mechanical fill level sensors?

"Liam" wrote: > Let's say we have an *unpressurized* open vessel with a capacity of

Yes. Probably the easiest is a float switch, but you could also do it with a pressure switch, an ultrasonic level sensor, a radar tube, etc. Check out

If your set points are always the same (full and "n") then you could use two binary sensors, like float switches. If you want continuos readout, you'd need something variable like an ultrasonic, pressure, or radar...radar probably wouldn't work in a vessel this small though. The last time I used it it was on a ~500 gallon tub.

Tom.

There are definitely multiple choices for your application. 250 ml is a relatively small volume, but can still be monitored with float valves. Some of the other sensors might be more appropriate depending upon the available space. I would start by looking at the optical sensors because they require less space and are highly reliable. I know Honeywell makes some, but there are multiple manufacturers of such items.

Dear Liam:

I concur with "Smitty Two". The application cries for a pair of "capacitive proximity sensors", if the "vessel" were non-conductive. Better still, a light beam could be used to probe the presence or absence of liquid, if the "vessel" were optically clear (this option can be really cheap, and extremely precise).

David A. Smith

Your own experience will bring to mind at least one device that serves that purpose on a bigger scale, and with cold water: the water cistern hooked to your sit down lavatory. This consists of a float that stops the water flowing through a valve at a given level. Make it smaller, and heat resistant.

Brian Whatcott

Yes. They're called level-sensors. :-)

It can be even simpler (i.e. no valve) given that all it has to do is to do is to switch a pump off and on at pre-determined levels. Obviously, a taller vessel will provide more accuracy... and the boiling liquid may cause some perturbation in the level resulting in phantom switching if the levels between off and on are too close. A side-tube (copper) with a magnetic float would solve most of the problems with the action of boiling.

A third magnetic switch can be placed at the level of the heating elements in case the pump fails to deliver.

I wonder if the "vessel" isn't insulated, if he couldn't just measure the temperature at the places he wants to do switching at...

David A. Smith

David, I am not sure how to reply to the underlying question, is the vessel non-conductive? Do you mean a heat conductor or an electricity conductor? The vessel is made of stainless steel or brass (something that can retain heat) and it holds water. There's a heating element on the outside rather than inside of the vessel, to avoid calcium and mineral buildup. 'Cheap' and 'precise' are excellent features. Are these devices fragile in any way? Can they take the heat? Liam

Brian Whatcott wrote: > Your own experience will bring to mind at least one device that serves

Mine never works. I'm always lifting off the porcelain lid. You're steering me towards optical :-) Liam

One of the design goals is to bring that 250ml of water rapidly to a near boil, so I was thinking a container wider than it is high. But perhaps there's an optimal height that lets the water come quickly to temperature but is tall enough to get accurate measurements. We would need a thermocouple to regulate temperature, right? A side copper tube is good if there's still bubbling that would affect accuracy. But this doesn't have to be very accurate. We just want to make sure the pump goes off. If there's a margin of error of a few ML when it goes back on again, that's fine. I was thinking ON again when down to 50ml. Pretty big distance between states.

How does the magnetic float communicate with the pump? Is there something attached to the pump that is tripped ON/OFF by the magnetic field?

Magnet gets five stars for retro aesthetic appeal :-)

Regards Liam

To retain heat, the vessel would be made of brass or stainless steel with a heating element beneath it. Would the heat radiate to all portions of the vessel surface? Regards Liam

////

I like this one. The fist sensor candidate that came to mind was the reed switch, but these have not won a great name for longevity. Even a little inrush current , or opening over voltage, is too much.

But an integrated hall switch or the like might do nicely.....

Brian Whatcott Altus OK

Dear Liam:

Electricity conductor. Capacitive proximity sensors cannot "see" through a conductor.

It'll still buildup, just not on the element. The calcium (etc.) doesn't vaporize, so it can't go anywhere else.

Fragile, no. Take heat... maybe not. That can be overcome by having a second "chamber" in parallel, acting as a sight glass. I have used a fitting and teflon (or any other kind of) tubing to remote the fluid level from the hot area. Any clear tubing can be used for a cheap optical sensor or two.

If Danaher Controls hasn't trashed their engineering department (like they have done to so many other aquisitions), Gems had some good sensors...

I've seen a sensor that had a body that was designed to straddle a clear tube, with an LED on one side, and a detecor on the other. All you'd need to do is strap it on... no penetrations required. Can't remember who it was that made it... want to say SMC, but that is probably not right. I'd check out offerings by Banner Controls, if optical seems the way to go.

David A. Smith

Yes. Reed switches. They look fragile in their glass cases but when mounted properly, will last a very long time. You're more likely to blow them up with too much current. You can use them to pulse a latching relay to provide power to the pump. I prefer a 555-timer chip to limit the maximum run time in case the float gets stuck below the top level sensor. The chip and support circuit are also cheaper than a latching relay vs ordinary relay with electronic.

I used reed switches and a magnetic float in another application... worked a treat for 5 years before the installation was no longer required.

As long as they're not mechanically stressed, they last a very long time. Mount one end rigidly and allow the other to "float" longitudinally. You must certainly prevent any bending forces being transmitted. With those provisions, the reed switch should last hundreds of thousands, if not millions of cycles.

You don't run the motor current through that switch. Use a latching relay or equivalent electronics to actually run the pump. Keep the control current through the reed switch between 10 and 100 mA and the switch voltage below 24V. A choke on the low-voltage supply will keep the current peak low when the relay is being energised and flyback diodes across the DC relay coils are advisable anyway.

The metal doesn't "retain heat", rather it conducts it well.

Only if you insulate the walls of the vessel, or make the vessel very thick-walled.

Are you boiling the water, or simply heating it?

Is there a microprocessor in this device for another purpose? I ask, because a fixed heater increases temperature on a small volume of water much faster than it does a large volume of water. Depending on a few things, you might get a single temperature probe to do everything.

David A. Smith

A single temp probe to do everything is attractive. But here are the constraints which seem to me to work against that approach, and they might also work against the magnetic switch/float approach too:

Vertical space is very limited. The vessel dispenses its water by gravity and the device which receives water from the vessel has a pre-defined location. And so, given those space constraints, the water-heating vessel has only about 3cm height. Therefore I was thinking

Water is being heated only to ~98C not to a rolling boil. But it must reach this temperature very quickly. ~1000W heating element is projected, more or less.

The thickness of the vessel walls is ~11mm. An overarching design goal is temperature stability. We need to keep the water at 98C +/- 1C. There would be a thermocouple to monitor water temperature and control the heating element.

Regards Liam

...

Single temp probe level control is out if the vessel is to be maintained at 98 oC for some time. Evaporation will be quite hgih.

This small height will prevent insertion of the temperature probe in any useful location.

I'd recommend you tap off to a separate open-top tube, and sense level on/in the tube. Any sensor will be fried by the 1000 watts, should the water supply fail for any reason.

There are hot water heaters that get installed directly to the cold water tap of a shower. No hot water tank required. Other than low dispensed volume, and low flow rate, could this be made to work for you?

David A. Smith

I have heard of that type of heater. But shower temperatures are much lower. We could eliminate the heating tank if a flash-heater is able to heat a much smaller amount of water, say 50ml, to 98C in ~5 seconds. Is that feasible? The entire 50ml would have to reach 98C, not the latter portion of it. The resulting 'dose' of water would have to be evenly heated to 98C. Do these little heaters have their own self-contained heating vessel?

How would the pump be told to cut off if such a flash-heater were positioned between the fresh-water reservoir-pump and the device which receives water from the now-expendable heating vessel? Or would the pump be set to move a defined volume and then cut off automatically?

How does the flash-heater 'know' when to turn on?

Regards Liam