I appreciate that while Kromann refers to a level senor, his application quite reasonably can be solved with high/low level switch control, if it can be realised economically. There is an attraction to fork vibrational level switches, mainly they are cheap and usually maintenance free.
However, I would be unhappy with a tuning fork sensor, (or any other invasive device) in this application without a proven history of success.
Sure, a clean sensor at the high level will show it is in air.
When the chocolate first reaches it, it will show the chocolate.
As the level falls, the fork may then be left coated. This may leave the device in a false high state or an alarm state.
If the fork is coated but otherwise exposed to air (or deprived of direct heating from the chocolate it was immersed in) the chocolate may set. This may cause the sensor to lose lock, i.e. the drive circuit cannot find resonance.
Insertion through the side of a jacketed tank would be a major undertaking and costly if appropriate insertion points do not already exist. The heating from the jacket may prove sufficient to help keep the sensor functional but I would be suspicious of this.
If the tank is not jacketed then the heat loss effects, especially as the sensor may act as a heat sink, will be noticable.
The option for jacketed tanks would be for top mounted fork (or ultrasonic) sensors on long stems.
When the chocolate level rises again the surrounding chocolate may
thaw out the chocolate coating and allow the switch to reset and resume function. However the chocolate level may have risen significantly above the switch by this time.
In the case long stem top mounted fork sensor(s) the sensors will now be out of contact with any direct heating. In fact conduction losses may even affect the immersed (low)sensor, sufficient for the chocolate in contact with it to set, despite any surrounding fluid chocolate, with the possibility of false alarms.
Also, dependent on how high up the stem above the fork the chocolate reached, the time the fork remains coated with fluid chocolate will depend on the drain time of the chocolate from the stem. More sustained false high signal and greater potential for some chocolate to solidify.
This same effect could cause the low sensor to remain in a high alarm condition even when the tank is otherwise drained (meaning the tank could empty, starving production).
One cannot assume the device will be sophisticated enough to remain functional.
Most fork level sensors have a fairly basic drive circuit. They are dependent on detecting a significant difference between resonant frequency from one state to the next. A tuning fork level switch costing $100 is not as sophisticated as a tuning fork density meter at $2-3000 and it would also be expected to have problems.
Float switches, tuning fork level switches or whatever intrusive device is considered, great care should be taken to discover if the device has a history of success in this application and under what circumstances.
Sometimes the lower cost of the device is negated by the added cost of the work-arounds. For example, in some applications where coating can occur, it is neceesary to include a spray nozzle just to clean the device, or a sensor heater. When this is the only sollution, it is a good sollution. When other technologies will do the job better, it is not.
The idea of a non-ivasive device is far to be prefered. One of the strengths of non-contact devices is in just such applications as these. The device may cost more to buy but pay for itself in no time at all. In this case while the application allows for high/low switches, a top mounted level measurement system such as radar, optical or ultrasonic may be preffered.
Eng-Tips: Pro bono publico, by engineers, for engineers.
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.