Examining a faulty handheld vacuum cleaner

[Jeff Layman]

We bought a Hoover Jovis+ handheld vacuum about 6 years ago. It was very reliable, and about 6 months ago I thought it was time to replace the filter, which had been cleaned many times. So I bought a pack of three filters. I should have known...

Today, after a couple of minutes' use, it just stopped working. Nothing. No lights, and no lights on placing in the charger. First challenge was how to get into it. I could see six obvious screws, and two more in an awkward position. I got them all out, but of course, the cleaner didn't come apart. It had some plastic snap clips, but they didn't seem to be the main issue. After 20 minutes, I found another couple of hidden screws, but even after removing them, it remained stubbornly in one piece. After another 10 minutes of brute force and ignorance, I found the plastic clips holding the main case on. Releasing those got me to the on/off switch and the led charge-state display. See <

> and < >. The transparent blue thing next to the switch seemed to be a fuse, but it was ok. I couldn't see the battery pack, but the only place it could be was under the motor.

The next step was to unplug the battery leads and see what the voltage was. It should have been 18V, but I was getting only about 0.3V. I could see a couple of small screws holding the battery pack cover on. They were a small torx size, but a narrow flat-bladed screwdriver fitted well enough to undo them. The cover still didn't come off! I then found another couple of the torx screws on the underside of the cover. After removal the cover came off to reveal five 18650 Li ion cells. There was also a circuit board of unusual shape which fitted between the top of the cells and the cover. I carefully pulled it aside as far as possible to see what was underneath. It was a rather complicated little board - see <

>. Testing the cells before they connected to the circuit showed they were fully charged at 18V. It was only after the

+ve lead entered the board that the voltage dropped. I examined the components with a 10x loupe, but there was nothing obviously faulty. The main active components were marked:

NCE30812K CG42 There were two of these. An NCE3080K appears to be a power mosfet.

S3G

3A diode

D5028A 502

28 pin IC. Controller for charging the cells?

So it's had it. I suppose that I can make use of the 18650s, and perhaps the 18V motor (supposedly rated at 120W!). Even if it was possible to get a replacement battery pack of that curved shape, I'm sure it wouldn't be far off the price of a new vacuum cleaner.

[Theo]

It sounds like it's tripped the protection MOSFET, so the load is disconnected from the cells. I wonder why?

One thing, sometimes the battery and the motor control are the same unit - so the unit is 'off' and they use the same MOSFET for protection as for on/off control. If you didn't have everything plugged in, or something went wrong with the trigger/etc, could exhibit that behaviour.

In normal battery land, the protection MOSFET could be tripped by over temperature (of the battery, but also the motor if that has its own thermistor): did your new filter cause the unit to get too hot? It's a common problem with Dysons when the filter clogs up: they run for a few seconds, but there's no airflow through the motor so it overheats and trips overtemp.

Or there was something 'wrong' with the load: overcurrent, cell undervoltage (ageing cells which drop out under load, causing an undervoltage lockout?) Or dodgy charging, although probably not in this case.

If it took 20 minutes to disassemble, probably a potential overtemp condition has cooled off. So I'm a bit surprised the controller didn't self-reset.

I think this is the chip, R5436T502BA-E2:

doesn't seem to have any nonvolatile storage, so it might be worth a try power-cycling it. Since power is coming from the cells, it would mean disconnecting the cell wires from the pack so power is removed from the chip.

Theo

[The Natural Philosopher]

Back in the day us aeromodellers would pick up any reduced price power tools purely for the battery cells.

The motors were occasionally used for flight but were rather heavy for the power. I would expect a 160W motor to weigh in these days at around 40g. a bit under 2 oz.

[The Natural Philosopher]

As do what you have there. probably worth replacing. Most power mosfets are similar - only slight variations in spec differentiate them

yep.

Yes. It is a battery protection chip seemingly.

Might be worth getting a spare.

Things to note. There seems to be a low value resistor R51 - check that hasn't died.

You don't show the motor - I assume it is a standard chinese can brushed motor? Have you checked to see if it spins up? wired direct to batteries?

I agree that the total loss of power after entering the board seems suspicious. But what does it mean 'the voltage dropped' - which voltage? - and is there an on off switch?.

[Jeff Layman]

As it had only been running for a couple of minutes, I don't know. I assume it's due to either the overdischarge detection voltage or discharge overcurrent detection voltage.

The motor "control" is on or off. I can't see any special state such as an "eco" mode.

There was no sign of overheating; the cells are in an enclosed plastic cover which would tend to insulate against heat loss, and although it took time to get into the pack it would still have been warm if it had overheated. There was a sub-miniature thermistor up against one of the central cells which is the overheat detection. I can't see any motor heating protection - there are just the power wires to it. We've run the vacuum for longer without problem, both with the original and new filter.

That's what concerns me. Overheat, overcurrent, etc should have allowed a reset after a while, but apparently not in this case.

Thanks for that. My internet search didn't turn up that chip, but it does look like the right one. Looks like the important pins are 5 (DRAIN) and 6 (DOUT) according to the circuit on p36, and the fets MDO and MDR. I've identified MDO, but there are no markings on MDR. Unfortunately, notes [13] and [14] don't give recommended types for the fets.

It's not easy to follow the circuit on the board; it looks like a multilayer PCB (in fact the "Power dissipation" measurement conditions on p44 refer to the board material being "Glass Cloth Epoxy Plastic (Four-Layer Board)"). The component designations are, of course different on the board than those in the circuit, which doesn't help.

Anyway, I unsoldered B+ from board, left it for 30 minutes, and resoldered it. It's still dead, I'm afraid, so looks like a new vacuum is on the cards.

[Jeff Layman]

This sort of thing is about 220g:

[Jeff Layman]

As the pcb appears dead to both charging and running the motor, I would think it's the main IC at fault rather than one or both of the mosfets.

Getting the chip probably wouldn't be a problem. Soldering it in definitely would be! I would have to cut the old chip out leaving its "stumps" to solder the new one onto as it's a multi-layer board. AFAICT, the chip's connections are mainly within the board. There are some on top.

That appears to be Rsens in the data sheet Theo found. According to that it's around 100mohms. My DMM won't measure that far down, but shows it is 0.4ohm, which is near enough for me.

Looks like a standard 18V motor. A search didn't find the serial number on it, but some of it may be hidden in the surround. It works perfectly when connected directly to the battery +ve and -ve terminals.

Just that it read 0.3V at the +ve output and should have been 18V, so something on the pcb was preventing the full 18V appearing.

Thanks for the comments. It's like many of the specialist circuits these days - if you've got the proper test gear it's probably possible to find where the fault lies. Whether it's worth it or not is another matter!

[Theo]

Is the trigger switch definitely working?

Does power go directly through the switch, or does it use these MOSFETs for switching with the trigger?

For a power cycle I think you'd need to desolder all the cells, otherwise the chip could be parasitically powered through the other cell inputs.

The only other thing I can think of is to check the inputs and outputs to the chip: are the cell inputs all good voltages, are the sense pins showing the right things. For example if the sense resistor or thermistor were to have a dry joint they could go open circuit, which could cause the chip to think there's massive overcurrent or a very cold temperature and go into protection shutdown. Or if the cells have got sufficiently out of balance and the board doesn't have the cell balancing transistors to enable it to rectify that (I can't see them in the picture, but they could be on the other side of the board).

It's possible the chip is just dead, of course, but it's also possible something in its environment is 'not right' and it's decided to take its ball home.

Theo

[The Natural Philosopher]

We used to get about half a horsepower out of those. Didn't last very long mind...

[The Natural Philosopher]

slim chance but worth it. I wonder if the diode is there to stop the batteries being overcharged. Bit actually drives the circuit when not on charge. If that goes then board will appear dead except for charge

[Chris Hogg]

Oh thank you for that! My little hand-held Dyson, bought a good few year ago, was doing just as you describe. I contacted Dyson and they suggested the battery contacts needed cleaning, which didn't make a lot of difference. Apart from that, I assumed the battery was failing. In fact I didn't realise that there was a filter - my late wife was the one who used it, and I 'inherited' it, as it were. Opened it up and the filter was heavy with gritty dust (which doesn't say too much for the cyclones' ability to remove it from the air-stream). Gave it a thorough rinse under the tap, dried it out in the airing cupboard, and it's working much better now - it even sucks!

[Jeff Layman]

Yes. If I bypass the control electronics but connect the batteries directly to the switch, as soon as I push it on the motor turns.

I didn't check, but I think all power goes through the fets as they are the devices which limit the discharge current.

Perhaps given the right test gear, several days, and - if I identify the fault and the spare(s) are in stock - and I get it working again, then that's fine. But it doesn't deal with The Management's requirement for a working vacuum cleaner for those little spills!

There are many new ones out there in Argos, Curry's, Amazon, etc. It's just a matter of finding the right one. Amazon have the same one, but "currently unavailable", and a similar one, but it is a 15.6V NiMH-powered model (

)! Strangely it seems to have been made available after the 18V Li-ion one. But then can you believe anything in Amazon's figures when both the Jovis+ vacs have a stated noise level of "1dB"!!!

[Theo]

:-)

On some (I have one built from a collection of DC59 and V6 parts) there are two filters: one in front of the motor (rod filter that goes down the middle of the cyclone) and there's also a post-motor filter. On the HEPA filter versions it twists off as a separate piece, on the non-HEPA ones it's just some foam inside the plastic casing where the air exhausts. If you didn't get the post-motor filter it's worth doing that one too.

The standard advice is to let it dry for 24 hours before reinserting, because apparently any dampness can cause motor damage.

I dislike the plasticy construction (easy to break) but the suction is very good, as are the accessories, and there are plenty of them out there that it's easy to get parts or even whole spare machines for not much money. Especially if you know the trick about shutdown due to a clogged filter as above :-)

(although they try and stop you interchanging parts from different models with little plastic tabs - but you can cut those off)

Theo

[Mark]

If the voltages of the 18650s are ok and you have tested the motor runs ok from an external power source just fit a new BMS

[Jeff Layman]

The problem is space and shape. As you can see from the third link in my OP the battery pack is curved, and the BMS PCB is curved to fit it. There is almost no space outside to fit anything, even something like the one you linked to or others on eBay.

[Dave W]

But did you examine the soldering on the thick wires? There is a condition where big components get shaken in the factory after being soldered but before the solder completely sets, because they retain heat. Years later, a circular crack appears round the joint, disconnecting it. It's suspicious that the fault only appeared after you had moved things about.

[Jeff Layman]

The vacuum cleaner had been working perfectly for over 5 years. It hadn't even shown the odd hiccup in use. As it happens, to test whether the battery pack worked directly with the motor I had to unsolder one of the thick wire connections (B+) from the PCB. I couldn't see anything wrong with it or B-, and, as the circuit went through the on/off switch, I assume all the soldered joints were fine as the motor worked.

I've got to look at it again to see if any of the BMS PCBs available on eBay or elsewhere might fit somewhere else in the case, but I doubt it.

[The Natural Philosopher]

They might prove a source of components to repair the one you have.

[Clive Arthur]

<snipped>

Get a cheap broken one from Ebay, eg...

[Jeff Layman]

That's unfortunately a corded vacuum cleaner. But I did find an 18V one of interest. Hoover gave the same name - Jovis - to li ion, NiMH, and even mains corded vacuum cleaners!