Choose a board to board high current connector

[nour]

Hi,
I am working on a small 4 layers board (50mm*50mm) for a Stepper motor driver. This board will be a daughter board which will contain only the controller and the Mosfets and other small packages(0603-0805) caps and res.
Rated operating current up to 10A and voltage up to 55v.
Also, there will be some control signals In and Out
I want to choose a proper board to board connector that can handle the rated operating voltage and current and at the same time has a small package size

After doing some search, I found this connector which may be a good choose



According to the datasheet, it is 0.5A rated and this is a 40 pins connectors so I suppose it can handle up to 20 amps!
I am going to use one for power supply, one for the driver output and another one with fewer pins for control signals.

I have chosen this one because I want all the connector to be at the same height

My question now, is this connector suitable for the task?

[BravoV]

According to the datasheet, it is 0.5A rated and this is a 40 pins connectors so I suppose it can handle up to 20 amps!

At what temperature ? Also are you sure your circuit distributes the current evenly ?

[nour]

At what temperature ?

According to the datasheet, the operating temperature up to 85C and I will make sure the board doesn't reach that high at all conditions(active cooling)

Also are you sure your circuit distributes the current evenly ?

Did you mean the connector pins, well I am not sure about it, but assumably yes!

[nour]

Any other great ideas?

[bob225]

Pin pitch and how many ways ? surface mount or through hole ??

0.5 A x 40 pins = 20 amps , at a low pitch the traces on the pcb wont be able to handle the current, take a look at the standard drivers and there construction

At 50x50mm pcb size your going to have major issues with heat soak

[Rerouter]

It says 30 milliohms of initial resistance, so lets say its 50 after a few cycles to be certain, per plug,

ok so half an amp per pin would put it at about 0.5W of heat per side for all 40 pins used, Thats going to need some level of heat dissipation, in the way of planes going into your board, and it will be somewhat hotter than the normal calcs due to being confined between 2 boards,

[nour]

It says 30 milliohms of initial resistance, so lets say its 50 after a few cycles to be certain, per plug,

ok so half an amp per pin would put it at about 0.5W of heat per side for all 40 pins used, Thats going to need some level of heat dissipation, in the way of planes going into your board, and it will be somewhat hotter than the normal calcs due to being confined between 2 boards,

Yes, you are right, but I don't think 0.5W is a big of a deal! and also the figures that I mentioned is the maximum operating conditions
I am planning on running the board within the range of 4 to 5A, so I don't expect this to be an issue.
The problem that I want the board to be a generic solution as possible(I am planning on making this an open source project after an intial running of the board) so it can run on the full range of power without a problem

[Zero999]

It says 30 milliohms of initial resistance, so lets say its 50 after a few cycles to be certain, per plug,

ok so half an amp per pin would put it at about 0.5W of heat per side for all 40 pins used, Thats going to need some level of heat dissipation, in the way of planes going into your board, and it will be somewhat hotter than the normal calcs due to being confined between 2 boards,

Yes, you are right, but I don't think 0.5W is a big of a deal! and also the figures that I mentioned is the maximum operating conditions
I am planning on running the board within the range of 4 to 5A, so I don't expect this to be an issue.
The problem that I want the board to be a generic solution as possible(I am planning on making this an open source project after an intial running of the board) so it can run on the full range of power without a problem

I agree 0.5W isn't much and t he data sheet for the connector, states the temperature rating of 85^oC, already includes the temperature rise, due to the power dissipated.

[T3sl4co1l]

Connectors are almost always rated with one pin carrying current, the rest unloaded.

The current capacity drops precipitously when more pins are loaded!

Last time I did that, I used Molex Mini-Fit connectors.  Big, beefy, reliable, and rated for the job.

If you need something more compact, consider blade type connectors.  They may be more expensive or hard to find (particularly the ones up to 60A+ per contact, not that you need that much!).

Tim

[nour]

Connectors are almost always rated with one pin carrying current, the rest unloaded.

The current capacity drops precipitously when more pins are loaded!

Last time I did that, I used Molex Mini-Fit connectors.  Big, beefy, reliable, and rated for the job.

If you need something more compact, consider blade type connectors.  They may be more expensive or hard to find (particularly the ones up to 60A+ per contact, not that you need that much!).

Tim

May I ask you about this phrase:
"The current capacity drops precipitously when more pins are loaded"
Why?

I have considered using those connectors with blades, but the problem I am facing that I need all the connectors to be at the same hight (there is still data connection)so it will be suitable for a board to board connection

[T3sl4co1l]

Connector ratings are all about heat.

Tim

[Rerouter]

its rated on 1 pin, (generally a center pin) sometimes with a plane to sink heat, sometimes not, all to inflate the ratings above a line in the sand they have been asked to meet while meeting other engineering requirements, e.g. amount of material, density, ease of connection.

In short the reason is marketing, if they can say it meets 0.5A max per pin under this list of very specific requirements, It gets found by more engineers than saying it can handle 0.2A on every pin up to 200 cycles, with minimal trace thickness.

sometimes, like in that datasheet (marketing fluff sheet) you have to dig elsewhere to find the actual info, like what conditions the rating was tested under, what the thermal resistance of a given pin is likely to be, and from there calculate what plane areas, and what approaches you need to take to meet your spec,

Rather than them being honest, and just choosing the next pin pitch up on the listing... 

[rsjsouza]

As others have mentioned, heat and aging will be your enemy.

If you can afford a wider distance between boards, check something designed for the task, such as:
https://www.samtec.com/connectors/rugged-power/power

[nour]

As others have mentioned, heat and aging will be your enemy.

If you can afford a wider distance between boards, check something designed for the task, such as:
https://www.samtec.com/connectors/rugged-power/power

Very expensive
I need like 6 of those connectors, they have the perfect combination but too expensive 

[nour]

I think I will change my strategy, I will use for signals a normal IDC cable and for the power connectors I will use this:



According to the datasheet, it can handle up to 8A may be I will add an extra couple of them and that's it
The real problem is the space on the board 
will see!

Datasheet:
https://www.tme.eu/en/Document/d4df6bb0f061af30c576bc5b5b8df3df/15EDGA-3.xx.pdf

[Zero999]

How much money and time do you have to develop this?

How many are you going to build?

If this is for mass production and you've got lots of time and money to develop it, then how about bench testing different connectors, to see what the power loss and temperature rise is at different currents?

[nctnico]

I think I will change my strategy, I will use for signals a normal IDC cable and for the power connectors I will use this:



According to the datasheet, it can handle up to 8A may be I will add an extra couple of them and that's it
The real problem is the space on the board 
will see!

You have to derate by a factor of two to be safe. If space is a problem you can also use screw terminals which you can solder into the board. These don't need to be located at the edge of the board.

[bob225]

50x50mm is a very small area to distribute the heat of the driver and the fet's

Off the shelf drivers are cheap and plentiful

eg. leadshine DM860I rated for nema 34, 20-80v and 7.8 amp


If your running servo motors with encoders it gets a bit more complex

I run Nema 23's on a dual width openbuilds C-Beam

[dmills]

Don't even think of buying Samtec from a full service distributor, you will cry at the cost, instead deal direct, MUCH more reasonable for a few tens on up.

Molex have similar things, or go for the 2mm hard metric stuff for the data and appropriate power modules for the power, generally also available in press fit which is far more reliable then soldering.

Regards, Dan.

[max_torque]

I guess one good thing about putting the same signal through lots of pins is that you can use a large power plane to connect to the connector, and that means good thermal; contact too! (don't use thermals, connect the entire plane to each pin).

It would be a quick test to get two copper clad pcbs, two connector halves, solder the connectors to each pcb, and run some current tests to see when things get out of control!