Tank Monitor replacement - digital

[RichHog]

This is what I'm talking about. If the whole thing gets replaced then might as well put a clock in it and Battery level meter

[Zeppelinium]

Originally Posted by RichHog

...M7 72 TW only has the water pump switch work, which is the only critical must have component, tank levels, battery level, panel light all shot and the plastic is a little ugly. So, how do I replace the whole gizmo, not just electronics behind the face plate?

That's the objective--replace the whole panel with one or two panels that fit exactly in the rectangle of the bezel that you have. My 70 Caravel and 72 Overlander have exactly the problem you have and the same bezel/panel. The general scheme is loosely depicted in this drawing:



But here's the rub. The early 70s tanks only have three level sensors (4 wires, one as the voltage to the tank and three return lines), while the later models, the ones with the meters, have 4 sensors. In addition, the early 70s Airstreams only have two tanks--no gray water holding tank, just black and fresh water tanks.

So there are three possible configurations of the tank level monitor:

1. Two tanks with three levels, for owners who want just a stock replacement monitor.

2. Three tanks with three levels, for owners who want to or have added a gray tank to the before-1974 Airstreams.

3. Three tanks with four levels, for owners who want to replace the bulky meter panel (which I'm going to do in my Sovereign and Safari--anyone want the old meter panel?). The meters really aren't that nifty, since they don't provide a continuous, eg, gas-gauge in your car, indication. It's more step-wise, as so (the meter indication is the reverse of this):



The battery volts and amps can be provided by commerical modules for about $20 each. These can be wired to the on-off-momentary switch for the tank monitor. Temperature indicators are available for $12-ish, but might be difficult to integrate into the panel--maybe the right solution is to provide a clip to mount it proud of the panel, but so that it sorta looks like it belongs.

Anyway, the above three situations can be handled with two versions of the monitor and two front panels. Speaking of panels...

Coming from the Audio/Video install arena, we often custom punch out metal plates and then engrave the faceplate. Is that the next step to make this work? Has anybody out there already done this? I for one would love to have a nice state-of the art display panel there with some sexy lights and LEDS.

What does a panel cost if they are done in onesey-twosey kinds of numbers? Can these be done in 0.060 or 0.125 aluminum or is there a plastic that would be suitable? We can take a detailed discussion off line, but I'm interested in whether the labels for the tanks, eg, BLACK GRAY WATER, would be engraved in the panel or would be a backlighted open window, etc. Right now my biggest issue with the panel is the spacing of the LEDs. They are 5mm diameter and I wouldn't want them to have more than about 2mm of space between them in a finished product.

I'm open to bartering a monitor ...

Zep

[Zeppelinium]

Originally Posted by 2xS

... My real question is: in your experience, did you notice the output voltages of the sensors to be more static or analog? Is it essentially just switches in the tank or is there a variable resistance to play with?
...

Here's data using the same resistor ladder and the same faux sensor board in three different glasses of fluids. It is interesting to see that there was almost no voltage variation in the black water as the sensor board was slipped up and down, compared to the other liquids. I was also somewhat perplexed that the voltage jumped when just the driven sensor hit the fluid. I wouldn't have expected any change until two sensors were in the fluid, since the circuit isn't complete until then. What have we here, a battery?



Even with the variability, the voltages are monotonic. As soon as the weather improves, I'm going to get real data from the tanks in the Overlander (water and black, no gray). I expect some difference will show up due to the tank size and sensor spacing.

I calculated, based on the data from the glass of fluid, that the equivalent impedence of the fluids are:

Water - 8.2K
Gray - 5.1K
Black - 3.3K

Zep

[Zeppelinium]

Yippee, finished! And much simpler, don't you think?



Zep

(apologies and thanks to XKCD)

[RichHog]

I would add a decent sized LED clock to the panel as well. The av ones are black metal with silver etched lettering usually 19" long to fit standard cases and equipment racks, also 1-6" high

[barts]

Cool - I've not yet dug into why our '71 Tradewind's sensors/indicator doesn't work; this has started me thinking... thanks for the info on how these sensors work.

One could also use a PIC microprocessor to convert the analog readings into digital values, and then control the output LEDs directly; this would let one tweak the settings via a PC. http://sparkfun.com is a fascinating supplier of interesting semiconductor stuff for hobbyists; the sensor devices are particularly sophisticated (GPS, carbon monoxide, accelerometers, ...) . The magnetic level sensor seems particularly useful....

One can imagine a small display cycling through the available data (voltage, battery current, tank levels, etc).

As far as making panels goes, 1/8" 6061 T6 is really easy to work in the home shop. Cut shy of the line, and file to final size; it's easy to hold a few thou w/ a fine file. Get a hold of some layout dye and layout the holes, etc, on the back side of the panel. Center punch the holes, and drill on a drill press. Square/rectangular holes are easily filed from round, too.

- Bart

[Zeppelinium]

Originally Posted by barts

... The magnetic level sensor seems particularly useful....

Do you have one? I can't tell anything about how it works, how to install, how to adjust for various depth of tank, etc. It would be great to have a continuous analog level reading!

Zep

[barts]

Originally Posted by Zeppelinium

Do you have one? I can't tell anything about how it works, how to install, how to adjust for various depth of tank, etc. It would be great to have a continuous analog level reading!

Zep

I don't have one, but from the description it's an self-adhesive potentiometer, with the wiper in the form of a magnet. You stick it to a tube that contains a float w/ a magnet inside of it, and the small magnet inside the pot. tracks the one inside the float that's inside the tube. Plumb the tube top and bottom to track the level inside the tank and you're set.

- Bart

[Zeppelinium]

Ugly monitor panel out...(avoiding any discussion of the "electronics")





New monitor panel in...





You can see the prototype is fairly far along. It copies the original angle and will do fine, but the face seems a bit too recessed and the angle could be somewhat less, say 12-15 degrees. For my standing height (the exact average man, circa 1975), the panel is looking over my head.

The volt and current meters are $7 units that require an isolated 9V battery for power. I was hoping to use power off the monitor board, but that got too complicated. The saving grace here is that it only takes two screws to remove the unit from the dome, so getting at the battery won't be too big a problem. Besides, the thermometer will also need a battery change.

Speaking of the thermometer (and always looking for the cheap solution) I modified this popular unit to get a flat face on it so it could be pressed up against the back of the panel. You have to sand off the built-in light lens and remove the button switches (these can easily be remoted outside the unit).



You can see where the thermometer (about $11) fits into the panel. It turns out the thermometer LCD display is the same height as the volt and ammeter displays. Lucky day. The final panel should look something like this (laser cut from 3.32 plastic (blue on surface and white inside, making white lettering). I've been told it will cost about $25, which is too low to believe.



Zep

[RichHog]

Looking Good!

[Zeppelinium]

The weather has actually acted like spring yesterday and today--motivation to get out of the lab and take some real data. Here's what the fresh water tank looks like. Note the visual estimate of where the sensors are versus the tank draining time (yes, I did sit there for 76 minutes, twice, to get the data). Ack. The good news is that the sensor performance and measured voltages were very well behaved. I may need adjust the resistor ladder a bit to get the voltages right smack dad in the middle of the detection ranges, but I'm pretty happy with the results.

I whipped up the test circuit so I could get the data without having to install the whole monitor panel. The test circuit exactly duplicates the input side of the tank monitor, except that the excitation voltage from the battery is 9.6V versus the nominal 9V from the monitor circuit. I normalized that difference out in the chart.





Note the odd spacing of the sensors (estimated positions are described in the boxes in the chart). It's difficult to get an eyeball in position to even see the suckers, much less measure them (probably could have thought more about this when I built the cabinets), so the positions are just eyeball estimates. The drain time may be a better indication of what level the sensors are really reporting. Need I repeat that I really think the drain diameter could have been better thought out when the tank was installed...

Black tank data tomorrow. It will only be two points, since the original early 70s black tanks only report two levels and I don't want to contemplate slowly filling the black tank...

Zep

[Zeppelinium]

Ready for Albuquerque, finally. The last big discovery today was that the water pump puts alot of trash on the 12V line and will fry unprotected ICs. So another day goes by not doing woodwork, but now the monitor has a fairly robust filtered power supply.

I'm going to try out some night lighting. The red is a little too "submarine" but it's fine for the prototype. The "production" model may have soft white or amber. These lights are any time the monitor is on, although it would be easy to put in a push button for on demand illumination.





The LEDs a in the 1/2" maple frame, aimed back at approximately a 60 degree angle.



Onward and upward...

Oops, one more thing. Gotta make a shunt for the ammeter for battery current. One foot of #10 copper and a couple of heavy duty wire cleats should do it. We'll see. What you want is a very high current resistor that provides one millivolt per amp.

Zep

[RichHog]

how many 6 packs of beer have gone into the R & D on this thing? Maple wood framing? Very cool!

[Zeppelinium]

Originally Posted by RichHog

how many 6 packs of beer have gone into the R & D on this thing? Maple wood framing? Very cool!

two-buck Chuck, my brother. I drank all the beer I ever needed back in SEA. Not to mention losing money playing poker.

The good (or should I say great) news this morning is the shunt works waay better than expected. Calibration is a bit of a challenge, since it's hard to find 1% precision power resistors in the 1-4 ohm range. Never-the-less, I think it's close, within about 5%. I used a 6 amp/12V power supply and two 4 ohm/50W resistors and two meters to get the best readings I could.

The shunt is pretty basic. If you failed to retain the fuse block from your old Univolt, which had a nice copper plate shunt on it, you can make one for $2.08. If you use #10 wire, the wire tables say you need 12.01 inches of wire for 0.001 ohm (roughly speaking). I wound up using 10.25 inches. If my calibration is off, the shorter wire means the meter will be showing less current than is actually flowing.



Used an scrap piece of plastic from my Sovereign tank welding experiment as the backing piece. A dry piece of wood would work fine, since the voltage across the terminals/lugs will always be less than 100 millivolts.

The electrical installation is a bit counter-intuitive for me--I'd have thought these things go in the positive lead of the battery. However, putting it in the negative (common) lead ensures that the wires from the voltmeter are essentially at ground potential. This floats the battery above ground just slightly.



The fuses in the voltmeter wires are problematic for me. It's ultra-safe, but it would take a short to ground in the wire on the high side of the shunt, close to the shunt, like 12" or less, in order to drive enough current through the small wire to be dangerous. The voltmeter wires have approximately 8 times the resistance of the #10 wire in the shunt, so the preponderance of battery current would contiue through the large wires (unless the battery common becomes open below the shunt). The original fuses are 8" down the wire. This 8" is the most "dangerous" part of the length, so the fuse on the other wire protects this short length. So, there are scenarios where the fuses could be helpful. Vanishingly small scenarios. But I left them in the circuit. You could be just as safe with one fuse in the low side wire (any other insight would be appreciated).

How well does the "amp meter" (the combination of the 200 mv voltmeter and shunt) work? Great! I was concerned that the long wires from the back of the Overlander to the front dome would be noisy. The meter is 3-1/2 digits, so it measures down to 0.1 mv, or 100 milliamps of current. With just a few minutes of observation, the meter jumps around a bit in the tenths digit (1 count only). It shows the water pump pulling 6.5 amps in open flow, ramping up to 7.4 amps as the pressure builds up to shutoff. It also shows turning the porch light on and off (less than 100 ma), and the radio (400 ma). When I get a long enough arm, I'll check what it shows for the hitch jack.

I just can't believe it's this stable and precise. The old analog meters were essentially useless.

Zep
(playing Meatloaf and Joe Cocker--happy happy)
saving the folk music for another day!

PS--I forgot to mention up in post #23 that the temperature gauge looks good--it's right up flat against the panel, looks like it was designed just for this.

[barts]

Originally Posted by Zeppelinium

How well does the "amp meter" (the combination of the 200 mv voltmeter and shunt) work? Great! I was concerned that the long wires from the back of the Overlander to the front dome would be noisy. The meter is 3-1/2 digits, so it measures down to 0.1 mv, or 100 milliamps of current. With just a few minutes of observation, the meter jumps around a bit in the tenths digit (1 count only). It shows the water pump pulling 6.5 amps in open flow, ramping up to 7.4 amps as the pressure builds up to shutoff. It also shows turning the porch light on and off (less than 100 ma), and the radio (400 ma). When I get a long enough arm, I'll check what it shows for the hitch jack.

I just can't believe it's this stable and precise.

Nice work!

The panel in our '71 Tradewind is mostly non-functional; I look forward to getting the basics on the trailer working so I can try my hand at this sort of thing.... I'm torn between something completely digital based on a Mimo touch screen and a steam-punk sort of polished brass panel w/ black Bakelite knobs & switches .

Oblig. geekishness:

The reason the digital meters are stable is that they use integrating A/D converters - they essentially average the voltage over each time interval. This is very handy, since no input filtering is required to prevent aliasing due to noise....

[Zeppelinium]

Originally Posted by barts

...The reason the digital meters are stable is that they use integrating A/D converters - they essentially average the voltage over each time interval. This is very handy, since no input filtering is required to prevent aliasing due to noise....

I should have thought of that. These things are slow, like 2-3 conversions a second. That's one heck of a low pass filter. I'm used to 12-bit A/Ds that do conversions in the low kHz range. Now all I need for the "production" version is an isolated DC-DC converter so I can get rid of the battery.

Zep

[barts]

Originally Posted by Zeppelinium

Now all I need for the "production" version is an isolated DC-DC converter so I can get rid of the battery.

Zep

You need a meter w/ a common ground like this one for voltage; this will work w/ your shunt.

Here's a larger one that has a back-light....

- Bart

[Zeppelinium]

BARTS, I appreciate your inputs. I'm refining the design before I modify my other trailers, so keep them coming.

I like these panel meters a little better (second one down). They have a smaller bezel, match the temperature unit in appearance and window size, are a bit cheaper (and use the same basic electronics) and draw only 1 ma.

The mechanical installation is the big issue for me at this point. I'm looking at ways to eliminate as many wires as possible and make sure the panel is easy to install and easily removable for maintenance, should that be necessary.

Zep

[Zeppelinium]

The cosmetically correct version is finally together.





As you can see from the back, the panel consists of several different modules, which means it can be configured as a full 5"x22" replacement for the orginal early 70s monitors, or can be physically rearranged into other shapes, even as small as just the monitor itself (in about 4"x4").

One-off panels are about $30, so making a different arrangement is not a big deal. The panel is 1/8" plastic that is laser-cut. I've added an 0.032 aluminum back, but it's still very flexible. It needs to be in a frame (see post #32 above) to provide stiffness and some protection.

Zep

[barts]

Nice work! very clean, professional and appropriate for a 'stream.

- Bart