Motorised 2 port valve; what has slowed it?

[MeldrewsMate]

@windbagUK , I accept fully your assertion that the replacement 2-port valve has made your E911 alarm go away, but you seem reluctant to accept that any sort of stress testing of the refurbished system would be beneficial i.e. suppose the Sika switch needs a water velocity of 1 m/s to switch and with the old valve in place the average velocity was indeed 1.0 m/s (varying between 0.9 and 1.1). The new valve may give the system an average velocity of 1.05 m/s, varying between 0.95 and 1.15 m/s; a small increase, but within a 'normal' range for this type of valve.

The control strategy, of which neither of us has full details, could monitor the Sika switch every 5 seconds. On each iteration of the 'flow proving' program the system may be programmed not to alarm (E911) unless, say, 3 consecutive underflow conditions are registered. With flow velocity varying between 0.9 (too low) and 1.1 (good) there will be more periods when 2 consecutive 'too low' signals are registered than with the new valve, thus the possibility of an E911 alarm being generated by 3 consecutive low signals will be significantly reduced with the new valve.

I am not suggesting that you remove and replace several zone valves in order to prove a point. I am suggesting that you have a curious mind and you wish to fully understand the reasons why the replacement of one component with an almost identical component would make a problem go away, and I am suggesting ways in which you can seek to reproduce the error message by altering one factor at a time. Once that factor has been established (such as deteriorating pump output) then action can be taken to increase reliability of the system, and using the example of deteriorating pump output, increase the flow velocity to between 1.0 and 1.2 m/s, such as to eliminate completely the possibility of 3 consecutive 'low flow' conditions. Conversely, and easier/cheaper to implement, is to restrict the flow slightly to significantly increase the possibility of the alarm triggering criteria being met. I hope this has made my point clear.

The facts and the variables are clear though.
1. The zone valve makes no signal to the control system - you have stated that the end-switch wires are not connected.
2. The only signal the controls receive is from a simple flow switch - the two conditions are GOOD FLOW and LOW FLOW.
3. The control strategy, and in particular that for the E911 alarm, is not known to us, though we can reasonably assume it is set in software - therefore it may have 'anti-bounce' and 'anti-spurious triggering' algorithms built into it; thus the impression given to the simple observer is that the alarm triggering time is considerably longer than the 10 seconds or so that the valve takes to open fully.

 

[windbagUK]

The 3-port, spring return valve, has always been the most unreliable part of my heating system. Which generally caused a failure every couple of years or so. Such was the regular failure, that I would always have a spare head ready to swap in. The spare would either be a brand-new head, or a head that I had refurbished myself, assuming it was a motor, or microswitch failure and I became quite expert at refurbishing them - though I would often find the mechanicals were often simply worn out, or the grease dried out. A major cause of the unreliability, I found, was due to the mechanical stress the parts were under, from the return spring and the motor being stalled, under power during heating demands.

Such was the poor reliability, that I had it in mind to attempt to devise a redesign of the head, which did away with the need for springs at all, but then - I came across a 3-port valve called a MOMO which didn't involve any return springs, or a need to stall the motor under power, which was a direct electrical swap for my spring return 3-port.

It is electronically slightly more complex internally, and requires a swap of both the valve and the head, because unlike the spring return - it only rotates in one direction, doing complete circles. Mechanically, it doesn't fight against the spring and the motor only runs to position then switches off, so it is under much less stress.

Early days yet, but so far the MOMO (Motor On, Motor Off) has been 100% reliable, though I do keep a spare head to hand, should it fail.

Thanks for that, Harry - interesting that others find such components cause issues with their system.
Seems to bolster the idea that they should be fitted with isolation valves so they can be replaced with minimal system disturbance.
Another significant advantage is that it should consume no power when open or closed.
One thing I don't understand with your description ; in my system, the power is cut to close the valve via the stored spring energy. I understand the MOMO goes full circle - but what provides the power if the control cuts it to the valve completely?
Does it have an extra permanent live feed to complete the other half cycle & then shut that off with the internal relay?

 

[windbagUK]

@windbagUK , I accept fully your assertion that the replacement 2-port valve has made your E911 alarm go away, but you seem reluctant to accept that any sort of stress testing of the refurbished system would be beneficial i.e. suppose the Sika switch needs a water velocity of 1 m/s to switch and with the old valve in place the average velocity was indeed 1.0 m/s (varying between 0.9 and 1.1). The new valve may give the system an average velocity of 1.05 m/s, varying between 0.95 and 1.15 m/s; a small increase, but within a 'normal' range for this type of valve.

The control strategy, of which neither of us has full details, could monitor the Sika switch every 5 seconds. On each iteration of the 'flow proving' program the system may be programmed not to alarm (E911) unless, say, 3 consecutive underflow conditions are registered. With flow velocity varying between 0.9 (too low) and 1.1 (good) there will be more periods when 2 consecutive 'too low' signals are registered than with the new valve, thus the possibility of an E911 alarm being generated by 3 consecutive low signals will be significantly reduced with the new valve.

I am not suggesting that you remove and replace several zone valves in order to prove a point. I am suggesting that you have a curious mind and you wish to fully understand the reasons why the replacement of one component with an almost identical component would make a problem go away, and I am suggesting ways in which you can seek to reproduce the error message by altering one factor at a time. Once that factor has been established (such as deteriorating pump output) then action can be taken to increase reliability of the system, and using the example of deteriorating pump output, increase the flow velocity to between 1.0 and 1.2 m/s, such as to eliminate completely the possibility of 3 consecutive 'low flow' conditions. Conversely, and easier/cheaper to implement, is to restrict the flow slightly to significantly increase the possibility of the alarm triggering criteria being met. I hope this has made my point clear.

The facts and the variables are clear though.
1. The zone valve makes no signal to the control system - you have stated that the end-switch wires are not connected.
2. The only signal the controls receive is from a simple flow switch - the two conditions are GOOD FLOW and LOW FLOW.
3. The control strategy, and in particular that for the E911 alarm, is not known to us, though we can reasonably assume it is set in software - therefore it may have 'anti-bounce' and 'anti-spurious triggering' algorithms built into it; thus the impression given to the simple observer is that the alarm triggering time is considerably longer than the 10 seconds or so that the valve takes to open fully.

I accept all that you say. The end facts are indeed a good summary, and yes, I am curious.
Looking at the Samsung manual, it describes E911 as :
Flow S/W OFF error
In case of flow S/W OFF in 30 sec during water pump signal is ON (Starting)
In case of flow S/W OFF in 15 sec during water pump signal is ON (After Starting)

So, it looks like it gives valves 30 seconds to "settle" when the system comes on (presumably main UF heating demand coming on), but only 15 seconds when swapping function (eg DHW triggering on) - as it was only failing when the DHW was requested by the program, not when the heating was being switched on. (perhaps if that 15 sec were changeable to 30 seconds, I may never have had a problem).

Given that, I am unclear about exactly what test(s) you think will help.

I can choke down the flow of the Wilo Yonos Pico 25/1-8 pump - but it does change the flow values on its own during running.
It is a 6-year old model that only has two "variable potentiometer" settings; one for "radiators", one for "underfloor".
It is set to "underfloor", which gives a "metres head" display on the LED. That LED value changes between 6.9 to 7.3 without any manual intervention, so it must be doing some internal control of its own - but on what basis, I do not know.
In that context, it is not obvious what any reduction of the "variable potentiometer" control would have in real flow-rate terms.
There are other isolation valves, but I have no feel for what flow change would be invoked by what "turn fraction" at any point.
My experience with such is that they are highly non-linear, much like the gas flame controls on my hob.

The only other trivial place to choke the flow would be on the control valve at the end of the sight-flow gauge, effected by a large flat screwdriver. But that is so crude that a tweak could change the flaw substantially. The snag there is that the flow-rate observed with that sight-gauge before and after the E911 occurred with the old valve was rock solid, so whatever was causing the issue was not observable by that method.

I think you might have been suggesting plugging/unplugging the Sika return from the control head when running ?
But the E911 only happened in that small time interval when DHW was being switched on, so there is not much time to deduce what the software is doing - plus having to do this with the top off the Samsung control unit live running.

Finally: you say:
The new valve may give the system an average velocity of 1.05 m/s, varying between 0.95 and 1.15 m/s; a small increase, but within a 'normal' range for this type of valve.
While I can believe that, it implies that it was a case of random luck whether a new unit would pass the relevant spec or not.
As so many folk are convinced that full valve replacement does rectify the issue, and attest that it does, it seems not a matter of such luck - and that there IS a fundamental difference in performance between old & new units.

But what IS that difference ?

I only have a single statistical observation - of one unit being swapped - but it is inline with the "Swap will work" strategy.

 

[Johntheo5]

Re the pump, I too have a Wilo Yonos Pico 6M, in UFH mode or any mode it constantly displays the pump power in watts W, it only displays the head in meters M while you are changing it incrementally in 0.1M steps and then reverts to displaying the power W. If it is displaying 6.9 to 7.3 W (watts) then it is pumping little or nothing.

 

[Harry Bloomfield]

Thanks for that, Harry - interesting that others find such components cause issues with their system.
Seems to bolster the idea that they should be fitted with isolation valves so they can be replaced with minimal system disturbance.

Usually it is the head which fails, jams, or wears out - most can be replaced without disturbing the valve.

One thing I don't understand with your description ; in my system, the power is cut to close the valve via the stored spring energy. I understand the MOMO goes full circle - but what provides the power if the control cuts it to the valve completely?

The 3-port is different to the 2-port, in that the 3-port only needs power to enable the mid-position, or CH position. With no power applied, they default to HW.

There are MOMO versions of the 2-port valve, for instance -

Sunvic Momo SZMV2323 2-port valve 28mm​

 

[MeldrewsMate]

...Looking at the Samsung manual, it describes E911 as :
Flow S/W OFF error
In case of flow S/W OFF in 30 sec during water pump signal is ON (Starting)
In case of flow S/W OFF in 15 sec during water pump signal is ON (After Starting)...

There is another interpretation of the manual; that the system allows 30 seconds for the flow to satisfy the Sika switch on start-up, but if that flow signal is lost whilst already running the system will alarm after just 15 seconds, this seems to be plenty of time for your previous and present valve to fully open, so I suggest discounting that as a possible suspect.

...I think you might have been suggesting plugging/unplugging the Sika return from the control head when running ?
But the E911 only happened in that small time interval when DHW was being switched on...

Let's move on to examine this new information about how the alarm criteria is met during the running period. I may have missed something earlier - are you complaining that the alarm only seems to coincide with the DHW demand?
Does it occur with both CH and DHW demand?
As the Sika signal must be lost for a continuous period of >15 seconds for the alarm to be generated then it is prudent to examine the nature of the Sika switch and its output signal.
You have stated that the input to the Sika switch is a paddle. What is the device which translates this mechanical movement into an electrical signal?
A micro-switch (mechanical)?
A glass reed-switch operated by a magnet?
A hall effect semiconductor?
Do you have access to the terminal where the Sika switch cable connects to the controls? What is the FLOW ON voltage? What is the FLOW OFF voltage at these terminals?
It's also worth checking these terminals for tightness if they're subject to vibration.
I'm asking for these data because the (assumed) micro-controller inside the Samsung unit will not monitor the signal continuously, but may look for it once per 100ms, once per second, or even once per 5 seconds...we don't know and it's not really relevant that we do. My suspicion is that the Sika signal is disturbed during a transition from CH only to either CH+DHW or to DHW only. If a magnet operated paddle switch over-travels it may disturb its output.

 

[MeldrewsMate]

Please post either a picture of the Sika switch, or a link to show the spec of your switch. The one I found suitable for insertion into copper pipework is VK325M0P10PD11, not the same as you quoted, but it seems all the Sika VK325 switches use magnetic actuation of a reed switch. The tech spec suggests that overtravel is unlikely, however reed switches are not immune from vibration so my suggestion, after the recommendations above, is to check the security of the output terminals, followed by pipe support in the area of the switch; the important part is to eliminate or significantly change the resonant frequency of the pipe at the connection point into the water system - a block of wood will do for now.

One further thing to consider (because a fault 'going away' is not the same as the elimination of that fault) is to again consider the Samsung control system. Does the ASHP flow 'switch' between DHW or CH output? This is equivalent to a 3-port valve with no mid position. When the CH valve closes under spring force the valve plug (which is a rubber ball on a lever) may vibrate under the force of the slowing water during the last few mm of travel. If the pump pressure (up to 0.8 bar for your pump) is set too high this vibration may (just may, nothing certain) disturb the reed switch at just the right frequency and phase to occasionally induce fault E911 (>15 seconds). Taking into account the type of contact anti-bounce software frequently incorporated into control systems the chance of this may be much greater than the pure maths may suggest!.

 

[windbagUK]

There is another interpretation of the manual.... ...this seems to be plenty of time for your previous and present valve to fully open, so I suggest discounting that as a possible suspect.

I agree

Let's move on to examine this new information about how the alarm criteria is met during the running period. I may have missed something earlier - are you complaining that the alarm only seems to coincide with the DHW demand?
Does it occur with both CH and DHW demand?

No, it only appeared with DHW demand when the system opened the old 2-port valve - and was usually happening overnight. It was not consistently failing. Sometimes it would run for three days without fault, other times it would fail three times in succession when deliberately invoking new demand for DHW. It was independent of whether there was already CH demand or not.

The CH demand is separately actuated via a similar 2-port valve, but that has not tripped the system into failure - yet.
Samsung technical seem to think it may start failing on that demand when that valve becomes "sluggish".
That valve is sited at a similar distance from the flow gauge/Sika switch combo.

As the Sika signal must be lost for a continuous period of >15 seconds for the alarm to be generated then it is prudent to examine the nature of the Sika switch and its output signal.
...
Do you have access to the terminal where the Sika switch cable connects to the controls? What is the FLOW ON voltage? What is the FLOW OFF voltage at these terminals?
It's also worth checking these terminals for tightness if they're subject to vibration.
I'm asking for these data because... If a magnet operated paddle switch over-travels it may disturb its output.

I'm following this post with some details of the Sika switch as requested.
We did look at the Sika before draining out the system - there was nothing untoward being loose. It is bolted to the plant room floor amid loads of other pipes, and there is no scope to alter that.
No we didn't look at any voltage the Samsung controller put down the switch line, but I'd be surprised if it were any more than a sub-12V DC logic or relay voltage. Putting down mains voltages would invoke extra safety considerations for routing that connection.
We did test the paddle-switching-open-to-closed-circuit when removed, and can say there was no no sign of anything loose, so have not attempted to dismantle the head connection as there was no indication of a fault there.
But:
As the fault has been totally eliminated by putting in the new motorised valve, I don't see how the Sika can be the source of the fault - unless there is some sort of exotic interaction between the two - but that still implies a feature of the motorised valve opening character that developed over time caused the issue.

 

[windbagUK]

Please post either a picture of the Sika switch, or a link to show the spec of your switch.

I only found a single Google find for the full spec of the label attached to our Sika switch: VK325M0SAMSU17
That result was here:
SIKA Flow switch type: VK325M0SAMSU17 - Set point 16.0 ± 1.5 l/min
This was the ONLY Sika switch website listing that not only displayed the set point, it was correct for our system.
Hopefully this won't tread on significant copyright toes:

That indicates it is a reed switch activated by a magnet on the far end of the visible paddle, as you indicate.
My "VK325M0SAMSU17" Google search only returned one other Sika switch, but only as far as the VK 325 M designation, and doesn't confirm the flow switch setting.

..., however reed switches are not immune from vibration so my suggestion, after the recommendations above, is to check the security of the output terminals, followed by pipe support in the area of the switch; the important part is to eliminate or significantly change the resonant frequency of the pipe at the connection point into the water system - a block of wood will do for now.

To repeat, it is bolted to the plant room floor amid loads of other pipes, and there is no scope to alter that.

One further thing to... consider is the Samsung control system... When the CH valve closes under spring force the valve plug (which is a rubber ball on a lever) may vibrate under the force of the slowing water during the last few mm of travel. If the pump pressure (up to 0.8 bar for your pump) is set too high this vibration may (just may, nothing certain) disturb the reed switch at just the right frequency and phase to occasionally induce fault E911 (>15 seconds). Taking into account the type of contact anti-bounce software frequently incorporated into control systems the chance of this may be much greater than the pure maths may suggest!.

Here, you may be right, and is consistent with there being nothing fundamentally "wrong" with the old valve, just an unfortunate interaction between a small valve aging process and the Sika switch mechanics. I've called that an "exotic" interaction earlier on.
In my book, that is more of a system design fault.
Not trivially predictable at the design/build/trial-install stage, but a fault nonetheless.
If so, then waving arms and saying the error is because the valve has become "sluggish" doesn't help in the slightest !

It may point to an alternative solution in slightly tightening the return spring tension to eliminate any "resonant" interaction.
Not that I'm going to do that in my current setup, but if the similar valve for CH demand starts tripping an E991 in a similar manner, I may try a bit of spring tweaking before swapping the motor head/replacing the whole unit.

Also , when others find their supremely clean valve being designated as "sluggish", they might find this thread and consider whether they might have strayed into a similar area.

Samsung say that current units use a more elaborate flow gauge which would eliminate such a paddle switch, so this "fault" has probably been addressed - so no intention at dissing Samsung ASHP units.

 

[Johntheo5]

Any metod of determining if the flow through this switch is lower when on DHW duty?

Set point 16.0 ± 1.5 l/min​

 

[Harry Bloomfield]

Any metod of determining if the flow through this switch is lower when on DHW duty?

Set point 16.0 ± 1.5 l/min​

I wonder if there is enough play in the drive mechanism, or spindle, to allow it to signal 'in open position', but with the actual valve not properly opened?

 

[Johntheo5]

Is this a settable setpoint? can it be set as low as 14.5LPM & as high as 17.5LPM, and if so (or not) what is the alarm hysteresis?.

 

[windbagUK]

Any metod of determining if the flow through this switch is lower when on DHW duty?

Set point 16.0 ± 1.5 l/min​

As in the thread, there was NO movement of indicator in the sight gauge when the DHW valve opened or closed

 

[windbagUK]

I wonder if there is enough play in the drive mechanism, or spindle, to allow it to signal 'in open position', but with the actual valve not properly opened?

It wouldn't explain why the E991 trip condition went away with the motorised valve swap-out, as the same Sika switch was put back in.

 

[windbagUK]

Is this a settable setpoint? can it be set as low as 14.5LPM & as high as 17.5LPM, and if so (or not) what is the alarm hysteresis?.

No, not settable. Again, the E991 went away with motorised valve replacement and the same Sika switch back in place.
Alarm Hysteresis probably not even known to Samsung technical.
Besides, "hysteresis" may be determined by firmware interrogation timing of the Sika switch.