This 6.8k resistor has been added to Box 1 to remove the cause of one of the problems causing the - - - - in the weather station's display.
Haven't been able to replicate this effect, however I agree with the general body of opinion that it happens fairly frequently, so I suggest trying the following -
This is something I and plenty of others have experimented with. I shone a bright light directly thru a bare sensor PCB to the underside of the surface mount - blobbed micro-controller IC U1. This caused the - - - - in the Weather Station's display. Remove the light, and it works normally again.
This is with both the inner and outer white plastic shells removed, and a fairly bright 160W outside PAR 38 flood lamp shone right onto it. It strikes me that the design of this sensor may need a re-think, as a number of people's sensors work at night or indoors, but fail outside in the daytime.
"Ok.... I thought I'd start a new thread for us all to LIST the Cures that we have found. For me the following has worked (3 weeks now with None)
1) Install the resistor Modification that Ian describes in his postings.
2) Install an aluminum foil tent over your sensor so that it hangs at least 1" below the sensor and is still open to the air at the bottom. I used a rubber band to secure mine.
3) Ensure that the Hygro sensor is OUT OF DIRECT sun light.
4) Avoid running the wires to the out side sensor near ANY electrical lines or wires that contain AC voltages.
6) insure that the Connection box is INDOORS or in a water proof housing.
I hope this helps some of you out. If needed, I'll post a Pic of my Aluminum Tent.
Bob
"Like others on this Board, I've experienced the unexcusable problem of the '----' readings from the outdoor Temperature/Humidity Sensor. My WM-918 worked fine during indoor testing but immediately started displaying the data error indication. After several months of studying the problem and getting no help from Oregon Scientific, I developed a simple solution that might work for some of you. The problem, I believe, is related to the design of the TH Sensor (THS), the component layout and distance between exposed traces and connections.
CONDENSING MOISTURE on the outdoor TH Sensor circuit board (PCB) seems to be at the core of the problem in my unit. The '----' occurred consistently at sunrise, increased into neat total oblivion in the middle of the day, and usually stopped at sunset. This occurred with the THS shielded from direct sunlight. I live near Chicago where relative humidities tend to be above 40% most of the time. If you live in 'dry heat' I wonder how much of this problem you'll see?
Based on this conclusion, here is my implementation of a simple fix among a dozen or so experiments, and at least three weeks of error-free data recording. NOTE: Three other fixes are added in case this alone doesn't do it for you.
TRY THIS FIRST (THE CLING-WRAP SOLUTION)
FIX 1: Remove the circuit board from the THS. Take about a 6x4 inch sheet of a plastic cling-wrap and place the PCB in the middle with the top of the plastic just below the T and H sensors (be gentle, lift the H sensor to about 45 degree angle off the board). Now make a tight fold (removing air pockets)across the PCB - like a sandwich with plastic on front and back sides. Then, roll tightly about four times. Use several strips of scotch tape across the middle, just beneath the T and H sensors, and at the bottom near the wires. Add a strip of tape 'between' the sensors across the top to effectively seal the entire PCB exposing only the T and H sensors.
Gently punch through the little guide holes and remount the circuit card. Reassemble the THS and set outdoors. See if that reduces or eliminates the '----' indications. Record your daytime readings using your PC and anyone of the Windows-based programs.
IF YOU STILL GET ERROR READINGS, NOTE the following add-on fixes:
FIX 2: I also took a 12" square sheet of metal window screen (get a roll at a hardware store, like ACE) and just folded it once around the THS. I stapled and trimmed the screen edges. I placed the sensor about 4 feet off the ground on the north-side of the house ( for indirect sunlight) and tie-wrapped the shielded THS to a large existing grounded pipe. NOTE: This fix, by itself, does NOT solve the problem.
FIX 3: JUNCTION BOX: I studied the signals at the indoor junction box (JBX) for several weeks and concluded that the THS seemed to exhibit a RESET problem when the '----' appeared; very likely condensation changed the power reset time constant - no shorts were detected. Every 5 or so seconds the 1.5 volt output line to the THS is toggled to 0 volts for about a second and then returned to 1.5v. The THS initializes, takes T and H measurements, transmits clock and data bits (some 48 bits per frame), and then enters into an infinite loop waiting for the next reset condition. There is a 10 uF cap on this line, at the output of the 1.5 volt regulator (SMD package). I replaced it with a 0.1 uF capacitor to reduce the rise time edge (don't do this if you are not comfortable with a 15 watt soldering iron and working on PCB components). This fix, by itself, does NOT solve the problem.
FIX 4: I added individual 0.01 uF caps from ground to the Clock and Data lines inside the JBX to reduce 60 Hz AC pickup. If you have a scope, you can see the induced AC signal during the reset period. Larger caps increase the rise and fall times of the signals and could create skew errors you want to avoid. This fix, by itself, does NOT solve the problem!
CONCLUSION
My WM-918 has run error-free for the last three weeks - just as it should have out of the box. The manufacturer is hereby advised to produce future TH Sensor units with a final varnish dip or insulating spray over both sensor front and back sides. A number of questionable solder fills were observed but these were not touched up. Remove your test pad - the gap on the pad is very very small and suseptible to moisture bridges.
SUMMARY: Try FIX #1 first - assume condensation is negatively affecting outside TH Sensor operations. Prevent condensation on the components by sealing the entire circuit card in cling-wrap plastic exposing only the Temperature and Humidity sensors. Lift the Humidity sensor a bit to make sure there is air flow around it.
Anyone offering a substantial reward for a good easy fix? Oh well, I'll have to wait for fame and fortune at some later date. I welcome your discoveries - if you haven't tossed the unit out by now. We engineers love a challenge - too bad we had to pay for it as well.
Good luck.
Victor"
Further thoughts were added at December 12, 1999 at 12:44:20 by Victor Nowakowski
"I thought I'd add a few more observations and conclusions to my previous comments and help further describe the operating environment and reliability of the TH Sensor (THS):
ENVIRONMENT CONDENSATION DYNAMICS
All you need for condensation to occur on any surface is either (a) warm moist air + cool surface or (b) cool moist air + warm surface: and both these condidtions can occur inside the TH Sensor each day.
Solids act like thermal capacitors that retain heat. With indirect radiation, it is primarily air (gas) that transmits heat to the solid. At dawn, the THS is in equilibrium with overnight cool air; but as the air warms throughout the morning, a temperature gradient can occur at the component surface thus creating the (a) condition. As the THS unit reaches daytime temperature equilibrium, the outside air begins to cool thus reversing the thermal gradient and thus creating the (b) condition at the end of the day. With enough air-borne moisture and temperature gradient, condensation occurs - exposed components and electrical traces are all affected surfaces. Without a moisture barrier over electronic and passive components, the TH Sensor will maybe, perhaps, sometimes be reliable. (Oh my!)
ENVIRONMENTAL THERMAL DYNAMICS
Aside from protecting the THS unit from the negative affects of moisture condensing directly on electronic components, test pads and signal traces, the plastic cling-wrap moisture shield has a secondary important effect: it also helps stabilize the temperature at the circuit board level.
One of the best temperature insulators is air, and the plastic cling-wrap effectively creates a moisture-proof thermal blanket. It remains a flexible insulator so it won't crack at extreme ambient temperatures. Likewise, the heat generated by the microprocessor and resistive components will diffuse from the center of the PCB out to the edges keeping most of the components from experiencing sudden or rapid "thermal shock." In delicate circuits, this indeed can become a significant problem affecting both accuracy and operational reliability. Was the TH Sensor tested in 0% to 100% humidity with simultaneous thermal changes representing ambient heating and cooling?
ADDITIONAL COMMENTS
Was it really necessary to put the electronics next to the sensors? I think this is the core of the problem - surely a better design to terminate these signals (using instrumentation amplifier design) is very common and ensures that the more sensitive components remain indoors.
All engineers are supposed to know that costs for commercial (0C to +70C) grade components are significantly less than for industrial (-40C to +85C) or military (-55C to +125C) grade specified parts.
The accuracy of the design is irrelevant if the overall function is unreliable.
NEW DESIGN RECOMMENDATION
1. Is it time to re-think the design? All the comments found on this bulletin board express a strong interest in the product concept, but also express a need for a solid, reliable low-cost design. I imagine that a relatively straightforward change, with minimal engineering effort, is to make a backward compatible solution that extract the electronics from the TH Sensor. Put that electronics in a dongle that plugs into Junction Box TH Sensor port and then plug the outdoor external T and H remote sensors into the sensor processing dongle.
This approach was going to be my next experiment if the "moisture/thermal seal" solution did not work produce reliable results.
2. Future models should integrate the functions into the main unit. In my opinion, I'd say that outdoor Temperature and Humidity, together with barometric pressure, rain and wind data, are fundamental to environmental monitoring products in the class of the WX200 and WM-918.
SIMPLE TEST SOFTWARE
For those looking for a simple Windows-based program, search the world archives for a program called 'SBWeather.zip' - older versions were free with donation requests from the author (Scott Baker) - I recommend it. I've written my own software (in Pascal) for lab integration into my experimental system platforms. Thanks, by the way, for those folks that worked to define the data frame protocols - I was half-way through that effort when I stumbled into this bulletin board.
CLOSING COMMENTS
A lot of expensive engineering talent (via "distributed open-source engineering") was and is still focused on this problem simply because this is our hobby. We have a strong personal interest in studying, monitoring and predicting weather. It's not just the money we spent on this apparatus that drives us, it's the need for affordable instrumentation that's keeping us involved. Our weather interests will outlast the existence of the vendors of these products and devices. But aside from building our own systems, we were willing to try to make these solutions do what they were supposed to do.
If this effort is translated into a company's budgets, senior management would take notice very quickly. That's why engineers usually don't make good business people - we love the technical challenge more than we love making money, and business people know that. Aren't technical people silly that way? But let's make sure the business people don't ignore our contributions either.
Hope all this helps. Discovery is often built upon previous revelations, so I give credit to all those that are independently working and sharing to make this a reliable solution - in spite of the manufacturer and their distributors. I hope the product vendors are doing more than just listening.
Anyway, have fun and remember: weather study is, after all, a hobby for some but it also is a critical decision variable for others. We are drawn in these directions as a means to understand, respect and protect ourselves from the dynamics of nature.
Happy Holidays to all."