Author Topic: Dyson AM08 fan repair - help and guidance please!  (Read 4043 times)

0 Members and 1 Guest are viewing this topic.

Offline RichyJTopic starter

  • Newbie
  • Posts: 8
  • Country: gb
Dyson AM08 fan repair - help and guidance please!
« on: July 08, 2019, 05:25:02 am »
Hi all,

great forum (and "Hi Dave" if by some chance you're reading this  8) ;D). This is quite a long post because I've tried to include all relevant info; if you read it all then many thanks, if you don't have time then thanks for looking at least!

I've got a problem with a Dyson AM08 fan and have been trying to track it down but without luck so far. I would be grateful for some guidance on how to go about troubleshooting and finding the problem - as much for my education as to fix the fan really. If anyone recognises the problem and can tell me exactly what to do then great, but I would also like to try to understand what happened - I'm curious like that...

I have some experience with electronics but nothing like some of you gurus. I'm getting into the hobby after many years of fixing bits and bobs (mostly on cars before) and developed a real fascination for it. I already had a Fluke 123 Scopemeter and some cheapie DMMs, but recently added a DE-5000 LCR meter (based on the reviews on this forum) and a FY6800 function generator to play around with. At Christmas my wife treated me to my first 'proper' scope - I had an old Hameg analogue scope but didn't use it much, preferring the portability of the Fluke 123 - which is a Siglent SDS-1104X-E. I've really enjoyed playing around and looking at waveforms of capacitors and inductors, etc. I'm a doctor, so (hopefully!) reasonably bright and quick to learn  ;)

The fan had a squeaky bearing in the motor and eventually wouldn't turn on - showing just an F4 fault. Dyson are no use here, saying to just send it back but it's well out of warranty so I "took it apaaart" and found the BLDC motor was sticky. I carefully injected small amounts of motor oil (after some research) using a fine needle and freed it up so it rotated beautifully, then plugged it back in. It did spin a couple of times, but then stopped and after this, the whole unit was dead - nothing at all on the LCD display which even before had always shown something. There was also a funny, high-pitched whining from the PCB but I couldn't pinpoint the exact component.

Using my shiny new DE-5000, I measured the inductance across the BLDC coils and found one was much lower than the other two - I assume the oil damaged the varnish and shorted some turns, leading to some damage on a PCB component. I have access to a couple of good boards and compared various resistances. I noted that the resistance between ground and the 5V pin of the voltage regulator and the microcontroller VCC pin was only about a steady 40ohm, as opposed to a rising value in kiloohms on the good boards. I took this to mean a capacitor was no longer being charged by the resistance test and was shorted. However, I've inspected the board under my (fairly decent) 10x stereoscopic microscope and checked ESR and resistances around the PCB but found nothing that enlightens me - just lots of low resistances in the digital section. My Fluke only measures to 0.1 ohm, so resistance tracing isn't really going to work. I can't get a 121GW past my wife yet. The DE-5000 can measure to 1 mOhm and I've done the lead upgrade, but they're still too big to get on small pins, etc. The tweezers are good but only open so far.

I did notice that when I power the board the voltage regulator gets very hot very fast (ask me how I found that out... :palm:). I also only meaure about 240mV to ground instead of the 4.98V on a good board. I desoldered the VR and injected 5V from my power supply (current-limited at first, but it topped out around 100mA anyway - about what the KY5050 gives out normally). I didn't find any sudden voltage drops suggesting a short, but may have missed areas. The LCD did light up like this but the segments were random and didn't show anything sensible.

I was able to get hold of a burnt out PCB and desolder everything from it so I could put it on the scanner and follow traces under chips etc. I also used a trick I saw Dave use of inverting one side and positioning it on top of the other side with variable transparency to line up vias etc. It seems to be a 4-layer board.

My questions are:

- Can anyone explain how the low inductance damaged the PCB - which component(s) in particular would be affected?

- Does my capacitor theory sound right or would an IC input burning out show up like this too (worried about the MCU as I can't replace the firmware on that!)

- Is the hot VR a problem in that component or a sign that it's feeding a short - replaced it with a similar spec one but that got hot too...

- Is there an obvious test I'm missing which would move me a step closer - I'm out of ideas now...  :-//

I've attached various scans and photos if anyone's interested! I haven't included the LCD/power button daughter board as it doesn't seem to change resistances etc when connected.

The main components are:
Varistors:              07K300 (blue) and BC1382 (yellow)
Optocoupler:         MOC3083
MOSFETs:             FDD5N50NZ (six)
MCU:                   104BAA
Buck converter:    LNK304DN
Hex inverter:        SN74HC14
Flip flop:              SN54HC74
Half bridge drive:  73833 (three)

As I said, although a quick solution would be nice if it helped me fix the fan, I'm really interested in learning about what went wrong after I oiled the fan bearing and how that led to the problems.

If you got this far, congratulations and thank you!! I would be very happy to get further readings/pictures if anyone is interested.
 

Offline MosherIV

  • Super Contributor
  • ***
  • Posts: 1530
  • Country: gb
Re: Dyson AM08 fan repair - help and guidance please!
« Reply #1 on: July 08, 2019, 08:12:37 pm »
Quote
I can't get a 121GW past my wife yet. The DE-5000 can measure to 1 mOhm
Even the very good 121GW is not going to be any better at measuring very low resistances.
The only dmms that are any good at very low resistances use the 4 wire system (Kelvin Clamps), usually only bench meters.

Quote
Can anyone explain how the low inductance damaged the PCB - which component(s) in particular would be affected?
Do not know about the inductance.
You cannot expect a transistor to drive into a short circuit indefinitely at high current without damaging the transistor. So, with the damaged coil in the motor, one of the drivers to the motor coil has probably burnt out the transistor of the driver.

Quote
Does my capacitor theory sound right or would an IC input burning out show up like this too (worried about the MCU as I can't replace the firmware on that!)
Nope. Complex circuits cannot be boiled down to simply measuring the input resistance. Semi-conductor are just that, they can conduct or not conduct. What effect this has when you measure their resistance when un-powered is difficult to know.

Quote
Is the hot VR a problem in that component or a sign that it's feeding a short - replaced it with a similar spec one but that got hot too...
VR is not a recognised term or acronym. Vreg or just regulator.
Well yes, it has been burnt out by something further down circuit, probably the motor driver.
Replacing it will only burn out another one unless the primary fault is fixed.

I probably would have given up when discovering the motor was seazed.
That would have been an indication that the problem was not electronic but electrmecanical.
Now you have a duff motor and burnt out controller board.

Is it fixable?
Not ecomonically. You have to replace the motor and the controller board.
It will be vrry difficult to repair the controller board to component level without the schematics.
 

Offline RichyJTopic starter

  • Newbie
  • Posts: 8
  • Country: gb
Re: Dyson AM08 fan repair - help and guidance please!
« Reply #2 on: July 10, 2019, 07:06:26 am »
Thanks MosherIV for the detailed reply!

I know you're right about the difficulty due to lack of schematics; they seem to be jealously guarded by Dyson and I've found nothing helpful despite using my best Google-fu...

I've lifted a few bypass capacitors and now have the Vreg (thanks for pointing out my mistake regarding VR!!) supplying a steady 5v and only running warm, however, compared to the good boards there are now some incorrect voltages at two of the MCU outputs running to the LCD connector, so I suppose it might be a goner. I take your point that complex circuits can't be simplified to simple resistances.

I did realise that this wasn't likely to be economically repairable. Really, I'm just curious and don't actually need to fix the board - even if I did fix it, I haven't got a spare motor to run with it! I'm enjoying the journey and learning plenty along the way and it doesn't matter if it doesn't work again.

Thanks for taking the time to post such a detailed and considered reply. If nothing else, perhaps the board pictures will help someone else searching for a fix in the future!

Cheers!

 

Offline ahakman

  • Regular Contributor
  • *
  • Posts: 87
Re: Dyson AM08 fan repair - help and guidance please!
« Reply #3 on: July 10, 2019, 10:40:29 am »
Did you measure the motor coil inductance with the motor connected to the drive circuit, or disconnected from the drive circuit?

It sounds like one of the motor drive FETs is shorted, and thus dragging the power supply down of the whole board. If you measured the inductance of the motor with it plugged into the control board, the shorted FET will affect the inductance reading.

2 possibilities:

1) when the bearing siezed, the driver board fed a lot of current into one winding, melting the insulation AND shorting the FET, or

2) the motor is fine, the driver board is bad. If you have multiple boards, plug another board into the motor. Oil is not going to change the inductance of the motor windings (unless your "oil" is more like acetone, or xylol, something that could potentially melt the varnish insulation on the windings)

Measure the motor disconnected from the control board - that will tell you something.

Also, measure resistance from the source to the drain of all 6 mosfets Q2 Q4 Q9 Q3 Q5 Q7 (the 2 terminals with the larger traces - the terminal with the small trace is the gate) - I'm guessing at least one of them is shorted (low ohms). With the board off, and the motor disconnected, you should have high resistance between the source and the drain of all the mosfets - high k-ohm to m-ohm territory.

Another trick to find what's causing the short, cover various parts of the board in rubbing alcohol - wherever the alcohol evaporates fast, that's getting hot, and is the cause of your short. That's basically the poor man's thermal camera.

You don't really need schematics to fix most things - intuition and experience will get you there in most cases. Schematics can help with the "tricky" cases!
« Last Edit: July 10, 2019, 10:48:33 am by ahakman »
 

Offline RichyJTopic starter

  • Newbie
  • Posts: 8
  • Country: gb
Re: Dyson AM08 fan repair - help and guidance please!
« Reply #4 on: July 10, 2019, 03:55:11 pm »
Thanks ahakman,

The inductance of the motor was tested disconnected from the board - ie via the pins in the motor connector - so I'm pretty sure the motor is knackered. I don't want to test it on a known good board because when I did that once before that board stopped working too and displayed these same symptoms as now.

I think the controller on the board can 'sense' the motor seizure (maybe from the Hall sensor) and shut itself down - it then gives an F4 or F2 message on the LCD. Not sure what the difference is and Dyson won't tell. When this has happened before I've often managed to get the fan up and running again by swapping in a good motor from another unit (yes, I've done a few of these now!!). I have also managed to relubricate a seized motor and have it running normally again (you can tell a motor about to seize by the fan's current draw on speed 10 - about 340mA seems normal).

The MOSFET tests showed all good - about 16Mohm on all, so I don't think it's those either.

I have done the freeze spray test before but it just showed a hot Vreg and nearby components, didn't see anything else. I might repeat that now I'm seeing 5V output (which I wasn't before).

I'm pretty sure the problem is somewhere in the digital logic part. All the other voltages - mains (yes, Isolation Transformer used!!), 12V and now 5V working, but a few inputs/outputs to/ from the MCU are different than on the good board measured under the same conditions (ie disconnected from motor, etc).

The abnormal pins are:

4 and 5 reading steady 5V instead of a pulsed waveform
17 and 18 reading 0V instead of a 50% duty cycle 5V peak to peak squarewave
25 reading 0V instead of a high baseline 5V with pulses to low
26 reading 4.4V instead of a low baseline with high pulses
27 reading steady 0.265V instead of 0.8V low baseline with high pulses.

Not sure what this all means - more detective work required!! Need to look again at the datasheet to see which are in and which are out.

May all be a wild goose chase, but I'm having fun!
« Last Edit: July 10, 2019, 03:57:24 pm by RichyJ »
 

Offline fzabkar

  • Super Contributor
  • ***
  • Posts: 2221
  • Country: au
Re: Dyson AM08 fan repair - help and guidance please!
« Reply #5 on: July 14, 2019, 08:31:08 am »
@RichyJ, U9 looks chipped. Was that your doing?

The first things to check in a microcontroller are:

1/ Power supply rails -- are they up to spec?
2/ Crystal (if it has an external one) -- is it oscillating?
3/ Reset pin -- is the uC being held in the reset state by some external logic?

Is the uC sensing your control inputs, ie can you see the switches or remote control sensor change state?

Can you identify the uC's ground and Vcc pins?


« Last Edit: July 14, 2019, 09:18:33 am by fzabkar »
 

Online NiHaoMike

  • Super Contributor
  • ***
  • Posts: 8973
  • Country: us
  • "Don't turn it on - Take it apart!"
    • Facebook Page
Re: Dyson AM08 fan repair - help and guidance please!
« Reply #6 on: July 15, 2019, 04:18:51 am »
Might it make more sense to just replace the motor and driver with a R/C BLDC and ESC? Or replace the whole fan assembly with a computer fan?
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.
 

Offline RichyJTopic starter

  • Newbie
  • Posts: 8
  • Country: gb
Re: Dyson AM08 fan repair - help and guidance please!
« Reply #7 on: July 18, 2019, 04:40:43 pm »
fzabkar -

Thanks for replying. No, I haven't chipped anything - the photos of the populated boards are lifted from a parts website as it's difficult to get good quality pictures (didn't feel like getting my SLR out!!).

To answer your questions:

1- power rail to the MCU gets 5v - same as a good board
2- can't find a crystal on the board; not sure how clocking is done although the MCU datasheet mentions an internal clock which can be piped out
3- RESET (pin 2) is at 5v - same as the good board (the datasheet has a bar over the RESET - get confused over whether this means reset is signalled by high or low!!)

The MCU receives the IR packet, as confirmed on the scope. I haven't yet found what happens in response on the good board.

BTW, are you the same fzabkar who posted tons on various hard disk recovery forums? If so, you helped me recover data from a 7200.11 with the BSY bug in firmware - thanks a million!!


NiHaoMike -

you're right that that route would probably work out easier. TBH, I don't really need the fan to work again, it's more just curiosity driving this now. I'm certainly learning plenty about various electronics topics and PCB design. It's the first time I've used the scope for anything other than just messing around too - bonus! I've been watching the MOSFETs high speed switching and the Hall sensor readings associated with this and learned quite a bit about PWM driving a motor (300V  :o - glad I got that isolation transformer!).


Sorry for the delayed reply - why does work always have to get in the way...?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf