EEVblog 1592 - Fluke 287 Multimeter Teardown

[EEVblog]

Fluke 287 Multimeter Teardown

[floobydust]

Ahhh very nice. The transparent plastic appears to be needed for the jack-detect, IR opto-sensor passes through two jacks if you look carefully.

The supercaps are discontinued and soon to be unobtainium. I believe it's the same supercap used in Fluke 189, 287, 289 - but not sure.
07/2023 Elna is phasing out making coin supercaps, they are NRND but some production and stock available.
Elna 0.2F 3.3V DSK-3R3H204T614-H2L new p/n is RSCSK2043R3D01004T. Mouser 555-DSK3R3H204T614H2
Elna 0.33F 3.3V DSK-3R3H334T-HL new p/n is RSCSK3343R3D02008T. Mouser 555-DSK-3R3H334T-HL (thicker 2.1mm vs 1.4mm- fits)

The second PTC RT2 might be for LowZ?

Dave, please taunt Brymen to up their rotary switch tech instead of this "good enough" philosophy 

[golden_labels]

Looking at the video, the clear plastic part completely covers the top side. Including the display. I suppose one wants to have transparent plastic over the display.

As for the purposes, I may only come up with two ideas:

• Dust and water protection.
• It was cheaper/easier to make another part than to mold the case in just the right shape to hold all parts together.

Dave, please taunt Brymen to up their rotary switch tech instead of this "good enough" philosophy 

And then we learn that this idea is already patented. Or — the other end — Brymen won’t implement it, because they couldn’t treat it as their own IP.

[Black Phoenix]

Weren't the supercaps Panasonic? My 2007, first from the line batch of 289, was Panasonic until it gone to Fluke support because of the Supercar issue. Didn't disassemble after it so I don't know if it was replaced by a new Supercap or deleted entirely.

Although new batch motherboards are equipped with lithium coin cell and holder.

[floobydust]

I think Panasonic was used but they discontinued their line up first, leaving Elna. I wonder what the power drain is, if a coin cell works.

The Fluke rotary switch does have a few patents. 'Digital multimeter having enhanced rotary switch assembly' US 8093516 and US 8946571 to show a couple.
Brymen rotary switch a copy of a 1989 Actron US patent US4876416.

Another function of the clear plastic insert is passing IEC 61010  they wrap it in tinfoil to see if an arc can jump out to your hand, at the seams, LCD, buttons etc.
You must have good HV isolation from the multimeter internals - PC board/battery holder to the outside case where your hand would be.

[SiliconWizard]

The additional application processor with all the RAM really looks like huge overkill for such a multimeter. Almost makes no sense.

[EEVblog]

Ahhh very nice. The transparent plastic appears to be needed for the jack-detect, IR opto-sensor passes through two jacks if you look carefully.

Also for the serial port.

The supercaps are discontinued and soon to be unobtainium. I believe it's the same supercap used in Fluke 189, 287, 289 - but not sure.
07/2023 Elna is phasing out making coin supercaps, they are NRND but some production and stock available.
Elna 0.2F 3.3V DSK-3R3H204T614-H2L new p/n is RSCSK2043R3D01004T. Mouser 555-DSK3R3H204T614H2
Elna 0.33F 3.3V DSK-3R3H334T-HL new p/n is RSCSK3343R3D02008T. Mouser 555-DSK-3R3H334T-HL (thicker 2.1mm vs 1.4mm- fits)

I was thinking of bodging in a replacement. I have a kit of various supercaps.

Dave, please taunt Brymen to up their rotary switch tech instead of this "good enough" philosophy 

What's their problem?
Didn't they perform fairly well in Jow Smiths switch testing?

[EEVblog]

The additional application processor with all the RAM really looks like huge overkill for such a multimeter. Almost makes no sense.

Maybe they had used it in some other product?
Works well enough from a low power point of view.

[SiliconWizard]

The additional application processor with all the RAM really looks like huge overkill for such a multimeter. Almost makes no sense.

Maybe they had used it in some other product?

Yes, it's probably used in a whole line of products, and reusing the design is understandable. The MSP430 apparently implements the base measurement functionalities and most likely doesn't have enough resource left for implementing the UI and additional features of these more advanced models, so they did maximum reuse here and separated the two.

It's perfectly understandable, but I can't help having a problem with products using a lot more resources than what's strictly needed to get the exact same features. It's just waste.
The big bloat problem, it's everywhere, and the justification is always the same - that helps the company (or at least they think it does) by lowering immediate development costs, but it's otherwise a burden to everyone else, in the grand scheme of things. So that's my rant on "the great bloat".

Works well enough from a low power point of view.

Well, no doubt it's perfectly usable and a nice product, but they claim a 100h min battery life, which is not horrible, but so-so. Sure could get significantly lower consumption with a less complex overall system by migrating what's done by the MSP430 in a recent, low-power 32-bit MCU and implement all the UI on it as well. You get the idea. Simpler BOM, simpler system, cheaper, much lower power.

Now some companies do much worse in terms of bloat. Fluke is definitely still reasonable. They haven't used a 3GHz ARM SoC with 8GB of RAM to control a DMM with a monochrome display.

[bdunham7]

Well, no doubt it's perfectly usable and a nice product, but they claim a 100h min battery life, which is not horrible, but so-so. Sure could get significantly lower consumption with a less complex overall system by migrating what's done by the MSP430 in a recent, low-power 32-bit MCU and implement all the UI on it as well. You get the idea. Simpler BOM, simpler system, cheaper, much lower power.

You're looking at an almost 2 decade old design for the meter and a bunch of obsolete chips.  They've respun the board for the current versions and I have not seen at teardown on a new 287/289.  I'd be curious, but not curious enough to peel the calibration sticker off of mine and take it apart.

[scopeman]

I still like my Fluke 37 and 77 series. 1000Hr life out of a carbon zinc battery. I would bet that a Lithium Iron battery like the Eveready Ultimate Lithium 9V would go a couple
of thousand hours or more.

Sam
W3OHM

[Black Phoenix]

I think Panasonic was used but they discontinued their line up first, leaving Elna. I wonder what the power drain is, if a coin cell works.

Well since it keeps date and time and some parameters and it is the only function of it, probably the same drain as a PC motherboard BIOS battery when 5vSB is not present.

[EEVblog]

Well, no doubt it's perfectly usable and a nice product, but they claim a 100h min battery life, which is not horrible, but so-so. Sure could get significantly lower consumption with a less complex overall system by migrating what's done by the MSP430 in a recent, low-power 32-bit MCU and implement all the UI on it as well. You get the idea. Simpler BOM, simpler system, cheaper, much lower power.

200 hours life quoted while data logging.

[EEVblog]

I still like my Fluke 37 and 77 series. 1000Hr life out of a carbon zinc battery. I would bet that a Lithium Iron battery like the Eveready Ultimate Lithium 9V would go a couple
of thousand hours or more.

[EEVblog]

You're looking at an almost 2 decade old design for the meter and a bunch of obsolete chips.  They've respun the board for the current versions and I have not seen at teardown on a new 287/289.

How do you know they re-spun the design?

[bdunham7]

How do you know they re-spun the design?

When I sent mine in for a battery draw issue thinking it would be the supercap, they informed me that the problem was actually one of the chips on board.  They didn't say which one, but the tech did say that particular part was completely NLA and the only fix was to replace the meter.  The new meter runs different firmware and the versions are not compatible between the original and new meters.  Others have reported that they now use a soldered-in coin cell for the RTC but I haven't seen for myself.  Also, I'm pretty sure the boot time is faster.

[SiliconWizard]

Well, no doubt it's perfectly usable and a nice product, but they claim a 100h min battery life, which is not horrible, but so-so. Sure could get significantly lower consumption with a less complex overall system by migrating what's done by the MSP430 in a recent, low-power 32-bit MCU and implement all the UI on it as well. You get the idea. Simpler BOM, simpler system, cheaper, much lower power.

You're looking at an almost 2 decade old design for the meter and a bunch of obsolete chips.  They've respun the board for the current versions and I have not seen at teardown on a new 287/289.  I'd be curious, but not curious enough to peel the calibration sticker off of mine and take it apart.

If they have made any significant modification of the architecture, I'd somehow highly doubt that would yield anything less bloated - I would actually expect more bloat as time goes by.
But the most likely is that they haven't changed much, at least regarding the part I was talking about. What bunch of "obsolete chips" have you identified?

[floobydust]

From Dave's pics:
Fluke28x REV 015
U26 - MC9328MXSCVP10, i.MX ARM920T, 100MHz, LCD controller, 2 32-bit timers (oh so many lol), RTC, USB etc. MAPBGA-225 discontinued 01-2021.
U28 - 640P30TF65 64Mb as 4MBx16 NOR FLASH BGA-64
U25, U27 - IS66WVC2M16ALL, PSRAM 32Mb as 2MBx16, 70ns VFBGA-54 DRAM

It does look like overkill ARM9 even if the LCD is super high res - ultimately the multimeter is reading only one or two quantities, what would a Cray do with all that?
I think the added costs - 8 layer board, BGA mix with all that the through-hole, big CPU and memory might have caused the product to be expensive?

I got burned once in a product design using that "mobile/cellphone" PSRAM as it has a very low power self-refresh mode - but they all got obsoleted by Micron and the Fluke shows an Intel part and we all know how their venture into mobile went...
The memory IC going obsolete, it scuttled my whole project. I would guess that necessitated Fluke board rev. but not sure what they went with or if power drain increased. Assuming the supercap only holds the RTC, not data logs.

[...] Dave, please taunt Brymen to up their rotary switch tech instead of this "good enough" philosophy 

What's their problem?
Didn't they perform fairly well in Jow Smiths switch testing?

It's single point, high contact pressure and seems to be finicky about production tolerances. Not sure what the outcome was in this thread BM786 Switch Issue.

[bdunham7]

But the most likely is that they haven't changed much, at least regarding the part I was talking about. What bunch of "obsolete chips" have you identified?

Since they kept the same model and presumably the same functionality, I wouldn't expect them to make any significant changes, just enough to accomodate parts.  I didn't look up everything but for example the cell-phone SRAM chips show obsolete if I looked them up correctly--no surprise there.

[SiliconWizard]

I haevn't looked up the exact parts, but PSRAM is relatively easy to source and replacing them is no problem. I bet they could even find the same pinout, as it's probably standard JEDEC stuff?

[Black Phoenix]

Well I've done an experience with my 289. I've been having it with Panasonic Eneloops on it. The charge is the same as it was last time I use it, so no consumption of battery as when it was sent to support where a pack of 6 batteries would be empty in less than a month.

But simulating a battery change, so removing the back cover, remove batteries and insert again gave me the message of needing to correct time and date.

So I would say that my fix was just the removal of the supercap from the PCB and was left that way.

Again I can't be sure because I didn't open the multimeter after repair and I can't be bothered to do it since it works well and the time and date is an easy fix of connecting to the pc and getting the correct time and date via the Fluke ViewForms sync function.

It doesn't affect saved data in the memory, only the RTC function.

[floobydust]

Cellular RAM is constantly obsoleted with no equivalent. It's either VCC or pinout/package as deal breakers, industrial temp range is also much harder to source. It makes designing an embedded system (expanded bus) high risk as you are tied to a part that 1 or 2 manufacturers make with a lifetime of a few years.
The IS66WVC2M16ALL PSRAM in standby at <0.15mA will auto-refresh.

I'll bet there is an operating system stashed somewhere in meter's firmware, that's why the main CPU is ARM9.

[bilaal.yousuf]

Does anyone have the schematic for fluke 287 and also for 179 please?

[EEVblog]

Does anyone have the schematic for fluke 287 and also for 179 please?

https://www.eevblog.com/forum/testgear/fluke-189iian-(289)-service-manual/?action=dlattach;attach=1964316