Neato replacement - horrible battery life - resolved

[ormandj]

Hi,

I just wanted to give my feedback regarding the battery life issue many people see with new/replacement Neatos.

I had an XV-11, it died on me a half week ago or so (wouldn't turn on, just made some odd numbers then shut back off). Talked with Neato, they shipped me a replacement. Received it yesterday. I fired it up, did 'new battery', let it charge during the day, then kicked off a run. It went back to charge after 28 minutes. 3 cycles, all around 20-30 minutes, never completed the bottom floor of my house. I let it charge, rinse and repeat, 3 times, 3 cycles, throughout the night. It never did more than 33 minutes on a run.

Fast forward to now. I pulled out my 3800mAh Amazon battery pack, plugged it in, did new battery, and let the Neato charge. First run: 63 minutes (and 9 seconds) until the brush/vacuum shut off and it returned to base. It completed 3/5ths of my home in one pass. I've seen it said before, but my experience seems to indicate that the battery packs Neato are using, are now terrible. I don't know if it's dead cells or what (I'm not going to tear the pack apart to find out until after I talk with Neato about this) but I've seen a lot of forum posts with the same problems, who get replacement packs/robots from Neato without any resolution. If you're in the low run-time boat, and are willing to experiment, grab one of the 3800mAh packs from Amazon. I think it was mislabeled as 3500mAh (still more than the stock supposed 3200mAh) and ran somewhere between 30 and 40 USD. I had gotten one for my old robot before it died, so had one laying around, and it completely correct my runtime problems.

The new-to-me Neato is Rev 113, running 3.1. My house is 75% relatively deep carpet, with the rest tile, and I am using the pet brush and filter. I also have a dog that sheds a dog every few hours, and due to not having a working Neato for near a week - I'm on bottom floor cleaning #4, and I still have to empty the dirt bin of hair every 15-20 minutes or so. Pretty much the worst case scenario, running the 3.1 firmware, and getting over 60 minutes of battery life. It wasn't even an overnight charge.

[vic7767]

You've pointed out what many of us have experienced with the newer model (XV-21) Neatos. There was a run of poorly built (lowest bidder) battery packs that Neato is replacing when the low run times are reported to them. I've built some custom batteries that are 4500 mAH that improved the run times somewhat but nowhere near an hour.

What model did Neato send you ?

[ormandj]

You've pointed out what many of us have experienced with the newer model (XV-21) Neatos. There was a run of poorly built (lowest bidder) battery packs that Neato is replacing when the low run times are reported to them. I've built some custom batteries that are 4500 mAH that improved the run times somewhat but nowhere near an hour.

What model did Neato send you ?

The model they sent me (which gave no better than 30 minutes life):

McNair
Ni-MH 4/3A3200mAh 7.2V

I'm not going to disassemble the pack to get the cell specifications, but my guess is they are either poor quality cells, or horribly balanced, though long-term charging didn't seem to fix it, so perhaps a dead cell?

The model I got from Amazon (in the vacuum running right now, so can't get the label) are listed as "Neato XV-11 Vacuum Cleaner Battery 945-0005 (7.2v 3500 mAh 25.2 Whr) Battery - Replacement For Neato Robotics 945-0005" , sold by East Coast Photo, and are mentioned as 'by' Synergy. They are listed as 3500mAh, but the label lists them as 3800mAh. Some of the reviewers have commented the same. Looks like they are $46 for a pair right now.

My current run is a second pass, and the Neato is just over 63 minutes right now, the battery is showing low on the display, but the Neato is almost finished cleaning my bottom floor (two passes for about 1500 ft^2, 75% heavy carpet, with pet brush/filter). I will say, after the first 60+ minute pass, it did take a good 4 hours for it to start the second pass. Not sure if this is because I selected "new battery" after installing the new packs, or just the additional time required due to the charging method used by the Neato. I'll know more once it's fully charged again and I kick off yet another round of cleaning (I've got a week's worth of dog hair which is going to take 5-10 passes to collect due to how much my dog sheds/how deep the carpet is).

Either way, this exact same XV11 with the same firmware was 30 minutes or less per pass with the included batteries. Something isn't right with them, and yes, I attempted 'new battery' and did overnight charging, neither made any difference. No matter how many passes I ran it, it was not improving.

[vic7767]

What model Neato did you install the 3800 mAH packs into ?

[a4refillpad]

the OP mentioned a rev113 neato running 3.1.

[glnc222]

The Prius hybrid car manual says their battery is never discharged below 30 per cent in normal use, and is warranted for 10 years etc. 70 per cent of the 4500mah batteries is 3150 close to Neato's original 3200. 70 per cent of 3800 is 2650; is that less than Neato's draw in use? If Neato draws 3 amps (not known exactly) and runs 50 minutes as observed on carpet, it uses 2500mah in the run -- within a well-margined 3800 battery. It would be interesting if the much cheaper 3800's can provide as much life as the 4500's. Special equipment is needed to measure battery life in reasonable time through some accelerated, intensive procedure -- more difficult than just determining capacity. It's interesting how battery sellers have very little data about their batteries and their applications, or at least never publicize it. Maybe they don't like to admit they run down at all.

[ormandj]

What model Neato did you install the 3800 mAH packs into ?

XV-11 Rev 113 running FW 3.1.

[ormandj]

With higher capacity packs, and assuming the Neato is charging based on overall voltage measurements and not just a counter (I'm almost positive this is the case, it would be nuts to charge based on a counter), the larger packs should be getting more mAh in the charging cycle. That should allow for longer run times before voltage droops below whatever level the Neato decides is 'time to charge'. That said, as the 'neato' user indicated there is a cut-off at 54W so you won't get double the run time if you manage to fit 6400mAh worth of cells into it.

That's why I feel the stock batteries I received must be faulty in some way, as it's obvious the Neato is fine, and also charging my packs fairly well. I just had a third run, 61 minutes on the first pass. We'll see how the second pass is.

Found a post by 'neato': viewtopic.php?f=20&t=16813#p113899

He explains it will utilize up to 54 Watts. At a 1.2v per cell nominal voltage, 12 cells total, that's 14.4v, ((54*1000)/14.4) == 3750mAh. That being the case, it sounds like you actually can get slightly more run time out of a larger battery pack. It would stand to reason you'd want the 4500mAh pack if you want to avoid as much of a discharge percentage, and theoretically longer runtime, as the 3800mAh pack is probably going to trigger whatever low voltage cutoff the Neato uses to determine the battery is nearly depleted (to account for weakened battery performance, for example) prior to achieving 54W of draw. If the 4500mAh packs aren't getting a full 3750mAh discharge (you could verify this by charging them with an off-board smart charger and metering the mAh applied during charging) perhaps the cells aren't rated appropriately. Wouldn't be the first time I've run into mis-rated cells.

The Prius hybrid car manual says their battery is never discharged below 30 per cent in normal use, and is warranted for 10 years etc. 70 per cent of the 4500mah batteries is 3150 close to Neato's original 3200. 70 per cent of 3800 is 2650; is that less than Neato's draw in use? If Neato draws 3 amps (not known exactly) and runs 50 minutes as observed on carpet, it uses 2500mah in the run -- within a well-margined 3800 battery. It would be interesting if the much cheaper 3800's can provide as much life as the 4500's. Special equipment is needed to measure battery life in reasonable time through some accelerated, intensive procedure -- more difficult than just determining capacity. It's interesting how battery sellers have very little data about their batteries and their applications, or at least never publicize it. Maybe they don't like to admit they run down at all.

+1 to your above point/question/thought. When I saw it was a 3800mAh battery pack, I immediately was hoping for more longevity out of it. Just having a working robot that can actually clean the floor is a good start. Beats the heck out of the stock packs I've gotten. My first Neato/stock packs (when new) managed to run 50 minutes or so, but this was in a different home with different carpet and a different set of obstacles. Now knowing there is a 54W cutoff, which is just beneath the total capacity of my cells, I really don't have a buffer. Based on my calculations above, that gives the 4500mAh about a 16.6% excess in capacity vs. potential total draw (assuming my math is correct).

The new packs were nothing like this, so what vic mentioned above sounds plausible re: bad packs. It appears they are still sending them out, at least with replacement units, as I received mine yesterday, the 23rd of Feb 2013.

[glnc222]

Thanks for your reports. Goes without saying there were bad packs. I look at other issues and may have confused readers.

From what I've read, voltage is not reliable as an indicator of battery discharge except as an indicator of full charge or full discharge, along with other data such as charging current. Tracking the battery during runs is supposed to require integrating current draw over time calculating actual charge withdrawal (similar to a watt hour household electric meter except voltage is constant). Neato indicated here recently the system has a maximum withdrawal figure it uses to make a run, a figure which is revised downwards if some voltage alarm goes off indicating full discharge.

The voltage battery discharge curve has a fast drop at the start, a long flat section rendering it unreliable in the middle, and a fast drop at the end -- when there is likely too little charge left to return to base. The New Battery option will reset the revised maximum charge (current x time) withdrawal allowed back to the default -- among other things not revealed. The maximum withdrawal limit means a bigger battery does not increase run time, which was one of Neato's reports. The maximum is coded in the program because it is impractical for Neato to measure the capacity of the battery; voltage cannot indicate that. My 4500 battery actually charges fully to a lower voltage than the 3200's, and runs the usual time.

When the regular battery wears out within a year of running daily, or within a year and a half as mentioned in the manual, the limited life is most likely due to deep discharging the battery to squeeze in enough work; hybrid cars achieve long battery life by avoiding damaging deep discharge. The same may be achieved with the Neato using a bigger battery. How big it has to be is a question. It is very likely a 4500mah battery has long life as such was specified for the VR100 using that size battery, though in LiOn. At that time this size may have been available only in that denser chemistry; now they have shown up in NiMh XV models can use. Whether the 3800mah hours are big enough to give longer life was my question. They very well may supply that. Too bad there is no precise data.

edit: report posted here at http://www.robotreviews.com/chat/viewtopic.php?f=20&t=16866 on wattage confirms Neato deep discharges the 3200 mah battery.

[mfortuna]

I have done a lot of work with NIMHs and NICADs when I used them for RC racing. The discharge curve is not flat in the middle. In fact it is very linear with a constant slight drop in voltage over time. I feel that voltage alone can be used to determine capacity.

[djos]

My XV-11's batteries have got down to 30mins run-time (from about 60ish a year ago) and I was going to send them to Vic for upgrading to 4500mAh but funds are a bit tight atm (bub #2 due shortly) and so I've ordered some 3500mAh Neato batteries from ebay seller "uk-expert". Unlike many of the other Neato battery sellers on ebay these guys seem to specialize in batteries so Im hoping they are worth the $50au (inc shipping).

Once I get them i'll report back on what sort of run-times they produce.

[Wing Nut]

Once I get them i'll report back on what sort of run-times they produce.

Great! I'd be interested to hear how it all goes. My 18 month old Neato gets run maybe twice a week, pre and post weekend, and although I have yet to notice a drop-off in run-times, I guess I should expect the batteries to start showing some signs of old age.

[djos]

Np, ours gets run 3 times a week and covers the entire 170m2 house - it used to do it in two runs but recently started doing it in 3 runs - just this week it stared missing half a room as It didnt have enuf juice on the 3rd trip to finish the house.

[glnc222]

I have done a lot of work with NIMHs and NICADs when I used them for RC racing. The discharge curve is not flat in the middle. In fact it is very linear with a constant slight drop in voltage over time. I feel that voltage alone can be used to determine capacity.

There may have been concerns with effects of temperature and current load on the voltage confusing its indication of charge level. Static environments and usage rates could render less noticeable.

[Osteo]

Hey OP, is this the battery, and if so was it a single or two pack? I'm out the door to go to work, looked quickly, and don't see mention of it being a single or dual pack of batteries... Thanks in advance!

http://www.amazon.com/945-0005-7-2v-350 ... s+945-0005

[ormandj]

Hey OP, is this the battery, and if so was it a single or two pack? I'm out the door to go to work, looked quickly, and don't see mention of it being a single or dual pack of batteries... Thanks in advance!

http://www.amazon.com/945-0005-7-2v-350 ... s+945-0005

Looks like it. I described it in one of my posts, and that matches the description. Came as a pair. Read customer reviews/description to be sure, I make no promises about something I'm not selling.

[Osteo]

Thanks for the reply. I did look around and the original url I posted would get you two batteries but the following flatly states 2 in the comments and the picture. Thanks for posting about 'em as my run time after a full year has really degenerated and was going to look for new batteries today, so great timing.

Here's the link:
http://www.amazon.com/Neato-Vacuum-Clea ... 269&sr=1-4

P.S. And by my run time I meant my Neato's, not mine, as after a few cups of coffee my run times are great!

[glnc222]

one of the 3800mAh packs from Amazon. I think it was mislabeled as 3500mAh

Perhaps the supplier cannot guarantee to always have 3800mah cells. Supply highly varied this industry.
There is also a question of how accurate the labeling is, whether hyped, and the distributor is covering risk.

[ormandj]

Received my second replacement, this one is revision 64. With the 3800mAh batteries I'm still getting 1+ hour on carpet with a fair number of obstacles, including 4 sets of angled dining room table legs (Neato's kryptonite). Very pleased! The r64 version seems to do a better job cleaning, I'm very happy with it. This one has firmware 3.1, as well.

[djos]

My XV-11's batteries have got down to 30mins run-time (from about 60ish a year ago) and I was going to send them to Vic for upgrading to 4500mAh but funds are a bit tight atm (bub #2 due shortly) and so I've ordered some 3500mAh Neato batteries from ebay seller "uk-expert". Unlike many of the other Neato battery sellers on ebay these guys seem to specialize in batteries so Im hoping they are worth the $50au (inc shipping).

Once I get them i'll report back on what sort of run-times they produce.

Arrrgh, 4 weeks later no new Batteries have arrived.

So I emailed the Seller who immediately refunded my money without me even asking them too - I was quite prepared to wait another week just in case Aus customs was to blame. Any was so happy with the instant excellent service I promptly ordered from them again! Hopefully the postman will behave himself this time!