A point often overlooked by many.
Even premade links and jumpers can have bad crimps where the crimp is made on the insulation. Makes contact for a while then fails. I think sometimes the operator doing the crimping things the crimps are insulation displacement type connections. They are not.
Do not discount the el cheapo breadboards either. They are often a push in once then forget it type. Especially if as so often happens the pins on header strips are inserted on say “breadboard friendly” things like Pico and the like. Sure they are on a standard 2.54mm grid but that is where it all ends. The pins seem to destroy the spring contacts on some of the boards floating around. One insertion of these pins and that is it. I have discarded 3 so far while still using the “WISH” board I have had for 12 or 15 years. So old the plastic is now yellow.
I have seen projects where the final product is constructed on one of these boards and I shudder when I look. Unless the plating is of extremely good quality it will tarnish over time and possibly become unreliable.
Cheers Bob
Hi again Dave, just for interest, I did a 45 min warm up of the sensor and went back numerous versions of my sketch and used a 12V DC switch mode power supply to the sensor (no Jaycar kit).
Here are the results, 1 sample per minute:
So, I’ll be using the Jaycar kit, an AC supply to the Jaycar kit and running median.
Time for me to finish the MQTT and HomeAssistant integration so I can finish the proof of concept and get this out into the field.
Those figures are terrible. What is the actual measured depth?
I’d like to know what the actual technology of the sensor is and be looking for a different type of sensor.
FYI some depth sensors need a ‘breather’, is there a small tube in the cable that should be open to outside air?
perhaps i’m clutching at straws here… is there any source of vibration anywhere? Pumps, traffic etc?
I’ve heard that vibration can seriously impede other types of tank level sensors. I’ve never used a pressure based one before but I’m curious if they are affected as well.
Hi Dave
It works OK on battery according to Matt.
Matt also has never said exactly what his original power supply was. Could be useful information. Switching supplies, particularly ones that have some left over switching artefacts like battery chargers which don’t care much, can do some funny things which is put down to “noise” ( a good all round term) but is usually due to the inductance of fairly long thin wires. This can be reduced by a) using wires as short as possible or b) using larger diameter wire or indeed both.
By his own admission Matt has done some research
and I would agree to that. And he has seen good results by using an AC input to that Jaycar kit which is exactly that. A linear power supply.
I still maintain that had he wired up that simple 7812 regulator I sketched he would have seen some results. That Jaycar kit is apparently configurable for a DC input and single supply. It seems pretty versatile but you have to assemble it to what you want. You can’t have everything at once. I think this kit configured for DC input and single supply would be basically what I have sketched.
The 7812 and similar regulators are very good at “cleaning up” power sources, that is why I suggested trying one. You can actually use a simple circuit using just 3 components which works. 1 transistor, 1 resistor and 1 zener diode. Cheap, cheerful and quick.
But Matt seems to have decided not to try my solution so really have not much more to offer at this stage so will leave all the running around to others.
Cheers Bob
Hi Liam, a running median filter works a treat……by itself….it seems to interfere with mqtt & HomeAssistant integration ……I’m struggling to successfully integrate the code. Might try another type of low pass filter.
Bob, I agree, but given the price segment we’re playing in, I’m thinking that the sensor is lower quality, the current to voltage converter is also lower quality, so there are probably a number of contributing factors here.
I’m happy with the bandaids and the associated risk.
Hi Matt
I have 1 last idea.
You stated previously the sensor is 12 - 36V. You have been using a 12V supply of some sort. When you used a battery you had a vast improvement.
Now if your battery was fully charged and was a lead acid type it could have been somewhere north of 13V. If your supply of a nominal 12V was a bit marginal I wonder if being slightly lower or on the border line of the sensor operational range would be the cause of your troubles.
Can you operate with a higher voltage like 15V or just use your 19V laptop supply and see what happens. Sometimes some queer things can happen when you are operating right on the fringe of the stated range with no “fudge factor”.
Cheers Bob
Thx again Bob. The Jaycar kit came as 15V as standard, plus I had the same thinking as you, so have stayed with 15V (and the 12V regulators are still in the plastic bag from Jaycar!).
Hi Matt
So that would seem to eliminate that idea.
How did you configure the Jaycar kit. Single supply @ 15V would be the idea. I still have not seen the schematic or the instructions for that unit. Did you actually measure 15V output.
I have been subscribing to the on line version of Silicon Chip for some years now but the original of that article was a bit before I commenced so I don.t have access to the complete issue, just a preview. Unfortunately I have had to discard all those earlier magazines so I have thrown that series out.
Cheers Bob
Hi Matt
Can you post a schematic of EXACTLY what you have. including the part numbers of all the bits.
There is something funny going on here.
Cheers Bob
