MVHR Monitoring and Control

[MJNewton]

I did wonder if this should be in the MVHR section but actually it's more about the technology side of things so figured Boffin's Corner might be best place... The reason for posting is that I'm always really interested in reading about what other people have done with all things home automation, electronics, DIY etc and figured that the very people behind all that might well be interested in what I've been up to too. I've also got a bunch of graphs which for some reason have always appealed to me and I'm sure I can't be alone in that! So here goes!

 

The follow graphic illustrates what it's all about, and that is essentially the 'command and control' of our Titon MVHR unit. Titon do sell an LCD touchscreen controller for it (although I'm not entirely sure if it is compatible with our particular unit) but it is pretty expensive and I figured I could do more for less, and have more fun in the process!

 

 

It is based around the following components:

 

• Raspberry Pi Zero W - This is the 'heart' of the control system and monitors temperatures (using DS18B20 1-wire temperature sensors) in all the ducts (supply, extract, intake and exhaust) as well as humidity in the extract (using an AM2302 humidity sensor) and (soon) air quality in the supply (using a Plantower PMS5003 particulate sensor). As I will elaborate further below, with all this information it can automatically boost on humidity rises, cut back or turn off the ventilation entirely if outside air quality suffers, report heat exchanger efficiency etc and likely more things in the future too (I've been thinking about putting summer bypass under its control and perhaps complete fan control too). It also creates the graphs and serves them up for viewing in a web browser.
• Shelly Uni Smart Relay - Used as the interface between physical boost switches in the house and the MVHR unit so as to increases their utility in terms of allowing different actions for short vs long presses, but more importantly it also provides an easy way for boost and setback to be triggered from other triggers whether that be from the Pi, widget buttons on our smartphones and voice control via Alexa.
• Shelly 1PM Smart Relay - Used to monitor the power consumption of the MVHR as it was something I was interested in understanding even before fitting the system, and it used to also serve as a means by which to disable the ventilation in response to a fire but I have since decided I don't actually like that and would rather the ventilation kept working. It is still connected to the smoke alarm circuit though and will alert me via app/SMS if the alarms do trigger.
• Raspbery Pi Pico W - Acting as a 'mobile' means to interface with a PMS5003 air quality sensor whilst I am still playing around with it, but when the sensor goes in the ducting I'll read it directly from the Pi Zero from then on.

 

So breaking down some of the functions further...

 

Heat Exchanger Efficiency
 

With knowledge of the supply, extract, intake and exhaust temperatures you can calculate heat exchanger efficiency in real time. The Paul Heat Recovery page on the subject describes two methods in more detail and the following graph illustrates what are fairly typical efficiencies figures (high-80/low-90%) I see from the unit:

 

 

Automatic Humidity-Related Boost
 

The Titon does have its own humidity sensor with an adjustable (but unmarked) threshold and whilst I've never actually tried it I've always assumed it would be susceptible to false positives (triggering when it shouldn't) or negatives (not triggering when it should) and perhaps more to the point was always going to be a bit of a pain to fine tune (and adjust seasonally) given that the unit is in the loft. So, given I am monitoring humidity (in the extract manifold) anyway if we look at a typical graph it is easy to see when boost needs to be triggered:

 

 

Those peaks correspond to showering and cooking (e.g. mostly the hob) and as I said are really easy to spot with a rate-of-rise calculation. The Pi takes measurements every 5 minutes and as a starter-for-ten I set it to trigger the boost (via an API call to the Shelly Uni) if the humidity rose by more than 5% within the 5-minute window. This turns out to have been a perfect threshold as it has never knowingly missed a shower (or pan of boiling water) or triggered when it shouldn't have in the several years it's been running now so I've not cause to make any changes.

 

Automatic Ventilation Setback
 

This is the most recent addition to the control system, and is still work in progress but as it's been so interesting to me I thought I'd share the not-quite-finished product...

 

Our neighbour (who thankfully we get in really well with, but even friends can be annoying!) bought a wood burner for their recently-built extension. We've been mindful of what we might be breathing in from it as it seems barely a day goes by when there isn't some article or other written on the subject. Indeed, on occasion over the winter we've seen smoke coming from it but more to the point we've been able to smell it through the MVHR too. Granted, this is mostly when I've seen the smoke and have jumped up on a chair to sniff the incoming air from one of the vents and it seems the planets have to be in alignment for this to occur e.g. a cold still day with no wind, the neighbours using damper-than-ideal logs etc but it's happened a two or three times now and whilst I've been convincing my wife how sensitive the human nose is and that we shouldn't assume we're being poisoned I did begin to really want to quantify how 'bad' the situation really was. Aside from anything else it was a good excuse to buy another sensor, produce more graphs and give another aspect of control. Like others whilst we normally run the MVHR 24/7 and enjoy the fresh air it provides but this of course is only effective if the outside air is indeed fresh so my goal is to detect when it isn't fresh and either turn the ventilation down (whether that be both supply and extract, or perhaps just supply I don't know yet) or perhaps even turn it off temporarily when such a situation arises.

 

The PMS5003 particulate sensor is a delightful little thing - the size of a small box of matches and costs around £15. You wouldn't expect much for so little but I've read several accounts now of people having compared them with air quality sensors costing thousands and them being a reasonable enough match, certainly good enough for taking measurements that are meaningful, consistent and reliable enough to feed into a home ventilation control system. The sensor is based around a laser whose light is reflected by the presence of particulates in its path from which the sensor can determine particle quantities in the 0.3-10μm diameter range which is exactly the region of interest when determining common particulate matter based analysis and determination of air quality. The following graphics illustrate the what the sensor can detect:

 

 

The sensor does everything for you in terms of detecting the particulates and then presenting summaries of particle count and mass of different particle sizes over a serial line that I am reading with a bit of Python code running on a Raspberry Pi Pico W (currently powered by a USB powerbank for mobility):

 

 

The Pico is configured to continuously read from the sensor and gets an update roughly once per second and, via HTTP through its wifi interface, I can request and parse these readings from a script running on the Pi Zero:

 

 

I capture the readings once per minute which are then stored and can be graphed to show particle count of any specific detectable size but I am finding the 'all detectable sizes' (>0.3μm) summary useful enough:

 

 

The particle mass figures are even more useful as they can be compared against Air Quality Index charts such as the UK's Daily Air Quality Index (DAQI) 10-level bandings shown below:

 

 

Now technically the DAQI levels, like most AQIs, apply to 24hr mean readings thus non-continuous spikes as seen above don't really up the level - it's the 24hr average (as shown in the table) that really counts if wanting to compare against the index.

 

The neighbours haven't been using their woodburner over the past week or so as far as I can tell and so it would've been rather boring to put the sensor in the supply manifold as ultimately intended so instead it's been sitting in our open-plan lounge/kitchen/diner and it's been quite an eye opener! As the following graph illustrates, cooking really does produce a lot of particulate matter:

 

 

I think the first spike was preparing some halloumi burgers on the hob, the second the bread machine in action and the third some sausage rolls in the oven.

 

What was really quite alarming was when my wife burnt some sesame oil in a frying pan... I was out at the time so have only got her word but she insists it wasn't that smokey (and to be fair when I got in later I could smell cooking but didn't assume anything more than that) and yet the particle count graph looked more like a nuclear bomb had gone off:

 

 

The spike doesn't look all that significant until you look at the y-axis scale - a peak of over 3000 particles/0.1L which is over ten times the peak we'd normally see when cooking! It was interesting to see the PM1, PM2.5 and PM10 counts against the DAQI for that episode too:

 

 

Whilst as mentioned it doesn't mean all that much shown like that given the DAQI is meant to be used with a 24hr mean (thus 13μg/m3 of PM2.5 in this case, which is still only Level 2 'LOW' on the scale and not the 195μg/m3, or Level 10 'VERY HIGH' as it might otherwise appear presented like this!) but it does make for an interesting perspective nevertheless. Some might say it also serves as a quantifiable measure of my wife's cooking skills but obviously I wouldn't...

 

So that's where I've got to really! Next steps are to buy another air quality sensor as it seems a shame to lock this one away in the MVHR ducting as there are too many experiments begging to be done - lighting matches, candles, perhaps the barbecue and maybe driving around with it too (the sensor, not the barbecue) to see what the air is like on a commute or in a city etc. Also with two sensors I can have one inside the house to automatically boost when required, and one in the supply duct to do the opposite and turn the ventilation down/off temporarily if the outside air quality is lower than desired.

 

I've written more than I planned to so well done - and thanks - if you've got this far. Hope it was of some interest, and if there are any questions or any way I can help you do something similar I'd be happy to oblige.

 

[dpmiller]

On the fire/ smoke side of things, will the Aicos all be wired interlinked or will you be having any on wireless? Do you have any need for CO monitoring? gas hob/ woodburner/ whatever?

 

Whilst we're on wireless, I've used an Ei414 fire alarm interface for two functions-

MVHR off during fire alerts

MVHR boost during CO alerts.

[Dave Jones]

very intersting, we are in the process of planning our own mvhr and your post has answered many questions.

 

We are looking at the Zehnder Q350, UV and hepa filtration with Comfopost CW6 air chiller for ASHP cold feed in the summer.

 

Did you look at alternatives before choosing the titon unit ? Any pro's cons ?

 

 

[JohnMo]

49 minutes ago, dpmiller said:

MVHR boost during CO alerts

If you are getting this type of alert you should be getting out of the house ASAP. Then figuring why you are getting high CO.

 

Your MVHR will not save you

 

[SteamyTea]

Looks pretty comprehensive.

The AM3202 is not known for reliability and accuracy, but I have used the 'sister' DHT22 with no issues.  I even calibrated a few with the salt brine in a jam jar method and found them to be pretty good, unlike the DHT11 (the blue ones) that are very poor.

9 hours ago, MJNewton said:

Raspberry Pi Pico W (currently powered by a USB powerbank for mobility)

How is that working out? My old powerbank was not delivering a high enough voltage, but it was very old and very, very cheap.

[dpmiller]

@JohnMo I never said anything about staying in the house...

[JohnMo]

3 minutes ago, dpmiller said:

never said anything about staying in the house...

Why automate the MVHR then...

[dpmiller]

why not tho? Seemed like a no-brainer as the wiring was going past a zero-volt relay contact...

 

/if it does ever go off at least I know the emptied property is being purged.

[MJNewton]

1 hour ago, Dave Jones said:

We are looking at the Zehnder Q350, UV and hepa filtration with Comfopost CW6 air chiller for ASHP cold feed in the summer.

 

Did you look at alternatives before choosing the titon unit ? Any pro's cons ?

 

I actually bought the unit off eBay and picked it up locally - it had been fitted as part of a development of houses but the wrong units had been installed so they were all removed and replaced with the next size up. Whilst they'd been commissioned they'd never seen actual use. I suspect this opportunity meant a short cut for my usually rather protracted research, planning and selection process but I've no regrets - it seems a very capable unit. Sure, limited in terms of fancy controls but it seems to work well, is quiet, has good quality Papst fans and Titon's technical support line have been great when queried about nuances of operation (e.g. I wanted to know the exact points at which the summer bypass kicked in and they were happy to elaborate by email). It seems I'm not alone in going down the new/nearly-new route through the likes of eBay and I'd do that again too.

[MJNewton]

1 hour ago, dpmiller said:

On the fire/ smoke side of things, will the Aicos all be wired interlinked or will you be having any on wireless? Do you have any need for CO monitoring? gas hob/ woodburner/ whatever?

 

They all happen to be wired, and with this being a retrofit all there was left for me to do was add a detector in the loft and the relay base.

 

Re CO we've only got a gas boiler and that's room-sealed so I figured the risk was fairly low. Not zero of course though so we do have a battery CO alarm collocated with it. Truth be told if the Aico CO alarms were cheaper (they're around £50 or so) I'd probably add one in because, well, why not?

 

Quote

MVHR off during fire alerts

 

As mentioned I ended up changing my stance on that one. Inspired, I think, by the fact that some other MVHR units have dedicated smoke/fire alarm inputs I originally figured it worthwhile emulating that function with my setup but recently I began to question exactly why. It's not like the MVHR will be pumping in loads of oxygen and fueling the fire in any meaningful way (or I assume it won't?) and so if anything I think I'd prefer the potential smoke clearing benefits from leaving the unit on throughout. Even if the fire goes out and we're in a state of panic outside there's surely some benefit in the MVHR continuing to clear the house.

 

The alarm interface to the smart relay is what I really appreciate as it does give me some peace of mind when away that the house likely isn't burning down - that's the sort of thing that casually wanders through my mind whilst sipping magarita's by the pool! Of course, completely different story should I ever get that alert!!!

[MJNewton]

50 minutes ago, SteamyTea said:

Looks pretty comprehensive.The AM3202 is not known for reliability and accuracy, but I have used the 'sister' DHT22 with no issues. 

 

Are they not one and the same thing?

 

Quote

How is that working out? My old powerbank was not delivering a high enough voltage, but it was very old and very, very cheap.

 

It seems to be working fine. It wasn't originally part of the plan but I had an issue with my phone and when the supplier sorted it out they also sent me a powerbank as a goodwill gesture. It just happened to arrive on the day I was began playing with all this stuff and so immediately allowed me to move the sensor around more easily.

 

The Pico does apparently cope with supply voltages down to 1.8v I believe which must surely help, but the sensor is a bit more sensitive and requires a minimum of 4.5v according to the datasheet.

[SteamyTea]

17 minutes ago, MJNewton said:

Are they not one and the same thing?

Yes they are.

I think the DHT11, the blue ones, got the range a bad name.

18 minutes ago, MJNewton said:

The Pico does apparently cope with supply voltages down to 1.8v

That is useful to know.  I wonder that the ESP8266, which I have a few of, works down to reliably.

[Radian]

13 hours ago, MJNewton said:

What was really quite alarming was when my wife burnt some sesame oil in a frying pan...

Reminds me of our over-hot-oven roast beef fiasco  (mmmm beef fiasco)

I have a nasty suspicion that breathing in the smoke that evening did all kinds of mischief to my body. The following day I had lots of pain in my lower back and it's been moving around my joints ever since. Left knee today, shoulder yesterday kind of thing. When you see the WHO reports on the dangers of air polution and the effects of PM2.5 it's quite believable that this might have been responsible. 

[SteamyTea]

39 minutes ago, Radian said:

beef fiasco

Beef and Dairy Network

 

[MJNewton]

1 hour ago, Radian said:

Reminds me of our over-hot-oven roast beef fiasco  (mmmm beef fiasco)

I have a nasty suspicion that breathing in the smoke that evening did all kinds of mischief to my body. The following day I had lots of pain in my lower back and it's been moving around my joints ever since. Left knee today, shoulder yesterday kind of thing. When you see the WHO reports on the dangers of air polution and the effects of PM2.5 it's quite believable that this might have been responsible. 

 

Thanks - just had a read - and yes, it has made me think about more about the whole air quality issue. I am also starting to think that so many articles seem to focus on what everyone else is doing that's causing you harm, and not perhaps what you might be doing yourself that might be just as bad if not worse. To be fair, that's exactly what I did - concerned almost entirely with what the neighbour's wood burner was doing to us, but now I am as much interested in what we're doing to ourselves! Just hope it doesn't become an obsession*! 😂

 

*Edit: Non-paywall version in case the link doesn't work for you

[Radian]

1 hour ago, SteamyTea said:

Beef and Dairy Network

 

Just watched the first Taskmaster with Mike Wozniak as a competitor. Funny guy.

[jack]

21 minutes ago, Radian said:

Just watched the first Taskmaster with Mike Wozniak as a competitor. Funny guy.

I'd never heard of him before, but he makes that series imo.

[gregh]

On 03/03/2023 at 13:06, MJNewton said:

 

Thanks - just had a read - and yes, it has made me think about more about the whole air quality issue. I am also starting to think that so many articles seem to focus on what everyone else is doing that's causing you harm, and not perhaps what you might be doing yourself that might be just as bad if not worse. To be fair, that's exactly what I did - concerned almost entirely with what the neighbour's wood burner was doing to us, but now I am as much interested in what we're doing to ourselves! Just hope it doesn't become an obsession*! 😂

 

*Edit: Non-paywall version in case the link doesn't work for you

Please do post what you find with the sensor when the woodburner is going. Looking to fit MVHR as part of a reno project and we seem to be surrounded by neighbours who love to burn rubbish in their gardens...

[MJNewton]

19 minutes ago, gregh said:

Please do post what you find with the sensor when the woodburner is going. Looking to fit MVHR as part of a reno project and we seem to be surrounded by neighbours who love to burn rubbish in their gardens...

 

Will do. As mentioned I decided to keep the original air sensor for permanent use inside the house so I ordered another from a different seller on Aliexpress. Unfortunately they sent the wrong one and, eventually, ended up refunding me. So I had to order another - in fact bought a couple this time - and they arrived today so will try and fit one at the weekend. Of course, this will trigger an acceleration into spring and that'll be the end of the wood burner for another 8 months so I might have to wait a while to see how effective it is! I am fairly confident though now that I've seen how sensitive these things are, and whilst waiting for the sensor I've rigged up a relay to be able to trick the MVHR to go into frost mode (basically by swapping in a fixed-value resistor instead of the thermistor to make it think it is -10C outside) which will cut the supply but keep the extract going.

[Ajn]

Very interesting section on your findings

On 02/03/2023 at 22:41, MJNewton said:

What was really quite alarming was when my wife burnt some sesame oil in a frying pan

with the use of the particulate sensor, I think I have just found a new project to add to our MVHR system.

Our MVHR (Domus HRXD) has been in just over a full year now, through two winters. I have a little issue with it's operation around temps below -1 and that the controller is a bag of poo The clock needs to be reset every time there is even a minor power failure. I have set the controller to a background mode so if my added on system or more likely me myself fails then the other half can use the original system and not my toys as she calls them.

I am a fan of the d1 mini micro controller, node-red, mqtt and the Raspberry Pi. So my system is built around these. The bathrooms and kitchen have d1's with humidity, temp, PIR and light sensors. There is a d1 in the MVHR control box that has a temp sensor in each of the four MHRV pipes (as well as temp sensors on all underfloor heating returns. It also looks after the co2 sensor that is in the room extract pipe. This same d1 also switches the MVHR fan speeds and inhibit function.

 

The control comes from a Pi3 running mqtt and node red. mqtt is the message hub for all the sensors while node-red provides all the control functions. For example, boosting the MVHR if the humidity exceeds a limit in any room, introduces a mild timed boost should a bathroom be used for a period but humidity remains the same, reminds us if the small bathroom light has been left on. Switches the MVHR to bypass mode dependent on the outside temperature, the internal temperature, time of year and time of day. The co2 monitor allows the MVHR to run at half the supposed level required. Should the co2 go over 400 it goes to a "normal" rate and boost if over 1000. Eight people in the building will push it to 400 and above. For something else to keep it busy it posts data to a database  on another pi. Things like, what fan speed and how long to calculate running cost along with the various temps to calculate savings. Not to mention it's use in control and monitor of the shed heating and ventilation.... all OTT but a good bit of fun... now, must order one of those particulate monitors. Can't wait to tell the wife exactly how many particulates are getting up her nose and are causing her sneezing... I can see lots of brownie points for that one... not.

[MikeSharp01]

1 hour ago, Ajn said:

d1 mini micro

Great devices - but I do worry how often they phone home! I am working my way through the Rpi Pico W toolchain, back to the 80s without a windows IDE, to wean myself off the D32 Pro I have built my ptype kit around. Alongside the Rpi MQTT node red etc. The particulate monitoring will be a great addition - but I am not sure I want to know with my cooking skills😁

[MJNewton]

Nevermind the neighbour's woodburner... it's the summer I should be more worried about! This ongoing hot spell has revealed pretty poor air quality at quite a number of times:

 

 

And this doesn't coincide by my neighbour and/or I wheeling our barbecues out (although they are detectable spikes!) as it is country-wide:

 

 

This has got me wondering: is all particulate matter the same in terms of the hazard to health? Or even if not exactly the same is it at least broadly similar i.e. is it more the size of the particulates that are the issue (small enough to enter the bloodstream?) or is the type (is, say, wood smoke worse than whatever it is the engulfs us over summer - if that's what's happening here)?

[Dave Jones]

very very bad year for pollen, my eldest is really struggling. the hepa+uv filter on the mvhr will be a godsend.

[MJNewton]

23 minutes ago, Dave Jones said:

very very bad year for pollen, my eldest is really struggling. the hepa+uv filter on the mvhr will be a godsend.

 

Ah, yes, and as a hay fever sufferer you'd think I would've thought of that! I am more of a tree pollen sufferer and the species that affect me most have been and gone now.

 

I believe pollen tends to be on the larger side and the graphs do seem to indicate that it is the larger particulate sizes that are highest so that could well be it.

Edited June 13, 2023 by MJNewton