Quote:
I think I did not understand what was happening and really appreciate your explanation. So am I right in now thinking that the ribbon is something that will be used during final measuring phases and that the Mil-Spec plug will be used in the production units? So the ribbon is like a small bridge between the boards?
Yes. Another way to think of the ribbon cable and those two boards is as an extension cable for the mil-spec connector. The normal M8 that we'll all be getting is mil-spec-to-mil-spec connector with no ribbon cable or adapter boards in between.
I don't know how much Michael wants to tell us about that mil-spec connector, but it looks like one that's designed to carry pretty high-speed digital signals (like 10s or 100s of megahertz). We use one that has a similar footprint (Samtec ERF8 and ERM8 series) to carry multi-gigahertz signals.
The ribbon cable is also probably not usable for M8s that have to work out in the real world. Depending on what Michael's transferring over that connector, he picked that connector balancing a variety of design constraints:
- packaging: the M8 has to be small, so the connector has to be dense and short. The ribbon cable is entirely too big and clunky.
- emissions: fast digital signals can radiate a lot of noise, and connectors that have shields (this one doesn't appear to have one, but this is not necessarily important), enough pins to use every other line as ground (this connector can be used in this way), and defined electrical characteristics like having a characteristic impedance can all help reduce radiated noise. When you reduce emissions you usually also help with susceptibility: the situation where an external source of electrical noise (like cellphones) can screw up your device. When you do the FCC and CE testing for electronic devices, you are expected to test both emissions and susceptibility.
- bandwidth: the ability to pass fast digital signals without distortion which could cause things like jitter if you're driving the clock through that connector, or unreliable communications between the boards. Many of the things that help with emissions also help here.
- durability: you don't want anything breaking or coming loose when it's handled roughly (UPS and FedEx in my experience as most users are pretty careful). Things that are plugged and unplugged also have a cycle rating that says how many times they can be plugged and unplugged before they become unreliable (not important here, but applicable to things like headphone jacks).
- cost: you don't want to use a more expensive part than is necessary to achieve your design and quality goals. There are some very expensive connectors out there for no other reason than that they are produced in very small quantities. Some connectors require handwork to assemble onto a board, and that can cost more money (for the extra human time for putting the board together) and decrease reliability. As much as possible, you want the assembly of the board to be automated: with little exception, it will be cheaper and more reliable.
- availability: available from different vendors is good so you don't run out of parts, or in case one vendor goes out of business
So you can see that picking a connector is a mix of design and manufacturability, and it's not uncommon to have to compromise on one or more of these things because there isn't a connector that fits your needs perfectly.
And that's just for one connector! There are lots of other things that go into the board along with the board itself that can affect the quality, cost, and reliability of the device.
![size]](http://files.head-fi.org/images/smilies/rolleyes.gif[size=13px][/size])
