FFC connectors in a T&M instrument, good or bad idea?

Some of us have restricted our use to coarse pitch choices, reducing some potential problems. I've been using six-wire 1mm pitch FFC, outside, exposed to the elements, for six months, w/o any failures. The FFC sneaks out of a bee-hive, between supers, midway up, and supports the weight of a PCB + jack, and a molded 6-wire cable. It's plenty rugged enough for the job, using Molex 522070633 locking connectors.

I've had some recent experience in low-volume T&M production with FFC cables. They work fine (as any LCD manufacturer can tell you) but they may require some extra assembly training, as no two manufacturers' FFC connectors seem to work exactly the same way. Some of them are downright unintuitive -- e.g., FH40-30S-0.5SV, which is 'open' when it is 'closed' and vice versa. Somebody ripped the entire connector off of one of our PCBs trying to remove the cable from one of these.

Also, not all FFC cables are created equal. Some of them are unnecessarily stiff, so they are prone to pulling out of their connectors during final assembly. Molex 0982660326 is an example, lots of problems with that one went away when we switched to Wuerth 687730200002. The latter is slightly longer but also much more flexible, and cheaper as well.

You need to order one of every connector and cable you intend to use, so you can evaluate its mechanical qualities in person. Not to mention the footprints and pinouts and clearances and whatever else.

-- john, KE5FX

My TDS460 has a few, using the 0.1" pitch, crimped kind I think. Seems fine.

More relevant, pretty much everything with keypads or LCDs has flex somewhere in it, and usually at a finer pitch (0.5mm being very common). Cellphones and some laptops are good examples, but cameras (heh, those that haven't been displaced by cellphones) contain some excellent examples of flex circuits and connectors. (Single-row edge connectors and dual-row plug style connectors are common on these.)

None of these examples are particularly unreliable. They aren't long lifetime products for the most part, but they are made in the millions, and you'd hear of a lot more failures if they were that bad. And the displays and keypads are just as typical of modern DSOs.

Heh, worth noting that micro-coax is common on a lot of these displays, because of the high bandwidth (usually something like 24 lane LVDS?). Same connectors (fine pitch, single or dual row, header or plug style) but not so much made with flex. :)

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Design Website:

Back story: We have a history of reliability problems in some of our instruments because of some flaky connectors. Most of it can be traced back to bad workmanship when doing the crimps. We've been working with contractors to increase quality and we also significantly increased QA inspections, but some defects still go through. Of course we use cable testers but the main problems are intermittent connections that pass the cable tester test.

So I'm thinking, the best way to avoid bad connector crimps is to get rid of the crimps altogether, right? FFC cables seem like a great solution. They also eliminate manual labor required to manufacture electrical harnesses and therefore reduce chance for mistakes and also reduce cost.

The main drawback I see is the limitation to point-to-point connections (can't do a Y harness for example), but it's something I can live with.

Another drawback is the perceived fragility of such connectors. It's actually the main concern of my colleagues when I brought up the idea of these connectors. Is it founded? I found some high robustness connectors for the automotive market (Hirose FH52) and I feel that these should be plenty robust for our use but I have no real-world evidence to provide.

There also seems to be a potential issue of inserting the cable slightly crooked. But tabs on the flex cable seem to be a good solution to this problem.

All in all I feel that FFC cables are a great option for internal wiring, but I'm having a hard time convincing my colleagues.

Any feedback on the use of these connectors is appreciated.

Molex makes some seriously cool ones, e.g. 5025982393, which use 1-mm pitch pads in three rows, with a FFC cable pitch of 0.3 mm. We use those to connect with cold plates.

Cheers

Phil Hobbs

Thanks everyone for the feedback.

Would you guys say that my assumption is correct that in theory, these connectors should be more reliable than crimped ones, simply because you remove the chance for a bad crimp?

In other words, is there any compelling reason NOT to move every/most internal harnessing to FFCs? Simpler (no harnessing drawing required), cheaper, more reliable. What's not to like?

My intuition is 0.1-inch MTA punchdown types would be more reliable. E.g., 3-640440-2

Depends on how often it needs to be taken apart. Those things are only rated for 20 cycles.

Similar situation here. Too many crimp failures causing too much flying around the world fixing things. The most reliable connections are the ones we switched to soldered type.

We use FFC for board-to-board connections that get attached only once. We ran into trouble here too, but once we got manufacturing straightened out, no more troubles.

Anywhere we used FFC for something that gets plugged/unplugged with any regularity turned into a disaster. From crooked cables, to broken connector latches, to creased cables...always a problem.

• Who sez can't do a Y?

Step one: Double-sided flex to duplicate wires at connection, plus feed-thru vias to connect the duplicates together.

Step two: Two single-sided flex with same pattern (which otherwise be a standard, normal connection).

Step three: Sweat solder each single-sided flex to one end and side of the double-sided flex (becomes the pick-off and "Y" place).

Done, but would be advisable to find a way to mechanically hold/fix each single-sided flex at/near solder places.

Thanks again everyone.

To address some comments, I'm thinking about FFC for things that are only c onnected once at production time. So mating cycles is not really an issue.

The MTA connectors suggested are indeed reliable in my experience but such large connectors are not always practical. We also need to deal with 100 Mb ps LVDS signals and I don't think MTA connectors are appropriate there, eve n if our bitrate is relatively low compared to the Gbps levels now commonpl ace.

Soldered connectors is not really an option. We prefer not to have to do an y soldering in-house as this requires more skilled labor. There's also the age-old debate about crimped vs soldered and it's pretty well established n ow that crimped is better than soldered (for discrete wires at least). That is, assuming that both are correctly done, which is exactly the problem he re. Both require good workmanship and quality of workmanship is the problem I'm trying to solve by getting rid of it altogether by using FFCs.

They're used widely in the Automotive world. That's a pretty harsh environment (and has some serious quality standards).