Skip to first unread message
seb303
Dec 12, 2016, 2:13:22 PM12/12/16
to loxone-...@googlegroups.com
I'm using the Miniserver to control a heating system, and am PWMing the heating pumps in order to maintain a low level output (for a passive house) and also to make efficient use of the pumps (i.e. obtain sufficient temperature drop from flow to return).
I am also PWMing the actuators (at a much slower rate) to obtain relative outputs in different zones sharing the same pump.
I was initially using a PWM period of 60 seconds for the pumps, but having looked at the Endurance Curve on the datasheet for the Loxone relays, it seems that this would result in quite a short relay life. So I've changed to 120 seconds.
Even so, I am anticipating having to replace these frequently used relays at some point. Has anyone had to do this for a Miniserver or Extension? Are there drop-in SSR replacements for the mechanical relays that Loxone use?
Would have been better IMO if Loxone had just used SSRs in the first place!
I am also PWMing the actuators (at a much slower rate) to obtain relative outputs in different zones sharing the same pump.
I was initially using a PWM period of 60 seconds for the pumps, but having looked at the Endurance Curve on the datasheet for the Loxone relays, it seems that this would result in quite a short relay life. So I've changed to 120 seconds.
Even so, I am anticipating having to replace these frequently used relays at some point. Has anyone had to do this for a Miniserver or Extension? Are there drop-in SSR replacements for the mechanical relays that Loxone use?
Would have been better IMO if Loxone had just used SSRs in the first place!
Andrew B
Dec 12, 2016, 7:33:25 PM12/12/16
to loxone-...@googlegroups.com
SSRs aren't panacea either. I'm facing the same issue, although I'm trying to set my periods much longer. Also have a set of snubbers that I'll get around to installing one day soon.
One thing I'm trying is to use an Elexol IO24 with some of their switch boards. These switch boards use 8 darlington transistors, plus the switch boards are relatively cheap and connected to the IO24 via ribbon cables. I'm using these to drive DIN mounted relays that switch the pumps on and off. Currently I have just a few relays controlled this way, and the remainder are directly connected to Loxone relays. I have another Elexol on the way, and will move to using it entirely (and mounting it in the mechanical room so that I'm no longer constrained by the dozen wires I had originally run between attic and mech room... 1 ethernet cable does all the signalling).
I question your need for high PWM. We've been in our passive house since May, and have found that its temperature is very stable and changes quite slowly. This would seem to argue for a much longer PWM.
seb303
Dec 13, 2016, 4:57:03 AM12/13/16
to Loxone English
On Tuesday, December 13, 2016 at 12:33:25 AM UTC, Andrew B wrote:
I question your need for high PWM. We've been in our passive house since May, and have found that its temperature is very stable and changes quite slowly. This would seem to argue for a much longer PWM.
We have 2 types of heating in the house:
- The ground floor has UFH embedded in the structural slab. This is very slow to respond and the temperature on this floor very stable. The main problem with a very long PWM period for the pump is that the return flow temperature then fluctuates widely which causes the mixer valve to modulate (which is itself a bi-directional stepper motor controlled by 2 Loxone relays).
- The upper floor has Skirting Heating under the large glazed areas. I have found that it feels more comfortable if there is a constant stream of slightly warm air coming from the Skirting Heating (to compensate for the cooler radiant surface of the glazing). Also, this system is quite fast to respond, so a long PWM period tends to create temperature fluctuations.
The other reason for having a relatively short PWM period for the pump is because the actuators are also PWMed as necessary (in order to have relative output between zones). This works better if the PWM periods are quite different. Currently have the pump period at 120s and the actuator period at 1200s.
I appreciate your comment that SSRs aren't a panacea. But I think for heating control, they are far preferable to mechanical relays. The Elexol kit looks interesting (although I'm reluctant to buy more hardware after spending so much on Loxone Extensions). But I don't think Elexol do an SSR output stage, so I'd still need some way to switch the (mains) pumps, actuators and mixer valves. Hence my original thoughts about swapping out Loxone relays for SSRs.
Simon Still
Dec 13, 2016, 6:33:35 AM12/13/16
to Loxone English
It may be a restriction of the control you have over your heat source but could you try a different approach? I'm in a similar situation to you - a newly built near-passive house with UFH in concrete on the ground floor and a more responsive system (overlay UFH under wooden floor) on upper floors.
Ours is controlled by the gas boiler manufacturers Weather Compensation controller.
The pumps run continuously (at a slow speed) during the heating period (4am to 10pm) with a very low flow temperature determined by the outdoor temperature (theres a heating curve - right now it's 11C outside and the concrete floors are running at 21C, the wooden floors at 22C. the air temperature in the first floor room I'm sitting in is 20C. From memory the flow temperatures increase to about 24C/28C when its around 3C outside).
Flow rate between rooms is controlled by the manifolds but using manual valves (ie it's was set on install and left) and there are no room stats. Because the floor temperatures are so low you don't get overheating (the delta between the room and floor drops so heat transfer becomes minimal) and the temperature is consistent between rooms.
The boiler modulates output to control the primary (higher) flow temperature (for the wooden floors) and there a mixing valve to blend down the lower floor temperature for the lower floors.
seb303
Dec 14, 2016, 4:18:19 PM12/14/16
to Loxone English
Hi Simon,
It's an interesting approach, and I'd imagine it works well in a highly insulated house with a large area of radiant heating. You have the constant electricity usage of the pumps, but I guess that's only 30W or so per pump, and ending up as heat energy within the building anyway.
My heat source is somewhat different: it's a 2000L thermal store, heated by biomass (logs) and solar thermal. In the winter it's almost all biomass because of lack of sun, and shading by the hill we live on.
So my aim is to try and get as long between lighting the boiler as possible, and to use the energy in the store as completely as possible between burns (because there is significant energy loss per burn due to the 90L capacity of the biomass boiler which is situated out in the cold).
The UFH downstairs is not such a problem, but the Skirting Heating upstairs needs quite a high flow temperature to work properly - at least 40C, and more on a cold day. The flow temperature is achieved by mixing water from high up the buffer tank with the return. The excess return water is then fed back lower in the buffer tank. So the lower the return temperature, the better the buffer tank stays stratified and the longer it contains useful heat nearer the top.
Unfortunately I purchased the system just before the new European regulations came in requiring variable speed pumps, so it has standard 3 speed pumps and the lowest speed is still much faster than needed to get a decent temperature drop.
I'm experimenting with a pump PWM period of 300s at the moment, but it's tricky to keep the flow temperature stable without frequent adjustment of the mixer valve (a different Loxone relay to wear out!)
So I may have to go back to something like 120s and just accept that I'll have to replace a couple of broken relay with an SSR at some point.
It's an interesting approach, and I'd imagine it works well in a highly insulated house with a large area of radiant heating. You have the constant electricity usage of the pumps, but I guess that's only 30W or so per pump, and ending up as heat energy within the building anyway.
My heat source is somewhat different: it's a 2000L thermal store, heated by biomass (logs) and solar thermal. In the winter it's almost all biomass because of lack of sun, and shading by the hill we live on.
So my aim is to try and get as long between lighting the boiler as possible, and to use the energy in the store as completely as possible between burns (because there is significant energy loss per burn due to the 90L capacity of the biomass boiler which is situated out in the cold).
The UFH downstairs is not such a problem, but the Skirting Heating upstairs needs quite a high flow temperature to work properly - at least 40C, and more on a cold day. The flow temperature is achieved by mixing water from high up the buffer tank with the return. The excess return water is then fed back lower in the buffer tank. So the lower the return temperature, the better the buffer tank stays stratified and the longer it contains useful heat nearer the top.
Unfortunately I purchased the system just before the new European regulations came in requiring variable speed pumps, so it has standard 3 speed pumps and the lowest speed is still much faster than needed to get a decent temperature drop.
I'm experimenting with a pump PWM period of 300s at the moment, but it's tricky to keep the flow temperature stable without frequent adjustment of the mixer valve (a different Loxone relay to wear out!)
So I may have to go back to something like 120s and just accept that I'll have to replace a couple of broken relay with an SSR at some point.
seb303
Dec 14, 2016, 6:48:31 PM12/14/16
to Loxone English
Having thought a bit more about what you said Simon, I am going to try the approach of modulating the flow temperature right down once the room temperature comes very close to target. It may be that this will provide a better solution than PWMing the pump.
Currently the flow temperature is determined broadly by the exterior temperature with some extra boost if the average room temperature happens to be well below target (which doesn't really happen).
Since it's all controlled by the Miniserver, anything's possible - although I've given up trying to do it with the standard function blocks and am now using a custom program. The standard function blocks are probably fine for a more conventional heating system, but I found they lacked the ability to really optimise.
Currently the flow temperature is determined broadly by the exterior temperature with some extra boost if the average room temperature happens to be well below target (which doesn't really happen).
Since it's all controlled by the Miniserver, anything's possible - although I've given up trying to do it with the standard function blocks and am now using a custom program. The standard function blocks are probably fine for a more conventional heating system, but I found they lacked the ability to really optimise.
Andrew B
Dec 14, 2016, 9:07:31 PM12/14/16
to Loxone English
I think you'll find this a worthwhile experiment. And I'm glad to hear someone else say that they think custom programs are important. Please submit a ticket with loxone to improve their custom program support. Right now it is piss poor, yet despite that I have about 2000 lines of picoC running in four programs... and I want to do a lot more. The functional block programming just doesn't scale and doesn't give enough flexibility. If they would provide s proper compiler and debugger, they could make the miniserver dramatically more powerful.
seb303
Dec 15, 2016, 4:51:08 AM12/15/16
to Loxone English
I have submitted a couple of tickets to Loxone about bugs in the custom program functionality. The impression I get is that they are not willing to put much resources into developing this feature. It's a shame as I find over and over that the large macro function blocks do not exactly fit my needs, whether it be for lighting, heating, etc.
And once one starts adding a load of auxiliary function blocks to bend the functionality as desired, then it becomes cumbersome to replicate for each room, etc.
Personally I would do much more with custom programs if it was more practical. The biggest pain for me is having to reload the whole miniserver every time I make a small code change. It would be a big step forward if the custom programs could at least be simulated within the config software. Debug visibility into the code would also be a great help.
I am also hitting the limit of 7 custom programs in my config (yes it is 7 not 8 - I have a ticket opened about this).
A compiler would be nice too, although I haven't had a problem with my custom programs using too much CPU - probably because I'm not doing much real-time processing of incoming events.
And once one starts adding a load of auxiliary function blocks to bend the functionality as desired, then it becomes cumbersome to replicate for each room, etc.
Personally I would do much more with custom programs if it was more practical. The biggest pain for me is having to reload the whole miniserver every time I make a small code change. It would be a big step forward if the custom programs could at least be simulated within the config software. Debug visibility into the code would also be a great help.
I am also hitting the limit of 7 custom programs in my config (yes it is 7 not 8 - I have a ticket opened about this).
A compiler would be nice too, although I haven't had a problem with my custom programs using too much CPU - probably because I'm not doing much real-time processing of incoming events.
Simon Still
Dec 15, 2016, 7:05:03 AM12/15/16
to Loxone English
On Wednesday, 14 December 2016 21:18:19 UTC, seb303 wrote:
Unfortunately I purchased the system just before the new European regulations came in requiring variable speed pumps, so it has standard 3 speed pumps and the lowest speed is still much faster than needed to get a decent temperature drop.
I'm experimenting with a pump PWM period of 300s at the moment, but it's tricky to keep the flow temperature stable without frequent adjustment of the mixer valve (a different Loxone relay to wear out!)
So I may have to go back to something like 120s and just accept that I'll have to replace a couple of broken relay with an SSR at some point.
What's the cost of replacing the pumps? It is going to be less hassle than the relays?
Could you 'off board' the relays you think are going to wear out - use analogue outputs from Loxone to drive an off board SSR relays?
Andrew Brownsword
Dec 15, 2016, 11:59:55 AM12/15/16
to Loxone English
The couple of times I had my miniserver open, it appeared that its on-board relays are soldered to the motherboard, not socketed. This is one of the reasons I have been using to a model which does not place as much wear on those relays (don’t switch big loads with them, and reduced switching frequency). I am further moving away from the built-in relays by adopting a more Elexol-centric approach (another IO24) — the Loxone controls those via Ethernet. The Elexol offers either darlington transistors (which I’m using) or their own relay boards, and either way they are on low-cost breakout boards. I’m using the transistors to control DIN rail mounted relays from Phoenix Contact — nice push-in connections, DPDT options, wide variety of coil and power side types, and trivial & cheap to replace the relay itself.
--
You received this message because you are subscribed to a topic in the Google Groups "Loxone English" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/loxone-english/Z5SsUSrx34Y/unsubscribe.
To unsubscribe from this group and all its topics, send an email to loxone-englis...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/loxone-english/5bc54a10-f1d7-48ba-be2a-8261e07eda4c%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
seb303
Dec 15, 2016, 5:04:52 PM12/15/16
to Loxone English
All the heating controls are running off 2 Extensions rather than the Miniserver iteslf, so perhaps a less dense board to de-solder from. I've got the tools & skills to de-solder the relays if necessary, but yes I agree, perhaps better to run some off-board SSRs. I have plenty of analogue outputs free.
Replacing the pumps is also a bit of hassle, what with the inevitable water/anti-freeze loss and potentially having to re-pressurise the system slightly (which needs a pump since our spring-fed water pressure is not high enough). I don't think the pumps are that expensive - probably less than £100. I haven't checked to see how slow a variable speed pump can run, but I guess it's pretty slow.
I'm first going to try modulating the flow temperature right down instead of PWMing the pump and see how that works.
Replacing the pumps is also a bit of hassle, what with the inevitable water/anti-freeze loss and potentially having to re-pressurise the system slightly (which needs a pump since our spring-fed water pressure is not high enough). I don't think the pumps are that expensive - probably less than £100. I haven't checked to see how slow a variable speed pump can run, but I guess it's pretty slow.
I'm first going to try modulating the flow temperature right down instead of PWMing the pump and see how that works.
Andrew B
Dec 15, 2016, 7:11:09 PM12/15/16
to Loxone English
Nonetheless, I encourage everyone who has had to do picoC development to submit their own tickets asking for improved "custom program" functionality. High on the list should be a compiler and a debugger. The more of us, the more they are likely to pay attention.
seb303
Dec 17, 2016, 5:29:55 PM12/17/16
to Loxone English
I've played around with modulating the flow temperature downwards to decrease the heating output as the target temperature gets closer and found the following:
Skirting Heating (rapid response, relatively high flow temperature, convection)
It works very well for Skirting Heating, and allows the room to be held at temperature, with a very low level of hot air being released over the colder surfaces (i.e. large windows).
I had to improve my control algorithm for the mixer valve (a stepper motor with 2 inputs that are pulsed for hotter or cooler), so that it settles quickly when the required flow temperature changes (e.g. when someone opens a door and lets the cold in) - otherwise I would just be wearing different relays! In any case, the power of the stepper motor is much smaller than a heating pump, so the relays should have a higher endurance.
Underfloor Heating (slow response, low flow temperature, radiant)
For Underfloor Heating, it didn't work at all. The UFH pipes are embedded deep in the 100mm structural slab of the house. What I found was that when the flow temperature was reduced, the mixer valve ended up at 0%, and the water went round and round with virtually no temperature drop - due to the very large thermal mass of all the warm concrete surrounding the pipes.
So I've ended up going back to PWMing the pump for the UFH. I'm using a PWM period of 30mins for the pump, and 2 hours for the actuators. This should minimise the wear on the relay contacts, and is perfectly fine for what is a very slow response system.
I'm just glad I've got PicoC to do this with as it couldn't be done easily (if at all) with the standard Loxone function blocks.
Skirting Heating (rapid response, relatively high flow temperature, convection)
It works very well for Skirting Heating, and allows the room to be held at temperature, with a very low level of hot air being released over the colder surfaces (i.e. large windows).
I had to improve my control algorithm for the mixer valve (a stepper motor with 2 inputs that are pulsed for hotter or cooler), so that it settles quickly when the required flow temperature changes (e.g. when someone opens a door and lets the cold in) - otherwise I would just be wearing different relays! In any case, the power of the stepper motor is much smaller than a heating pump, so the relays should have a higher endurance.
Underfloor Heating (slow response, low flow temperature, radiant)
For Underfloor Heating, it didn't work at all. The UFH pipes are embedded deep in the 100mm structural slab of the house. What I found was that when the flow temperature was reduced, the mixer valve ended up at 0%, and the water went round and round with virtually no temperature drop - due to the very large thermal mass of all the warm concrete surrounding the pipes.
So I've ended up going back to PWMing the pump for the UFH. I'm using a PWM period of 30mins for the pump, and 2 hours for the actuators. This should minimise the wear on the relay contacts, and is perfectly fine for what is a very slow response system.
I'm just glad I've got PicoC to do this with as it couldn't be done easily (if at all) with the standard Loxone function blocks.
Reply all
Reply to author
Forward
0 new messages