Tank Monitor Project at ECR

[Darryl&Rita]

Plumbing for Waste Tanks

This may be a little less conventional, but it's working for me and seemed to be the easiest--I just drilled into the ABS drain lines upstream of the gate valves and used a 1/8" pipe thread tap to thread the hole. This seemed a lot easier than trying to install a tee and adapt all the way down to 1/8".

This is probably a better idea than T and bushings as far as cleaning the sensor face. I won't stick my neck out and offer any guarantee on accuracy life, that's a fool's game. But I will say I'd install with a dry sensor face, rather than wet.

[Jim Gell]

David,

 

From my naive perspective, it sounds like there will be computer related and technical aspects of the project that I will need a lot of help with. And there would also be plumbing and wiring tasks that will take time, but can be done using skills I already possess.

 

Your challenge is to teach to the lower skill level of the group without disappointing those who have higher skills and knowledge going into this project. In that light, I don't want to tell you what to do, but it is possible that most if not all the participants are going to have basic plumbing and wiring skills like I do. If you are not already, you may want to spend a majority of your time with us at the rally focused on the technical aspects such as programming options and interfacing with displays and controls. It's hard for me to visualize how much there is to cover and how long it might take trying to teach us old dogs new tricks. It is more of a general suggestion, than a knowledgeable specific one.

 

If you have not thought of this, you might consider setting us up so we have "lab partners". Have the participants who come with a better understanding of technical aspects paired with folks like myself, who don't have a clue. You could teach a section and then walk around and observe and assist as needed while the partners discuss what they just learned to be sure everyone is up to speed before you go on. This would encourage partners to carry on discussing and learning after the class too. That would be better than those hard of learning like myself getting overwhelmed and giving up, or bogging the whole class down while you try to bring the slowest learners up to speed.

 

Jim

[Nuke-E]

Hi Jim,

 

I don't plan to spend much time at all on how to tie in to the plumbing, or how to get 12V power from the fuse box to the board--in fact, part of the reason I'm starting with a variable resistor as the "sensor" is to take all of that out of the picture until everyone has gotten familiar with the software and programming side of things. I'll be bringing code that's "ready to load", but plan to spend a fair amount of time going through it so that everyone understands how to tinker beyond what we do in the seminar. I also realize that not everyone wants to monitor holding tanks, so the approach starting with a variable resistor would be just as relevant to someone wanting to measure temperatures with a thermistor.

 

The good news--and the reason for starting with the Photon as opposed to some other controller--is that it's pretty easy to pick up programming the board. Once you have the board connected to a WiFi network, all you'll use is a web browser. Between the reference guide and other people's projects, there's a mountain of code floating around to use as a starting point for something custom.

 

To make sure I'm not losing the crowd, we'll start with a really simple program that just blinks the on board LED. That will demonstrate the commands to set up and write to a digital output, and cover the basic structure of any program on these boards (which is essentially identical to that of an Arduino). Then we'll hook up a variable resistor and have the LED blink in response to the resistance, and then look at the cloud-based "stuff" to read from the board and publish/log those readings.

 

At that point, even if you go no further, you would be able to monitor the level in a holding tank via anything with a web browser. Obviously, a permanent display has some appeal, and I plan to cover at least a little bit about how to do that, along with some of the basics for switching high power (i.e. >mA) loads.

 

As far as someone to help you out, I think you'll be fine. I'm not planning on just standing and talking--you won't be left behind.

[MsChrissi]

Just looked up the sensor, it is perfect. We recently completed a project and are ready to put it into production, it is a small color OLED display in a bezel with rotary encoder knob, push to select LED and beeper, this runs off an Arduino. It's function at the moment is to take the 0-5v signal from two capacitive fuel sensors in aircraft and give a graphic display of fuel level as well as fuel in individual tanks and total. It has a timer as well that would not be needed in this application. What is nice is it has a very good calibration routine we wrote where it takes in the total then 5 measured amounts an gives a smooth display even with irregular shaped tanks. It also has alarms. I could rewrite the code for water, black and grey tanks. I was going to clone the SeeVue system but this seems like it would do the trick. We are going to make some 24" capacitive fuel probes and finally have an accurate reading on the Volvo, the Volvo fuel gauge in our VNL is crap.

Don't have any pictures of the display right now, crazy busy here but will have to do some. So many RV projects on Red Flyer. We put in an uber cheap three camera system with a video switcher and it worked like a champ, you will love these itty bitty cameras. The switcher is set up with a monitor in the cab, a monitor in the back, each has a three way switch to select left, rear, right views. So simple.

Don't worry, I'll redo this thread o it's own with some pictures and links =)

[Nuke-E]

Sounds great!

 

Your timing is funny--I just updated the code for the tank monitor yesterday to save calibration values in non-volatile memory instead of putting it in the code and flashing it. Since that was with a web-connected board, the simplest (hardware wise) calibration routine is browser based--but I had been contemplating a display project for the national rally! Any chance you'll be there?

 

I'm guessing you're using a cubic spline fit? Or at least that's what would seem to make the most sense, though linear interpolation would be almost no effort to code.

 

We ended up with quite a few people building their own setups. I'd like to do a project that builds on the skills from the first seminar, but also figure out a way to incorporate everyone who's interested that wasn't at the ECR. Maybe have some PCBs made that the Photon, tank sensors, and a display will plug in to (probably an Arduino display shield, just to keep it simple). Still a long ways off at this point.

 

I tried the capacitive sensors (like SeeVue), and even made my own using an Arduino to read the capacitance. It all worked great on the fresh tank, but fell apart on the waste tanks. Too much steel structure for a reliable reading--the sensors were picking up the capacitance to the steel much better than the capacitance in the tank.

[MsChrissi]

Hey Nuke,

Cubic spline would be neat but linear interpolation is the most commonly used and works fine. In aircraft we calibrate 0,1,2,3,4,5, 7, 10, 15. and full because you are only critical during the last 5 gallons each side, our tanks are 22.5 each but the instrument can handle much larger.

For an RV waste tank you are not so much interested in quantity as you are in percentage so maybe calibration points at 0, 25, 50, 75, 100% and linear interpolation between? I also use smoothing/averaging to dampen the readout. For water would it may be better to go with gallons?

I like the pump relay option, wish I had that on my boat way back when.

The 1.5" display goes into a 3D printed bezel and housing, the Arduino has it's own custom interface shield that includes a ribbon connector for the display and screw terminals for I/O, the Arduino and shield go in a 3D printed enclosure for mounting.

I was having good results with the non-contact capacitive experiments but was concerned I'd have the same issues that you did. Though perfect for an all composite airframe. Still thinking these pressure transducers might be nice to try though with the G's pulled and 0-18k ft range the readings might be all over the place. Most pilots fly 0-10k ft and +/- 1G. There is one scheme out there that uses two ports to compare tank readings to a static reading but that only compensates for altitude.

Randi does not want to go to the time/expense of the national rally until the RV is done but I am thinking maybe an overnight in the back of the minivan might do the trick, we'll see.

...Chrissi

[Darryl&Rita]

Curious that the capacitive sensors wouldn't work. I've done them in all steel, poly-lined steel, & poly tanks surrounded by steel. Used both contact & non-contact varieties. They're not plug-&-play, but they're the most robust we've found,

[Jack Mayer]

Chrissi,

 

I understand about the Rally - you guys tend to stay pretty "busy" with your projects. I'd love to see pics of the truck progress if you ever feel like posting any.

 

I'd be a buyer of the Volvo tank readout if you ever want to produce another one - once you get it done. Although I tend to fill before I get to 1/4 tank, it would still be nice to know what I "really" have in there, without using a flashlight and guessing.

[Nuke-E]

If you're just doing a readout, linear interpolation is perfectly fine. I'm working on a few things (like estimates of time to full/empty and calculations of flow rates) where derivatives are part of the calculation, and the discontinuities in the first derivative with linear interpolation can sometimes cause undesirable numerical effects. Eventually, I want to tie the tank level sensing (in a flow metering sense) into logic to detect certain activities--such as taking a shower or doing laundry--to make the water heater a little smarter.

 

I wouldn't worry about altitude in the plane (30,000ft is only about 0.3% change)--acceleration due to the plane's motion though would definitely matter.

 

As far as 3D-printing, what are you using? Ultimaker (with a heated aluminum bed) and PrintrBot in the RV here.

[Nuke-E]

Curious that the capacitive sensors wouldn't work. I've done them in all steel, poly-lined steel, & poly tanks surrounded by steel. Used both contact & non-contact varieties. They're not plug-&-play, but they're the most robust we've found,

 

I'm assuming you're referring to capacitive proximity sensors, where you have a single sensor that either detects water or doesn't. I think the large foil tape area and the steel being electrically coupled to the ground electrode is what causes trouble for the tank monitoring systems.

[Darryl&Rita]

Nope, full on level transmitters. As long as 20' of sensing area, as short as 6". Also set up lots of single point switches, into the sides of vessels.

[Nuke-E]

Nope, full on level transmitters. As long as 20' of sensing area, as short as 6". Also set up lots of single point switches, into the sides of vessels.

 

I think to do that is definitely possible, just not in a sub-$200 package (for 3 sensors) like what's out there now. I thought my real simple DIY sensors (charging and discharging using an Arduino, calculating time to voltage) wasn't sophisticated enough, and tried the commercial version, but it was pretty erratic. I'd also be curious what voltages you were using--24V possibly?

 

To be perfectly honest, between trying my own and the RV-targeted product, then having the light bulb go on as far as reading hydrostatic pressure, I never really looked at industrial sensors. It'd be interesting to see if any of them could be implemented at a price that would be reasonable (as compared to looking at the FW tank or down the toilet).

[MsChrissi]

Jack, we're hoping a few might be interested in a real readout for the Volvo tanks. We ran out once, don't want to do so ever again.

 

Nuke, we're using a Rostock MAX delta bot from SeeMeCNC. we needed the potential large build area and speed plus needed it in a kit form. Great kit, great support. I'd never have bought one to print cutesy things, we do not need Yoda heads, police call boxes or tacky plastic jewelry but we really needed a way to prototype, make parts, enclosures etc. It does the trick nicely.

 

We're also running the elevator with an Arduino, gives safety interlocks etc.

 

...Chrissi

[Darryl&Rita]

 

I think to do that is definitely possible, just not in a sub-$200 package (for 3 sensors) like what's out there now. I thought my real simple DIY sensors (charging and discharging using an Arduino, calculating time to voltage) wasn't sophisticated enough, and tried the commercial version, but it was pretty erratic. I'd also be curious what voltages you were using--24V possibly?

​Industrial environment, 24 volt DC. Probably $200 of set-up costs. Either 4-20 mA output, or 1-5 volt

 

To be perfectly honest, between trying my own and the RV-targeted product, then having the light bulb go on as far as reading hydrostatic pressure, I never really looked at industrial sensors. It'd be interesting to see if any of them could be implemented at a price that would be reasonable (as compared to looking at the FW tank or down the toilet).

[Nuke-E]

Nuke, we're using a Rostock MAX delta bot from SeeMeCNC. we needed the potential large build area and speed plus needed it in a kit form. Great kit, great support. I'd never have bought one to print cutesy things, we do not need Yoda heads, police call boxes or tacky plastic jewelry but we really needed a way to prototype, make parts, enclosures etc. It does the trick nicely.

 

We're also running the elevator with an Arduino, gives safety interlocks etc.

 

...Chrissi

Sounds like you're having fun! That's a nice printer--I'd be curious how well it does printing taller objects in ABS without a heated environmental chamber. I have all of the stuff to build my own CNC router table and printer, but it's been in boxes for about a year already--I've wanted to reconsider a couple things on my design to try to keep large parts warm better as the print progresses.

 

On the fuel tank thing, are you guys having that much trouble with the stock gauge or do you guys push it more than I do? Or is there something else going on, like water in the tanks? I mean, I've gone a good 20-30 miles past 0 miles to empty a few times...

[MsChrissi]

Our Volvo gauges say empty at about 3/4 full, not worth fussing with at this point when we have a design that can interpolate a cylindrical tank closely enough. We used to stick the tanks and compare to a spreadsheet I printed out. With 250 gallons of greasil on board we can just barely make the round trip from StL to Lakeland Florida. Just barely. We found out the hard way that if we run the Rigmaster much it can tip the scales from just barely to improbable so we stop in Paducah on the way home. StL fuel prices are always cheaper.

 

People address the ABS printing enclosure with full Ikea closets or plexiglass (with 3D printed brackets) or in our case I put a wrap of saran around the lower part just to keep drafts from messing with me. We have one guy up in WA state printing casting patterns of main landing gear parts for a group restoring a Spitfire in Australia, big parts.

 

So far we've CNC'd a 9x20 lathe, a CNC Sherline mill, retrofit a Bridgeport CNC mill with PC based CNC controls, converted a Grizzly Tools 12x36 gunsmith lath to CNC with a 6 position tool turret, bar puller and pneumatic collet (Microkinetics conversion) and just finished cutting up metal for a CNC plasma cutter based on a Precision Plasma HD gantry, Hypertherm 65Pro w/CNC interface to CandCNC turnkey electronics package with Linux controls. We have run out of room so are building the plasma table like a travelift you would use to lift yachts. We can raise the table, move it into the barn and park the Kabota and motorcycle under it =)

[Jack Mayer]

So far we've CNC'd a 9x20 lathe, a CNC Sherline mill, retrofit a Bridgeport CNC mill with PC based CNC controls, converted a Grizzly Tools 12x36 gunsmith lath to CNC with a 6 position tool turret, bar puller and pneumatic collet (Microkinetics conversion) and just finished cutting up metal for a CNC plasma cutter based on a Precision Plasma HD gantry, Hypertherm 65Pro w/CNC interface to CandCNC turnkey electronics package with Linux controls. We have run out of room so are building the plasma table like a travelift you would use to lift yachts. We can raise the table, move it into the barn and park the Kabota and motorcycle under it =)

See, David, they ain't messin' around.....

[Nuke-E]

See, David, they ain't messin' around.....

 

No, they aren't! CNC stuff (back when the controls were in big cabinets, not PCs) was were where I got my EE start--spent a lot of time with one of these control panels. I remember when an Allen Bradley 8400 control was fancy new stuff! My first "real" electrical engineering project was getting a gantry machine set up to run with (back then) a really new PC-based controller. Full-length ISA card in the PC, 100-pin ribbon cables, etc--and mating it to motors and drives that it hadn't been used with previously (so neither manufacturer knew the other's product). That's all been replaced with software motion controllers, and it's far, far, simpler now for sure.

 

If I were to start from scratch, there's so much open source hardware and software that effectively does for free what would have been >$10K not that many years ago. Stuff like LinuxCNC, Smoothieboard, and all of the Arduino-based stuff (e.g. RAMPS) make it pretty cheap and easy to walk in to at least a stepper-based multi-axis motion controller.

[NeverEasy]

Whew! You folks are way out there on this stuff. Since coming home, I started working on the Tank Level Indicator with the Photon. Calibrated it to the variable resistor that David gave us at the rally from CCW being near zero and CW being max at 3.28 vots.. Had problems with the calibration as the Photon was seeing analog voltage levels different than was being seen by a meter on pin A7. Figured that out by publishing "tnk1" to the Dashboard Log. It was seeing 4092 mv. ??? Once I put that in tnk1full calibration was easy.

 

Then I set out to light a series of 10 LEDs. I used D0-D7 for the first eight which worked fine. Turning the pot up and down shows one light at a time for 10% voltage increase.

 

Now the problem: A0-A6 will not digitalWrite HIGH. The specs say that they can be used for GPIO. The simple Blink AN LED1 code would not turn on an LED either. Is there something I am doing wrong or I am going to have to live with less LEDs? I wan to reserve one GPIO for an audible alarm at 80%.

 

Help!! Chet

[MsChrissi]

David,

Smoothieboard, now we're talking! I only have so much time to play with this stuff but if we ever get bored that will be the next upgrade to the deltabot, after we install the E3D hot end upgrade that has been sitting on the bench for over a year..ugh so many projects so little time and they call this retirement?

[Ronbo]

I found that I am busier in retirement than when I worked. Much more so now that we are running to radiation and chemo.

[MsChrissi]

I found that I am busier in retirement than when I worked. Much more so now that we are running to radiation and chemo.

Sorry to hear about the treatments taking up your time, that's the state a very close friend is in now and she says what time she has is spent either very uncomfortable or sleeping. Projects keep my mind from turning to mush. I always struggled with programming when I was younger so at 60 some years ago I decided to take it up and master it so that I could do various control related projects. Best thing I have done to exercise the brain muscle.

[Nuke-E]

Whew! You folks are way out there on this stuff. Since coming home, I started working on the Tank Level Indicator with the Photon. Calibrated it to the variable resistor that David gave us at the rally from CCW being near zero and CW being max at 3.28 vots.. Had problems with the calibration as the Photon was seeing analog voltage levels different than was being seen by a meter on pin A7. Figured that out by publishing "tnk1" to the Dashboard Log. It was seeing 4092 mv. ??? Once I put that in tnk1full calibration was easy.

 

Then I set out to light a series of 10 LEDs. I used D0-D7 for the first eight which worked fine. Turning the pot up and down shows one light at a time for 10% voltage increase.

 

Now the problem: A0-A6 will not digitalWrite HIGH. The specs say that they can be used for GPIO. The simple Blink AN LED1 code would not turn on an LED either. Is there something I am doing wrong or I am going to have to live with less LEDs? I wan to reserve one GPIO for an audible alarm at 80%.

 

Help!! Chet

 

Hi Chet,

 

The numbers you were publishing are the analog reading on a 12-bit scale, i.e. from 0 to 4095. That corresponds to 0 to 3.3V, so a reading of 1V would correspond to 1241.

 

As for the analog outputs, you should be able to make it work with analogWrite. Just be conscious of how much total current you're pushing to the LEDs.

[NeverEasy]

Thanks, Dave. Clear as mud. I have a lot to learn but this is fun.

 

After thinking about it and having read MsChrissi's post where "In aircraft we calibrate 0,1,2,3,4,5, 7, 10, 15. and full." I think I will do something like that for the fresh water tank and the reverse for the gray and black. That will limit the number of DigitalWrite pins and make the rest available for other things.

[Nuke-E]

One thing you might take a look at if you want to cut down on the I/O for display is an OLED panel--I think MsChrissi is using something similar (they're really common, and super cheap bare) to this, but with your photon and the display mounted on the board, you don't have to do nearly as much work. There are also holes for soldering in leads for your sensors. It'd be really small, but you could display a graph from each tank at the same time.

 

My project for the weekend is linking the thermostat and the photon that's controlling the lithium battery system--so I'm charging when the A/C and water heater aren't running, and won't really need 50A power.