Dispensing an Accurate Amount of Liquid from Two Tanks

Not very large, about the size of a bottle of whiskey.

Sorry Doorman,

Only clicked now on the joke lol, thought you were refering to the tank size!

Doorman's on the right trail. His approach is the simplest, at first look.

Using your drawings and one flow switch and valve won't work, without at least a check valve in each tank outlet, otherwise, the tanks will cross flow and reach the same level.

I'd go with a three way valve and a low level sensor (or flow meter) on tank 1.

Hi Guys,

Thanks for the response. I am unable to put a float switch into the tanks as they are fully enclosed except for the outlet. Once they are empty they are discarded, I know this all sounds very weird. The medium is not actually water, I was just using that as an example. The only way to measure here is when the liquid leaves the tank/container/whatever.

Your help is much appreciated here. Thanks again.

Robotech,

Sorry did not read all the threads.

What you can do, is use a screw on type attachment (Like when you fill a butane tank or lighter) as they have internal valves for keeping things in.

Below these connections would be a glass (vertical sight glass) section that would have a volume allowance for switching over via the three way valve and a small bleed valve interlinked across the two connection points for interventing between the two bottle connections, as a changeover is done.

The three way valve will still allow cross filling to take place as well.

Look at the Dover/ITT quick disconnects (need the liquid/gas ones) for ideas of mounting and dismounting your fluid bottles.

<...accurate...>

Does this mean weight or volume?

The volume of the liquid varies with temperature, as does the volume of the vessels.

The purpose of a tank is to allow the level to go up and down. With the tanks toggling like that, there is actually no point in having them; one might as well replace them with a flowmeter.

This is actually petrol/gasoline dispensing equipment technology.

Hi PWSlack,

Thanks for the reply. This machine will have an operator that will change out the small tanks periodically, the reason I would like the first one to go empty first is so that the operator can see the one is empty by means of a pilot light and change it out. This way the will be no break in the dispensing process. The tanks are a predefined size that I cannot change.

The equipment's customer can, though. All the customer would do is buy a petrol pump instead. Far simpler, and inherently safer, particularly as the fluid is other than water; "What you don't have can't leak" - with apologies to Dr. Trevor Kletz, prominent process safety guru and author.

Capacitive sensors there are some designed to provide level sensing through the tank wall. You didn't say what the tank material is.

Also the heading says accurate amount of liquid. You using a metering pump? It has a displacement which should be accurate. That can be used to run programing to let you know when the tank is near empty to switch to the other. But only if each tank has a specified volume in it. That doesn't stray too much.

Thanks ozzb,

The liquid volume and tank sizes are different from manufacturer to manufacturer unfortunately. I am not using a metering pump due to cost and the large footprint it has, this is a very compact machine. I am going to use a low rpm pump and time it to get the desired amount dispensed. This is why it is imperative to have no air gaps in the line.

Thank you for replying, it is much appreciated.

Robotech,

what you are doing is done everyday using various methods;

• I assume the pump you show is your actual metering control device?
• If so then the best way is eliminate the two valves you show and utilize either an electrically or pneumatically operated positioning actuator onto a three way ball valve (look at the PBM range).
• So as tank one reaches its lowest point using a level sensor. The valve will slowly close off the 1st tank flow, but because the internal design of the Ball segment. The 2nd tank flow is introduced without any stopping of low or air introduction.
• You will have to time the switchover speed of the actuator, for both flow directions when the 1st tank has been refilled and the level switch is actuated to reverse from tank 2 to tank 1.
• Now I hope the pump you are using is VFD driven to allow you to adjust your metering flows against the fluid viscosity and you have sized all the piping for this viscosity- keep all connections as thru-bore as you can and any elbows to a full flow 3D type.

Hope this helps and look at the PBM three way ball valve catalogue- note: you will need Teflon cavity fillers around the ball segment (several others do not), if you do colors etc. is easier to clean and PBM do make a sanitary version with a CIP capability.

Hi Geoff Daly,

Thank you for the response, I do like your life statement, very true!

The problem I have here is that I cannot put anything in the tanks, I also cannot put anything on the sides of the tanks to measure the level. All I can logically think of is a flow switch that can be placed right at the outlet of the tank. I also cannot waste any of the liquid in the tanks, it must all be used. I have this feeling though that this is going to cause issues though with air getting in the line.

Thanks again.

There is a solution for the problem of air introduction:

Use a two part, zero-dead volume ball valve connection from tank(s) to line. Depending on which part of your situation is more important, you should purge the ball valve slightly if air is the problem, and not is the value of the liquid is the problem. If this is medical device, then you must allow for purge to prevent introduction of air bubbles into the patient, and collect any spillage during purging if the medication happens to be a radioactive tracer solution or other nuclear medicine. Or use the three way valve on an actuator to switch over when A is low (use some sort of scales to measure the weight remaining so that when the weight is low enough a switch takes place. You could still use the zero dead volume (small or large) valves to change over the vessels from "empty" to full, fresh. Why not stay on B until it is "empty"? A similar weight control scheme could be used on this volume as well.

If these are totally enclosed tanks, how does the liquid drain out? Something has to force it through; are the tanks pressurized? What is replacing the liquid in the tank as it empties, and is the liquid incompatible with air?

The reason for asking is that sensors have to detect a presence or a change of state in the process to function. If you literally want to drain every last drop from these tanks you will need to allow air or some form of replacement gas into the tank. Assuming they are not pressurized and the liquid is not extremely viscous or aggressive I would simply suggest a vertically collapsing buffer tank (think accordion) mounted between the tanks & before the pump, so that the fully collapsed height is lower than the tank outlet. Set a proximity sensor or similar to detect when it reaches the height you determine as empty, since the buffer tank should maintain equilibrium with the active feed tank. If you provide some form of bleed mechanism you may even be able to drain the feed tank into the line for complete evacuation, as the buffer tank would still be feeding the pump in the meantime. If evaporation or reaction with the air is not a problem the just use a large diameter sight glass instead- again, make sure that there is capacity below the empty level of the feed tanks.

P.S. Your post #4 sounds fine to me!!

Hi JNB, Thank you for your response, this sounds just like what I need! I was thinking of a small tank under the other two, however what if the operator changes the type of liquid being dispensed? This idea would be ideal for float switches and the likes. Your help is much appreciated!!!

Well, if the fluid is changed and float switches are used, then the different SG of each liquid will affect the volume point where the float switches operate to detect "vessel empty", altering the volume dispensed from liquid to liquid.

Is this equipment static or in some sort of moving vehicle?

Well, you could place a 3 way solenoid valve at the Tee in your picture. The line from tank 2 to the Tee would then be blue (no air).

Drill a hole in the tube between tank 1 and the Tee and insert an ultrasonic liquid level switch.

Wire the switch to the 3 way solenoid valve and a power supply.

Now whenever bottle 1 empties, the valve switches to bottle 2 automatically. Whenever a new bottle is placed at location 1, it is automatically engaged.

Is the liquid transparent, translucent or impenetrable? is the container transparent? I was wondering if the reflection or refraction of a laser beam could be made to serve as a level indicator.

Answering your question:

It will not work: You did not provide a refill way. For the rest, I read some good answers.

Add a "buffer" tank downstream from both tanks to handle the material until the tank is changed. It would need to be of adequate size to handle the time it takes to change a tank. It's like having a battery to power things when the wind stops!

Hi Guys,

Thank you for all your input, your advice has been extremely helpful and has given me a lot to think about. Hope you are all well, take care.

-Rob

What you have in mind is very similar in principle and methodology of delivering fluids to patients in a hospital environment application. The device they use are known as infusion pumps. Fluids are measured and delivered at any predetermined rate and must be without any Air. Depending on the required accuracy of fluid deliveries, the most common methods are done using volumetric pumps, and screw type pumps, all driven by stepper motors. Piggy backing, similar to your drawn illustration is common and for lesser acccuracy may even utilize gravity flow. Fluid containers can be rigid as in vented bottled solutions, or contained in plastic bags that can either be pressurized or not. It will help you to visit a hospital for observation and visualization.. Most of your questions and concerns will be answered...Good luck

Maybe you should do a patent search before you invest too much time, effort and money reinventing the wheel. The process you describe is already being done in agriculture.

If you are just doing this because you are sitting around being bored, then that is a different matter.

When I noticed the word accurate in your question I thought of something interesting. You could hook up a bottle of gin to one side and a bottle of dry vermouth to the other and dispense the vermouth and gin mix at a ratio of 11 parts of gin to one part of vermouth and have an automatic dry martini maker.

A good martini, no doubt, but what if he is attempting to make binary explosive? Is some vetting of the OP in order?

Depending on what you are wanting to dispense, it's already made. Bag-In-Box beverage systems use them everyday, very cheap. Google bag-in-box parts. -- JHF

Having trouble with my cad, hope you can see the sketches.

Regards JD.

Hi JDRetired,

Thank you for taking the time to do this. This is very interesting, would I need to get this made up or is there something like this that I can buy. Thank you for all your help here.

-Rob

Hi Robotech; Looking at the sketch again I realised the fluid head acting on the area above the weight negates the float action, therefore it will not work, if the weight is removed then tank two will lift the float? Sorry

Regards JD.

PS could have a sliding counter weight on the reset lever that balances the weight when tank one is in service, but set so that it slides down the reset lever at the change over position, biasing tank two into service. or a micro switch at this position to assist?

Regards JD.

Upgrade two? Have modified the reset arm with a rolling counter weight that should ensure that tank two will come into service, and the float will not reverse it. I have also moved the outlet down, because if you are using this to dispense spirits that would require an ingress of air back into the tank, then tank two would not have worked. There will be some air above tank two outlet but it will not effect the total flow from that tank, on reversal the air will go up into tank one as normal, in either case quantity should not be affected.

Regards JD.

Hi JDRetired,

Thank you so much for this information, this is a fool proof system and requires no electricity to run saving on I/O. Is there anything like this available on the market or would I need to get one machined? Thank you in advance, hope you are well.

-Rob

Not on the market, you would have to develop it yourself. I am still not happy with the inclusion of the weight with liquid acting on it? this I think will stop the system from working? therefore you would need to remove the weight from the design and also move the rolling counter weight from the right side of the reset lever pivot point to the inside. It still needs a lot of work to get it right.

Regards JD.

Further to my above post, a further update as described. I have only done one sketch as the others should now be obvious. This one I think should work ok?

Regards JD

Jdretired-Robotech,

Basically your concept is like a three way shuttle valve you see on hydraulics/air operations.

The simplest is still my answer in thread 9- use a PBM three-way sanitary ball valve, these are off the shelf and non-electric.

Both can be controlled so no air flows into either side if slowly actuated. Other than jd's needs to be constructd and the PBM is off the shelf.

Go to www.pbmvalves.com and look at the three way Series 9 Sanitary range of valves - T-port 180 degree code 06. (look at this Youtube video

http://youtu.be/eEN4w5HLx9I shows the flow pattern)

Hi Geoff and JD,

Thank you very much for all your input, this has really helped me out and is much appreciated. Hope you are all well.

-Rob