Ballast Fill Level Monitoring

[freeheel4life]

ID, this is so great and I am going to have to draw some stuff out to explain what I've done to 2 boats now and hopefully snag some pics when they come back for a 25 hour service. Have definitely been dealing with the ballast sensors in the rear but I havent seen the hose sensor...unless it's on the line just after the ballast sensor, which would make sense( bags full and starts venting thru the ballast sensor).

Also want to nerd out and talk about multiple CAN pairs and why its beneficial (or not). Still havent fully read this thread yet as theres a lot of fat to chew, and digest. At least for a nuts and bolts guy that likes his meter and is trying to keep up

[IDBoating]

What you’re saying is exactly right. I totally get it. If I still had my old 21v this is exactly the kind I project I would take on. Much harder to do on a newer boat though. Just because it’s a new boat.

It's not harder on a new boat unless you insist on trying to integrate it with the existing CAN/helm system. In fact, using the existing CAN network is pretty simple - network negotiation takes care of the modest amount of extra traffic quite nicely. The challenge there is knowing which wires in the harness are CANH and CANL. There are accepted psuedo-standards for those wire colors but the marine industry was using those colors for other things long before CAN showed up, so across brands (and models within brands!) there's no standard at all.

The more difficult challenge on "new" boats is trying to integrate into the existing screens, as discussed earlier. Again, there's no standards, everything is proprietary, and Murphy/Medallion have zero interest in helping anyone crack into their systems. Even if you did, their protocols could change next year and suddenly you're right back where you started. It's a cat-and-mouse game that never ends. That's why I think a one-size-fits-all standalone solution is the way to go.

[IDBoating]

I hope I might be able to get the specs output for ID's sensors to make a testing model to use with my design.

We'll be happy to share the specs for our sensors and you're welcome to experiment as much as you like! However, keep in mind that our sensors "speak CAN" so you'll need to have a CAN interface on your end to interact with them. Not difficult, in fact if you're using an off the shelf platform like Arduino or something there are lots of ready-to-use CAN interfaces that plug right in and have prewritten API's for you to use.

I'm a little discouraged, since ID's. research showed this type of system did not work well.

I didn't mean to discourage you. It is possible to measure ballast with pressure sensors. But there are significant disadvantages. Here's a short excerpt of a report we wrote regarding pressure sensors and ballast.

Dynamic Range: The Key to Accuracy and Repeatability
Dynamic range is the difference between the minimum and maximum values that can be measured. For example, the dynamic range of a one foot ruler is 12 inches because it can measure values between zero and 12 inches.

Dynamic range is vitally important. When a sensor's dynamic range matches the parameter being measured, those measurements are more accurate and repeatable – and the effects of noise and interference are minimized. A yardstick is the wrong tool when you're measuring the thickness of a human hair; likewise, you need the right tool to measure ballast.

Water's own weight exerts a pressure of just 0.433 PSI per vertical foot. For a two-foot-tall ballast compartment, the total dynamic range from empty to full is thus (2 x 0.433 =) 0.866, or less than 1 PSI. At sea level and normal temperature, this means a pressure sensor must detect 0-100% in a PSI range from 14.7 (empty) to 15.6 PSI (full), so the easy (and wrong) answer is that you need a pressure sensor that can measure a dynamic range of 1 PSI based at 14.7 PSI.

But ballast compartments are vented to atmosphere, which means altitude affects their pressure. Lake Tahoe CA is over 6000 feet, dropping the empty pressure to just 11.7 PSI and the full pressure to 12.6 PSI. Now the pressure sensor must accommodate 11.7 to 15.6 PSI. That's a dynamic range of 4 PSI, more than 4X what is actually being measured. {IDBoating comment: Note this means you need a ~16 PSI pressure sensor when you're only measuring a 1 PSI range. In other words, the 0-100% range of the pressure sensor is just 1/16th of its total, or 6.25%. And THEN you want to divide that into 1% increments. That means you're trying to use a 16 PSI sensor to accurately measure steps of (0.0625 / 100 =) 0.000625 PSI! Each step is less than one-thousandth of a PSI. I can promise you that is tough to do even in a lab environment, not to mention the electrically noisy environment of a boat.}

It is clear that a pressure sensor must widen its dynamic range to accommodate all anticipated conditions, yet it will only use a small fraction of that dynamic range to actually measure the ballast. This is exactly the wrong thing to do if accuracy and repeatability are desired, and results from measuring a second-order effect (pressure) instead of the actual, desired value (amount of water).

By measuring the actual water, our Ballast Sensor is insensitive to changes in elevation, pressure, and temperature. Furthermore, our sensor matches its dynamic range to the ballast compartment, maximizing sensitivity while minimizing the effects of noise and interference.

Our advanced electronics module with its dual microprocessors allows us to accommodate any size ballast compartment by simply changing the length of the sensor tube, thus applying 100% of our dynamic range to 100% of the ballast for maximum performance. Our electronics even enables a single length ballast sensor to be used for multiple sizes of ballast compartments by clever installation techniques, thus minimizing the number of different SKU's that must be stocked while reducing cost through economy of scale. And in every case, the full dynamic range of the ballast sensor is utilized to deliver optimal accuracy and repeatability.

It seems hard to believe that there has not been an aftermarket setup yet.

I suspect it's because no one has cracked the ballast sensor problem - until now! The display side is relatively straightforward. But you need good sensors because otherwise "garbage in, garbage out".

[IDBoating]

ID, this is so great and I am going to have to draw some stuff out to explain what I've done to 2 boats now and hopefully snag some pics when they come back for a 25 hour service.

I'd love to see some drawings and/or photos.

Have definitely been dealing with the ballast sensors in the rear but I havent seen the hose sensor...unless it's on the line just after the ballast sensor, which would make sense( bags full and starts venting thru the ballast sensor).

Strictly speaking, putting hose sensors after ballast sensors is redundant - but some have discussed doing it as a belt-and-suspenders backup. Frankly I think it's unnecessary. Hose sensors for ballast make sense only where you don't have the physical room for a true ballast sensor, which is the case for many center and bow ballast compartments. Otherwise, use a true ballast sensor.

Also want to nerd out and talk about multiple CAN pairs and why its beneficial (or not).

Ready to chat anytime!

[UNSTUCK]

It's not harder on a new boat unless you insist on trying to integrate it with the existing CAN/helm system. In fact, using the existing CAN network is pretty simple - network negotiation takes care of the modest amount of extra traffic quite nicely. The challenge there is knowing which wires in the harness are CANH and CANL. There are accepted psuedo-standards for those wire colors but the marine industry was using those colors for other things long before CAN showed up, so across brands (and models within brands!) there's no standard at all.

The more difficult challenge on "new" boats is trying to integrate into the existing screens, as discussed earlier. Again, there's no standards, everything is proprietary, and Murphy/Medallion have zero interest in helping anyone crack into their systems. Even if you did, their protocols could change next year and suddenly you're right back where you started. It's a cat-and-mouse game that never ends. That's why I think a one-size-fits-all standalone solution is the way to go.

I didn't really mean harder as in hard to install, I meant harder as in harder to bring myself to do it. I have a hard time bringing myself away from stock on this boat. I'd worry about potential value loss by making some changes and hacking into the ballast system and potentially needing to permanently mount another screen just shouts loss of resale. As nice as it would be to make a ballast fill system that actually is smart, I don't think I have to have it, in this boat.

I sure have a hard time taking my eyes off SC and all their (your) innovations.

[Bamer]

SC? Skiers Choice?


Sent from my iPhone using Tapatalk

[freeheel4life]

Yup

[IDBoating]

I'd worry about potential value loss by making some changes and hacking into the ballast system and potentially needing to permanently mount another screen just shouts loss of resale.

The way I handled that in the early days of AutoWake development was with Ram mounts for tablets. They have some very strong suction-cup mounts that stuck right to the side windshield next to the helm. Worked perfectly and made zero changes to the boat (in case the AutoWake concept didn't work out!).

Folks have the same concerns about fancy cars that you do about your boat, and there are lots of products that address those concerns. I think many of those solutions would work well in boats to allow mounting a phone or tablet at the helm.

[oldwakedoc]

ID, my current plan is using a CAN network.

Did you guys do any research on using flow sensors? It seems the issue of accuracy without restricting flow becomes a limitation.

Thanks you for sharing your result on the transducer option. Those are great points and I see why you decided not to consider this option. Now considering returning the transducers vs. using them to design my system then swap out for a better ballast level sensor .

[IDBoating]

ID, my current plan is using a CAN network.

Well, that makes it a lot easier! Bit rate of 250K is standard on wakeboats, by the way, and on our sensors.

Did you guys do any research on using flow sensors?

Sure. And measuring weight. And every other parameter we could think of.

Flow sensors have three big problems: 1) Cost (good/accurate/reliable ones are surprisingly expensive), 2) maintenance (they have moving parts, must be drained during winterization to protect them from freezing, etc.), and 3) contamination (their moving parts are exposed to lake/ocean water which means non-biological and biological contamination is a huge, Huge, HUGE concern - there's a reason people put hideously expensive antifouling paint on their hulls!).

There's an important concept called "first order effect". Here's another excerpt from an introductory writeup for our Ballast Sensors:

-----
Properly measuring ballast levels is not easy. Over the years many manufacturers have tried many approaches, only to abandon them due to cost, production, or maintenance challenges. For example, Nautique has tried both flowmeters (which are fooled by air bubbles) and pressure sensors (which are fooled by a lot of things). Recently some tried pressure sensors beneath the ballast compartment to weigh the water, but could not resolve cost and production problems. Everyone eventually settled for timers despite their known errors and problems because no better solution could be found.

We approached the ballast problem with a fresh perspective: What everyone wants to know is "How much water is in the ballast compartment", so we will measure that. Not time, not pressure, not voltage, not current, not flow. Just measure the water. Period.

The amount of water is a "first order effect". We are not measuring some second-order effect (such as time, pressure, flow, etc.) and then estimating how that compares to water in the ballast compartment. We directly measure the parameter in question, which means we are insensitive to air bubbles, temperature, pressure, altitude, battery and alternator voltage, and all the other gremlins that have frustrated previous attempts to solve this problem.

The WakeTouch Ballast Sensor leverages our success with noncontact water sensing to deliver an inexpensive, reliable solution. This does not mean doing it is "trivial". Indeed, each Ballast Sensor electronics module has TWO separate microprocessors to handle the necessary operations and mathematics. This processing power allows us to compensate for the "lake effect", detect and report various error conditions (such as low battery voltage) for the benefit of other on-board systems, and deliver linear response over the sensor's full dynamic range.

Our Ballast Sensor communicates on the CAN bus like all other modern marine electronics. Unlike old-fashioned voltage and current outputs that are prone to interference, and which need an analog input somewhere, our electronics module already includes CAN via an industry standard Deutsch connector.

Our electronics includes LED's that indicate proper operation and provide a rough indication of ballast level even when no other systems on the boat are working. As long as the sensor is powered, the LED's provide a ballast level indication even if the the ballast compartment itself is obscured from view. This is also convenient during R&D on new boats when the helm electronics and/or firmware is not yet functional or the CAN network is not running. Just provide power, and our sensor starts helping you to get other systems working.

We recognize that everything has production variations, so we calibrate each of our Ballast Sensors by actually filling them with water in a custom-built calibration fixture. This means every Ballast Sensor is fully calibrated and tested, with actual water, in the actual way it will be used in the field, before we ship it. This is a lot of extra work, but we do it to insure consistent and repeatable ballast measurements.

Bottom line: The WakeTouch Ballast Sensor uses very advanced electronics technology to deliver a ready to go, plug-and-play solution to a problem that has plagued the marine industry for decades.
-----

The point is, a flowmeter is just like a timer. It's measuring a second-order effect. You're trying to measure the *flow rate* past a certain point in the plumbing, and then mathematically integrating that rate to estimate the water volume. That is not the same thing as actually measuring the water.

What if there are entrained bubbles? The flowmeter will still spin (the bubbles still take up volume in its volute) but once the water is in the ballast compartment those bubbles can precipitate out, leaving a lesser volume of non-entrained water. Now what's your ballast percentage?

Flowmeters are bad, just like timers are bad, and for the same reason. Like flowmeters, timers estimate the *flow rate* that the ballast pump "should" have and integrate that to estimate the water volume. But what if the battery voltage is higher or lower? The pump runs at a different speed, but the timer doesn't "know" that. Now what's your ballast percentage?

What if the engine is running? The alternator voltage is higher than normal battery voltage so the pump runs faster, but the timer doesn't "know" that. Now what's your ballast percentage?

How about if you fill while the engine is running (higher voltage) but drain half of the ballast with the engine off (lower voltage). The pump ran at different speeds, moving more water during fill and less water during drain even though you ran the pump for the same amount of time. Now what's your ballast percentage? You add and remove ballast during the day, sometimes with the engine on, sometimes with the engine off. Sometimes the hull is moving (causing a venturi effect on the throughhull and a subsequent reduction in flow) and sometimes it's not. After a few cycles, how much error has accumulated in a timer (or flowmeter) estimate?

Furthermore, what if you have a leak? What if you have a syphon? Timers and flowmeters have no way to know about any of these things. They just keep stupidly reporting their estimate even if you shank the fat sac and drain all the water into the bilge. Why? Because timers and flowmeters estimate a second-order effect (flow) instead of measuring what you actually want to know (the amount of water).

Sorry to ramble on, but hopefully I'm making my point! There's a right way to measure ballast: By actually measuring the ballast, not some secondary effect.

Thanks you for sharing your result on the transducer option. Those are great points and I see why you decided not to consider this option. Now considering returning the transducers vs. using them to design my system then swap out for a better ballast level sensor .

When you're ready, let me know and we can discuss the CAN protocol to talk to our sensors....

[skiguy]

I agree. 3rd party integration with integrated helm systems would by a tech and customer support nightmare. I like the idea of a separate stand alone system for support of all the older boats our there. I would look closely at a system that cost me less than $300. Any more than that and I would stick with the current manual "open hatch" monitoring system :-)

I think WakeTouch is in a good position to bring something like this to market - I understand the corporate focus has been on manufacturers but to me it seems like you're leaving a big market under served.

[IDBoating]

I think WakeTouch is in a good position to bring something like this to market - I understand the corporate focus has been on manufacturers but to me it seems like you're leaving a big market under served.

We simply can't chase every opportunity in front of us. We're behind on several pending products and have many more we'd like to consider. We don't want to do a half-baked job of an aftermarket ballast control system only to damage our reputation. So we'd be happy to work with someone else who wants to do it, but must politely decline to do it ourselves.

[oldwakedoc]

ID, sorry it has been crazy busy and no time for boat stuff Just getting back to this.
We discussed earlier about the CAN network your sensor broadcasts on. I am currently putting together a CAN 2.0A network.

Does your sensor work on a 2.0A or B? (wanted to know before I get to far with current hardware.)
Are there limitations in length of the vertical tube you are using with the sensor? (can't be too short or too long)
Can the vertical tube be shortened after manufacturing and still function, or is the length fixed once it is made?

If you have previously answered these questions, I apologize for missing it.

[IDBoating]

We use CAN 2.0B, which includes extended identifiers of 29 bits. This is basically a requirement since J1939, an automotive standard, uses 29 bit extended ID's anyway and virtually all messages on marine CAN networks are extended ID's. I haven't seen an A-only CAN chip in a very long time, so I suspect anything you choose today will support either standard - usually simultaneously, you just set a flag for each message (for transmit) or filter (for receive) indicating whether you want to use extended ID's.

If you haven't yet, read up on PGN's (Parameter Group Numbers) to get a feel for how J1939 organizes its CAN headers. This is a common pattern for all extended ID's.

There are no limitations on the length of the tube, other than you probably won't get good resolution under 12 inches tall. No, the tube cannot be "cut to fit"... well, I guess strictly speaking you could cut the bottom of it but you risk shorting the electrodes within the assembly and you'd also have to cut the electrode stack back away from the bottom end so you could attach new fittings for connection to the ballast compartment. Realistically, you should presume the tubes cannot be cut in the field. It's just not practical.

Hope this helps!

[oldwakedoc]

ID, I will confirm the options of my current hardware relative to the CAN network.
Are the height of your sensor tubes custom made lengths for each manufacturer and each boat, or do you offer manufacturers standard lengths?
Since there are no post modification option, you would have to make "universal" lengths to work for most after market applications.
I assume you make large runs for specific lengths and custom one off orders would be significantly more expensive.
If you use a tube that exceeds the height of the ballast bag, can you calibrate the tube to use a smaller function length or does the height have to match to height of the bag?

thanks