Arduino Autonomous Vacuum Cleaner

This is from the old read-only forum. I don't know how to import the topic, so I'll just continue here. Hope I am on the right sub-forum. http://arduino.cc/forum/index.php/topic,22648.0.html

Basically, the problem is charging the LiPo battery without removing from the system.

The god members including sir AltairLabs contributed much on this topic, and the updated partial schematic follows.

To those who can answer the question, please bear with me, for I am a novice in application of electronics.

Questions: 1. Sir AltairLabs suggested I connect the LiPo directly to the D2 diode, bypassing the FET. If the outside source is to be connected, will the current from both sources merge on the node after the diodes? Will this affect the charging of the battery since the loads will "effectively" source from both the sources, wrong? Or shall we revert to the original circuit, connecting positive terminal of LiPo to FET? 2. Switching regulators will be replaced by common linear regulators, since the deadline of this project is in two weeks time; the planned switching regs cannot be delivered in 2weeks, so we dropped that plan. I have read a lot about the disadvantages of linear regs, but we have no choice. So we opted to continue this project using linear regs. The problem is, the power dissipation of the regs. We computed, and dissipation can reach up to 2.5Watts burst, 1Watt continuous. If we are to add heat sink, can this stabilize the regulators? 3. Application circuits from the datasheet suggested we use .33uF before and .1uF after the regs, is this really enough? I thought the higher the capacitance, the better?

Thanks in advance!

@AltairLabs, hopefully you will read this. ^^

Sounds fun.

I made an hoover bot (see attachment). 6V screwdrivers for motors and a dust-buster on top. Arduino controlling it.

It did a random walk and did a fair job of avoiding getting stuck, but was pretty useless at hovering really. Still I enjoyed making it.

I just went for the plug it in approach rather than charging contacts and a beacon.

I don't understand what the MOSFET is for in your circuit. What is the source supposed to be connected to?

Si: I don't understand what the MOSFET is for in your circuit. What is the source supposed to be connected to?

It was originally intended to be a switching device, disconnecting the path between LiPo and regulators when external source is connected, thereby freeing the battery from the load, so the charger can charge the bat efficiently. It was suggested by sir AltairLabs. ^^

Do you reckon not using the mosfet will provide the same functionality? Sir AltairLabs suggested I connect the LiPo directly to D2. I just don't quite understand how it will free the battery once external source comes in. Could someone explain this to me?

Hi,

I can't really tell from your circuit, what you are trying to achieve, but given you seem to be on a tight deadline, I would just simplify things as much as possible.

As it stands, when you connect the battery charger, it will also supply current to the two voltage regulators. If one of your issues is avoiding this, then why not just put another switch in to disconnect the positive side of the battery from anything except the charger.

Si: I can't really tell from your circuit, what you are trying to achieve, but given you seem to be on a tight deadline, I would just simplify things as much as possible.

Yeah, that's why we opted to use linear regulator instead of the original plan of switching reg, since the ICs won't be available in 3weeks. But then, linear regulators dissipate too much power, so we will use heatsinks for each. About 2W burst and 1W continuous power dissipation if the calculation is correct, do you think common heatsinks will do?

Si: As it stands, when you connect the battery charger, it will also supply current to the two voltage regulators. If one of your issues is avoiding this, then why not just put another switch in to disconnect the positive side of the battery from anything except the charger.

That is what I am thinking of the circuit. I just don't understand why sir AltairLabs suggested I connect the positive terminal of the battery directly to the diode, bypassing the mosfet, which was originally there to switch. What do you think of this? The switch you are talking above, is it mechanical? Since we want to automatically remove the load connection from the battery once the external source is connected, do you think we stick to the original plan of using the fet?

Thanks for your input. ^^

Si: Sounds fun.

I made an hoover bot (see attachment). 6V screwdrivers for motors and a dust-buster on top. Arduino controlling it.

It did a random walk and did a fair job of avoiding getting stuck, but was pretty useless at hovering really. Still I enjoyed making it.

I just went for the plug it in approach rather than charging contacts and a beacon.

I don't understand what the MOSFET is for in your circuit. What is the source supposed to be connected to?

That image has to be the very definition of the word "contraption" - great job! XD

@zeitcheist the FET is only useful when the charger is onboard. Since diagrams (pix worth kilowords) have revealed youre intent to NOT have charger onboard then you should NOT use the FET.

AltairLabs:
@zeitcheist the FET is only useful when the charger is onboard. Since diagrams (pix worth kilowords) have revealed youre intent to NOT have charger onboard then you should NOT use the FET.

Sorry, I quite misunderstand this point. What is the difference of the charger being onboard and not? Either way, the constant current will still charge the LiPo, and if I will not disconnect the path from battery to loads, then effectively the battery is supplying current to the loads. Am I wrong? Please elaborate on this. I don’t fully grasp the point. Sorry for being so stupid at this matter… ^^

Si: Sounds fun.

I made an hoover bot (see attachment). 6V screwdrivers for motors and a dust-buster on top. Arduino controlling it.

It did a random walk and did a fair job of avoiding getting stuck, but was pretty useless at hovering really. Still I enjoyed making it.

I just went for the plug it in approach rather than charging contacts and a beacon.

I don't understand what the MOSFET is for in your circuit. What is the source supposed to be connected to?

Definitely my favorite picture of the month.

zeitcheist:

AltairLabs:
@zeitcheist the FET is only useful when the charger is onboard. Since diagrams (pix worth kilowords) have revealed youre intent to NOT have charger onboard then you should NOT use the FET.

Sorry, I quite misunderstand this point. What is the difference of the charger being onboard and not? Either way, the constant current will still charge the LiPo, and if I will not disconnect the path from battery to loads, then effectively the battery is supplying current to the loads. Am I wrong? Please elaborate on this. I don’t fully grasp the point. Sorry for being so stupid at this matter… ^^

Any ideas? ^^

  1. Sir AltairLabs suggested I connect the LiPo directly to the D2 diode, bypassing the FET. If the outside source is to be connected, will the current from both sources merge on the node after the diodes? Will this affect the charging of the battery since the loads will "effectively" source from both the sources, wrong? Or shall we revert to the original circuit, connecting positive terminal of LiPo to FET?

D1 and D2 make up a diode switch. All the load current will go through which ever diodes cathode has the higher voltage. As +15v is higher then +11,1 volts (+12.6 really for a fully charge Li-Po 3 series cell battery) All the load current will flow through the D1 diode. This leaves the battery isolated from the load and if a charger is connected directly to the batteries + and - terminals it will be able to except a constant current charge. One caution to check is to make sure your constant current charger doesn't output more then +15 volts max at any time of the charging cycle, or else it will then be also forcing current through the D2 diode because D2 anode voltage would then be higher then D1 anode voltage, forcing the charger to see both the battery and the regulators as a common load.

  1. Switching regulators will be replaced by common linear regulators, since the deadline of this project is in two weeks time; the planned switching regs cannot be delivered in 2weeks, so we dropped that plan. I have read a lot about the disadvantages of linear regs, but we have no choice. So we opted to continue this project using linear regs. The problem is, the power dissipation of the regs. We computed, and dissipation can reach up to 2.5Watts burst, 1Watt continuous. If we are to add heat sink, can this stabilize the regulators?

Should be able to. Most all linear voltage regulators will just shutdown if too much current demand or if the device gets too hot, self protected. So you can even try it will out heat-sinks and only add them if you actually see thermo-shutdowns of the regulators.

  1. Application circuits from the datasheet suggested we use .33uF before and .1uF after the regs, is this really enough? I thought the higher the capacitance, the better?

The regulators assume you are wiring a rectified and filtered DC to them. The caps before and after the regulators are just to prevent the regulators from breaking into oscillations under certain load current conditions. If the datasheet says they are the proper values to, who are we to argue with them? Lefty

Definitely my favorite picture of the month.

Thank you zeitcheist!

I don't think it will win any prizes for aesthetic design, or for that matter hoovering. But it was fun, and as a bonus, confused the hell out of the cat.

First thing, thank you very much retrolefty! And AltairLabs as well. You guys are a big help to us.

retrolefty:
D1 and D2 make up a diode switch. All the load current will go through which ever diodes cathode has the higher voltage. As +15v is higher then +11,1 volts (+12.6 really for a fully charge Li-Po 3 series cell battery) All the load current will flow through the D1 diode. This leaves the battery isolated from the load and if a charger is connected directly to the batteries + and - terminals it will be able to except a constant current charge. One caution to check is to make sure your constant current charger doesn’t output more then +15 volts max at any time of the charging cycle, or else it will then be also forcing current through the D2 diode because D2 anode voltage would then be higher then D1 anode voltage, forcing the charger to see both the battery and the regulators as a common load.

So as long as the voltage across D1 is greater than D2 at all times when external source is connected, then the diodes act as switch, eliminating the need for mosfet. This you made very clear, thanks. However, we thought of adding series resistor to the external 15v source so as just to drop the voltage to 9v, freeing part of power dissipation from the linear regulators, lesser heat. As your point stands, we should not do this series resistor, since that would make the D1 lower than D2 voltage. And as second thought, the only component needing power is the brain, which is the arduino, which keeps track of the voltage between + and - terminals of battery ( so that if the voltage read is about 12v, we disconnect and remove the robot from dock ). This would not consume much current, and therefore will dissipate less power, so less heat, even though the drop is about 7v. Am I right?

Another thing I want to clarify is the terminals of the LiPo we have. There are total of 6 terminals.

  1. Positive 11.1v
  2. Ground
    3.,4.,5.,6. Pins for charging.
    We tap the regulators thru the +11.1v. When charging, the dock has another 4strips for the 4pins of LiPo, and we will have regulated 15v as the external source as viewed by the robot. Will this be effective?

retrolefty:
Should be able to. Most all linear voltage regulators will just shutdown if too much current demand or if the device gets too hot, self protected. So you can even try it will out heat-sinks and only add them if you actually see thermo-shutdowns of the regulators.

We have attached heatsinks to the linear regs just for sure. We don’t want the regulator to shut down in the middle of the cleaning process. Anyway, your suggestion is not bad when testing, thanks.

retrolefty:
The regulators assume you are wiring a rectified and filtered DC to them. The caps before and after the regulators are just to prevent the regulators from breaking into oscillations under certain load current conditions. If the datasheet says they are the proper values to, who are we to argue with them?

Ok. Problem is, we can’t find .33uF locally, so we opted the next higher value, which is .47uF. Do you think this will affect the performance?

Thanks again! ^^

Si: Thank you zeitcheist!

I don't think it will win any prizes for aesthetic design, or for that matter hoovering. But it was fun, and as a bonus, confused the hell out of the cat.

Hmm. It is not I who posted that, jezuz did. Anyway, your bot is really great. ^^

Ok. Problem is, we can't find .33uF locally, so we opted the next higher value, which is .47uF. Do you think this will affect the performance?

It's not a critical value, .47 ufd will work fine.

Another thing I want to clarify is the terminals of the LiPo we have. There are total of 6 terminals. 1. Positive 11.1v 2. Ground 3.,4.,5.,6. Pins for charging. We tap the regulators thru the +11.1v. When charging, the dock has another 4strips for the 4pins of LiPo, and we will have regulated 15v as the external source as viewed by the robot. Will this be effective?

Not sure I understand your charger/battery setup. There are Li-po batteries that have so called "balanced charger' connectors where the charger can make sure each cell of a series connected multicell Li-po battery is monitored during charging and can bypass charge current around any specific cell until all cells are fully charged equally. It's called a balanced charger if it has that feature and proper connections to the battery. If that is what you have and your charger is compatible with your Li-po battery then it's just a plug and play situation and of no concern, other then you still must insure that the D1 Anode Voltage is always higher then the charger/battery terminal voltage (same as D2 anode voltage) at all phased of the charge cycle or else the constant current goes to too many places and is sure to confuse the charger's logic.

Yes, thanks. We are now working on the power supply.

I will update you as soon as possible. Thanks for all the help!

Best regards,

We are now finished with the power supply. Vacuum pump control deferred until we have fully created the brain.

If anyone here can contribute to this, please do so, we will be much thankful.

I have successfully coded the brain to search the home base (using the IR Beacon Transceiver pair from Pololu). Problem is, once "near" the base, the robot still tries to "find" it, since I have not included the algorithm that checks if it is indeed at the home base. The very reason is that I can not think of a good logic for this. The beacon from Pololu only outputs HIGH or LOW, so distance measuring is impossible.

Any suggestion on how we can implement this? We need the robot to dock at base and charge, much like Roomba does.

I'm stuck, please help.

Thanks in advance! :sweat_smile:

Any ideas? ^^

zeitcheist: Any ideas? ^^

How close does the IR beacon get you? It sounds like you need another sensor once you get close. It's pretty easy to do line following with a light sensor if you can set up an area where you can paint a pattern on the floor.

DCContrarian: It sounds like you need another sensor once you get close. It's pretty easy to do line following with a light sensor if you can set up an area where you can paint a pattern on the floor.

That sounds fine. Problem is, painting on the floor will make the home base fixed in an area. What if the user wants to reposition the home base?

DCContrarian: How close does the IR beacon get you?

About three inches rough estimate.