当放入200 V/m辐射时，25LC128中数据损坏的可能性

你有没有理由指出是电压场导致了腐败？我想这可能是你的软件的问题，并试图访问EEPROM，当电源电压超出规格。



      以下为原文

   
Do you have any reason to indicate it is the voltage field causing the corruption?
I would guess it's more likely a problem with your software, and trying to access the EEPROM when the supply voltage is out of spec.

ANS1：我没有理由问我为什么要问这个评级？到何种程度的辐射EEPROM工作？在数据表中，我没有得到任何与此相关的信息。（因为我正在使用1.2兆赫频率，10千瓦的电源）我需要知道这个参数。ANS2：相同的软件正在运行另一个板，但没有任何问题。因此，我们的思维从软件转向硬件。这就是为什么我们正在考虑辐射。ANS3：我会检查电源电压，并让您告知供应情况。此外，我认为电源电压不是原因，因为DSP连接到相同的电源水平（3.3 VD）。它有非常严格的电压调节。关于，Maulik



      以下为原文

    Ans1: I don't have any reason that why i am asking about this rating? up to which level of radiation EEPROM get work? I did not get anything related to this in datasheet. (Because i am working with 1.2 MHz frequency, 10 kW power supply) I need to know about this parameter.
 
Ans2: Same software is running with another board but dont have any problem. So, our thinking divert from software to hardware. that's why we are thinking about radiation.
 
Ans3: I will check the supply voltage and let you inform about supply condition. Also, I feel that supply voltage is not the reason because DSP is connected to same supply level (3.3 Vd). And it has very tight voltage regulation.
 
Regards,
Maulik
 

有什么更新吗？你能提供最大的辐射水平（EEPROM可以安全地保存数据直到那个特定的水平）参数吗？带着问候，Maulik



      以下为原文

    Any update?
Can you please provide maximum radiation level (EEPROM can hold the data securely till that particular level) parameter from your end?
 
With Regards,
Maulik

如果Microchip没有提供您在数据表中寻找的数据，那么数据就不可用，您需要进行自己的测试来确定设备的易受性。



      以下为原文

    If Microchip do not provide the figure you are looking for in the datasheet, then the data is not available and you will need to do your own tests to ascertain the device susceptibility.
 
 

嗨，正如你在你的1帖子中所说的，你不确定原因，也就是说你不确定腐败是由于E域还是H-field还是…？简单地说，在我的职业生涯中，几乎所有的情况下，EEPROM的腐败问题都是由于固件或不适当的硬件设计。在硬件或固件中没有问题吗？你的PCB有多少层？你多久写一次EEPROM？你的系统一直在运转吗？当做



      以下为原文

    Hi,
As you said in your #1 post, you are not sure about the reason, i.e. you are not sure if the corruption is due to E-field or H-field or...simply SW.
In almost all cases I have seen in my career, EEPROM corruption issues were due to firmware or improper hardware design.
What let's you think at this stage that there is no issue in your hardware or firmware ?
How many layers has your PCB ?
How often do you write the EEPROM ?
Is your system always ON ?
Regards
 

你的EMR场似乎相当激烈，你已经描述了你的电源为1.2MHz和10kW。你能确认附近没有电离辐射源，比如X射线吗？一个200伏/米远场似乎没有足够高的能量直接影响存储在ErOM的位单元中的电荷。近场似乎足以破坏I2C信号线。您可能需要尝试一些附加的屏蔽和低压电源滤波来抑制任何传导干扰。EMR可以破坏电子系统的方式太难表征，因此制造商很少公布单个部件对低频率的敏感性规范。电子病历。有几类零件具有对电离辐射的耐受性的规范。这些装置在卫星和宇宙飞船中最常用。



      以下为原文

    Your EMR field seems fairly intense and you have describe your power supply as 1.2MHz and 10KW.
Can you confirm that there are no ionizing radiation sources, like x-rays, near by?
 
A 200v/m far-field does not seem to be high enough energy to directly affect the charge stored in a bit cell of an EEROM.
 
The near-field seems enough to disrupt the I2C signal lines. You may need to try some additional shielding and low voltage power supply filtering to suppress any conducted interference.
 
The way EMR can disrupt electronic systems is too hard to characterize so manufactures rarely publish a specification of the susceptibility of individual parts to low frequency EMR. There are a few class of parts that have specifications for tolerance to ionizing radiation. These kinds of device are most often used in satellites and space craft.

只是关于你的问题的一些想法（我不是一个芯片设计专家）：我同意DAN1138。一个200伏/米的E场相当于200个UV/UMA典型的EEPROM单元在1～2μm量级，阶数为1V。典型的pn结宽度为~1UMI，因此估计在一个单元中的E场是1V/1Um＝1000000 V/m的顺序，那么外部E场将小于-5000倍，然后小于本地芯片Enfield。因此，你可以认为这太小以至于不能影响一个单元的工作。但是我的估计可能错了。我认为你应该看看EMC问题。声明你的电压调节器是一个好的不是一个坚实的观察。EMC问题通常是在你的HW设计（PCB轨道）的布线中引起的，并且会导致EEPROM控制/供电线路中的杂散信号（在这种情况下，不管你的调节器有多好，你仍然可以）。从你的逆变器获得共模和差模峰值。逆变器是已知的，由于开关拓扑而产生大量的EMC问题。您可以尝试用金属接地盒（E-field）和/或Mu金属盒（用于H字段）屏蔽EEPROM，并看看是否有帮助。如果不是，那么它可能是EMC问题（或SW），通常是扩散。邪教和昂贵的解决。



      以下为原文

    Just some thoughts about your question
(I am not an expert on chip design):

I agree with dan1138.
An E-Field of 200V/m is equivalent to 200 uV/um
A typical EEPROM cell is in the ~1 um^2 order with voltages of order ~1V.
Typical PN junction width is in the order of ~1um
I would therefore estimate that the E-field in a cell is in the order of 1V/1um = 1000000 V/m

The external E-field would then be ~5000x smaller then the local chip E-field.
You could therefore argue that this is too small to influence the workings of a cell.
But I could be wrong with my estimates.

I think you should look at EMC problems.
Stating that your voltage regulator is a good one is not a solid observation.
EMC problems are usually induced in the wiring of your HW design (PCB tracks) and can result in spurious signals in EEPROM control/supply lines (in this case it does not matter how good your regulator is, you still could get common mode and differential mode spikes from your inverter). Inverters are known to generate a lot of EMC problems due to there switching topology.

You could try to shield your EEPROM locally by a metal grounded box (for E-field) and/or mu-metal box (for H field)
and see if that helps.
If not, then it probably is an EMC problem (or SW) and that is usually difficult and costly to solve.

我首先要把责任归咎于电磁兼容问题。你可能在这个频率上处理100s甚至Ka，加上不可避免的谐波，使它变得更糟，这是一个重要的领域！在考虑芯片是问题之前，想想任何痕迹都是天线。我会尽可能地保护一切。带开关电流的区域/卷和带内存的板。我建议测试一下你的心境：用一些已知的数据写一个或多个松散的内存芯片，并把它们保存在你的工作内存中，然后再读回。如果有腐败，那么我相信电磁场会影响芯片。



      以下为原文

   
I would first put the blame on EMC issues. You may be handling 100s or even kA at that frequency, plus the unavoidable harmonics to make it worse, that's an important field!. Before thinking the chip is the problem, think of any trace as an antenna. I would shield everything as much as possible. The areas/ volumes with switching currents and the board with the memories.
I suggest a test for your peace of mind: take one or better a few loose memory chips written with some known data and keep them by your working ones for some time, then read them back. If there is corruption of those, then I would believe the EM field does affect the chip.

它可能需要一些事情来解决你的问题，所以不要期待一个神奇的子弹关闭。首先，你要处理近场条件，除非你的系统是一个城市街区的大小，所以你离你的源头很近。在1.2兆赫，铝的皮肤深度是微小的，所以最小的屏蔽努力，得到金属板周围将是一个很好的基础步骤。在10KW，假设你正在驱动一个50欧姆线匹配负载V＝707V，I是14 A。现在你的电力电子器件的排水板将取决于该技术的不同。电源设备的峰值/峰值RF电压可以是2X直流电源。得到一个接地的金属墙，在电源和任何低压电器上。直接交流耦合到电路板上的痕迹将影响任何高Z线在电路板上。这可以用一个探头和盒子里的短导线来证明。现在，你所有的H耦合会导致不必要的电流流过它会干扰的东西。看看板上的所有电线，它们将一个公共或屏蔽连接到电路板地面/数字地面，并看它们来自何处。连接到你的板和地面到你的系统中其他地方的两个电缆的每一个组合都是一个潜在的环形天线，通过形成一个回路，包括电缆的某些部分和系统底盘的一些部分，除了电缆本身。电流通过你的板会引起接地和错误。如果你有一个在1.2 MHz范围内的电流探测器，你可以测量每个电缆中的电流。在这一点上，感谢你的干扰是一个连续的射频源，而不是一个18“HV的火花，只发生一次或两个小时……至少你可以做出合理的测量。下一步是开始隔离信号接口。认为光耦合器，变压器，簧片继电器，和差分输入。如果没有DC路径进入，则是孤立的。在某些情况下，整个电缆上的铁氧体环形芯可以很好地隔离。共模电流的阻抗随匝数的平方上升。选择一个在1.2兆赫具有特定电阻阻抗的铁氧体，它将花费一些努力找到它们，但不要费心把VHF铁芯像在你最近的打印机后面找到的一样。你的板应该以一个单点接地，所有非隔离信号和电源接地到单点离开电路板。下一步，忘掉你所知道的关于接地和屏蔽的一切，去做一些研究。考虑每一个在你的点之外的地面作为敌方，并且潜在地是不想要的电流和干扰的来源。现在好好打猎，或者租一个火腿收音机OP。



      以下为原文

    It may take getting a number of things right to solve your problem so don't expect a magic bullet right off.
First off you're dealing in near field conditions so unless you system is the size of a city block so you are pretty close to whatever your source is. At 1.2 MHz the skin depth for aluminum is tiny so a minimal shielding effort that gets metal around your board would be a good base step. At 10kw, assuming you are driving a 50 ohm line to a matched load v=707v and I is 14 a. Now the drains/plates of your power electronics will be different from that depending on the technology. The peak/peak rf voltage at the power device could be 2x your D.C. supply. Get a grounded metal wall up beteeen the power stuff and any lv stuff. Direct ac coupling to traces on board will affect any high z line on the board. This can be demonstrated with a scope probe and a short wire in the box.

Now, all of your H coupling will result in unwanted currents going through stuff that it will interfere with. Look at all of the wires on the board that connect a common or shield to the board ground/digital ground and look at where they come from. Every combination of two cables that interfaces to your board and grounds to somewhere else in your system is a potential loop antenna by forming a loop that includes some part of your board and some part of your system chassis in addition to the cables themselves. Current through your board will cause ground shifts and errors. If you have a current probe for your scope that's good at 1.2 MHz you can measure the current in each cable.

At this point be thankful that your interference is a cw rf source and not an 18" HV spark that only happens once or twice an hour...... at least you can make reasonable measurements.

The next thing is to start isolating your signal interfaces. Think opto couplers, transformers, reed relays, and differential inputs. It's isolated if no dc path in to out. In some cases a ferrite toroidal core on the whole cable can be good enough isolation. The impedance to common mode currents goes up with the square of the number of turns. Pick a ferrite that has a specified resistive impedance at 1.2 MHz, it will take some effort to find them but don't bother puting a vhf core on like is found on the back of your nearest printer.

Your board should end up with a single point ground with all non-isolated signals and power grounded to a single point off the board.
Next step, forget everything you know about grounding and shielding and go do some research. Consider every ground beyond your single point as hostile and potentially a source for unwanted current and interference.

Now good hunting, or hire a ham radio op.