Home made Roomba Battery

[ksander]

I´m trying to make a home made battery to my Roomba 530. I put inline 12 rechargeable stacks (AA - 2300mah - 1.2v). Plus the capacitor of the original battery. And put in my roomba. But it does not work, it stay recharging a long time and do not run. If try to command it to start cleaning it walk 10 cm and stops.

https://www.dropbox.com/s/t2gjfxshdp5hio6/2013-01-06%2016.06.56.jpg

https://www.dropbox.com/s/gp1tg1s0vae8rqt/2013-01-06%2016.08.12.jpg

[vic7767]

What voltage does your battery hack have ? Have you performed a load test on your hack ?

Also there are no capacitors in a stock Roomba battery only a polyfuse and a thermistor.

[mfortuna]

Others have tried AA batteries with limited success. They just can't put out the current required to last a long time. But they should at least charge OK. Did you put the thermistor and the polyfuse in the new pack? It sounds like you missed one. Without the thermistor the battery will not get charged and you will see no pulsing LED.

[ksander]

I have put something I guess was a capacitor but now I think this is the thermistor, you can see it on the second picture. The voltage of open circuito is 15.6v.
I dont have this two things(thermistor and polyfuse), only one of then that I dont know what it is.
I can see the orange LED pulsing forever.

[ksander]

Oh! I find it, you can see here https://www.dropbox.com/s/l8lvz8m2y8ywe ... .54.08.jpg But I guess I have dameged it.
What is the parameters of the thermistor and polyfuse to me buy a replacement of then?

Thanks guys.

[vic7767]

If your Roomba Clean button is pulsing orange then the charging firmware detects the thermistor you have connected to the two small terminals. I suspect that the long charging times are due to high resistance between cells within the 3 cell holders. Have you attempted a load test yet ? Normally a fully charged stock iRobot battery will have a no load voltage of 17 to 17.5 vdc

[Gordon]

...but now I think this is the thermistor, you can see it on the second picture. ...

The item in this picture (that you added clips & wires to its ends) is probably the old battery's thermistor:

You can verify it to be the thermistor by testing as follows:

1) disconnect that device from Roomba.
2) Study this Thermistor Resistance vs Temperature graph, and print a paper copy of the graph.

{Source post: viewtopic.php?p=75740#p75740}
3) Measure the room temperature in which that thermistor has been for more than an hour, write that temperature on a sheet of paper.
4) Draw a vertical line on the graph from your room temperature reading upwards to cross through the curve.
5) Draw a horizontal line from the intersection of the curve an the line drawn in (4) over to the left scale (Resistance scale). If your room's temperature is hotter than 25C, the intersecting resistance will be a little less than 10,000 ohms. If your ambient temperature is cooler than 25C, the intersecting resistance will be a little larger than 10,000 ohms.
6) Next, heat your kitchen oven to an indicated 55C. Place your most accurate thermometer in the oven and, as needed, adjust the temperature control knob until the thermometer reads a steady "55C".
7) Connect your presumed thermistor to an ohmmeter via heat resistant insulated wires and place the thermistor and a portion of its test leads into the oven. When the ohmmeter stabilizes it should be reading about 3,000 ohms (plus of minus, say, 1,000 ohms).

If those two observations (in (5) & (7)) are as expected, you have a functional battery thermistor.

[Gordon]

... it stay recharging a long time and do not run. ...

If the robot was actually applying high-rate charging to your test-battery, I think there is a good chance that it got way over-charged; and, way too hot!

Did you happen to touch any of the AA cells after charging had been in progress for a couple hours? They might have been very hot to the touch!

R3-Roombas have several means (I think) to terminate high-rate charging. The most often used method is to cut off charging when the thermistor temperature reaches a specific value. I don't think we know that exact number, however I ran three hi-temp cut-off trials on either a 510 or 560 main_PCA and determined that high-rate charging is halted when thermistor temperature reaches the mean temperature of: <66.4 C>. General thinking is: That number is about 20 degrees Celsius higher than is safe for NiMH cells!

Your test cells were not thermally connected to the thermistor, so the thermistor remained cool while the cells could have gone way high in temperature. I believe that R3-Roomba's have a time-out feature that come into play when none of the planned means to terminate charging are invoked.

Now then, I think I have an answer to this posed problem:

... it ... do not run. ...

Roomba does not run because it senses a dead battery as soon as its motors become powered.

When Roombas sense a battery voltage of 12 volts, or less, it stops all motors and signals a need to re-charge. "How can the 530 'think' the battery is dead? I measured its no-load terminal voltage as 15.3V !", you might say.

The following is what is very likely going on:

A) First, you need to grasp several of Roomba's operating values:

1) You have just learned that Roomba stops the Clean Mode when battery voltage is sensed to be nominally 12-volts or lower.
2) You also must recognize that Roombas draws between 1.5 to 2 amperes dc current from its battery while sweeping floors.
3) It is also useful to know that high-rate charging current of a depleted battery begins at about 1.3A (for a 1.25A, nominal charging PSU). As battery charge rebuilds over the charging period, charging current diminishes.

B) You have chosen to use "battery-holder" to grasp and serially connect your 12 cells. This choice is the core of your operational problem. vic touched on it, but detailed the effect for charging (as I recall), not for Clean Mode. There is no fault with the holders mechanically grasping groups of cells, but there is a huge problem with the holder's contact resistance to each cell terminal. This is how it goes:

1) Each cell terminal has two terminals, and there are 12 cells, hence 24 terminals, 24 contact points, altogether.
2) We can't know the value range of contact resistance for your cells and cell-holders without actually measuring each one, but as I recall the number can be on the order of tenths of an ohm per contact for a spring-force electrical contact. Let's use 0.1 ohm as a trial.
3) The overall series resistance attributable to contact-resistance is then: R_cntcs ~= 24 * 0.1 = 2.4 ohms.
4) I think we now have everything we need to calculate the total voltage drop within your test-battery when it tries to support Roomba's Clean Mode current. I presume you are familiar with the Ohm's Law expression: E(volts) = I(amps) * R(ohms), or, for this case, I can say: E_terms ~= I_cln_md * R_cntcs; followed by applying values from (A,2) and (B,3), to get: E_term ~= 1.5A * 2.4 ohms = 3.6V.
5) Finally, you must appreciate that as Roomba is switched into Clean Mode the 1.5 to 2A current will be drawn from the battery and the initial battery terminal voltage, E_ini_term, of 15.3 will fall by the E_term voltage dropped across the summation of contact resistances. IOW, E_BAT ~= 15.3 - 3.6 = 11.7V; and, it is then that Roomba declares battery voltage is now less than 12V and halts operations.

===============
Similarly, battery charging is impacted by the sum of contact resistances. The net result is battery terminal voltage is a few volts lower than its could be if contact resistance was made closer to zero by spot welding metal tabs between cells.