) selected should achieve (as near as possible) the head & flow required when all zones etc are open, so one must have some idea of that requirment.

The PP works this way (IMO), the pp curve is set in the microprocessor, in the case of the 8M UPS2 on PP1 2.2M=0LPM & 6m=71.7LPM so that gives one curve (blue), now because I assumed that the flow is 33LPM @ 5M, then because Flow is proportional to the sqrt of head, then the actual flow @ 2.2M= sqrt(2.2/5)*33, 21.9LPM & at 6M= sqrt(6/5)*33, 36.1LPM., this gives the actual flow (red) curve. When changed over to PP mode the pump starts ramping up and the pump power will start rising with it, when the pump power and the set PP power are equal then the pump will stop ramping and thats it!. In practice, I think the pump just compares or is comparing the two “numbers” continuously and just ramps up and down as required to reach the equilibrium point.

The assumed head&flow was 5M@33LPM, PP1 achieves almost 30LPM@4M, if the required flowrate was halved to 16.5LPM@5M the PP1 will give [email protected] but as flow demand decreases line losses decrease as well so the PP setting will in practice give a increase in flow for that same head.

PP Control

"Radiator heating loop
This is the most common hydronic heating system. A boiler provides the heat, a set of distribution pipes runs through the whole house and the radiators branch off of the feed pipe and return cooler water to the return pipe. To make the system efficient, the radiators are equipped with thermostatic valves. These valves are the reason for the wide variability of the hydraulic resistance in such a system. Simply put, on a nice afternoon in March, with spring in the air, only some rooms on the Northern side of the house may have open thermostatic valves while the vast majority of the house is warm enough. The system resistance will be very high and the required flow of water low. On a cold December morning however, you have the opposite: all rooms are calling for heat, the valves are open and the system resistance is extremely low while a lot of flow is required.

To serve such systems best, the industry has come up with a control scheme called proportional pressure control. It starts with the assumption that about half of your pressure loss in the system will be in the distribution pipe while the other half is lost in the radiators. Consequently, the pump is controlled such that it will respond to a decrease in flow with a reduction of its head and that at zero flow, when all valves are closed, it will provide half the head pressure it has at maximum flow.

So how do you set such a pump? The pump needs to be strong enough to provide heat for the whole house, and therefore you should set it to reach maximum head when all valves are open. If you know your maximum design flow, you can pick that point out of the pump diagram. If not, you open all the thermostats in the house fully (assuming hydraulic balancing has been done), and then increase the power setting slowly until you see that the head pressure is not increasing anymore.

What happens if your setting is off? On the low side, you may experience boiler cycling and insufficient heating. On the high side, you may get “whistling” thermostatic valves. Whistling is surely inconvenient, but cycling means less efficiency – so, if you err, do it on the high side."

 

[decomplexity]

Ah ha - now understood. Thanks John.
When I have re-balanced the system I will make some more experiments.
A system such as mine that has (for a domestic installation) quite a lot of TRVs as well as two DHW cylinders sounds ready-made for PP. With the previous smaller UPS2 and previous (21KW) boiler, I tried using all three PP settings but even the largest of the three took ages to get the house to an acceptable temperature - the pump just hummed along to itself almost inaudibly.
When I have another go with the new system I will have first either to replace the ABV by a gate valve or similar as the former will close at low flow (= low differential pressure for PP) or continue to use my towel rail as a bypass. However, as I think you noted in an earlier post, my big UPS2 still has a decent 'zero-flow' head (2, 3 and 4M for the three PP settings respectively) and should be able to circulate enough water on pump over-run to cool the boiler (the smaller UPS2 with its zero-flow PP heads all of around 1M might have had an issue).

 

[doitmyself]

"Radiator heating loop
This is the most common hydronic heating system. etc etc"

Can you provide a link to the source of your quotation?

 

[John.g]

How to choose the correct speed control for heating systems - Lowara - International

 

[doitmyself]

The main design difference is of course that even on the lowest constant-pressure speed, the larger pump maintains a constant 6m head to around 2.2m3/h flow which is well beyond what I would want to expose the boiler to!

Can you explain what you mean by the part in red?

 

[decomplexity]

My comment was based on Viessmann's Technical Data Manual for the Vitodens 100W where the normal maximum flow rate is given as 1.4 m3/h and where anything greater needs either a low-less header or closely-spaced tees. The larger UPS2 in constant pressure mode has a constant head at 6m up to a flow of 2.2 m3/h, at 7m up to 2.8 m3/h and at 8m up to 3.2m3/h.

 

[Millsy 82]

Just turn your heating off let it cool down shut all but the 2 smallest radiators and shut the bypass.

Turn heating on get 2 rads nice and hot (shouldn't take long) slowly open bypass until you can feel water passing through it.

Turn temperature of boiler right down open a few of the rads to take the heat away.

Shut all radiators turn boiler to max

Drink coffee wait and make sure flow and return is no more than 20c difference but you don't want it too close.

Sometimes I think people lose themselves with figures and numbers they forget common sense.