Anyone build the Plant Watering Monitor from Feb Q&

[Khan505]

Hi All,<p>I am a newbie. I picked what I thought was an easy project to help build my skills but I am being thrown for a loop here and can use some help. Any input is appreciated.<p>In the Feb Q&A section of Nuts and Volts there is a project for a Plant Watering Monitor. Build the project as is turns up poor results for me. I have rebuilt it from scratch 4 times and had two people check my breadboarding to ensure I have done that correctly. The problem? The "thirsty" light is the only LED to flash and the flashing times are backwards from what the text indicates should happen (4 secs On and 1/3 sec off instead of 4 sec off 1/3 sec on). Connecting the probes, varying the 100K V-resistor, or doing several combinations doesn't change a thing between the "thirsty" or "happy" LEDs. I have check all parts to ensure they are good. At this point it looks like the design to me.<p>I swithed the postions of the resistors and added a diode across pins 2 and 7 of the 555 timer to produce the desired pulse but I cannot figure out what is wrong with the caparator wiring.<p>Has anyone sucessfully built this project? If so please contact me so I can try to fix this myself. I believe this to be a worth my time and trouble to troubleshoot to learn from this.<p>Thanks in advance.<p>Khan

[dyarker]

I haven't built it, but refering to the National spec sheet for the LM555, Applications Information, Astable Operation and Figure 5:<p>Connect as in the schematic in the magazine; except that the 10K (brn blk org) should go from supply to pin 7, and 100K (brn blk yel) should go from pin 7 to pins 6 and 2.<p>The output (pin 3) low time ratio is set by RB (pin 7 to pins 6 and 2). Output high time ratio by RA plus RB. RA and RB are reversed in the magazine.<p>Depending on which way you added the diode, it probably only made the timing worse. Leave it out. It could even blow the pin 7 transistor because there would be no current limiting while discharging the capacitor.<p>Now the 555 output should be about 0V for 3.5 sec, and almost supply for 0.35 sec.<p>---------------------------------------------
Thirsty LED only is a separate problem.<p>With the soil probes open (not in soil and not touching), the Thirsty LED should blink.<p>With the soil probes shorted (touching), the Happy LED should blink.<p>If the reverse happens, then the inverting and non-inverting inputs of the left LM339 section need to be swapped.<p>If still only Thirsty LED, double the potentiometer connections. If still no joy, then somebody (maybe me unless someone else is faster) will need the LM339 spec sheet and a calulator. But by quick eye ball check, the magazine schematic looks okay. Let us know.<p>**********************************
added after Philba's post<p>I missed the lack of common reference on the probe cicuit. And testing without the 555 section is a good idea, they are separate sections; wish I'd thought of it.
************************************
Cheers,<p>[ February 26, 2005: Message edited by: Dale Y ]</p>

[philba]

I recall looking at that circuit thinking it doesn't make sense since there's no ground reference for the inputs - how do you get a differential voltage to compare against? Basically, the 4 10K resistors and the 100K pot are supposed to be forming voltage dividers to create 2 differential voltages to compare.<p>When I built a moisture sensor, I used two probes as a variable resistor (infinite if dry, about 200K, if wet) as part of a voltage divider to feed the comparator. One probe connected to ground, the other probe to the input of the comparator and pulled to +V via a resistor. The reference voltage into the other comparator input was set up to be adjusted in between the voltages of the two cases (wet and dry). <p>Try something like this.
<p>This presumes that the resistance of wet soil is fairly low. When the soil is dry and probe resistance is infinite, R4 pulls the "+" input to +V and above the reference on "-". When the soil is wet, the "+" input will be pulled towards ground - Vprobe = V*Rprobe/(Rprobe+100K). Rprobe is the resistance between the two probes. Vprobe is the volatage on the "+" input. I'd measure the resistance of wet soil and pick R4 to be around that to make the wet voltage be 1/2 of your supply voltage.<p>edit - take the 555 out of the mix for testing purposes. connect the rail that comes off pin 3 to +V (4.5 if using 3 batteries). once you get it working, put the 555 back in.<p>[ February 27, 2005: Message edited by: philba ]</p>

[Edd]

WOW…when I saw this post Fri eve and went down to the LRC to see the article, it was just sitting there with 0 replies, now it has drifted waaay down the list..getting two assists on its way. I’m just putting what data I had put on flash drive and he can make use of any of our info….upon return. <p>NV_Plant_Moisture_ Post_02252005_1120Z<p>Chau om, Khan nam tram nam... a viet ?<p>In response to your query on having problems with the unit that you constructed I took a quick look at the info in the N&V issue in our LRC and took quasi-mental note of its circuitry. The right three sections of the 339 are being uitilized in a window comparator configuration with the bottom sector referencing to an above ground reference for the battery depleted function. The outputs of the other 2 comparators feeding LED indicators for referencing their high state(s). The sole 339 unit left, being used over at the moisture probes input interfacing circuit.
To easier troubleshoot the circuit that you already have built up, initially how about bypassing the pulsing B+ power supply provided off # 3 of the 555 circuitry by lifting that wire and tacking to +4.5VDC for a steady state voltage . Then, looking at the output of the first 339 and its 10 k pullup resistor, lift the wire at that point that feeds signal over towards the comparator trio to the right. It normally would send a voltage sample to the input of the comparator cluster to be responded to in their outputs. Instead lets subject the cluster to a rising voltage sample to initially confirm that the comparators LED driving circuitry is properly responding and at what thresholds, or even responding at all, in case of a def component or miswire on your construction or failure initiated by swapping around of wiring/parts, (if that has been done).
Being totally blind at this end, in the respect of not knowing what facilities/components you have, two suggestions for doing this. If you have a variable DC power supply 0-5VDC (minumal current capability) its just a matter of placing a protective/current limiting 1K resistor in series with the DC output lead and it then will then connect to the floating wire terminating at the comparator inputs while the power supply will go to circuit ground. Then you would slowly start at O Vdc level and bring up the sample voltage until LED activation and log in the voltage level present at the comparator input stage at which the activation occured. At the 4.5 Vdc level start slowly decreasing the sample voltage and log in the thresholds of change at the decreasing voltage applied.
Considering your not having a variable power supply, get a 1 K potentiometer at RS and it can be used to divide down your 4.5 Vdc voltage for the sampling voltage, as less than a ma will be required to inject to work against the set threshold levels. Viewing from the shaft end of the pot, its center term will connect to the 1k c-l/prot resistor that provides the sample to the comparator cluster. The CCW most terminal goes to circuit gnd, while the CW most terminal goes to +4.5Vdc. In this manner the applied sample voltage can gradually increased with a clockwise shaft rotation. Go thru the aforementioned sampling procedure in the same manner as would have been done with the use of a variable power supply unit to
establish the the switching threshold voltages. Since the voltage trip thresholds are now apparent, the next step would then be the checking out of the first comparator stage to confirm that its level shifting is happening when the probes are subject to moisture/dry conditions. This might be simulated with a one meg pot, using its center terminal and its ccw most terminal, by the connecting of those terminals across the sensor electrode connections. HOWEVER, there is one disparity, as the schematic is drawn, that should only produce a variance of full supply voltage down to ~4.25Vdc at the reference and will not be of the proper level to initiate a comparator level shift. I believe that a schematic error has crept in and that the electrode shown connected over to the comparator reference terminal that the adjustable trim pot is connected to should instead be disconnected from that point and then rewired. That electrode will be connected to a series isolation resistor of 10K (if keeping in the order of bias isolation afforded by circuitry pull up resistors)and that resistors open end goes to circuit ground. The other electrode that is connected to the comparator input that is without the pot stays as is.
If you will alter the electrode to that configuration and then simulate degrees of variable moisture conditions by making a slow 1 meg downwards resistive sweep with the pot across the electrode connections, there should be a comparator level shift on its output at ~ the 300k ohm res range of the pot..... all in accordance to the voltage reference worked against at the setting made by the 100k trim pot . You will have to experiment with setting the threshold of detection in accordance to the prescribed minumal degree of dampness your plant/planting medium will tolerate. Not being near the tech info, I forget if any mention was made to the probes construction .There would be several variables on their conductivity. The fact of whether they were rods or metal strips (surface area), interelectrode spacing, potting material type, presence of fertilizing salts and minerals would present different conductivity samplings. The most of those could be minumized by the use of limited exposure/ fixed dimension electrodes mounted to an insulative plexiglas spacer.
If there was a proper level shift of the front end comparator section of the 339 with a resistive shift applied across the probes, now the wire that ties the input stage to the comparator cluster can be reconnected and the pots resistive sweep across the electrodes reinitiated to see if the unit now responds in the proper manner to simulated moistness. Passes? then remove one cell to run on 3 Vdc to see if the low voltage presence is indicated by the L/V LED ......or possibly ??? this might still be too high of supply power and the use of the/a variable 4.5Vdc adj power supply might be needed to supply system power and creep down to that threshold. All ok ? then restore the derivation of system power back to the pulsing output of the 555 at pin 3....considering that circuit to be working OK..... seems like u said that the LED’s were blinking initially. <p>
73's de Edd
[email protected] ..….....(Interstellar~~~~Warp~~~Speed)
[email protected]…......(Firewalled-Spam*Cookies*Crumbs)
<p>I intend to live forever - so far, so good.

[philba]

Interesting post - kind of dense so its a bit hard to read.<p>My curiosity got the better of me and I put it into a SPICE simulator. I could not get the first comparator to work the way the circuit is drawn. However my circuit (posted above) did work in the simulator. By work I mean infinite resistance (ok, 20M ohms) caused the first comparator to output around 4.5 V and 50K resulted in a little under 300 mV. The original circuit with the same probe values (20M and 50K) resulted in an output of 4.5V for both cases. The circuit as written doesn't work.<p>[ March 02, 2005: Message edited by: philba ]</p>

[tarnall]

It would be nice if someone could redo the original N&V schematic so it is working right with explanations for setting up and testing.<p>73
Terry - KB6EBZ