flexible ribbon "PCB" to regular PCB

[Winfield Hill]

Looking for good ways to connect / solder a flexible ribbon PCB to a regular
PCB, without using a connector. We need to be able to pull hard on the link,
without the flexible ribbon coming out. Maybe using Molex Premo-Flex, etc., to
avoid having to fab the flexible ribbon? We only need four wires, for an I2C
connection.


--
Thanks,
- Win

[TTman]

I've seen them soldered directly to a PCB- it was very unreliable so
don't go that route (on an LCD to PCB backlight)

---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

[Winfield Hill]

TTman wrote...

>
> On 01/05/2019 12:56, Winfield Hill wrote:
>> Looking for good ways to connect / solder a flexible ribbon PCB to a
>> regular PCB, without using a connector. We need to be able to pull
>> hard on the link, without the flexible ribbon coming out. Maybe using
>> Molex Premo-Flex, etc., to avoid having to fab the flexible ribbon?
>> We only need four wires, for an I2C connection.
>
> I've seen them soldered directly to a PCB- it was very

> unreliable so don't go that route (on an LCD to PCB backlight).

OK, good advice. I can use the recommended connector, and after
mating the ribbon, seal the entrance and cover it all with epoxy.
I see 1.25mm six-wire is only 9 mm wide, but I assume it's strong.


--
Thanks,
- Win

[Tom Del Rosso]

Pass it through a slot in the board first to give it more strength.
Epoxy it there too.

[tabb...@gmail.com]

On Wednesday, 1 May 2019 12:56:54 UTC+1, Winfield Hill wrote:
> Looking for good ways to connect / solder a flexible ribbon PCB to a regular
> PCB, without using a connector. We need to be able to pull hard on the link,
> without the flexible ribbon coming out.

then you're out of luck, solder isn't that robust.
You can feed it through 2 slots to improve things, but it still won't be long term tuggable. Maybe 8 slots :)

NT

[Winfield Hill]

Tom Del Rosso wrote...

Awesome idea!


--
Thanks,
- Win

[TTman]

I'd prefer to use clear silicone- a bit of flexibility/resilience

[Winfield Hill]

TTman wrote...

Yes, but it peels off easily. Hmm, hot glue is slightly
flexible, and doesn't peel quite so readily.


--
Thanks,
- Win

[Tim Williams]

"Winfield Hill" <hi...@rowland.harvard.edu> wrote in message
news:qacjq...@drn.newsguy.com...

> Yes, but it peels off easily. Hmm, hot glue is slightly
> flexible, and doesn't peel quite so readily.

Depends on type, of course. I would also recommend 3M VHB tape. Maybe
that, inbetween two slots, with the connector on the exit side, tape on the
opposite side and the cable routed between all three in a zig-zag fashiion,
would be best.

Incidentally, mind EMI. I2C is not "inter-printed circuit" for a reason.
Even with a ground plane on that cable, don't expect much performance!

Tim

--
Seven Transistor Labs, LLC
Electrical Engineering Consultation and Design
Website: https://www.seventransistorlabs.com/

[Winfield Hill]

Tim Williams wrote...
>
> Winfield Hill wrote,

>>
>> Yes, but it peels off easily. Hmm, hot glue is slightly
>> flexible, and doesn't peel quite so readily.
>
> Depends on type, of course. I would also recommend 3M VHB tape.
> Maybe that, inbetween two slots, with the connector on the exit
> side, tape on the opposite side and the cable routed between

> all three in a zig-zag fashion, would be best.

God ideas, thanks!

> Incidentally, mind EMI. I2C is not "inter-printed circuit"
> for a reason. Even with a ground plane on that cable, don't
> expect much performance!

We'll be using a slow data rate. With a choice of pin
order, how would you arrange +3.3, SDA, SCL and GND?


--
Thanks,
- Win

[Tim Williams]

"Winfield Hill" <hi...@rowland.harvard.edu> wrote in message

news:qacls...@drn.newsguy.com...

>
> We'll be using a slow data rate. With a choice of pin
> order, how would you arrange +3.3, SDA, SCL and GND?
>

+3.3, SDA, GND, SCL, GND. Or actually, putting 3.3 in the middle and GNDs
at the ends may be better for hot-plugging or cable-wearing purposes, but
not relevant electromagnetically (as long as there's a bypass cap from 3.3
to GND at each end, which there usually is; for the same reason, always
connect all GNDs and supplies at the connector, whether you need the current
capacity or not).

I forget if 5-pin cables or connectors are common. Maybe you're stuck with
4 for space, or 6 for availability?

Low data rates may not be enough, because of noise generating false clocks.
I2C is generally tolerant of fast noise, but a cable exposes a much broader
bandwidth than PCB traces do; it can pay to add ferrite beads (for damping
and filtering) and small caps (as many pF as you can afford, usually
100-470) at both ends of the cable. Uh, beads facing the cable, that is.

[Tom Del Rosso]

Tim Williams wrote:
> "Winfield Hill" <hi...@rowland.harvard.edu> wrote in message
> news:qacls...@drn.newsguy.com...
>>
>> We'll be using a slow data rate. With a choice of pin
>> order, how would you arrange +3.3, SDA, SCL and GND?
>>
>
> +3.3, SDA, GND, SCL, GND. Or actually, putting 3.3 in the middle and
> GNDs at the ends may be better for hot-plugging or cable-wearing
> purposes, but not relevant electromagnetically (as long as there's a
> bypass cap from 3.3 to GND at each end, which there usually is; for
> the same reason, always connect all GNDs and supplies at the
> connector, whether you need the current capacity or not).
>
> I forget if 5-pin cables or connectors are common. Maybe you're
> stuck with 4 for space, or 6 for availability?
>
> Low data rates may not be enough, because of noise generating false
> clocks. I2C is generally tolerant of fast noise, but a cable exposes
> a much broader bandwidth than PCB traces do; it can pay to add
> ferrite beads (for damping and filtering) and small caps (as many pF
> as you can afford, usually 100-470) at both ends of the cable. Uh,
> beads facing the cable, that is.

Beads on a kapton cable? (Or "flexible PCB" as it was described.)

[Tom Del Rosso]

Or make it wide enough for a hole or 2 and screw it down. Glue below
and washer on top.

Or if PCB slots or washers are too sharp on the edges, and might cause
it to break, a few large guage wires with insulation, can be mounted
like jumpers closely in parallel. Then you can zigzag around them, and
glue. That should give it the least stress.

[Tim Williams]

"Tom Del Rosso" <fizzbin...@that-google-mail-domain.com> wrote in
message news:qacn7r$kvg$1...@dont-email.me...

>> Low data rates may not be enough, because of noise generating false
>> clocks. I2C is generally tolerant of fast noise, but a cable exposes
>> a much broader bandwidth than PCB traces do; it can pay to add
>> ferrite beads (for damping and filtering) and small caps (as many pF
>> as you can afford, usually 100-470) at both ends of the cable. Uh,
>> beads facing the cable, that is.
>
> Beads on a kapton cable? (Or "flexible PCB" as it was described.)
>

Ah, on the signals that is. :)

You would be able to put a clamp bead over the cable though, which is also a
good idea if you find noise to be a problem.

Did basically that, five years ago, on a device with a mainboard and a HID
board, connected by a 6" ribbon cable carrying I2C. It failed radiated
susceptibility at ~200MHz (basically the resonant mode of the two boards
joined by the cable). A clamp bead fixed it (damping the resonant mode -->
less V dropped across the cable at F_res).

That same project, had a number of issues that we strongly recommended
against, but which the customer demanded regardless. The elephant in the
box was they wanted USB over unshielded cables... Apparently it managed to
pass radiated, but EFT knocked it out. We were forced to use shielded
enclosures, cables and connectors (instead of nice cheap plastic automotive
connectors), nearly doubling the product cost...

[whit3rd]

On Wednesday, May 1, 2019 at 4:56:54 AM UTC-7, Winfield Hill wrote:
> Looking for good ways to connect / solder a flexible ribbon PCB to a regular
> PCB, without using a connector. We need to be able to pull hard on the link,
> without the flexible ribbon coming out.

One possibility (that i've never seen used) is to put holes in the
flex, to engage pins, for mechanical positioning, then apply a
clamp over a flex/ zebra-stripe/PCB sandwich. This isn't much of an improvement
over 'using a connector', though, it has loose parts to be
assembled. With the right plastics manufacturer, it'd be two; zebra
elastomer and snap-over plate (holes in PCB and flex engage the plate's
plastic snaps).

[Hul Tytus]

Win - ribon cable, ala Belden, etc, is a "golden oldie" for such
applications. Typically, though, a ground lead for each signal lead is
required.

Hul

[Winfield Hill]

Hul Tytus wrote...
>
> Win - ribbon cable, ala Belden, etc, is a "golden

> oldie" for such applications. Typically, though,
> a ground lead for each signal lead is required.

Thanks, but this time we need the super-thin aspect.
It sneaks between two "supers" = bee-hive sections.
We only need four I2C wires, but the smallest
stocked Molex 1.25mm Premo-Flex has six.


--
Thanks,
- Win

[Lasse Langwadt Christensen]

does it have to be molex and 1.25?

https://uk.farnell.com/pro-power/pp001497/ffc-jumper-cable-1-25mm-4core/dp/2776622

or

https://www.ebay.com/itm/253348078506

[Winfield Hill]

Lasse Langwadt Christensen wrote...

No, it could be either of those, but then I'd have to
be confidant of the mating connector. We do need super
thin. Haha, I'm not really sure of the Molex Premo-Flex
thickness spec. I can see it's really thin.


--
Thanks,
- Win

[gnuarm.del...@gmail.com]

They make a flat cable termination that is intended for soldering to a PCB. It has pins for through hold mounting, very secure. I want to say the individual pins in the body crimp to the flex, but I'm not certain of that. It might need to be made part of the cable. Do you not see anything like this in premade flex catalogs?

--

Rick C.

- Get a 1,000 miles of free Supercharging
- Tesla referral code - https://ts.la/richard11209

[Winfield Hill]

gnuarm.del...@gmail.com wrote...

>
> They make a flat cable termination that is intended for
> soldering to a PCB. It has pins for through hold
> mounting, very secure. I want to say the individual
> pins in the body crimp to the flex, but I'm not certain
> of that. It might need to be made part of the cable. Do
> you not see anything like this in premade flex catalogs?

Yes, thanks, I remembered seeing something like that,
so I looked for it, but didn't find it.


--
Thanks,
- Win

[k...@notreal.com]

On 1 May 2019 04:56:38 -0700, Winfield Hill <hi...@rowland.harvard.edu>
wrote:

Use "rigid-flex" board? ...basically flex FR4 laminated with rigid
FR4.

[Jasen Betts]

I understand that a hot bar is used to solder flex pcb to rigid pcb.

if you're going to epoxy the joint the weakest part will be where the
flex comes out of the epoxy,

--
When I tried casting out nines I made a hash of it.

[Jasen Betts]

On 2019-05-01, Winfield Hill <hi...@rowland.harvard.edu> wrote:

> Thanks, but this time we need the super-thin aspect.
> It sneaks between two "supers" = bee-hive sections.

So, open ended boxes, rebated to stack securely, or just stacked;
but softwood. Drill a 1/4" hole and put a flex with 3.5mm plug
through it. bees won't mind a 1/4" hole, if it bothers them
they'll plug it with propolis.

> We only need four I2C wires, but the smallest
> stocked Molex 1.25mm Premo-Flex has six.

The bees are going to glue that to both boxes with propolis, the
next time the hive is opened it'll snap depening on which side of the popolis snaps
first.

How many bee keepers are there on your team?

[Winfield Hill]

Jasen Betts wrote...

>
> On 2019-05-01, Winfield Hill wrote:
>
>> Thanks, but this time we need the super-thin aspect.
>> It sneaks between two "supers" = bee-hive sections.
>
> So, open ended boxes, rebated to stack securely, or just stacked;
> but softwood. Drill a 1/4" hole and put a flex with 3.5mm plug
> through it. bees won't mind a 1/4" hole, if it bothers them
> they'll plug it with propolis.

Have you tried the Broodminder Temp+Hum sensors,
that you can locate in the center of the hive?
They have a 1x1-inch bluetooth PCB in one end of
folded clear plastic, and a strip of the plastic
sneaks out of the hive between supers. The bees
leave the plastic assembly alone, no propolis.

But we have trouble getting their data into our
hive-monitor system, prefer a wired connection,
plus I'm placing eight sensor parameters (!) at
the middle-of-the hive location. But we'd like
to keep convenient place-it-between supers, move
it up as the hive grows and gets more supers, etc.

A six-inch piece of thin flex should work fine,
using the strain-relef solutions suggested here.

> How many bee keepers are there on your team?

Hah, sadly a bit under populated at the moment,
but hey, bee hives are best left alone, right?
Some experience beekeepers have offered to help.
Then there's me for elec and PCB design, plus
an excellent math programmer, plus a 3rd-year
EE summer student to help with cables, sensors.
An informatics post-doc for database software.
Plus an Institute technician one day a week.
And the project's PI, who's doing 3D printed
housings, etc. OK, that's a pretty good team!


--
Thanks,
- Win

[Jasen Betts]

On 2019-05-02, Winfield Hill <hi...@rowland.harvard.edu> wrote:
> Jasen Betts wrote...
>>
>> On 2019-05-01, Winfield Hill wrote:
>>
>>> Thanks, but this time we need the super-thin aspect.
>>> It sneaks between two "supers" = bee-hive sections.
>>
>> So, open ended boxes, rebated to stack securely, or just stacked;
>> but softwood. Drill a 1/4" hole and put a flex with 3.5mm plug
>> through it. bees won't mind a 1/4" hole, if it bothers them
>> they'll plug it with propolis.
>
> Have you tried the Broodminder Temp+Hum sensors,
> that you can locate in the center of the hive?
> They have a 1x1-inch bluetooth PCB in one end of
> folded clear plastic, and a strip of the plastic
> sneaks out of the hive between supers. The bees
> leave the plastic assembly alone, no propolis.

> But we have trouble getting their data into our
> hive-monitor system, prefer a wired connection,
> plus I'm placing eight sensor parameters (!) at
> the middle-of-the hive location. But we'd like
> to keep convenient place-it-between supers, move
> it up as the hive grows and gets more supers, etc.

> A six-inch piece of thin flex should work fine,
> using the strain-relef solutions suggested here.

so long as the bees really leave it alone, and the bee-
keeper manages to miss it when unsticking a stuck super
it now sounds like a good solution.

For strain releif zig-zagging through a PCB works
well with rubber or PVC insulated wires, but yeah
for something hard and smooth some glue will help

[speff]

On Wednesday, 1 May 2019 07:56:54 UTC-4, Winfield Hill wrote:
> Looking for good ways to connect / solder a flexible ribbon PCB to a regular
> PCB, without using a connector. We need to be able to pull hard on the link,
> without the flexible ribbon coming out. Maybe using Molex Premo-Flex, etc., to
> avoid having to fab the flexible ribbon? We only need four wires, for an I2C
> connection.
>
>

> --
> Thanks,
> - Win


There's the heat sealed connections (conductive heat sensitive adhesive) used in high volume. Or you could consider rigid-flex and then the board would come with the flex incorporated. But I don't think either one is all that suitable for small quantity unless cost is not a concern.

I think your connector + epoxy is probably best but I would keep the epoxy away from the connector lest it affect the reliability.

--Spehro Pefhany

This stuff is interesting- kind of a heat-sealed version of the zebra elastomeric connectors (anisotropic conductivity so you just have to line the pads up on either side) but again, probably unsuitable for small quantities:

https://www.dexerials.jp/en/products/a7/cp20000.html

[Winfield Hill]

speff wrote...

>
> There's the heat sealed connections (conductive heat sensitive

> adhesive) used in high volume. ...

>
> This stuff is interesting- kind of a heat-sealed version of the
> zebra elastomeric connectors (anisotropic conductivity so you
> just have to line the pads up on either side) but again,
> probably unsuitable for small quantities:
>
> https://www.dexerials.jp/en/products/a7/cp20000.html

Yes, very interesting. Can you suggest other heat-sensitive
adhesives (doesn't need to be conductive if I use a socket).


--
Thanks,
- Win

[Hul Tytus]

Win - for thinness, you might look at Parlax's "flat cable". Digikey's
catalog shows one with a thickness of .011 inch and a current rating of
1.5 amp. There are others with a lesser rating...

Hul

Winfield Hill <hi...@rowland.harvard.edu> wrote: