Thank you so much for all that write up, brilliant collection of info for research, sizing own home and validating installer proposals.
Question on heat pump power usage, 500w is impressively low. I know you have put the Heatgeek example as 5 kW, but even during their highest heat demand days since their record began, I see less than 2 kW power draw. Is this their typical power draw on a typical days? The higher rating is just head-room.
I’m looking at home battery first, so just trying to size the right inverter. 5 kW should do the trick if heat pump typically pulls no more than 2 kW and we are careful with high power appliances.
I raised the question of how to control flow rates on Vaillant ASHP at Introducing HeatpumpMonitor.org - a public dashboard of heat pump performance - #28 by Zarch which @Zarch suggested was more approriate here. Mick pointed out the relevant installer level configuration settings of the VWZ AI controller at:
Conf. heat. build. pump
Conf. cool. build. pump
Conf. DHW. build. pump
On my system this controller is built into the uniTower, rather than being a separate unit, but settings are identical. Values for the settings are “Auto” or manual from 50-100%PWM: %PWM is % Pulse Width Modulation which is the method used to control the pump motor voltage (hence power). Defaults on the uniTower are “Auto” for heat and cool but 65% for DHW.
At the factory, the nominal flow is automatically achieved by volume flow regulation. This volume flow regulation allows for efficient operation of the building circuit pump because the pump speed is adjusted to the hydraulic resistance of thesystem. Vaillant recommends that you retain this setting.
So the reason for my wondering about this was that my 3.5kW Arotherm+ is installed with a secondary pump in series on the flow of the external unit. Manual monitoring of the rates shows that during heating cycles the flow rate is at the maximum 860l/hr that the 3.5kW boiler plate specifications allow (currently target FlowTemp=32C). During DHW, the flow rates are 1186l/m (target 50+15=65C) which is way beyond the maximum flow rate specified. As Conf. DHW. build. pump is 100%PWM on my system it seems like there may be need for some playing around!
Hi Steve, I’ve been in deep discussion with Vaillant on this. There is some suggestion that they will invalidate warranties if you follow this route, but I suspect that monitor only with the ebus3 would be OK. I’ve asked if they would support anything done by the Open Energy Monitor community to develop something for ebus, but have had no response.
As an aside, was the default ebusd configuration respository sufficient, or did you have to produce custom config files for your system?
Currently looking at eebus (not ebus) which is suported on the more recent Vaillant gateways. This is an open specification and, in theory, should allow access to, control of and optimisation of the ASHP parameters via another eebus device/control centre. Will chase further once over COVID, but probably beyond my capabilities!
Before we start, we have to remind ourselves of the following
Heat Output = Specific heat capacity x flow rate (l/s) x (flow temperature - return temperature)
Specific heat capacity (SHC) of water is 4.2 and it’s 3.8 for Gylcol.
Flow Rate conversion m3/h to l/s = (m3h * 0.277)
Flow Rate conversion: l/hr to l/s = (l/hr / 3600)
So 860l/hr = 860 / 3600 = 0.238 l/s
3.8 x 0.238 x 5DT = 4.5kW
No if we go back to our ‘gold’ tech document, whilst it doesn’t list the 3.5kW model, it lists a 3kW one.
Which suggests that even that model can kick out between 4kW and 6kW of heat depending on outside temp and flow temp combinations.
During my hot water run, i’ve got my flow rate fixed at 550l/hr and I still output almost 8kW of heat cos my DT is around 10 on hot water runs.
VWL 35/6 (ASHP) and VIH QW 190/6 (uniTower) are the units. The ASHP has rated output of 3.4kW in UK spec sheets (as in Czech), and you are right it is the 3kW output curves in your golden find. My sloppyness stems from the 35 in the device name: sorry.
I mirrored your calcs yesterday but included density as it is not 1g/cc for a glycol mix, and got same (ish) results. Currently monitoring dT is a PITA as I only have the installer test outputs to go by. There is a limit to how often I’m going to push buttons during a heating cycle! At the end of a heating cycle dT is about 3C. I have ordered EmonTx4 which I hope will turn up soon so I can get the temperature monitoring sorted. It really bugs me that Vaillant don’t allow access to all this data via their access point!
I’m struggling to answer re the additional pump: I knew nothing about ASHP when the installer put it in and he was an experienced Mitsubihsi installer with all the right paperwork. However, this was the first Vaillant he had done (and it showed). I have been trying to learn about heat pump optimisation since then because I have never been happy with the install. Initial wiring and controller setup issues were sorted out by getting an accredited Vaillant ASHP installer in, but there are plenty more issues to deal with. System noise has always been one such issue, due to excessive flow rates through the emitters: it seems worse this heating season and, given the need to spend time at home with COVID, I’ve been digging further.
I don’t think there is any need for the second pump in order to get the flow rates @ dT5 I calculated to be necessary to meet the design loads in individual rooms. All loads are distributed via 20mm MCLP from a 10 way manifold which is 1m away from the internal unit (28mm copper). My ability to determine the total head loss across that lot is not there (yet?). However, at 860l/hr rated maximum, the ASHP pump has residual head of 5.7m, which the uniTower reduces by 0.6m in heating mode. This leaves a bunch of head for the rest of the system!
So I’m guessing that the installer played safe.
Once I can monitor the dT I am going to test with and without the additional pump. As it is in the flow, it may cause a big head loss if the impeller is not powered. If not needed, there is a great opportunity to replace with an 1-1/4" heat meter I guess!
Yeah, all the pipework sizing sounds more than adequate for the heat pump, which is a good sign.
I’ve played a lot with my install since it was put in, most of the changes driven by monitoring data.
I know a heat meter and associated kit is expensive, but i’m not sure how you could determine things are all working without it.
Your data is fascinating, for me at least! The power curve is all but identical in shape, cycle duration and cycle frequency to what I see. Not really what I would have expected in two different installs with differently rated units. My design load is only 1500w at -3C external. Do you know yours?
Is this the “default” 65%PWM pump setting? Looks very close. Will be interested to see what dT on the DHW cycle is here with 1186l/hr.
You can read all about my design and how we sized it here:
and a shorter summary here:
I will be very interested to see the heat meter trace from the 3.5kW unit. Just to see ‘how low’ it can go, in terms of electric in and heat out.
The 3.5kW is the smallest unit on the market I think. but I have heard it’s just a downrated 5kW unit. The ‘golden’ doc suggests there isn’t much difference in the ‘floor’ of both 3kW and 5kW units. It’s more than that the smaller unit can’t reach as high.
Yeah, it’s on 65%
I had to leave it here as I had issues with my heat meter when I ramped the flow rate up.
I have at last worked out what controls the Vaillant heat pump cycle timing; this may have been obvious to all but, just in case, here is my understanding in the hope it might help others. Thanks for contributors on What is Energy Integral - Air Source Heat Pumps (ASHP) - BuildHub.org.uk for the penny dropping moment, which I followed up with an hour watching the energy integral in Live Monitor on the Vaillant lcd and the pump power draw feed in Emoncms.
On the VWL35/6 at least, this is what happens:
Compressor starts at Energy integral of -60degmins (default setting, can be changed).
Every 30 seconds the difference:
(Flow temp) - (calculated target flow temp)
is calculated and added to the previous degmin value. So, if flow temp is 2C lower than target, -60degmin changes to -62degmin after 30secs.
The degmin value continues to get smaller (more negative) at an ever decreasing rate until the flow temp exceeds the target temp. At this point, the compressor modulates back to 25% (which appears a hard coded minimum) and, because the sign of the temperature difference has changed, the energy integral begins to increase.
Once the energy integral reaches 0degmin, the compressor goes into standby mode, with the energy integral continuing to drop every 30 secs.
Once the energy integral reaches -60degmin, the cycle starts again.
So the absolute value of the increment to the energy integral is clearly dependent on the balance between the rate of heat energy added by the compressor and the rate of heat loss in the emitter system. The balance of heat loss versus the modulated heat input controls how long the “on cycle” lasts, while the heat loss alone controls the length of the off cycle. Intuitively obvious once written!
Looks like the “heating start from” value may be an interesting tuning parameter, but that is a thought to follow up once my monitoring is more automated!
Hi Andrew, I was using stock ebusd config files … which are auto loaded as ebusd discovers devices on the bus. A few weeks ago I decided to try and get the necessary feeds working to populate the emon ASHP application (I needed a live Yield value) and as a result now have the config files locally (using a modified GitHub branch from mwildbolz) .
Good luck with Vaillant - but I suspect the reason ebusd exists is because they wouldn’t make the data accessible