I have a question about my Electrical Engineering Stack Exchange post: I need to transfer 10 signals and power between devices, is there any recommended solution?

I checked the relevant rules carefully, but I was at a loss. It seems to me that this question is quite open and not limited to asking which exact type of cable should be used, any scheme, structure or algorithm that can transmit the relevant signal is welcome. Or are questions such as "USB cable", "D-SUB cable" and "RS486 cable" all forbidden to answer as "specific products or places to purchase them"? At the same time, I also browsed other questions, and it seems that such as "Is there an IC that can do this?" is allowed or even encouraged, which in my opinion is in line with "specific products or places to purchase them" of.

Then again, the premise of this question is indeed not clear enough, but I don't see what it has to do with "specific products or places to purchase them".

I looked at some other questions, and it seems that "specific products or places to purchase them" is being used indiscriminately as a "fuzzily-defined offences", and some irrelevant questions are tagged with this label and closed.Such as this:Closure reason doesn't make any sense to me


1 Answer 1


EE.SE (Electronics Engineering Stack Exchange) is a Q&A (Question-and-Answer) engineering website. Unlike other sites, discussion and open-ended "projects" are a poor fit here. Everything must be a question/answer. A solid, definable question gets accurate answers. Hopefully; that is the goal anyways.

Partly this is because of the complexity of electronic engineering. There is far too much physics involved in something as simple as sending a signal over a wire to explain every detail involved. Such an explanation would take several years of college classes at least. Therefore, engineers break the (overall) problem down into sub-problems, and tackle each one methodically and individually.

For the askers, asking questions here becomes challenging after a number of years, because chances are very high that the question (or a variant thereof) has already been asked.

For the answerers, answering many questions after a number of years, they see lots of the same things over and over. They lose patience with newcomers whom do not understand or follow the rules. They also get bombarded with tons of (dumped on them) homework questions and other spam daily. So there is a great impetus to close questions from "new" users, primarily because most of that is spam, or posted by those expecting a quick answer to a $10,000 problem.

So you have just joined EE.SE (a reputation score of 1) and posted a question... great! But then it gets closed as being a "shopping" question. I think that is not the real reason it was closed - really, it is too broad, unaware, and project-like. But those are none of the four choices the reviewers get to choose from for closing the question, so the closest one they could agree on was "shopping."

Closing a question isn't a permanent thing here - they are often edited and reopened. Once you get enough reputation, you can see that there are thousands of closed questions here. Eventually they may go to the null bin, but not after years of inactivity.

So instead of editing the question to get it re-opened, you instead created another, identical question. This is specifically frowned-upon here as it means even more work for the moderators, and just proves that you have not read and understood the EE.SE Tour. This introduction is shown to every new user to the site. Unfortunately, it only has one line about closing questions: "Questions that need improvement may be closed until someone fixes them." So that isn't very clear at all. Even I didn't read it fully when I joined years ago, and my first question had to be helped (a lot!) But I learned from that, and moved forward.

In that Tour, it explains that you can ask about things like:

  • a specific electronics design problem
  • the theory and simulation of electromagnetic forces
  • a communication scheme
  • the writing of firmware for bare-metal or RTOS applications

It also states that things like these are not a fit here:

  • Shopping or buying recommendations
  • Consumer electronics such as media players, cell phones or smart phones, except when designing these products or modifying their electronics for other uses
  • Programming software for a PC
  • Anything else not directly related to electronics design
  • Questions that are primarily opinion-based
  • Questions with too many possible answers or that would require an extremely long answer

While I feel that the basis of your question was on-topic for this site (a question about wiring), I feel that an answer is not possible due to the ambiguitiy of the question. An answer attempt would take the equivalent of four years of college classes to fully explain. So it fits under the "questions that would require an extremely long answer" category.

Lets take a look at the question:

I am designing two handheld devices, which need to be connected to each other to transmit at least ten high-frequency (tens of megahertz) digital and analog signals and power supply,some of these analog signals are pulsed and some are analog level signals that may change rapidly. The accuracy of the analog signal is very high, and the external electromagnetic environment is also very bad, so I'd take care to prevent crosstalk between the wires and interference from the outside, which means (I guess) individual shields for each transmission line. If you use 10 coaxial cables, the equipment will be very bloated and messy, and I am not sure whether the general-purpose D-BUS cables can meet the above requirements. The overall cable length doesn't need to be very long (within 1 meter), so is there any recommended solution?

This is why I think the question was closed:

  • "Tens of MHz" = requires knowledge about signaling, dielectric and insulator properties, and transmission line theory. If you do not already know these, an answer may be useless for all but this one problem. How willing would you be to answer a question, knowing that your answer would only help this one person, in this one particular question, ever? That answer must be 42 pages long (minimum.)
  • "The accuracy of the analog signal is very high" = "Very high" is not an engineering unit. How accurate? Over what frequency range?
  • "external electromagnetic environment is also very bad" = ditto. How bad? What is the source of the EMI? What are its frequency components?
  • "crosstalk... means (I guess) individual shields" - Engineers do not guess; they are not called Guessineers! This one question is a huge can-of-worms, with many possible answers, because it depends on EMI, which was ambiguous.
  • "[length] ... (within 1 meter)" = ambiguous. Your "tens of MHz" and "very accurate" signals will be unequally degraded by 10cm, 100cm, and 1000cm of wire. How much delay, attenuation, and noise can your systems tolerate?

To design something to actually work, you have to know the details first. All of the details. You have to put units and limits on everything:

  • "Tens of MHz" --> 25.3MHz fundamental, with >-40dB overtones to 278.3MHz (analyze the signals with an oscilloscope and signal analyzer to ensure that what you are saying is accurate.)
  • "Analog Signal" --> 1vpp, DC-20kHz (saying the application can also help, because the answer could be different for life-critical EKG signal versus a top-10 radio playlist.)
  • EMI --> 100V/m environment (or whatever you do know, such as "inside automobile dashboard" or "right next to 100HP 480VAC motor" or "100m from 50kW WKRP broadcasting tower" etc.)
  • Length --> decide on a length, don't guess or assume. Length will affect delay and attenuation.

For the other aspects of the question, there are many sub-questions inside this question or things you need to think about, identify, and handle individually first.

Yes, the overall project may be digital and analog signals over some wires, some distance, in an electrically noisy environment. But the quickest way to get frustrated is to lump everything together and that's exactly what this question is. Nobody will touch it because they get a headache just looking at it.

Instead, break the problem down in to manageable chunks. Define exactly what digital signals are needed. Define what analog error is acceptable. Define the length of the wire, so the delay can be calculated. See if that delay affects anything. Define what is causing the EMI and how much of it (and what frequencies) are involved. Then shield each of these adequately. This may mean making several prototypes, some of which may not work well or at all. It may even mean throwing a lot of hard work away and going a different route, like optical fiber. But if you persevere, you can tackle one hurdle, then the next, then the next.

Also consider getting some books on all of these subjects. And if you can, take some classes. These will all help.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .