I give up!

Inside the Roomba and Scooba and more, Cool mods, Repair and Upgrades - including the all new iRobot Create Kit. Let's void that warranty baby!
glo69
Robot Addict
Posts: 521
Joined: February 25th, 2007, 8:18 pm
Location: Bakersfield, California

Post by glo69 »

Hello Pierre

Luckily, so I don't have to do a lot of research and attribution, you aren't really disagreeing with my statement. You are just asking a question that goes beyond the parameters of my statement. I don't see that you have actually questioned whether less depth of discharge = more cycles.

I do understand what you are asking/saying. I talked about the same idea in a post to Alice in this thread http://www.robotreviews.com/chat/viewtopic.php?t=5859 However, there is a major difference between her situation and yours. In her case she was doing 2 runs with her Scooba, using about 85% of the battery mah and then doing a partial 3rd run, intentionally using the other 15%.

In your case you are, necessarily?, using 100%. I'm going to do a post to Gordon going into deeper detail about NIMH voltages and Roomba shutdown next, but my short answer to you would be, it depends. If the Roomba is directly in front of the home base when it reachs the voltage that triggers it to go home, it probably won't damage the battery very quickly. If it takes it a while to find home, or runs down to the shut off voltage without finding home, then IMHO it could damage your battery quickly, and could just kill it someday.

Whether the cycle difference is straight line is, I think, unknown, unknowable, and improbable. My (educated?) guess is that it is not. However, with any batteries, the battery chemistries are so affected by the impuritiesi n the components that little can be said for sure. That is the reason we'll never know how an individual Roomba battery will last in any given situation. I could have a battery that I do everything, supposedly, right, and you could have one that you mistreat in many ways, and yours might last longer. Batteries are like gambling. If you play by the odds, you will come out ahead in the long enough run if the odds are in your favor, but that has no effect on the short run. Thus Casinos make money in the long run, no matter what happens any given night.

All you can do is play the odds, and the odds aren't good on using cheap batteries (iRobot) down to deep discharges on a regular basis. However, if it is much handier for you to do the Max runs, and you don't mind replacing your battery every 10 mos. instead of 14 mos, have fun. I would probably do the Max runs. To me the convenience would be worth an extra $10 or $15 a year. :) :)
Gordon
Robot Master
Posts: 4304
Joined: April 6th, 2005, 2:02 am
Location: Santa Ynez, CA USA
Contact:

What is V_batt when the Power-LED goes RED?

Post by Gordon »

Gary, I have a red-light datum! This is one measurement of battery-voltage at the discharge-point when a 4210-Discovery's MCU decides to switch ON the red Power button LED. The datum is 12.2Vdc.

The *whole nine yards* follow.

Empirical-Goal:
Measure a Roomba's battery voltage at the instant its controller decides no further cleaning will be done until the battery has been re-charged. Constrain the operating conditions to exclude any influence by a Home Base.

Preparations:
1) Prepare a Roomba battery to have a mid-level charge condition (this is a matter of 'convenience', to minimize time to reach cut-off).

2) If battery has been partially charged in the Roomba, do a power-cycle RESET.

3) Connect a DVM to Disco's SCI-jack (V_batt_pos and V_batt_neg pins in that Mini-DIN connector may be identified by using info found in iRobot Create O-M).

4) Strap the DVM and its test-cabling to the top shell of Disco.

Operations & Temporal Data:
1) Log V_batt_stby, the slightly loaded case, with Disco in standby-mode: V_batt_stby = 16.4Vdc (initial).

2) Place the apparatus on carpeted floor (this carpet is defined here), and Press 'Start', then 'Clean'.

3) Periodically log the DVM's indication as 'V_batt', and add any comment:
a) V_batt,(V) = 15.7, just after robot began cleaning (@t=0)..
b) V_batt,(V) = 15.2, @t=+3 mins.
c) V_batt,(V) = 14.8, @t=+7 mins.
d) V_batt,(V) = 14.4, @t=+11 mins.
e) V_batt,(V) = 13.9, @t=+15 mins; Pwr_GRN is getting less green.
f) V_batt,(V) = 13.2, @t=+19 mins; Pwr_LED color ==> Pwr_WHTish / YELlowish.
g) V_batt,(V) = 12.2, @t=+20 mins; Pwr_YELlowish ==> Pwr_RED, Status to RED, robot halted, and Pwr_RED began flashing. Game over!

Note: The error in reading these dynamic voltages is easily on the order of: ?0.1V, and could be said to be twice that estimate.

Power was then switched OFF, but the DVM left ON. About four minutes later I noticed battery voltage slowly climbing; it had reached a V_batt_stby = 14V (post_run) level.

I wish I had a data-logger, say about the size of a pack of cigarettes! Such a device would make this sort of data capture soooooo much easier and yield better accuracy.

With such a gadget at hand, I would be more inclined to include measurements showing a Home Base's influence on the end point(s)!
User avatar
mfortuna
Robot Master
Posts: 5853
Joined: February 5th, 2006, 9:35 am
Location: NH

Post by mfortuna »

RadioShack sells/sold a DVM with built in serial port. You can log whatever measurement the DVM is set to.

I have one I used to graph battery voltage over time during discharge.

Mike
glo69
Robot Addict
Posts: 521
Joined: February 25th, 2007, 8:18 pm
Location: Bakersfield, California

Post by glo69 »

Hello Gordon :)

I don't have that elegant of a solution, but I figured out a way to get a close idea of V when the Max run decides to look for home base and at shutdown. I'm still testing and will see how the numbers match up, probably tomorrow.

I also intend then to add a little of what Buchman says about these voltages and NIMH batteries. A lot of what Mike had to say fits right in. I still don't understand the discharge to 0V at 30A and left shorted til next use though, but I think it would have to be on an individual battery basis, on high quality batteries, and not to a pack.

One thing I'll mention now that you might find interesting is that my testing has convinced me that the Roomba MCU not only uses voltage to end a Max run, but tracks mah used and will send Roomba to the Home Base when the mah indicate, no matter the voltage.
Gordon
Robot Master
Posts: 4304
Joined: April 6th, 2005, 2:02 am
Location: Santa Ynez, CA USA
Contact:

Post by Gordon »

mfortuna wrote:RadioShack sells/sold a DVM with built in serial port. You can log whatever measurement the DVM is set to.

I have one I used to graph battery voltage over time during discharge. ...
Right you are, Mike; I too have one, its their Cat. No. 22-805. If yours is that same unit, then I think "sold" is the operative verb.

I can no longer use that logging function, even on a bench set up. My new DT PC will not run the old 22805-S/W, nor does it have the required I/F connector (cable is DB-9, each end)!

The logger I would like, does not have to display any data, it just needs to write it to a memory card, that can later be read via a USB adapter. Notice too, that I can't be following Roomba around for an uncertian cleaning-mission period, with, say, a laptop connected to the 22-805. Now, IF I could adapt a memory stick to that DMM, all would be cool!

BTB, thanks for posting your NiMH info. That is a wonderful post! Not to worry about its length; I like a post packed with details!
User avatar
mfortuna
Robot Master
Posts: 5853
Joined: February 5th, 2006, 9:35 am
Location: NH

Post by mfortuna »

Thanks Gordon!

I haven't used the DVM in quite a while but it did work OK with a USB to serial port adapter on my dell inspiron laptop.

Right after I posted I said to myself a serial cable wouldn't be much use on a moving roomba.

I have played around with a basic stamp and bluetooth module on a roomba (see the hacker section). I bet it would be pretty easy to measure the battery voltage and send it back to a PC.

I've kinda put the whole roomba bluetooth control thing on the backburner. I need to learn how to set up a GUI under windows to get to were I want to go. I did get to the point were I can read status, turn motors off and on, and drive the roomba forward. Sorry to get off topic!

Mike
Gordon
Robot Master
Posts: 4304
Joined: April 6th, 2005, 2:02 am
Location: Santa Ynez, CA USA
Contact:

Post by Gordon »

mfortuna wrote:...I haven't used the DVM in quite a while but it did work OK with a USB to serial port adapter on my dell inspiron laptop.
Aye! At the back-face of my old Inspiron-8000 notebook, which is now disbled, there is a DB-9 ("10101") connector that the RS-DMM plugged into; and it is that system I had used to log data!
...I have played around with a basic stamp and bluetooth module on a roomba (see the hacker section). I bet it would be pretty easy to measure the battery voltage and send it back to a PC.
Thanks, I'll search for it.
User avatar
mfortuna
Robot Master
Posts: 5853
Joined: February 5th, 2006, 9:35 am
Location: NH

Post by mfortuna »

This should be moved to the hacker forum but you can use SCI to initiate a cleaning cycle and then use SCI to read battery status while the roomba is cleaning. The only concern I would have for my setup is mounting the stamp/bluetooth module so it doesn't get torn off going under things.

I think it would be neat to measure voltage, current, temperature, and distance travelled once a second and store the data on the PC. Sounds like a possible labor day project since my wife is working.

Mike
Gordon
Robot Master
Posts: 4304
Joined: April 6th, 2005, 2:02 am
Location: Santa Ynez, CA USA
Contact:

Post by Gordon »

mfortuna wrote:...I think it would be neat to measure voltage, current, temperature, and distance travelled once a second and store the data on the PC. Sounds like a possible labor day project since my wife is working. ...
Better you, than *me*! :-D It would take me all the time between *this* Labor Day, and the *next* one to get the job done!

For that project, and assuming Roomba is in a remote room doing its data-generation while you are doing other work on or near your PC, keep in mind these side-issues:
a) There needs to be a warning tactic to alert you when an unexpected event occurs--such as scraping off the Stamp & BT H/W, which you mentioned.
b) Some means should be provided to log when the Pwr-LED changes to RED (so the corresponding V_batt gets time-tagged).

Go for it!
User avatar
mfortuna
Robot Master
Posts: 5853
Joined: February 5th, 2006, 9:35 am
Location: NH

Post by mfortuna »

I was too busy working on an old Italian sportscar to do this last Monday, it will probably be a bad weather day project some weekend.

The roomba SCI doesn't supply LED status, it just lets you set the LEDs. So I'm not sure how to detect the LED being set to red.

I figure the first test run will be in a room with little danger to my hardware setup.

I'll record data until I lose contact, which would imply fully discharged. I'm using a 9V battery for the stamp and BT module so I won't need the roomba's power.

Mike
Post Reply