Topic: build your own temp/hum sensor for vp2

[jerryg]

I read down in the tech corner thread about replacing the sensor on the davis board when it goes bad and got excited about trying it myself. I have a homebrew dual louver stevenson screen that i use for my davis sensor i removed from the davis shield and i did not need to remount my new sensor on their board. I bought the better speced sensor which is the sht15 and all i needed was a six conductor cable with a connector on one end which i had and a capacitor to decouple the supply voltage which i had, so the sensor cost me 31.56$ shipped and i have it running now on my backup station to see how it compares to my primary station. If everything ages out ok in a few days i will have a more accurate sensor at a third of the cost of the davis sensor. I had a sensor that was potted with clear stuff and i could see the wire colors and where the traces went to on the sht15. One thing the connections on the sensor are very small pads and close together and would be hard for anyone without the proper skill to hook up, but they make a sht75, same sensorl, but mounted on a board with pins to plug into a socket and only cost 5 bucks more. If i did another one i think i would spend the extra dollars and save alot of agony hooking up to the solder pads lol. Will report back in a few days after seeing how it compares to my primary sensor.

[Yfory]

I am also experimenting with temp and humidity sensor replacement. In my case I am testing a long cable, in place of wireless, to give me 24 bit resolution into my PC with low cost 24 bit ADC to USB PC interface.

I think we are all finding how nice it is to understand weather station circuits, to succeed with a better replacement, and to improve weather data quality.

Congratulations to jerryg!

Steve

[DanS]

This topic also applies to other brand stations as well. I have done the same with a Lacrosse 28xx station, improving it's accuracy using a SHT11 though. You're right about those connections being so tiny and close together! Like doing micro surgery.

[DeKay]

Nice job, jerryg.  The prices you have quoted are excellent.  Where did you get this stuff?  The SHT15 plus breakout is $42 before shipping at Sparkfun.  Can you provide a link or two?

[jerryg]

The price i talked about was just for the sensor.It came from the same place. I like the breakout  board, alittle pricey but has everything you need for a sensor and still at half price. The board i was looking at just has pins on a pcb board with the sensor, looks like would have to add the 100n cap. Here is the one i was looking at. http://www.newark.com/jsp/search/productdetail.jsp?SKU=18M2988

[jerryg]

Just an added note if anyone needs the pinout connections for the rj plug and the sensor i have it written down and will be glad to post it.

[Yfory]

On the soldering issue...I have a Weller WP25 ( I like a hot iron) and Weller tips are available (on EBay) that are needle sharp and quite long compared to standard tips.

I use a magnifying head set (2x, 3x, 10x, 25x like a jeweler uses) needle soldering tips to solder this type of board - it is now much easier. No mater your age and vision, 5 to 10 x really helps. I actually use 25x to check for the mistakes (like tiny bridges) I make at 5x!

I did not realize how much 10x helped until I soldered an SOIC op amp onto a SOIC board. I suspect the boards you are discussing are about the same size and difficulty with trace separation.

[jerryg]

The pads are super small as is the sensor itself. I used a fine tip weller iron and some hair size solder to avoid blobs. I used some hi power reading glasses for the old tired eyeballs lol. So far the sensor is working fine and tracking with my main sensor. We are supposed to get some fog in the early am and i want to see how it looks on the high end. Then when a norther comes in and the humidity drops low will see how it looks on the low end. With the specs being alittle better than the one davis uses should see a small difference, how much is to be seen. If it is just alittle then could save alittle money by getting the 11.

[dogdad]

jerryg -

You were kind enough to offer to post the pinouts and wiring, and I'd appreciate seeing them.  A neighbor gave me an old Davis mounting arm with the shield and sensors still attached but the wires had been cut about a foot long, and I'd like to put it to use for water temp.

If that sensor is bad, then I'd go ahead and follow in your steps with the sht11/15 chip.

TIA

Ron.

[AWL]

jerryg -
You were kind enough to offer to post the pinouts and wiring, and I'd appreciate seeing them.
Ron.

I would appreciate seeing these also. Wonder if this is the "bigger one" mentioned earlier in the posting?
robotshop.com

[SLOweather]

I purchased my SHT11s from Newark for $26.94:
http://www.newark.com/sensirion/sht11/humidity-sensor/dp/69C4848

The also have SHT15s for $33.07:
http://www.newark.com/sensirion/sht15/humidity-sensor/dp/18M2986?in_merch=Popular%20Products

[DeKay]

I did not realize how much 10x helped until I soldered an SOIC op amp onto a SOIC board. I suspect the boards you are discussing are about the same size and difficulty with trace separation.

I linked this video from one of my blog posts a little while ago for soldering fine pitch devices.  SOIC is easy with the techniques presented here.  I've successfully done TQFP parts with a Radio Shack firestick soldering iron that I bought umpteen years ago.  I really need to get myself a magnifier, though.

[jerryg]

Ok here is the how the connector looks, with the front facing you, that being the side without the latch, starting from the left with pin 1 blue 2 yellow 3 green 4 red 5 black 6 white  and Davis uses only four of them probably made it a 6 pin connector so it would not fit in any of the other sockets. The green lead is the ground. The yellow lead is the + voltage. The white is the clock. The blue is the i/o data pin. On the sensor lpin1 is green pin2 is blue pin3 is white and pin4 is yellow. You need a 100n capacitor between 1 and 4 on the sensor.

[AWL]

Question: on a VP2 wireless, with this setup, does this sensor read °F with no further modifications?

Thanks for the info, Doug

[jerryg]

I am not sure what you asked, the sensor will read f or c depending on what you tell the vp2 you want to read.

[AWL]

I am not sure what you asked, the sensor will read f or c depending on what you tell the vp2 you want to read.

You answered my question. I think I will order one today.

Thanks again, Doug

[jerryg]

Just a follow up on my sensor. I am well pleased with the results so far. The readings are really close to the metars i compare to and comparing it to the sensor i was using. I am still waiting for a strong cold front to come in an lower the humidity down below 20 percent to compare it with the davis sensor. I have 5 davis sensors with my stations and 4 of them top out at 98% and one at 100% to compare to. The thing i notice so far is the 15 readings are really linear, this am we had dense fog and it was showing 99% then 100% and back and forth, then when the humidity went down it decreased at a rate that was the same as the metars and is now showing the same reading. I have to use the slope setting to get davis sensor to read close when the humidity gets low, i have it to offset a -2 around 50% which gets it close. With the 15 i have not had to use the slope setting so far. If you look at the specs on the 11 and 15 you can see the difference. Here is the link to the specs http://www.sparkfun.com/datasheets/Sensors/SHT1x_datasheet.pdf   go down to figure 2 the graph chart and look at how the two sensors compare at the high and low end. The temps compare pretty close with the davis speced at 1 degree f and the 15 is speced around .5 degrees f i really did not expect much difference anyway. Maybe davis should offer an option to get the better speced sensor.

[belfryboy]

Great work, I can feel another build coming, Cheap replacment T/H senors anyone? I could quite easily do these to the same dimensions as the Davis sensor for use as a direct replacement.

[johnd]

For anyone thinking of a DIY digital (.166) T/H sensor, it's perhaps worth remembering that most of the issues with the T/H sensor have not been to do with the sensor element itself AFAICS, but with the way that the element is interfaced to and mounted on the daughter board. We're now on Rev J of the .166 T/H sensor, with progressively more and more weatherproofing being added around the sensor element and the board layout itself evolving over the past few years.

For anyone planning to use the sensor in a damp/rainy/windy climate I think that you're going to have to pay equal attention to the mounting details, as Davis are doing, in order to get a T/H sensor that lasts.

[jerryg]

I agree about the weather proofing and i do have the capability to weather proof the connections, but this is just a trial to see how much difference the better speced sensor makes. So far it looks good and even if it lasted only a year or two without weather proofing at the price you could afford to get another one or two. Now if i could just get some low humidity to check the bottom end readings i would be a happy camper lol.

[SLOweather]

OTOH, that doesn't explain why we experience so many failures with current sensors along the Pacific coast, and why replacing the SHT11 fixes the problem.

For anyone thinking of a DIY digital (.166) T/H sensor, it's perhaps worth remembering that most of the issues with the T/H sensor have not been to do with the sensor element itself AFAICS, but with the way that the element is interfaced to and mounted on the daughter board. We're now on Rev J of the .166 T/H sensor, with progressively more and more weatherproofing being added around the sensor element and the board layout itself evolving over the past few years.

For anyone planning to use the sensor in a damp/rainy/windy climate I think that you're going to have to pay equal attention to the mounting details, as Davis are doing, in order to get a T/H sensor that lasts.

[johnd]

OTOH, that doesn't explain why we experience so many failures with current sensors along the Pacific coast, and why replacing the SHT11 fixes the problem.

What do you mean by 'current' sensors though? Are you including rev 'I' and rev 'J' in that? (The letter should usually be shown as the final character on the label on the daughter board eg 7346.166J). My experience is that I and especially J seem to be lasting substantially better than some previous revisions, though it is relatively early days as yet for J. I don't have the exact dates to hand, but I think that I possibly came in sometime during the 2009 builds, maybe towards the end of 2009 but really can't remember exactly.

[Yfory]

This is a good discussion.

I have been working with a high quality Honeywell humidity senor for a few weeks now and I can verify your comments about taking care to weather proof the humidity sensor itself - as odd as that sounds.

I am experimenting with the Honeywell HIH humidity sensor series and Honeywell is very specific about the conditions the sensor will experience.

If a sensor is to be exposed directly to water droplets or wet environments, they suggest a very special "covered" version which resists water droplets but still measures air or other gas humidity. Obviously the sensors can not be sealed but the problem of wet (industrial) environments are very difficult to monitor for humidity without permanently damaging or at least changing the output of the sensor by getting the actual sensor surface wet!!!

It is the technology of the "cover" that is tricky.

The HIH series has an analog output which happens to interest me. The HIH series must be digitized to work with most weather stations.

[AWL]

For what it worth: I just built a steveson screen for my wireless VP2. Got everything mounted this past weekend. As I was doing this I also needed to replace the fan that came with the kit which was 2 years old. I did this with a $12 40mm(1.57") 12 volt case fan from Newegg . I hooked this up to a 9 volt dc adapter and it is purring like a kitten 24/7 now. To finally get to the point of this saga....I am also going to build the temp/humidity sensor which is being discussed here. I got to thinking there has got to be a better way to extend the current sensor than cutting the wires. Went to the local Radio Shack. They did not have a 6 wire RJ11 cable but they did have a 6 wire inline connector just like the 4 wire connector that comes with the Davis anemometer extension cable. I found an IT guy who had the cable, a straight through 6 wire RJ11 6' cable. After I got everything mounted, I unplugged the sensor cable, plugged in the connector and 6' cable, both the temp and humidity came back to the exact same reading, to the tenth. So for $8 for the connector and $5 for the cable I have now extended the cable length and will only need a PC board RJ11 connector and will be able to build the new sensor and never cut a cable.

Doug

[jerryg]

Well i finally got a strong cold front in that dropped the humidity low enough for a comparison of the two sensors. I have 3 davis sensors and the homemade sensor in the same screen side by side so i could get a comparison with the same conditions. This is not scientific just eyeball comparison watching on wu site and the two metars that i am between. One metar is 8 miles nw and the other is 12 se of me so get a good idea of the humidity when they are both reading the same. I was surprised how close the 3 davis sensors track each other. The temps of all sensors where withing a tenth or two of each other. All my davis sensors top out at 98% and the new one at 100%. I had a period where the hum. was holding around 75% for some time and they all showed that reading which is the normal calibration point for the sensors. When the humidity went up they all tracked about the same until around 90% when the new sensor slowly went up as the fog moved in and the davis sensors started to fall slowly behind. Just to show how dense the fog was my rain gauged showed .02 inches of rain from condensation on the gauge, really really wet stuff lol. When the davis got up to 96 the new sensor showed 98 and then went to 99 while it took longer for the davis to get to 97 and the new was on 100 for sometime before the davis topped out at 98. Going the other way from 75 they all tracked close until in the low 40s then the davis started to fall behind until the new was at 19 and the davis were at 21 to 22. I can offset the low end difference with weather displays slope adjustment for humidity by setting it at -1 it works pretty close from one end to the other except at 90 percent then it takes awhile for things to even out. I am well pleased with the new sensor and the davis sensors with the correction.