Sudden Death Answer

[fotografr]

Advertisement (gone after registration)

As some of you know, my M8 has been in Solms to be analyzed for a third "Sudden Death" failure. Jesko von Oeynhausen, who is the head of Quality Assurance, had a team of technicians work on my camera in an attempt to finally figure out what was going on. I received his answer yesterday, as follows:

 

"We spend the whole day to analyze your camera, and we found a defect in an electronical component (transistor). These transistors are of course put through the same tests upon delivery that all electronical components undergo (random sample tests) before being built into the camera's circuitry.These parts had passed the mentioned tests without objections. Even so, we later noticed that some of them must have been flawed and later failed."

 

I then emailed back to say there has been much speculation about batteries causing the problem, and asked whether this could be causing the transistors to fail. I received the answer to that question today:

 

"I read about someone (you?) who measured a very low currency with his battery that was used in a camera that died. We don't think that this made the camera or the transistor fail. It must be otherway round. If the camera hangs up it can be in a undetermined status and it can suck the battery down to a level where it normaly switches of before. If you normaly charge the battery again you will see that it it works properly. When the battery is empty, the camera switches off and there will be a currency of about 3,3 V."

 

Obviously, this is only one camera and we can't necessarily draw conclusions about the others from this. However, each shutdown that occurred with my M8 was pretty much the same sequence of events. At this point, I am optimistic that the problem has been isolated and can be prevented in the future.

[Jamie Roberts]

Brent, that is fabulous news!

 

Imagine a few stray transistors (out of a power supply board or chip set) going awry!! Talk about a needle in a haystack

 

Hope this helps everyone who's had this problem too....wouldn't that be great?

[farnz]

Thanks for reporting, Brent, and very encouraging news! "One swallow does not a summer make" so we should absolutely not jump to the conclusion that this has solved the riddle.

 

But, hey, my fingers are crossed!

 

Pete.

[colorflow]

Great news indeed. At least solves one of the SDS problems. Question remains with those sudden deaths that revive after a few hours or a few days. A hardware/component failure usually would stay dead and not revive, unless it is a marginal component that fluctuates with temperature?

 

Alan

[brucek]

I do hope that this doesn't require another trip back to Solms for my camera...

 

Bruce

[Guest guy_mancuso]

Brent and I have been communicating on this for the last couple days and actually this is pretty good news although bad but at least there is a pinpoint as to what is causing the issue, so for the end user this takes a lot of guess work out of the equation like a faulty battery or charger. Also i talk to Jesko all the time in e-mails and he is right on top of the issues and also all my reports go to him. So if Andrea's or one of the mods maybe put this up as a sticky and make this more official for the end user this maybe a good idea. i feel a little better about this to be brutally honest, this has been a HUGE issue with me and Leica so i am glad to see a pinpoint target at this cause.

 

I do think there is more testing on this that needs to be done also and on the battery stall situation although it has not happened to me in quite awhile. As most of you know I am beta testing the firmware and it has been very clean so far but i also get no info on what they have done so i have no idea if certain internal workings are being fixed they say the are and like I said things are looking far better. Keep the faith troups

[colorflow]

Advertisement (gone after registration)

Guy and Brent:

 

Would you kindly ask Mr. von Oeynhausen whether this transistor problem could cause a intermitent problem that can revive after a while?

 

Thanks in advance,

Alan

[jrc]

Mr. Oeynhausen I believe is referring to my battery which had the extremely low voltage when I first measured it after a Sudden Death. As he says, when I recharged the battery, it came back up to exactly 4.15v and has stayed there since. So for my camera/battery, what he says is apparently exactly right. This is very encouraging, and if it's only this small isolated problem, I'll buy another body.

 

JC

[fotografr]

Guy and Brent:

 

Would you kindly ask Mr. von Oeynhausen whether this transistor problem could cause a intermitent problem that can revive after a while?

 

Thanks in advance,

Alan

 

 

I'll ask, Alan, but there won't be an answer until next week because this is a holiday period for the factory.

[fotografr]

This is very encouraging, and if it's only this small isolated problem, I'll buy another body.

 

JC

 

I think a lot of us will. I'm feeling very optimistic about the situation right now--for the first time in months.

[Guest guy_mancuso]

Mr. Oeynhausen I believe is referring to my battery which had the extremely low voltage when I first measured it after a Sudden Death. As he says, when I recharged the battery, it came back up to exactly 4.15v and has stayed there since. So for my camera/battery, what he says is apparently exactly right. This is very encouraging, and if it's only this small isolated problem, I'll buy another body.

 

JC

 

I think this is related to the sudden stop issue too becuase the battery and camera mismatch to how much battery power there is and until you either replace or reinsert the same battery than the camera comes back. I think the 2 are somewhat related being the Sudden stop and the Sudden Death Syndrome. I am sure there testing this out now

[scott kirkpatrick]

Jesko von Oeynhausen, who is the head of Quality Assurance, had a team of technicians work on my camera in an attempt to finally figure out what was going on. I received his answer yesterday, as follows:

 

"We spend the whole day to analyze your camera, and we found a defect in an electronical component (transistor). These transistors are of course put through the same tests upon delivery that all electronical components undergo (random sample tests) before being built into the camera's circuitry.These parts had passed the mentioned tests without objections. Even so, we later noticed that some of them must have been flawed and later failed."

 

 

It's dangerous to overinterpret an English translation from a non-native speaker, but I think he is saying that random sampling (test some of the incoming components, and if they are OK, assume that all of them are OK) allowed some bad components to be used in the power circuits, So there may still be other dangerous cameras out in the field which will have to come in for repair when discovered. If you replace "currency" with "voltage" I think he is describing jrc's post. While the over-discharged battery came up to nearly full voltage, past experience suggests that it could have suffered damage at the voltage which was reported.

 

So don't start cheering just yet. let Leica decide what steps must be taken to keep the problem from occurring in cameras which have been shipped already. I trust them to keep it from happening in the future in new cameras. See the thread "Explanation of SDS" for a suggestion of measuring your camera's ability to discharge and shutoff properly. If it does this, it sounds like you are OK. And if it doesn't shut itself down in about 6 hours, turn it off yourself and see if your battery has discharged to less than the 3.3 volts in the email from Solms. In my tests, I haven't seen less than 3.4 volts.

 

scott

[fotografr]

I think that's good advice, Scott. Still, I have the feeling we are closer than we've been to figuring this out now. There are almost undoubtedly still cameras 'out there' with the bad transistors.

[harrisfoto]

Is there any one on the list that is current with electronic components purchasing and assembly???

Many years ago, When I was buying electronic components for use in medical equipment, the prices varied, according to the estimated QC rate on a given component. That is, eveyone knew that there would be a certain percentage of the components which would be bad.

We, as purchasers, could either assembly the product and then QC test, or test each component, before using it in the system. We could buy at various reject rates.

A "random test" only confirms the quality of what the Mfg sold to you. It does not QC the final product.

If you are putting a high value component in the product, like a Kodak sensor, then one needs to QC the all of the components, prior to assembly, because the cost of rejecting the the final product, based on QC is too high. So is the cost of reworking the failed product.

It sounds like Leica is using mechanical QC procedures instead of electronics component QC procedures.

Thus, the high final product failure rate, and high customer dis-satisfaction rate. 30 years ago this was called value engineering 101.

Most likely a lot of the problems experienced with the 8M did not occur with the DMR, because the DMR went through another mfg for final testing, who already had set up QC procedures for high-value, digital imaging products.

[barjohn]

Steve,

 

What you say is true. In addition, I might add, that the component tolerance specification may have been to low. Thus a component that was on the margin would be more likely to fail under stress. Specifying a higher tolerance or capacity component could have allowed for a greater margin of error in manufacture and still produced a reliable unit. It sounds to me like the current capacity for the specified transistor was not adequate. Hence, a QC inspection would not reveal the problem since it would be working when they tested it (unless they stress tested to 10% over and under or some other figure of merit). Subsequently, when the battery voltage dropped, the current load had to increase to maintain the same power level (power = voltage x current) to drive the motor and other electronics and the current level probably exceeded the transistors handling capacity thus burning it out.

 

The bad news is that the only real fix would probably be a recall to replace the under tolerance component(s) and thus avoid a sudden failure in the future for units that are currently in circulation. The fact that they haven't failed yet is no guarantee that they wouldn't at a future date when the circumstance(s) that have been causing the others to fail materialized.

[ho_co]

Most likely a lot of the problems experienced with the 8M did not occur with the DMR, because the DMR went through another mfg for final testing, who already had set up QC procedures for high-value, digital imaging products.

Harris--

Your points are generally well taken, but as I recall the history of the DMR, Leica originally tested each unit on receipt at Solms. But doing that, they found that there were too many defectives too late in the chain. They then set up their final control point within the manufacturer's facility. The feeling a lot of us had was that the slow deliveries stemmed from Imacon work that wasn't as good as Leica needed.

 

And since you mention the Kodak sensor: a couple people on the forum have reported a dead row of pixels, though Kodak specifically states that such is not allowed by their QC.

 

I don't think we can fault Leica completely at this point, though they do bear the ultimate responsibility for the M8.

 

--HC

[marknorton]

Scott's theory that the camera does not switch off when the battery voltage falls to the lowest working voltage seems very plausible; that limit is set both to protect the battery and ensure proper operation of the electronics. It's a critical value - set it too high and the battery life of the camera will be compromised. Set it too low and you risk operating problems and over-discharging the battery.

 

The speed you can run digital electronics at depends on the rise time of the signals - the time taken to transition from binary 0 to 1 - and the power supply voltage is used to "pull up" the signals and accelerate the transition. As the battery voltage reduces, the signals get sluggish and ill defined and the thing stops working.

 

It's common, for example, to be able to run digital chips at reduced supply voltage if you slow them down and the M16C processor used in the camera is one example. Its specification allows a reduction in working voltage range if the clock speed is reduced.

[ho_co]

Mark--

Just curious: Where did you find info about the M8's logic circuitry? Is it published?

 

(Referring to "the M16C processor used in the camera is one example.")

 

Thanks.

 

--HC

[marknorton]

HC, you can see references to the M16C processor in the firmware if you open it with a hex editor. The M16C is made by Renesas, previously called Mitsubishi.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[jrc]

Steve,

 

The bad news is that the only real fix would probably be a recall to replace the under tolerance component(s) and thus avoid a sudden failure in the future for units that are currently in circulation. The fact that they haven't failed yet is no guarantee that they wouldn't at a future date when the circumstance(s) that have been causing the others to fail materialized.

 

John,

 

But wouldn't this also mean that there might be some user techniques that would protect against a crash, such as examing your shooting technique, and then discovering the number of shots you can safely take before you need to recharge the battery? That is, learning how not to stress the components?

 

I personally plan to check the voltage on my batteries before I insert them, and then perhaps shoot ~200 shots and then recharge; and never again fully discharge them within the camera (at least, not until I'm specifically told that it is safe to do so.) Better to have to buy a new battery once in a while than send the camera back for repairs...

 

JC