Fluke 289 does not turn on -> blinking green power button

[ElecSeb]

My first post on this forum, but after reading a lot time to make an account yeeeaaahh.
However the reason for posting is not fun

I have a fluke 289 which does not turn on. I measured the powercap and it was too low (0.8V), so replaced it with a new one (now 2.0V) http://www.digikey.nl/product-search/en?keywords=338-3477-nd with this part.
However after replacement it still wont turn on. When inserting new batteries and pressing the ON button I see a green flashing pattern (no screen). I can press whatever button I like without result. Also turning off in not an option (only taking batteries apart).

Pattern of flashing: 5 quick flashes followed by a period of nothing and then repeat.

Any thoughts. I have had a look on the board and found no burned elements.

[AndyC_772]

Don't these things have a lifetime warranty? Just send it back...

[ElecSeb]

Not if you are not the first owner.

Update: Just found out that there is some sort of shortage between Cathode and anode of the diode C79 AND L1 (that area.
Plus and GND are connected.
Already desoldered some parts but so far without luck. Unfortunately not much time 2day and tomorrow on vacation. Needs to wait a week.

[retiredcaps]

Did you buy this 289 as working or non working?  If it worked, did it just stop working one day?

[ElecSeb]

Bought it non working. First started with the supercap.
By accident is measured in that area (where the plus sign is on the board) togteher with ground and found there is a short.
Should be in that region I suppose. Do not have an high res image at the moment since I am on vacation.

Would have started with looking up the IC which is close to the supercap.

[retiredcaps]

Bought it non working. First started with the supercap.

My understanding of the supercap is that it only keeps time and date and does not affect the functionality of the meter itself.

[ElecSeb]

That is what I read as well, however currently the + circuit from the battery is making a short with ground. Probably causing the meter to go bogus at this moment.
Will try to find a high def motherboard picture and draw in it where I measure zero resistance (single color).

[say]

That is what I read as well, however currently the + circuit from the battery is making a short with ground. Probably causing the meter to go bogus at this moment.
Will try to find a high def motherboard picture and draw in it where I measure zero resistance (single color).

Did you have any success with resolving this issue? I ran into the same problem today.

[ElecSeb]

Hi Say, at this moment I am still busy with fixing this meter (and 2 others with similar problems).
Currently I am on vacation and have no real access to the physical meters, but would you be able to share some high-res pictures of the board (bare board).

Furthermore do you have removed the supercap before measuring and did you clean the area AND did the supercap leak?

Please share this info and pictures and I will see if I can help you (based on my current knowledge of this meter).

For your info, I have ordered an MSP programmer to program the MSP430 chip and have a buspirate for the NXP/Freescale.
With one I am still working on the torec PWM chip, because it seems to leak. Once that is fixed they need to be washed and I can start with rebuilding.
(And new pictures will follow)

[say]

@ Elecseb

Here is a hi-res pic of the board. If this doesn't serve the purpose,  I can try for more pics of particular areas of the board. Taking a photo of the entire board has its own limitations.

The supercap had corroded top and it was obviously kaput as the meter didn't hold date and time when batteries were removed. It did not have any leakage as you can see in the second pic. So I did not feel the need for cleaning the board. Troubleshooting a multi-layered board is a daunting task. The available info on the internet doesn't show even a single success story

Happy vacations!!!!!!!

[ElecSeb]

Since I am on vacation I will try to explain which measurements you can take to see where it is possibly broken.
What I would do:

1. Remove the corroded supercap!

2. Voltage (when power terminals powered, but device is OFF):

• TP03 to GND
• TP17 to GND
• C67+ to GND
• C68+ to GND
• C89+ to GND
• E41 to GND

3. Resistance (ohm):

• TP03 to GND
• TP17 to GND
• C67+ to GND
• C68+ to GND

Btw, in my case it was most likely that the NXP chip was lost... I removed it from the board and my short was solved...
As I said earlier, I will continue after my vacation.

[say]

Thank you ElecSeb for great tips. I'll be on vacations too in couple of days. Will get back to this on return. However, I couldn't locate the NXP chip. Is it one of those on the other sid of the board?

[Electro Detective]

FWIW, mine has the crusty supercrapacitor still in there, and because of conflicting repair suggestions  and the fact the meter works fine (so far) with no apparent unreasonable battery draw,

I have left it as is and do a 6 month check on the stupidcapacitor.

Can anyone verify any possible ill effects of removing the cruddy supercapacitor and leaving the space blank? 

I don't mind resetting the clock on battery swaps etc

i.e. can the logging feature, software/firmware and or main board/s go weird if there is no capacitor device to charge ?

You would think FLUKE would have posted an addendum sheet by now addressing this ISSUE on their flagship meter   

[coromonadalix]

Unless mistaken Fluke never issued a recall for the supercap issue,   my 189 worked fine with no supercap while being ordered

Only time feature  / clock  did not work  until settled,  the clock will start a zero  .....  no problem ?

[Electro Detective]

If cap is removed and the clock needs to get set every time on a fresh set of batteries,

and the logging timer functions still work ok on both 189 and 289, and records/logs ok

then TBH I would prefer to have no supercrapacitor in those two meters, so it's one thing less to think about 

-----------------------

I can only assume that once the clock is set, the recorded dates and times during logging should in theory be kept in some internal memory with the trend info, (which should only apply to the 289 ?)

regardless of the condition or omission of the supercapacitor


So if it's only the current clock that resets when swapping batteries, with no ill affect on 189 and 289 meters,

my brown bearded caps will see the battery recycle bin asap,
replaced by a swabbing of isopropyl alcohol and air ! 

[ElecSeb]

Without supercap the meter should work just fine.
If you remove CR17 & CR18 (dual diode) it is clear that when using battery power (9V) the clock should remain its values (not tested since I do not have a working 28x).

However when 9V removed and supercap removed the clock settings are reset to the beginning of time.
The reason to remove the rusty supercap is to eliminate a possible short on the supercap.
Since we are not at this specific meter we are unable to judge (even if it is possible) to see if it is the supercap.

Furthermore, since it is corroded it is best that after the unit starts working it gets replaced by a proper replacement (Panasonic of some sort).

It is true however that I have not seen many repairs (0) where such problems are solved. 9 out of 10 it is an FET which is broken (see other topics). However this issue blinking LED is never solved and I spent many hours searching online. This is the reason why I am trying to figure the thing out myself.

Thank you ElecSeb for great tips. I'll be on vacations too in couple of days. Will get back to this on return. However, I couldn't locate the NXP chip. Is it one of those on the other side of the board?

Correct, the square black BGA chip U26 (this is the heart of the 28x and e.g. controls your LCD), the MSP430 U13 (seems to do all measurements like in all other fluke products), they are talking to each other through a bus (not yet found/defined). The MSP has special JTAG higher frequency and needs to be programmed with appropriate programmer and the NXP is normal JTAG.

[Electro Detective]

There are quite a few Fluke 289 owners at EEVblog to support a repair thread 

especially for non original purchasers,

and Ebay gamblers willing to take a shot at fixing the no go meters    


LOL, we may find out more about the units internals than Fluke does after a few user/repairers exchange of troubleshoot data.

AFAIK Fluke just swap out the flakey boards and perhaps contribute them to the nearest landfill ?   

[say]

@ElecSeb

Here is the requisite data:

1. Remove the corroded supercap!

It has been removed

2. Voltage (when power terminals powered, but device is OFF):

* TP03 to GND   --> 0.000V
  * TP17 to GND --> 3.232V
  * C67+ to GND --> 0.000V
  * C68+ to GND --> 0.000V
  * C89+ to GND --> 0.000V
  * E41 to GND.   --> 0.000V

3. Resistance (ohm):

* TP03 to GND.  --> 43K
  * TP17 to GND --> 628K
  * C67+ to GND --> 7.83
  * C68+ to GND --> 4.25K and increasing as it should.

I also detected shorts across C46 and C92 but can't trace where these caps are connected. It appears that there is a problem/short in LCD driving circuit.

[say]

I disconnected the 1.8V supply to NXP chip area that brought back most of the normal voltages on TPs except the 5 and 2.5 volts. These two voltage regulators are still disabled by MSP 430 because it cannot communicate with NXP chip. The blinking green power light also stopped blinking and is permanently ON now.

What might be causing a short in NXP chip or one of its tiny bypass capacitors is a question I might not be able to find out. Their micro size deters me from removing them for individual checking. Lot of tedious work ahead

[ElecSeb]

I did not have a lot of time (and will not have untill after Tuesday).
What you described with the small C sounds like what I measured. I have on 2 boards the NXP removed so I can check if they still have it, OR not. Furthermore I have 1 with NXP in place and can test that one. Probably this afternoon.

I have measured all C's of you described and was not able to find a defect. What you can (Thinking out loud):

1. Measure 1 C (C46 or C92), when not defect.
2. Connect a PSU with expected V and enough A.
3. Search with IPA or Flir for a heated element.

My conclusion so far is that MSP and NXP are delicate and easy to break. Not sure how, but I have 3 defect MSP and 2 defect NXP.
Still waiting for my MSP programmer...

[say]

@ ElecSeb, thanks for the tips. Perhaps someone could measure the resistance on C46, 55 & 92 with respect to the -ve battery pad. As well as on C79 + side. That will greatly help in isolating the fault.

[ElecSeb]

@ElecSeb
I also detected shorts across C46 and C92 but can't trace where these caps are connected. It appears that there is a problem/short in LCD driving circuit.

All 3 boards I have measure short (0.6 -> which on my old meter means short) for these 2 C (46 and 92).

Perhaps someone could measure the resistance on C46, 55 & 92 with respect to the -ve battery pad. As well as on C79 + side.

With respect to negative battery pad (for all 3 boards):

Code: [Select]

C46  = short -> 0 ohm
C55  = short -> 0 ohm
C79+ = +30kohm
C92  = short -> 0 ohm

Hope it helps.

[say]

It sure does help. I'm reading 8.7  ohms on C79+. I removed all the tiny caps reading same value but that didn't resolve the problem. Without the schematics it's hard to know what to look for now. Wish we could somehow find a schematic !!!!!

[ElecSeb]

Bin here:
http://hmpg.net/
And back, but no SM or schematic...

[say]

 [emoji106]   

[GeoffreyF]

Your link goes to something my computer detects as a virus.

Mind explaining what is there at your link?  Why should anyone click it?

[ElecSeb]

Just a normal page regarding end of internet, is around for ages (and often used)...

Page can be opened with any browser and tested with, KIS & Norton, so kinda curious what virus scanner/internet security you have...

[giosif]

Hi all,

I was just pointed to this thread, as it seems I am in the same situation as everyone else here: all three caps, C46, C55 and C92, measure short across their terminals (and are all tied to the battery negative - let's call it "ground" from this point onwards).
First, as I commented in the other thread, I find it strange for capacitors so spread out on the board to all exhibit this issue (not a an issue with the caps themselves, but an issue affecting them) and also in most cases with a dead Fluke 289.
Maybe they are all connected to the same power rail or data bus (and that somehow gets shorted whenever the NXP goes bad)?
Someone here mentioned they removed the NXP from the board and the short went away.
Can you please confirm that, with the NXP removed, the non-"ground" terminal of each of these three caps are linked to one another?

Also, is it just me but, having a really quick read of the NXP datasheet, it doesn't seem to have an internal storage?
If so, could we not just remove the dead NXP and replace it with a brand new one?

Thanks!

[Per Hansson]

Also, is it just me but, having a really quick read of the NXP datasheet, it doesn't seem to have an internal storage?
If so, could we not just remove the dead NXP and replace it with a brand new one?

Thanks!

I agree, it seems that the RAM is in separate IC's U25 & U27 and Flash is in separate IC U28, here is a review with high-res photos:
https://lygte-info.dk/review/DMMFluke%20289%20UK.html

[giosif]

I agree, it seems that the RAM is in separate IC's U25 & U27 and Flash is in separate IC U28, here is a review with high-res photos:
https://lygte-info.dk/review/DMMFluke%20289%20UK.html

Hmm... I suspect the challenge here is the reballing of the replacement NXP chip.

Nice, comprehensive review, BTW.

[mzacharias]

Hi all,

I was just pointed to this thread, as it seems I am in the same situation as everyone else here: all three caps, C46, C55 and C92, measure short across their terminals (and are all tied to the battery negative - let's call it "ground" from this point onwards).
First, as I commented in the other thread, I find it strange for capacitors so spread out on the board to all exhibit this issue (not a an issue with the caps themselves, but an issue affecting them) and also in most cases with a dead Fluke 289.
Maybe they are all connected to the same power rail or data bus (and that somehow gets shorted whenever the NXP goes bad)?
Someone here mentioned they removed the NXP from the board and the short went away.
Can you please confirm that, with the NXP removed, the non-"ground" terminal of each of these three caps are linked to one another?

Also, is it just me but, having a really quick read of the NXP datasheet, it doesn't seem to have an internal storage?
If so, could we not just remove the dead NXP and replace it with a brand new one?

Thanks!

I know this is crude, but it often works to feed a power supply at an appropriate voltage, and current limited as necessary, into the short, and simply wait a bit and see what gets hot.

If it's a true 0.0  Ohm short, then sure, no heat can be generated, but that's actually fairly rare.

I've found numerous bad microprocessors, chip caps, shorted tantalums, etc this way.

[giosif]

I know this is crude, but it often works to feed a power supply at an appropriate voltage, and current limited as necessary, into the short, and simply wait a bit and see what gets hot.

If it's a true 0.0  Ohm short, then sure, no heat can be generated, but that's actually fairly rare.

I've found numerous bad microprocessors, chip caps, shorted tantalums, etc this way.

Very good point, thanks!
Funny thing I've seen this done several times but, for some reason, it didn't really cross my mind in this instance (let's blame it on tiredness  ).
I will try that.

[giosif]

How much current am I supposed to run through this part of circuit?
I set my power supply to 1.5V and, initially, about 0.8A.
I then gradually increased current all the way up to 3A, but still no signs of heating on the board.
Heck, at 3A, the cable I soldered to one of the short points started to heat up, yet, I could not detect any hot areas/components on the board.

[giosif]

So, I bit the bullet and did what I have been considering for a while (but wanted to exhaust all other options before that): I've reflowed the NXP chip (i.e. U26).
This is because the seller of this meter mentioned they had dropped it, thus ending up with a broken screen.
So, I thought the impact may also have affected the connection between U26 and the board.

I am happy to report that, following the reflow, I am seeing a positive evolution: the 5V rail is now coming up and, consequently, the 2.5 reference voltage is present too.
Moreover, on pressing the power button, the LED beneath it turns on for a while, then turns back off, with the meter still drawing some current - from the videos I watched on Youtube, this is the expected behaviour.
Also, if I press the power button once more, the meter turns off (very little current draw from the power supply it is connected to).

The other good news is that, with the rotary switch set to Resistance, if I measure the current between COM and V terminals, I see very exact current being generated.
Moreover, if I press the Range button repeatedly, the current changes to very specific values (e.g. 100uA, 1mA, etc.).
So, at least some part of the meter is alive now.

Now, the not-so-promising observations (all this is done on the blind, as I have no working LCD):
- if I switch to Continuity test mode and short COM and V terminals, there is no sound generated (and I checked outside of circuit that the buzzer is working).
- if I set the rotary switch to DCV, for instance, and insert a lead into the u/mA or A jacks, the meter should beep to warn me I got the lead in the wrong place - however, I'm not hearing any sound.
- (still not sure if this is really an issue or maybe a red herring) the shorts on caps C46, C55 and C92 are still present.

[giosif]

It's ALIVE !!!

What I've done since my previous post was to test the IR connectivity (thankfully, I had the IR-USB cable for it).
And the connection worked:
- I issued the "ID" command and got the correct response back
- I started issuing some "QM" commands in various settings of the rotary switch and the meter displayed at least the correct function
- Suspecting the meter might have some custom configuration settings from the previous owner, I went ahead and issued a "RMP" to reset to factory defaults.
And, surprise-surprise, the meter beeped when I turned it off. 
So, of course, I then tested the input jack detection and I got the alert when the lead was inserted in the wrong jack.
Also, continuity testing is working now as well.
- Finally, I tested DCV readings from my power supply and they are pretty much spot on.

So, thanks to everyone here, it seems I've now got myself a meter in a fairly decent state (except for the LCD, of course). 
Thank you all for your help!

Now off to finding a replacement LCD for it...

[say]

Congratulations on excellent results.

[giosif]

Thank you, say!
I wish it gives a boost of hope to yourself and everyone here in a similar situation with their meter.

[giosif]

Just a quick update on this: I can confirm that the shorts on caps C46, C55 and C92 are NOT indicative of a problem.
My resurrected 289 measures short across those caps, yet it works just fine.
Moreover, I've recently got my hands on another working 289 (this one is the newer revision, with the coin-cell battery instead of the supercap) and the shorts across these caps is also present.

[niccowahyudi]

Can you tell me how you do it..?

[giosif]

Amm... Not sure if the question is for me and, if it is, what it is about (i.e. how I did what?).

[Black Phoenix]

Unless mistaken Fluke never issued a recall for the supercap issue,   my 189 worked fine with no supercap while being ordered

Only time feature  / clock  did not work  until settled,  the clock will start a zero  .....  no problem ?

No but if you mentioned to your AD that the DMM would have an abnormal consumption of batteries when turned off, they would ask you to send it back, even if you weren't the original owner (happen with me with my 289 that I am the original owner, and with one 187 that was bought used from a colleague).

At least in Portugal, they would service, exchange the cap/PCB for one fixed and even calibrate free of charge. From what I remember, mine doesn't now have consumption of battery but at the same time it doesn't keep the time when the batteries are removed (can't be sure, I don't have it with me currently to test).

[coromonadalix]

the 289  first revisions had a super capacitor who gave some problems,  normally  it should keep the clock ticking for a while ... the latest revisions had a coin cell if i recall correctly,  you could have a bigger current cosumption because of this

[Black Phoenix]

Never noted or even tested to be sincere after being service. I realised that the batteries were not being eaten at a fast rate as before - When it had the faulty Cap it was at a rate from full to half battery indicator in less than 2 weeks turned off.

Back in 2017 I tried with batteries a full 4 months in standby, no depletion of battery (I always take out the batteries if I use Alkalines instead of Eneloops).

Also worth mention that my 289DMM is one of the first ones to be manufactured, lower 200s serial number.

[niccowahyudi]

It's ALIVE !!!

What I've done since my previous post was to test the IR connectivity (thankfully, I had the IR-USB cable for it).
And the connection worked:
- I issued the "ID" command and got the correct response back
- I started issuing some "QM" commands in various settings of the rotary switch and the meter displayed at least the correct function
- Suspecting the meter might have some custom configuration settings from the previous owner, I went ahead and issued a "RMP" to reset to factory defaults.
And, surprise-surprise, the meter beeped when I turned it off. 
So, of course, I then tested the input jack detection and I got the alert when the lead was inserted in the wrong jack.
Also, continuity testing is working now as well.
- Finally, I tested DCV readings from my power supply and they are pretty much spot on.

So, thanks to everyone here, it seems I've now got myself a meter in a fairly decent state (except for the LCD, of course). 
Thank you all for your help!

Now off to finding a replacement LCD for it...



how you did this what i mean.. i new in this forum.. i want to ask you but i dont know how directly ask you,  i have same situation as you, my fluke 287 blinking, how yo do it with usb ir.. pls help me..

[giosif]

Hi,

Not sure I am reading your post correctly, but I think you are looking to fix the meter using the IR-USB cable.
If this is indeed what you intend to do, it is not how I fixed my meter and the part you are quoting was mainly to do with the fact I did not have a working screen for the meter (in addition to the original issue).
The way I fixed the meter was by reflowing the main IC (U26) - i.e. I used a hot-air station and flux to heat the IC and the board area around it until the IC would move ever so slightly (and then spring back on its own), when pushed (very gently) to the side with tweezers.
At that point, I let the board and IC cool down, put the meter back together and tried to power it on (using a bench power supply).
And, this time, it looked like the meter was powering on properly (i.e. the led would turn solid green for a few seconds, then turn off, while the meter was still drawing a few miliamps from the power supply - indicating it was not turning itself off).

Hope this helps...

[niccowahyudi]

As you discusion you try test that any component short.. i have same situation to.. my meter have short component to i try to check capacitor short and nothing capacitor short.. and my fluke gone blinking and it cant be turn off only with with cut off the baterry, i want to test to do like you are doing to..

[nixxon]

I just bricked my Fluke 289 that previously "only" had two issues with a non working LCD circuit and resistance measuring.

After connecting my oscilloscope ground to battery negative and probing the LCD driver pins one by one (the breakout board pins), the 289 entered the well known bricked mode (5 flashes of the power button). I have searched the internet, but it seems that nobody has figured out how to exit this 5 flashes of death mode.

It is a shame, because everything (other than the screen and resistance measuring) was working perfectly before I messed with the LCD driver pins. I could do regular logging and communicate via the IR-USB connection.

Has anyone figured out how to exit the 5 flashes bricking mode?

[Le_Bassiste]

the 289  first revisions had a super capacitor who gave some problems,  normally  it should keep the clock ticking for a while ... the latest revisions had a coin cell if i recall correctly,  you could have a bigger current cosumption because of this

to whom it may concern: did a supercap replacement on my 289 (ser. 22160xxx) long time ago.
current consumption is:

Code: [Select]

OFF     : 25 µA
ON      : 25 mA
Bright_1: 36 mA
Bright_2: 55 mA

HTH

[BravoV]

the 289  first revisions had a super capacitor who gave some problems,  normally  it should keep the clock ticking for a while ... the latest revisions had a coin cell if i recall correctly,  you could have a bigger current cosumption because of this

to whom it may concern: did a supercap replacement on my 289 (ser. 22160xxx) long time ago.
current consumption is:

Code: [Select]

OFF     : 25 µA
ON      : 25 mA
Bright_1: 36 mA
Bright_2: 55 mA

HTH

Thanks  , btw how did you measure the cap's current ?

[Le_Bassiste]

self-adhesive copper foil and kapton make for an excellent current spy.
" alt="" class="bbc_img" />