High failure rate of Maxim Integrated DS1245YP-100 chips on TDS7000 scopes

[Jwalling]

This is really starting to bug me. Over the last few years, I've repaired 10 or so Tek TDS7000 series scopes that had a bad Maxim Integrated DS1245YP-100 chip. This is the surface mount part with the PCAP snap-on battery. Some would fail the R/W test on the PPC board boot diagnostics, one failed the R/W test during self test when the scope was booted up, and  at least 4 now I have seen would draw excessive current when the scope is powered off. On my bench now is a TDS7104 whose DS1245YP-100 chip draws 100uA when power is off. Since the battery is rated at 130mA, that would give about 1300 hours or 54 days of battery time.

Worse yet is the price of these POS chips keep going up! I bought one on Dec 16th 2016 for $38.65 USD. Now they are up to $46.65 just 4 months later from Digikey!

I hate Maxim. You SUCK! If it wasn't bad enough that (some of) your chips are unreliable and your prices keep going up, you EOL parts with no suitable substitute - just look at the guy in this thread:

https://www.eevblog.com/forum/repair/repair-of-tek-tds3014b-need-a-source-for-a-ds1742w-nvram/

Not to mention the DS1486 which is in all the Tek TDS700 and some TDS600 series scopes. If you need a replacement, you better be ready to buy from shady sources, or be willing to do some surgery on the chip to wire in a new battery.

Grrr... 

Oh sure, I could buy this for $16:
http://www.ebay.com/itm/1x-IC-DS1245YP-100-/162118885220?hash=item25bf09df64:m:m8H-_pMo49zXYGK9mk5EgNQ

But I'm not so sure I would trust it. He's sold 11 of them so far, so it looks like other folks are seeing these fail as well.

Is anybody else seeing these chips fail more than what would be considered normal?

EDIT: Typo. I always make typos when angry!

[Pinkus]

I hate Maxim. You SUCK! If it wasn't bad enough that (some of) your chips are unreliable and your prices keep going up, you EOL parts with no suitable substitute ...

This is the reason why i try to avoid Maxim-only products and many other developers do too. I also learned this the hard way.

[Jwalling]

I hate Maxim. You SUCK! If it wasn't bad enough that (some of) your chips are unreliable and your prices keep going up, you EOL parts with no suitable substitute ...

This is the reason why i try to avoid Maxim-only products and many other developers do too. I also learned this the hard way.

 
Good! I wish every engineer would stop using their crap!

[nctnico]

It could be interesting to replace it with a small PCB containing an FRAM. This shouldn't be hard to make and the 128k * 8 FRAM ( FM28V100-TGTR from Cypress) costs around $20 in single quantities.

[Tomorokoshi]

Dallas Semiconductor Novrams with the built-in batteries, now owned by Maxim, are known to have been counterfeited. Are the Maxim ones a rebranding of the old Dallas parts? Perhaps the counterfeit ones also got that labeling.

[Jwalling]

Dallas Semiconductor Novrams with the built-in batteries, now owned by Maxim, are known to have been counterfeited. Are the Maxim ones a rebranding of the old Dallas parts? Perhaps the counterfeit ones also got that labeling.

Anything is possible I suppose, but I would imagine that Tek only buys from authorized distributors.
So no one else is seeing failures in these parts?

[dxl]

I also had two CSA8000's where the Dallas chip was using excessive current during standby.

[Tomorokoshi]

It looks like the TDS7000 series came out in 2007. This is before the run of counterfeit Novrams from Dallas. There was a time around 2010, 2011 when those things were hard to get. If Tek subcontracted, got gray market parts before people were more aware of the problems, etc., then it's possible some got in. Who knows. Do you have the parts that came out? Compare the silk screen, date code, etc. to the datasheet for the part. You could then also contact Maxim to see if it's a valid production silk screen and code.

The thing about the counterfeit Dallas parts was that the silk screen was better than the original. Very sharp, lots of ink, etc. At closer inspection there may have been subtle differences from the proper silk screen, but they were not noticeable to someone with only a casual interest in the parts.

The counterfeit ones "worked" for a couple years, then failed. Proper ones should be 10+ years once activated, and more if the equipment is powered.

[james_s]

I haven't really had any trouble with the Dallas chips. If it's a serious issue it would not be hard to make an equivalent part. Carrier PCB with a SMT low power SRAM, lithium coin cell, dual schottky diode for isolation and a circuit to hold write enable inactive when power is lost. If the real deal costs $40 now then a non-potted clone starts to look like a feasible hobby project.

[Kjelt]

You blame Maxim but you should blame Tek for using that part that has a fixed lifetime of 10 years after production till failure and they did it on purpose because they want you to buy a new scope by then and trash the old ones.
It is this short term environment unfriendly production way of thinking that is to blame IMO.

[Jwalling]

You blame Maxim but you should blame Tek for using that part that has a fixed lifetime of 10 years after production till failure and they did it on purpose because they want you to buy a new scope by then and trash the old ones.
It is this short term environment unfriendly production way of thinking that is to blame IMO.

Err... You don't get it. This is the snap-cap version with a replaceable battery, not the PDIP part.
And these are semiconductor failures, not the battery in any case.

More failures are documented here:
https://www.eevblog.com/forum/testgear/tek-csa7404-repair-project/msg884291/#msg884291
https://www.eevblog.com/forum/testgear/tek-csa7404-repair-project/msg984925/?topicseen#msg984925
https://www.eevblog.com/forum/testgear/tek-csa7404-repair-project/msg572469/?topicseen#msg572469

When the new part comes in and I remove the old one I'll take a picture of the old chip.

[wraper]

You blame Maxim but you should blame Tek for using that part that has a fixed lifetime of 10 years after production till failure and they did it on purpose because they want you to buy a new scope by then and trash the old ones.
It is this short term environment unfriendly production way of thinking that is to blame IMO.

Err... You don't get it. This is the snap-cap version with a replaceable battery, not the PDIP part.
And these are semiconductor failures, not the battery in any case.

BIG difference...  . Either way your scope fails, data lost. Actually they are worse, you can easily pull out the DIP part from the socket and clone it. No such luck with snap-cap SMT parts.

[Jwalling]

You blame Maxim but you should blame Tek for using that part that has a fixed lifetime of 10 years after production till failure and they did it on purpose because they want you to buy a new scope by then and trash the old ones.
It is this short term environment unfriendly production way of thinking that is to blame IMO.

Err... You don't get it. This is the snap-cap version with a replaceable battery, not the PDIP part.
And these are semiconductor failures, not the battery in any case.

BIG difference...  .

Yes, it is. 

Either way your scope fails, data lost. Actually they are worse, you can easily pull out the DIP part from the socket and clone it. No such luck with snap-cap SMT parts.

You can't do SMT rework?

It's actually easier to replace the SMT part. You don't have to remove the PPC board from the scope.

[wraper]

You can't do SMT rework?

SMT rework is easy. But you cannot pull them out to clone or backup contents unlike those in DIP package. And battery will fail in either case, sooner or later.
BTW, my guess is, you, probably, had not done even 1/10 of SMT rework I've done, including large BGA, micro BGA.

[Jwalling]

You can't do SMT rework?

SMT rework is easy. But you cannot pull them out to clone or backup contents unlike those in DIP package. And battery will fail in either case, sooner or later.
BTW, my guess is, you, probably, had not done even 1/10 of SMT rework I've done, including large BGA, micro BGA.

 

You're probably right. I usually delegate that sort of work.

[wraper]

Ok, I'll tell a story of one little ST SNAPHAT M4T32-BR12SH6 battery, which is snapped on SOH-28 IC. "A good type" in your opinion. Happened with me a few months ago. So one 10yo device which weights a ton failed. I traveled from Latvia to Ukraine, 7 hour flight including transfer + 5 hours by car to the destination. So I figured out that Eurotherm 2704 controller failed. I had a user config backup I made few years before that but no factory config backup. So after a config restore, controller worked only partially. Factory config is dependent on modules installed, and part number (config) is long and with many variants. So there is no chance restoring it. I took the controller and traveled back to Latvia. Sent it to Eurotherm/Shneider Electric in Germany to get factory config restored. We paid EUR 800, received the controller back, restored user config and repeated my travel to Ukraine. What a f...ing bless. All of the time that device was down.

[Kjelt]

You blame Maxim but you should blame Tek for using that part that has a fixed lifetime of 10 years after production till failure and they did it on purpose because they want you to buy a new scope by then and trash the old ones.
It is this short term environment unfriendly production way of thinking that is to blame IMO.

Err... You don't get it. This is the snap-cap version with a replaceable battery, not the PDIP part.
And these are semiconductor failures, not the battery in any case.

Ok I have a Tek scope with the sealed versions for RTC and BB-SRAM , there it was a valid argument.

[Jwalling]

Getting away from the painfully obvious "batteries eventually go boom when powering stuff"

The bad chip has a date code of 0042GH (2000) and a lot code (?) of 120737.

The replacement chip draws 20nA.

Since my original post I made a kluge adapter that would allow me to power the NVRAM externally and monitor the current draw when powered down so I won't have to solder wires to the the Dallas chip.
I don't think I'll ever trust these parts again so will probably make this test SOP on incoming TDS6000, TDS7000 or TDS/CSA8000 scopes or any other equipment that uses them if the battery is dead.

The DS9034PC snap-cap battery replacement is $10.18, once again, from Digikey. It contains a BR1632 130mA lithium battery. What a great deal for $10, when the battery is $2!
It has 2 solder tabs 180 degrees apart:
https://www.digikey.com/product-detail/en/panasonic-bsg/BR-1632A-FAN/P022-ND/273642
(No stock)

Doesn't seem to be a common part - Mouser has none as well.

Could probably schlock this one in:
https://www.digikey.com/product-detail/en/panasonic-bsg/BR-1632A-HAN/P023-ND/273643

[swperk]

My TDS7404 failed to boot after having been in storage for several years.

The PPC console output showed that the DS1245YP-100 was failing the walking 1's test. The snap-cap battery was dead, so I replaced it, but there was no change.

I discovered that the DS1245YP-100 was putting a high drain on the snap-cap battery and that it had the same date code and lot number as @Jwalling's. I ordered a new DS1245YP-100 from Mouser, installed it, put on a fresh snap-cap battery, and now all is well.

Chalk up another failed TDS NVRAM!

[james_s]

It could be interesting to replace it with a small PCB containing an FRAM. This shouldn't be hard to make and the 128k * 8 FRAM ( FM28V100-TGTR from Cypress) costs around $20 in single quantities.

It might be worth trying, but none of the gear I've tried retrofitting one of those into has worked with it except for a pinball machine. They are hardware compatible with standard NVRAM but they require toggling of one of the control lines prior to each write to latch the address or something like that which standard SRAM does not require.

There are other parallel NVRAM devices out there though that can be used if the Dallas parts are experiencing a high failure rate. Personally I haven't had any reliability problems with them but they are expensive. Given what these scopes cost when new though I can see why Tek wasn't bothered about using an expensive part.