let the data begin...

jpf321

<strong>1st</strong>, my "FAR RAD" datalogger is a bit under the weather .. so don't believe what you see.



<strong>2nd</strong> .. each thin vertical line is 15min increment. beginning on 11/16 each thick vertical is 1hr increment.



<strong>3rd</strong> .. I do not have oil consumption data here .. I'm still figuring out the best firing rate .. today I reset to factory defaults on the nozzle. Since Friday(11/12?) I've burned 16.04gal. Most of the consumption was due to testing and tuning the fire.



<strong>4th</strong> I do not have Heating Degree Day Data here .. I will figure out how to add it somehow.



<strong>5th</strong> I've played around a bit with the "threshold temperature" which denotes "ON" state, currently:

 STOVE>200,

 HEADER>190,

 MAIN END>150 and

 FAR_RAD>100



<strong>6th</strong> my MAIN END temperature probe is not under insulation as my HEADER probe is. Hence, my main end seems to "shut-off" early. But I think what's most important is when things go "ON" rather than "OFF" (at least right now)



<strong>7th</strong> these samples were every 30s .. I've readjusted moving forward, the samples to be every 60s.



I'm sure this is a work in progress, and as I said, I've been tinkering alot as this data was being collected so draw no conclusion from the data in this post .. as the season sets in, I will be tinkering less and learning more.



11/09 - <a href="http://screencast.com/t/VFJTn9gre">http://screencast.com/t/VFJTn9gre</a>

11/10 - <a href="http://screencast.com/t/OxTu7OmcFu0">http://screencast.com/t/OxTu7OmcFu0</a>

11/11 - <a href="http://screencast.com/t/PdCWMiUE">http://screencast.com/t/PdCWMiUE</a>

11/12 - <a href="http://screencast.com/t/iznQbtsleZQg">http://screencast.com/t/iznQbtsleZQg</a>

11/13 - <a href="http://screencast.com/t/Jf9XbtPK">http://screencast.com/t/Jf9XbtPK</a>

11/14 - <a href="http://screencast.com/t/MXdUP5w5FE">http://screencast.com/t/MXdUP5w5FE</a>

11/15 - <a href="http://screencast.com/t/pGQp1PfwC">http://screencast.com/t/pGQp1PfwC</a>

11/16 - <a href="http://screencast.com/t/heH7ar7NmeUu">http://screencast.com/t/heH7ar7NmeUu</a>

jpf321

this is all excel...

but I'm happy to learn of other options

lutorm

Here's mine:

Nice job. I got my setup working too, here's some data from this morning:







(sorry for the too-wide graph). This is with a totally home made setup. An Arduino microcontroller on the boiler is taking data from a number of Dallas 1-wire temperature sensors and sending it up to our server with a pair of XBee radio modems. After a bit of tweaking, it now seems to work reliably. The temperature probes are very accurate (.05C resolution) but only work up to 125C.



A couple of interesting things I noted:

1. The "stove pipe" sensor seems to work well for detecting burner on/off. Now I just have to add a filter that looks if the temperature is increasing or decreasing and that should be set.
   
   
2. The boiler case and header top out at 88C. This is much lower than I expected. I'd suspect the probes weren't getting a good signal if they weren't showing exactly the same value. Is boiling at

lutorm

****?

ah, forum can't deal with less-than signs... that caused the post to be cut off. Here's the rest.

1. The boiler case and header top out at 88C. This is much lower than I expected. I'd suspect the probes weren't getting a good signal if they weren't showing exactly the same value. Is boiling at less than 100C a sign of a dirty boiler? Both boiler and the pipes are insulated.
   
   
2. The leaking Gorton main vent is clearly visible, when the boiler pressure builds at around 07:10, it gets much hotter. This is expected because with a bit of pressure it's leaking steam.
   
   
3. The pressure-boiling temperature dependence is clearly visible when the burner is pressure cycling. The temperature of the boiler goes up and down with pressure.
   
   
4. When the thermostat finally turns off, I think you can see the system going from positive to negative pressure at 07:32. The main vent starts cooling down a lot faster, and the boiler water slower. Then at 07:35 there is a kink in the boiler water graph, and it (and maybe the main vent) starts cooling down slower again. That would be when the vents opened (the rad vents open before the main vent at the end of the cycle), and from then on the boiler is just cooling passively.
   
   

Can you tell I'm excited to get it working? ;-) Any comments?

jpf321

some new and perhaps interesting charts

How long I fire (in hours) per HDD (seems to be settling at around 0.3hrs / HDD)

http://screencast.com/t/DqAKRBCdK2K BURN/HDD



How long all items are over threshold against HDD (for rad to be hot 24hrs, it's projected to require a 59HDD (5degF avg daily temp))

http://screencast.com/t/nm01JNNYX HRS vs HDD



How long the FarRad is hot vs. Burn (%) (the rad stays hot about 150% as long as I fire that is, fire for 1hr .. rad is hot for 1.5hrs)

(I got a new FarRad datalogger, all data moving forward should be accurate, some data to this point was corrupt.)

http://screencast.com/t/oRJtlBc0qJvy FarRad/Stove



Highest and Lowest Recent HDD Cycle charts

(each thick vertical =1hr, thin vertical = 15min, starting from 00:00)

http://screencast.com/t/s6ZiqYhSu Highest HDD so far = 24

http://screencast.com/t/PYtxutEn3 Lowest HDD recently = 7



Recovery Cycle from Deep (12degF) setback

http://screencast.com/t/5qz9RRivAUze 58degF --> 70degF recovery = 6hrs burn.



Keep in mind that this year, my T-stat will be set at 70 with no programmed setbacks .. if I set it back for some reason (away for weekend or something), I will note it specifically.



Everything is in Excel, I can provide someone the .xlsx file if they are interested. I have a cycle chart for everyday.

HDD data taken from degreedays.net using KLGA as station, heating, 65dF, daily.

Mike Kusiak_2

Nice graphs

What's causing the burner cycling between 6:40 and 7:15? Appears to be cycling every 10 minutes or so. What type of thermostat and how is the anticipator set?



I would expect the burner not to cycle at all until you hit the pressure limit and start cycling on pressure, which seems to begin around 7:20.

Mike Kusiak_2

Stove?

What does the stove graph represent? Is this the flue gas temp sensor?

jpf321

yes .. stove = stovepipe .. flue temp

yes .. stove = stovepipe .. flue temp .. an indication of when fire is burning. 

FJL

What Is This Stuff?

Am I correct that you are measuring the temperature of various parts of your steam system?  Is that what all of this is about?  I'm not familiar with whatever equipment you are using; that is why I ask.

lutorm

It's the LWCO...

It cycles the burner off for ~ 90s every 10 minutes.

Mike Kusiak_2

Dallas sensors

How are you coupling the temp sensors to the pipe? I have found pretty poor thermal conduction with thermocouple sensors unless you use some thermal compound (heat sink grease) between the sensor and pipe. I also use a small piece of copper foil tape over the sensor to give it more surface area for conduction. Poor contact might explain the lower than expected header temp.



What Dallas sensor are you using and how do you interface to it?



I believe you have an atmospheric boiler with a vent hood. You can get a much faster response time if you place the sensor up in the hood near the exhaust stream but not actually touching the metal. If you dont put it directly in the stream you shouldn't exceed the sensors max temp.

Mike Kusiak_2

Image

Cant seem to get the attached graph image to work in below post. Trying it again here.



Found the solution. Filename cannot contain a #. Thanks JPF!



(By the way, this is a graph of flue temperature of a hot water boiler recovering from a warm weather shutdown on a 55 degree fall day.)

lutorm

DS18B20

I'm using DS18B20 1-wire sensors. The pipe and vent ones are glued with thermal adhesive. The others are simply lying between the insulation and the metal. The lower temp could be caused by poor thermal contact but since they are inside the insulation there should not be large thermal losses so that seems less likely.



I thought about mounting the sensor in the exhaust itself but even if the sensor can take 125C, the wire insulation is only good for 105C. At least this way I know that the insulation is taking no direct heat, and it seems good enough to pick up the boiler firing within the 15s data rate on the batabase.

jpf321

my measuring and analysis...

is a bit more low-tech than lutrom's .. i am using Lascar Temp USB Dataloggers (best deal if from katsci.com (http://is.gd/i70L8) .. the use type-K themocouples rated for -200C - 1350C ..



I have 4 units deployed. Stove Pipe, Header, End of Main and Far Radiator. They are all time-sync'd. I retrieve the data every now and again, it takes me under 5mins to dump the logger and redeploy. They need to be physically dumped via USB. With 1min sampling they can hold 11days of data.



I also have 2 ambient units, one which I used to keep outside, one which I sat near the T-stat. But now for consistancy, I get my Heating Degree Day Data from degreedays.net.



Because I can't read real-time temp from the loggers, I am using a Lacrosse Technologies wireless thermos http://is.gd/i74Kh (I currently have 1 station with 2 wireless sensors but asking for more for X-Mas) .. I use that to monitor the real-time temps in several rooms, especially the ones I have TRVs in right now. So far, all rooms are max'ed (via TRV) at 69.something. I am pleased. The wireless thermos are not as precisely calibrated as I might like, but they are all within a degree.



Lastly, I do have TRVs installed in Bedrooms (4), Kitchen and Foyer (rad + riser there usually overheats without TRV). Sometimes my boiler will pop on and once the mains fill, I go right to 10oz. because half my load is TRV'd off. I have actually been standing at gauge (not surprising) when I have witnessed a TRV opens and the gauge begins to drop right off to 2oz or so.



FYI, I have no problem reading the pipe temp using the Lascar K-Type thermo probe. I have it laying on the pipe lengthwise and I have it taped on with Aluminum Tape, and that is all under my 2" insulation. That is also the only Datalogger that I have with LCD display. In the middle of a cycle, I show 215degF which is right on the money :-). In my mind, as long as all your probes are mounted similarly, and consistently, that's all you need since you are really looking for timing, patterns and thresholds, not exact temps.



Thanks for reading. I'm happy to answer questions or share my excel sheet. I'd like to make the excel a bit more user friendly, but for right now it will suffice. I may be bold and dump everything into MySQL with excel to handle the graphing, but then again, there's always next year :-) .. I currently have 43000 records and it's only Dec 2. Excel starts to bog with that many records and charts for each day. I'd prefer to automate the queries from a MySQL channel and have a parameter for what day I'd like to chart. As well as have MySQL handle some of the simple analysis via query rather than have a 13MB excel sheet. (I used to be a PHP/MySQL guru, so I should be able to manage it, it's the motivation which is lacking .. perhaps if I felt more people had similar desires for such a "product", I'd be a bit more motivated.)



I know this post is getting lengthy .. but I do have a cool Excel formula for turning the Lascar date fields (Euro D/M/Y) into US Date (M/D/Y) Format (the date field must be imported from the .txt file AS a TEXT column) .. here is the formula which assumes your "DATE to be in column "B" =DATE(VALUE(MID(B2,7,4)),VALUE(MID(B2,4,2)),VALUE(LEFT(B2,2)))+TIME(VALUE(MID(B2,12,2)),VALUE(MID(B2,15,2)),VALUE(RIGHT(B2,2)))

then "Custom Format" the cells as "m/d/yyyy hh:mm:ss"



OK I'm done blabbering ..

jpf321

actually

I need to update my signature. I'm firing at factory specs now. how long had your house been in deep setback temp? mine was longer than 24 hrs so the entire contents was thoroughly cold..that's a great deal of mass....and there was quite a bit of pressure cycling which may not be reflected in the chart.

arches

more info on arduino/xbee setup

Hi Lutorm,

I'd be curious to hear more details about how you set up the arduino and xbee to remotely monitor temps. I've been thinking about doing something similar to monitor temps in each apartment (or at least several apartments) in a 25 unit co-op apartment.  Which xbee device did you use, and what kind of range can it cover (distance wise).

jpf321

you may be interested in....

http://www.datanab.com/Solutions/BarioNetWirlessGtwy.htm

a bit pricey for me..but maybe your budget is bigger..they also have 1-wire solution

lutorm

Arduino setup

I'm using two series one 1mW xbees with chip antenna (http://www.sparkfun.com/products/8664). The range isn't great, they are separated by maybe 8 meters and I get periodic dropouts in the comms (though that might be the fault of a faulty xbee). The Arduino is hooked up to the 1-wire probes and reads them periodically, then sends a data chunk and checksum through the Xbees to our file server, where a little Python code is running to decode it and add it to the RRD.



This works decently because I only need one Arduino. I've been thinking of getting temps from the different rooms, but that requires long 1-wire connections. If you want several remote sites, with this setup you'd need multiple Arduinos, Xbees and power to each. It would be a bit more complicated. In that case it might be easier to try to design a setup where you use a temp probe that can be read through the XBees ADC. They can operate headlessly and just report the outputs of the ADC. Then you would only need an Xbee and a temp sensor at each location. (Kind of the way this works: http://www.ladyada.net/make/tweetawatt/)

arches

thanks guys

I'll check out the xbee and datanab ideas...both look interesting. The alternative is to upgrade our existing HeatTimer to include an internet communications module...which then allows for wireless apt temp monitoring (with the purchase of additional devices)..but the upgrade is quite pricey.

jpf321

in cooperation with datanab

www.buildingmonitor.com

Alex265

jpf321

May be I missed but I can't find among your charts data for Therms/HDD or CCF/HDD. Aren't those the most direct indicators of how efficiently (in terms of cost) the system operates?

jpf321

I run oil

so I tend to think in GPH. and that chart is there I think, perhaps by way of stove hours/degree day. since the chart would be the same if the two multipliers of gph and the convert to therms..just ignore vertical scale and consider hours as therms.

Alex265

Oops, sorry, missed that you are running oil

I'm just looking to get an idea of what those numbers can be... OnTime/HDD is good, but not with a modulating burner, I think.

jpf321

correct

modulating burner would require actual consumption data to accompany the other data points

jpf321

a few more days of data ....

i've been pluggin away at my data ... below are a few recent charts...



Coldest day Recntly HDD=32: http://screencast.com/t/nylLmkjUaOj

Warmest Day Recently HDD=12: http://screencast.com/t/2Vstqpku7D3



I looked back at my notes an realized I stopped playing with the boiler on 11/17 .. so I cropped all data before that date ..

Hours On/HDD: http://screencast.com/t/etwxHWpP25z

I have also displayed a 10-period Moving Average .. seem to be settling in at around 0.27hrs/HDD



Hours vs. HDD: http://screencast.com/t/4FsiFlb3RB

I'm really trying to understand why there's such a strong correlation between HEADER and FAR RAD as well as STOVE and MAIN END. Any thoughts? Header is insulated, main end is not. Maybe it's just dumb luck.



Far Rad Hours vs. Stove Hours: http://screencast.com/t/r9TlEhpnV6

jpf321

data update

1) Between 12/10 and 12/13 I had pressure set too low .. I was short cycling like crazy .. on 12/13 I set it back to 8oz.

2) I turned the T-stat down 1degF over the weekend .. now constant 69  (was 70) with no setback .. the change showed up in the following 2 charts:

Burn HRS/HDD: http://screencast.com/t/hYQ8Fmzpok

FarRad vs. Stove: http://screencast.com/t/t8i04ikZnz



My theoretical "24hr Far Rad" changed since the last chart .. went from 13deg daily avg to 11deg daily avg .. may also be due to T-stat change.

component HRS/HDD: http://screencast.com/t/DC1IkIvI





coldest Day Recently HDD=42: http://screencast.com/t/Hkzu3WpkkP

warmest day recently HDD=15: http://screencast.com/t/Tvd4uCirJq

jpf321

data update ...

Burn/HDD: http://screencast.com/t/onqj10Vem

FarRad/Stove: http://screencast.com/t/LFkCp5JPc

Hrs/HDD: http://screencast.com/t/RRhJmxILCC

haaljo

Interesting datalogger info

I don't understand dimensions.

.~1/3 gallon/HDD 

If 4K HDD/heating season you expect to burn about 1300 gallons? Is thatGot?

HRS VS HDD. Are you showng radiators "on" for 20 phours/day

Good stuff

jpf321

you almost have it...

that's ~1/3 hr / hdd.. during at a factory rate of 1.08 gph.



hrs/hdd is the various parts "on" vs hdd...the highest i see is about 18 hrs on the far rad.



thanks