发 帖  
原厂入驻New
发烧友10周年庆典,全网超值优惠来袭!千元现金券,下单抽奖赶紧参与》》
[问答] 无法通过JTAG引脚与FPGA通信
63 xilinx JTAG FPGA
分享
嗨,
我们目前已经制作了一些有Spartan 6的海关委员会,并且不断损坏它们。
产品编号XC6SLX150
所有电路板都工作正常,但是非常间歇性地我们实际上是在破坏斯巴达人,有时甚至会破坏平台电缆USB(程序员)。
我们看到的问题是我们无法再通过JTAG引脚与FPGA通信
在某些情况下,如果我们实际上从台式电源为3.3V(损坏后)供电,我们现在也可以看到它的电流消耗增加了几百mA。
我们相信FPGA的损坏部分存在泄漏路径,从而产生更高的电流。
请注意,FPGA仍然可以通过外部板载SPI闪存进行编程,因此功能有点。
当我们将USB编程器插入到供电的系统中时,总会出现问题。
由于问题是间歇性的,因此很难调试。
但我们的理论是系统是浮动的,当插入连接到接地工作站的USB编程器时,我们正在创建一个ESD(当前转储)事件,从而损坏FPGA。
当从定制板的接地到Gnd测量时,我们看到最初的高压降(90V)最终会流失。
当从USB编程器的接地测量到地时,我们看不到电压降,因为它们处于相同的电位
作为我们设置的一部分,我有多个定制板,其中一块板我使用笔记本电脑对Spartan 6进行编程。由于笔记本电脑也在浮动,我们从未损坏过这块板。
它始终是连接回接地工作站的电路板。
两块板的电源部分和两块板上Spartans的连接基本相同。
任何建议都会受到高度赞赏,因为这个问题导致了我们的主要问题。

以上来自于谷歌翻译


以下为原文



Hi,

We have currently made some customs boards that have a Spartan 6's on them and are continuously damaging them. Part No. XC6SLX150

All the boards work OK, however very intermittingly we are actually damaging the Spartans and sometimes even damaging the Platform Cable USB (programmers).

The problems we are seeing are that we can no longer communicate to the FPGA via the JTAG pins. In some cases IF we actually power the FPGA 3.3V (after been damaged) from a bench power supply we can now also see that its current consumption has raised by a few 100mA. We believe there is a leakage path from the damaged section of the FPGA that is creating the higher current. Note the FPGA can still be programmed from external on board SPI flash so therefore somewhat functional.

The problem always occurs when we are plugging in the USB programmers in to a system that is powered.

As the issue is intermittent it is very hard to debug. But a theory we have is that the systems are floating and when plugging in my USB programmer that is connected to a grounded workstation we are creating an ESD(current dump) event that in turn damages the FPGA's. When measuring from earth to Gnd of the custom board we see an initial high voltage drop (90V) that eventually bleeds away. When measuring from the Ground of the USB programmer to earth we see no voltage drop as in they are at the same potential

As part of our setup I have multiple custom boards, one of these boards I use my Laptop to program the Spartan 6. As the Laptop is also floating we have never damaged this board. It is always the board that is connected back towards the earthed workstation. The Power supply section of both boards and connections to the Spartans on both board are essentially identical.

Any suggestions would be greatly appreciated as this problem is causing us major problems.
0
2019-6-14 12:33:25   评论 分享淘帖 邀请回答
5个回答
嗨mcorriga,
几个问题:
1.什么是电源架构,即什么是正常的电源(PoE,墙壁疣等),以及
如何在船上修改(LDO,切换台等)?
2.除电源连接外,该板还有其他外部世界连接吗?
3.由于外部连接的隔离要求,电路板逻辑是否设计为浮动,或浮动是电源设计的无意副产品?
4.当你说90V“最终流血”时,你的意思是,例如,在几秒钟内,
并且作为测量探针的结果?
或者通过其他方式“最终”流失了吗?
5.您是否有任何理由反对为电路板接地提供单独的接地连接,即使只是
为您的开发环境?
法案

以上来自于谷歌翻译


以下为原文

Hi mcorriga,
 
A few questions:
 
1. What is the power supply architecture, ie., what is the normal source of power (PoE, wall-wart, etc), and
how is it modified on board (LDO's, switcher's, etc) ?
 
2. Are there any other external world connections to this board besides the power connection ?
 
3. Is the board logic designed to be floating due to an isolation requirement of the external connections, or is the floating  an unintentional byproduct of the power supply design?
 
4. When you say the 90V "eventually bleeds away", do you mean, for example, in a few seconds,
and as a result of the measurement probe ?  Or does it bleed away anyway "eventually", thru some other path ?
 
5. Is there any reason that you are opposed to providing a separate grounding connection to the board ground, even if just
for your development environment ?
 
Bill
 
 
2019-6-14 12:45:17 评论

举报

男,
在连接JTAG或任何人之前使用共同的地面!!!!!
这听起来像是一个坏的或反向的交流电源连接(热和中性,或两个0.01uF电容器在仪器中分离到地面(绿色接地线和公共接地之前的常见做法))。
不要做!
你正在吹灭ESD保护,因此也就是设备。
太多的能量,太多的电压:坏,坏,坏。
任何设备如此连接,你认为是好的,可能会很快在某些时候变坏。
即使没有明显的损坏,这样的部件也会损坏部件,部件可能很快就会失效。
我曾经在工作台上发现了6 VAC的差异:在建筑物的主要配电中错误的接通意味着6伏特具有几乎无限的电流能力:它可以熔化1英寸宽的铜带!
我们让一个许可的电子技术人员重新安装了工作台,并让当地公用事业公司正确地将电源分配到建筑物上。这6个VAC吹掉了我们工作的相当数量的电路板,直到我们发现问题为止,
虽然电压很低,但是电流是不可想象的。当时任何5V IC几乎都是立即进行烘烤。这是6 VAC的峰值负摆动,反转偏置衬底会破坏器件。
Austin Lesea主要工程师Xilinx San Jose

以上来自于谷歌翻译


以下为原文

m,
 
USE A COMMON GROUND BEFORE CONNECTING JTAG OR ANYTHING ELSE!!!!!
 
This sounds like a bad or reversed AC power connection somewhere (hot and neutral, or two 0.01uF capacitors split to ground in an instrument (common practice before green ground wires, and common grounds)).
 
DON'T DO IT!  You are blowing out the ESD protection, and hence the device.  Too much energy, too much voltage:  bad, bad, bad.  Any device so connected that you think is good, will probably go bad at some point, fairly soon.  Even though there is no obvious damage, such a zap has damaged the part, and the part will probably fail very soon.
 
I once found a 6 VAC difference in grounds on a workbench long ago:  a wrong tap in the main power distribution to the building meant the 6 volts had a nearly infinite current capabilty:  it could melt a 1" wide copper strap!!!!  We had the bench re-wired by a licensced electircal technician, and had the local utility correct hte gound tap on the mains distribution to the building.  This 6 VAC blew out a considerable number of boards we worked on until we discoved the problem, and fixed it.  Although the voltage was low, the current was unimaginable.  Back then any 5V IC was toasted almost instantly.  It was the peak negative swing of the 6 VAC that reversed biased the substrate destroying the device.
 
 
 
Austin Lesea
Principal Engineer
Xilinx San Jose
2019-6-14 12:55:38 评论

举报

1.什么是电源架构,即什么是正常的电源(PoE,墙壁疣等),以及
如何在船上修改(LDO,切换台等)?
电源结构是Wall-wart,然后在板上我们有SMPS和LDO。
(已经看到3.3V LDO实际上是因为它无法获得当前的电流)
2.除电源连接外,该板还有其他外部世界连接吗?
是。
3个板组成一个系统。
一块板有一个HDMI连接器,来自一个也没有接地的源。
其他人则在板与板之间使用屏蔽电缆菊花链。
因此,该系统具有相同的潜力。
只有当我插入处于地电位的JTAG电缆时,我们才能吹出零件。
3.由于外部连接的隔离要求,电路板逻辑是否设计为浮动,或浮动是电源设计的无意副产品?
它是浮动的,因为电源的非共同副产品
4.当你说90V“最终流血”时,你的意思是,例如,在几秒钟内,
并且作为测量探针的结果?
或者通过其他方式“最终”流失了吗?
测量探头导致流血
5.您是否有任何理由反对为电路板接地提供单独的接地连接,即使只是
为您的开发环境?
我们从现在开始将地上的插头和电线连接到我们的系统中进行最佳实践。
如果我在一周内不再损坏任何部件,那么我们肯定会说这就是我们的问题。
我确定我们不能成为第一个这样做的人,因为有很多电源不能连接它们所连接的系统。
我想说我们遇到的另一个问题是我们正在使用一个小连接器将Jtag连接到我们的系统,并且非常难以直接使用它。
所以我们将它从边到边逐渐加入。
我们的输入JTAG连接器只有一个接地。
(在连接器的mddle中被激活)
感谢您的回复。

以上来自于谷歌翻译


以下为原文

1. What is the power supply architecture, ie., what is the normal source of power (PoE, wall-wart, etc), and
how is it modified on board (LDO's, switcher's, etc) ?
 
The power structure is a Wall-wart and then on the board we have both SMPS and LDO's. ( Have seen the 3.3V LDO actually blow as it cant source the current)
 
2. Are there any other external world connections to this board besides the power connection ?
 
Yes. 3 boards make up a system. One board has a HDMI connector from a source that is also not Earthed. Others have a shielded cable daisy chaining from board to board. Therefore the system is all at the same potential. It is only when I plug in the JTAG cable that is at earth potential that we seem to blow parts.
 
3. Is the board logic designed to be floating due to an isolation requirement of the external connections, or is the floating  an unintentional byproduct of the power supply design?
 
It is floating because on unintentinoal byproduct of the power supply
 
 
4. When you say the 90V "eventually bleeds away", do you mean, for example, in a few seconds,
and as a result of the measurement probe ?  Or does it bleed away anyway "eventually", thru some other path ?
 
Bleeds away as a result of the measurement probe
 
5. Is there any reason that you are opposed to providing a separate grounding connection to the board ground, even if just
for your development environment ?
 
We are from now going to take a plug and wire in the ground from this into our system for best practice's. If i dont damage anymore parts in a week then we can definately say this is what our problem is/was.
 
 
Im sure we cant be the first one to do this as there are alot of Power supplys that do not earth the systems they are connected to. I would say that one of the other problems we have is that we are using a small connector to connect the Jtag to our system and to get it in straight it very hard. So we end up edging it in from side to side. ALSO we only have one ground on our input JTAG connector. ( which is loacted in the mddle of the connector)
 
 
Thanks for your reply.
2019-6-14 13:03:09 评论

举报

嗨奥斯汀,
谢谢你的回复,我们计划在系统的地面。
另一方面,无论如何都要恢复我们的USB编程器加密狗。
注意我还没有时间调试这些,所以任何指针都会很好。
希望有时间看看这是在未来
谢谢,
中号

以上来自于谷歌翻译


以下为原文

Hi Austin,
 
thanks for you reply, We do plan in groundin the systems.
 
On another note is there anyway to recover our USB programmer dongles.
 
note I have had no time to debug these yet so any pointers would be good. Will hopefully get time to look at this is in the future
 
thanks,
 
M
2019-6-14 13:08:47 评论

举报

男,
90 VAC意味着可能存在1.414 * 90的峰值负电压。
如果加密狗不再工作,我想你需要获得一个新加密狗。
没有可识别的“sacualicial”组件,因为它从未被设计为忍受这种滥用。
在连接任何东西之前,请务必检查地面,
Austin Lesea主要工程师Xilinx San Jose

以上来自于谷歌翻译


以下为原文

M,
 
90 VAC means that a peak negative voltage of 1.414 * 90 could have been present.  If the dongle doesn't work any longer, I think you will need to obtain a new one.  There are no identifiable "sacraficial" components, as it was never designed to endure that sort of abuse.
 
Always check your grounds before you connect anything,
 
Austin Lesea
Principal Engineer
Xilinx San Jose
2019-6-14 13:28:26 评论

举报

只有小组成员才能发言,加入小组>>

38个成员聚集在这个小组

加入小组

创建小组步骤

关闭

站长推荐 上一条 /10 下一条

快速回复 返回顶部 返回列表