Re: high curent PCB connector

On Sun, 16 Feb 2025 19:32:56 +0200, Tauno Voipio
<tauno.voipio@notused.fi.invalid> wrote:

On 16.2.2025 7.11, Bill Sloman wrote:

On 16/02/2025 3:17 am, john larkin wrote:

On Sat, 15 Feb 2025 14:13:20 +0000, piglet <erichpwagner@hotmail.com>
wrote:
>

On 15/02/2025 12:20 am, Phil Hobbs wrote:

piglet <erichpwagner@hotmail.com> wrote:

john larkin <JL@gct.com> wrote:

Does anyone have a favorite high-current PCB connector?
>
I'd like to get 20 wires into a pluggable connector, to go on a
module
like this:
>
https://highlandtechnology.com/Product/P948
>
We need at least 7 amps per contact.
>
That litle green Phoenix connector is cool. Wires screw into it
without tooling, and it's easy to mate and unmate. But it's only 5
pins.
>
I was planning to use four of them, with two on a baby board, but
that
idea wasn't popular.
>
>

>
Phoenix  do make 20 pin version (as two rows of ten) with 10 amp
per pin on
200 mil pitch.
>

>
Do they actually make them m, as in somebody has them for sale, or
are they
just listed in a catalogue?
>
Connector catalogs are full of things that they would love to make
for you
if you want to order 100,000 pieces and can wait six months.
>
Cheers
>
Phil Hobbs
>

>
But JL's customers are aerospace so will be used to exotic, hard to find
connectors - would they respect him for using an easily buyable
connector :)
>
piglet

>
The nice thing about the Phoenix connectors is that you don't need a
soldring iron or pins and crimp tools to terminate wires, just a small
screwdriver.

 
That's not nice. It's just cheap. At Cambridge Instruments the argument
for going over to crimp connectors was that you found a lot less duff
connections when you were putting product through final test.
 
The parts and tools were more expensive, but fault-finding and fault
correction were expensive enough to tip the balance.
 
<snip>
 

>
Crimps can also be dud, and they can be difficult to find.
>
Decades ago, we had a Data Products line printer which occasionally
went beserk with paper feed, throwing half a box of chain forms at
speed through the printer. The paper feed was a DC servo built by
the book, with acceleration and deceleration controls done with suitable
integrators. After a long hunt, it was seen that the occasional insanity
came from a missing tachogenerator signal. The culprit was one of the
tacho signal wires crimped partially on top of the insulation.

I rode a LASH ship from San Francisco to San Pedro, a lovely cruise,
to try to find out why it would go berzerk now and then. Turns out it
was an intermittent connection from the speed feedback tach.

I had designed the control system some years before, and I limited the
influence range of the RPM feedback, so the runaways weren't the full
32,000 horsepower.

Nowadays we do most control systems in uP or FPGA code, so we can have
lots of protections.