[Strawbale] Moisture sensor redux (was Re: cement stucco problems?)

[Mark Bigland-Pritchard]

It seems to me that various things are required of a moisture sensor.  
Which option you choose depends on how you rate these parameters:
(1) accuracy of the sensor itself, across the full range of likely 
humidities and temperatures
(2) cost of the total system
(3) robustness
(4) size of sensor + associated protective layers (the bigger it is the 
more it will affect wall hygrothermal conditions - especially if you 
have a permanent tube going most of the way through the wall - and so 
the less reliable the reading will be)
(5) [where appropriate] ability to give a reading that can be 
interpreted by a data logger for continuous monitoring purposes
(6) anything I've missed?

Now, the wood block resistive method is good on (3) and (if the cheap 
meters that Rob mentioned before are reliable) very good on (2).  It 
could probably be made good on (4), which makes me wonder why all the 
assemblies I've seen, possessed or seen piccies of involve a plastic 
tube going most of the way through the wall.  But it's useless on (5), 
and not very good on (1).  Changes in temperature will result in errors, 
because the electrical conductivity of wood approximately doubles for 
each temperature increase of 10 deg C.  And the variability of wood 
samples means that either each sensor has to be individually 
recalibrated or else there is going to be a significant loss of 
accuracy.  (Somewhere on the web there is a copy of the USDA Forestry 
Service's "Wood Handbook - wood as an engineering material".  If you 
look at section 3, page 21, you will see a graph showing the variation 
of timber electrical conductivity with moisture content.  It shows 
considerable variability among wood samples.)

The ceramic sensor capacitative method is excellent on (1), good (+ with 
a bit of design flair could be made excellent) on (4), and good on (5).  
(3) is a problem, but it can be overcome with careful installation.  (2) 
is becoming quite good - I don't have a north american price as yet, but 
RS components in Britain are selling high-quality Honeywell ceramic 
humidity sensors for under GBP20 each.

Personally I think it's worth paying a bit extra for the accuracy.

As regards the issue of r.h. readings vs moisture content readings, the 
two parameters are correlated for any given material.  In other words, 
straw moisture content depends on local air relative humidity (local 
meaning at the relevant point inside the bale), not on absolute 
humidity.  [Yes, there is some temperature variability for the sorption 
isotherm, but, as Don Fugler's 1996 report puts it, /"There may be some 
influence of temperature on response, but for the range of temperatures 
studied (13 C to about 28 C) the influence was small and difficult to 
determine."/]  So Rob's discussion on this is - if I understand 
correctly what he's saying - a bit of a red herring.  I don't see an 
inherent reason to favour rh values or to favour mc values.  But in fact 
there is one reason of scientific convention why rh values should be 
preferred: mycologists always express moisture thresholds for fungal 
growth using "water activity", which is just the local equilibrium r.h. 
expressed as a decimal instead of a percentage.

And yes, Rob, I agree that r.h. readings are useless without temperature 
readings.  So are m.c. readings.  I wouldn't, even in the simplest 
monitoring exercise, consider measuring bale moisture without also 
measuring bale temperature.  This is not an extra expense, it is a 
necessary expense whichever way you do it.  Furthermore, mould growth is 
temperature- as well as moisture-dependent, so we need to know the 
temperature to know where the moisture levels are in relation to the 
risk threshold.

As regards power inputs to sensors, you're going to need it for a 
continuous monitoring job anyway.  If you're doing something simpler, 
the Honeywell HIH series sensors just need a 5V dc supply (and will work 
with anything from 4 to 5.8V).  A circuit to get a consistent 5V output, 
either using mains or a battery as the original source, is quite simple 
and inexpensive to make.  (Even I can do it.)  And the reading takes 
less than a minute to stabilise when you switch on, so you only need to 
draw power for short periods of time.  Power _output_ then can be read 
with a standard multimeter.

So on this one I think there are very good reasons to review a CMHC 
decision made "a number of years" more than "over ten years ago".  
Things have changed since then - not least that the cost of accurate 
capacitative r.h. sensors has come down considerably.

Mark



Rob Tom wrote:

> On Sun, 15 Oct 2006 03:11:58 -0400, Mark Bigland-Pritchard  
> <mark at lowenergydesign.com> wrote:
>
> I'm guessing that Mark's "smallness of capacitative sensors" query has 
> to  do with a comment I made recently to HW on the SB-r-us list; 
> something to  the effect:
>
> "Why bother futzing trying to re-invent the wheel with moisture 
> sensors  when the CMHC-style wood block sensors do the job just fine ?"
>
>
> But first to answer Mark's question on "size":
>
> The wood blocks themselves (in the CMHC-style sensors) are actually 
> very  small -- a wood disk the size of a 25 cent piece in the case of 
> the "first  generation" type for which Habib Gonzalez provided 
> fabrication notes -- or  a wood plug that can be smaller than the 
> first segment on one's baby  finger, in the case of the "second 
> generation" type for which Rob Jolly  provided fabrication notes.  
> [Both of the preceding sets of notes with  images are stored in the 
> FILES section at SB-r-us]
>
> If the wood blocks were any smaller, I suspect that most baleheads 
> would  have difficulty getting their fingers to behave during DIY 
> home  fabrication of the sensors.
>
> Even at their comparatively "bulky" size (when compared to a 
> capacitor) it  requires a drastic change in headspace from the 
> everyday (unless one is a  watchmaker or such-like) to develop the 
> fine touch necessary to drill the  tiny wood blocks, insert the two 
> even tinier bolts/nuts/washers and  connect the tiny small gauge 
> telephone wires.
>
> Any bulk the sensors may have is due more to the protective sheath 
> around  the sensor itself, the sheath being made from 
> readily-available, stock  plumbing parts that in many cases, people 
> would already have, as a result  of home plumbing jobs that they've 
> done .
>
> Even so, the entire sensor assembly including protective sheath can be 
> as  small as the cap on a fountain pen (actaully, a bit shorter in 
> length) ,  which, in relation to the size of the bales into which it 
> will be  inserted, is negligible IMO.
>
> What is more relevant (again IMO only) is that (from my cursory  
> understanding of them) capacitative sensors can only yield humidity  
> readings whereas the CMHC-style wood block sensors give direct 
> readings of  moisture content.
>
> Humidity readings are by nature (again IMO) useless without the  
> corresponding temperatures at which the humidity readings were taken.
>
> Further, humidity  readings can only give one a picture of the  
> instantaneous and transient conditions unless one has some means of  
> continuously logging a series of temperature and humidity readings 
> over a  long period of time and that usually means additional 
> gizmology which in  turn means a source of power to run the gizmology 
> which in turn means that  if for some reason the power is interrupted, 
> data is lost and you have  nothing.
>
> [ Interestingly, almost as if on cue, the grid-supplied electricity to 
> my  home just blacked out as I wrote and re-wrote the above.]
>
> And humidity *is* relative.  ie 90% relative humidity at minus 10 
> degrees  C represents very little actual water in terms of actual 
> %moisture content  and would pose no threat whatsoever to bales.    
> OTOH, 90% relative  humidity at 30 degrees C represents a great deal 
> more actual water in  terms of %moisture content and could be a threat 
> to the health of the  bales in one's wall.
>
> I say "could be a threat" because it would depend upon the duration 
> of  those high humidity/temperature conditions and if there was any 
> sort of a  regular pattern of those conditions over the long term. 
> Small spikes over  the short term are to be expected with plastered 
> walls exposed to wetting  and not really anything to get worked-up about.
>
> OTOH, with the CMHC-style wood block sensors, the reading that is 
> yielded  gives you an instantaneous reading of the longer-term health 
> of the bales,  which is, what I suspect, most SB homeowners are 
> interested in when  monitoring their walls.
>
>
> PS: The "redux" descriptor in the subject heading refers to the 
> moisture  sensor development discussions that were carried out on the 
> CREST SB list  long ago and I didn't fully realise how "long" ago that 
> was until  yesterday when I went for a walk in the woods with Don 
> (Fugler) (and Don  chased a poor widdle porcupine up a tree and then 
> got us lost).
>
> He (DOn that is, not the porcupine) reminded me that the last time we 
> saw  each other was over 10 years ago and it was a number of years 
> before that  when I met with him the time prior, to discuss the 
> moisture sensor  evaluation study.
>
> ===* ===
> Rob Tom
> Kanata, Ontario, Canada
> <archilogic at chaffyahoo dot ca>
> winnow the chaff from my edress in your reply
>
>