Author Topic: High current on PCB backplane (~65A) - Experience?  (Read 2356 times)

0 Members and 1 Guest are viewing this topic.

Offline DavidTopic starter

  • Frequent Contributor
  • **
  • Posts: 280
  • Country: gb
High current on PCB backplane (~65A) - Experience?
« on: August 31, 2018, 07:46:30 pm »
Hi all,

I am in the process of designing a 19" rack based product at work with very little time to get things up and running  :palm: Ideally I need to run high current through the backplane (approx 65A continuous). My current design actually runs the high current through wires connected straight to the back of the DIN41612 connectors on the board. However for several reasons it would greatly simplify things to run these on the PCB. I can't go to heavy on the copper weight as unfortunately there are some TSSOPs on the board and the track/gap limitations will become a problem. I don't really have any financial limitations but whatever solution I go with must be readily available and not rely on anything exotic.
My current thoughts are either to run the power on multiple 2oz. Planes or to somehow attach some busbars to the board with press-fit or similar bolt fixings (e.g. Ernie power taps). Has anyone had any practical experience with this kind of problem? Hints/tips/tricks?

Cheers,

Dave
David
(United Kingdom)
 

Offline dmills

  • Super Contributor
  • ***
  • Posts: 2093
  • Country: gb
Re: High current on PCB backplane (~65A) - Experience?
« Reply #1 on: August 31, 2018, 09:26:00 pm »
A common problem in things like large SDI router frames, there are a couple of solutions I have seen in that space.

Your ERNI (Or Wurth, see to do very similar things) press fits with a copper bar bolted to them, works but can be a bit of a pain to assemble.

Solder in busbars, essentially a piece of tin plated copper with pcb pins on a spacing of your choice, also stiffens the board quite nicely which can be a win depending on exactly how ridiculous your stackup is anyway.

A water ject cut copper sheet laid over the high current trace and bolted thru the PCB at intervals, those smt style 'Pem' nuts are nice for this (But do use a disk spring under the screw), and 65A (Assuming you are doing point of load regulation) does not take all that much, a 2mm thick, 10mm wide copper bar will not even get warm.

In any of these remember to deal with the return current the same way! Who me?

I would also remind that fusing the individual card sockets on the backplane is a good idea, those 2mm sockets do sometimes take damage and it is better to blow a fuse on the backplane then take the whole thing down in smoke and fire as the connector burns out!

Regards, Dan.
 

Offline DavidTopic starter

  • Frequent Contributor
  • **
  • Posts: 280
  • Country: gb
Re: High current on PCB backplane (~65A) - Experience?
« Reply #2 on: August 31, 2018, 10:38:39 pm »
Hi Dan,

Thank you for your reply, some very good suggestions, I have been toying with a couple of those ideas. Not sure how quickly we could get a custom though-hole busbar made (or where?) so maybe bolting down strips is the way to go.

I have attached a diagram showing the basic idea. Each of the 16 white rectangles represent a DIN41612 connector. Each one of these has several signal lines and three "high current" nets; two supply inputs (only one used at any one time) and an output (not referenced to supply return, thus two wires each). The supply input current is only around 5.5A and the outputs only ~5A. The blue rectangle represents the location I want to route all the 16 outputs to. This is where external wiring will mate to the board. The two supply inputs (red boxes) will come onto the board roughly where shown via some threaded boses or similar. I am wondering if I can get away with a single power plane for each of the supplies and simply routing the outputs across the board? The data signals passing between connectors on the board are all referenced to the supply return.  :-//

Edit: The 65A figure is the input current with some diversity applied.
P.S. Backplane is approx 400x400mm
« Last Edit: August 31, 2018, 10:46:30 pm by David »
David
(United Kingdom)
 

Offline max_torque

  • Super Contributor
  • ***
  • Posts: 1320
  • Country: gb
    • bitdynamics
Re: High current on PCB backplane (~65A) - Experience?
« Reply #3 on: September 01, 2018, 01:40:46 pm »
https://e-fab.com/products/pcb-stiffeners/

Something like that ^^ ought to be easy to get laser or water jet cut these days to suit your design
 

Offline DavidTopic starter

  • Frequent Contributor
  • **
  • Posts: 280
  • Country: gb
Re: High current on PCB backplane (~65A) - Experience?
« Reply #4 on: September 01, 2018, 03:00:59 pm »
That could be an option, getting it tin plated would be the key to that I think.

I've been thinking more about just using planes...If I use an IPC trac current calculator (albeit they extrapolate above 35A I think?), then with a plane of 400x400mm, 2oz. internal and a modest temp rise it appears I could easily meet my current requirements? Hmmm  :scared:
David
(United Kingdom)
 

Offline dmills

  • Super Contributor
  • ***
  • Posts: 2093
  • Country: gb
Re: High current on PCB backplane (~65A) - Experience?
« Reply #5 on: September 01, 2018, 03:52:03 pm »
Watch the current density near the power input terminals!

One interesting possibility I have not played with is the "Wire in board" stuff that Wurth offers, basically a wire pressed into a milled slot in the board and wither buried or plated over, it might be worth exploring.

Pretty much any plating company should be able to plate tin onto copper.

Regards, Dan.
 

Offline DavidTopic starter

  • Frequent Contributor
  • **
  • Posts: 280
  • Country: gb
Re: High current on PCB backplane (~65A) - Experience?
« Reply #6 on: September 01, 2018, 07:54:22 pm »
That buried wire does look interesting, although probably a bit unique for this project.

Good point about current density, those Wurth Redcube terminals claim upwards of 100A capability...I wonder what assumptions they make around current density to claim those numbers?
David
(United Kingdom)
 

Offline dmills

  • Super Contributor
  • ***
  • Posts: 2093
  • Country: gb
Re: High current on PCB backplane (~65A) - Experience?
« Reply #7 on: September 01, 2018, 08:14:08 pm »
4 Oz copper or greater I would bet, possibly on several layers!

I tend to see those numbers with the same eyes I use when reading page 1 of a mosfet datasheet!

Regards, Dan.
 

Offline JohnG

  • Frequent Contributor
  • **
  • Posts: 580
  • Country: us
Re: High current on PCB backplane (~65A) - Experience?
« Reply #8 on: September 01, 2018, 09:07:11 pm »
4 Oz copper or greater I would bet, possibly on several layers!

I tend to see those numbers with the same eyes I use when reading page 1 of a mosfet datasheet!

Regards, Dan.

I second this. If cost is not the main constraint, you can do 1 to 2 oz on the outer layers and 3-4 oz for inner layers, where you don't have to mount parts. Another option is to use more 2 oz inner layers. 12-14 layers is not uncommon in brick-type power supplies.

John
"Reality is that which, when you quit believing in it, doesn't go away." Philip K. Dick (RIP).
 

Offline DavidTopic starter

  • Frequent Contributor
  • **
  • Posts: 280
  • Country: gb
Re: High current on PCB backplane (~65A) - Experience?
« Reply #9 on: September 01, 2018, 10:17:37 pm »
It looks as if I can get away with 4oz. copper and still be able to route my TSSOPs. I'm thinking of a stackup along the the lines of the attached image.

David
(United Kingdom)
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 27570
  • Country: nl
    • NCT Developments
Re: High current on PCB backplane (~65A) - Experience?
« Reply #10 on: September 02, 2018, 12:31:09 am »
IMHO you should be able to do fine with a 6 layer board using a single power plane for each 65A supply. Be sure to derate the connectors you use by at least 40% and use the higher current (use the D or E style DIN41612 connectors). Also think about whether the current of the load has to go through to 0V plane. Given that you seem to have dual supply voltage tells me you can have a common grounding point near the connectors which means almost no current is flowing through the ground plane of the backplane. Maybe a 4 layer board is doable as well. I think no matter what build-up you choose the biggest problem will be getting 65A into the board and you'll probably need multiple connections for that.
« Last Edit: September 02, 2018, 12:40:43 am by nctnico »
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22182
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: High current on PCB backplane (~65A) - Experience?
« Reply #11 on: September 02, 2018, 12:53:03 am »
Tons of width, I don't see what's wrong with a conventional 4 layer board for this (assuming you can route all your signals that way -- 6 layers or more may be desirable, in which case power is a slam dunk).

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline max_torque

  • Super Contributor
  • ***
  • Posts: 1320
  • Country: gb
    • bitdynamics
Re: High current on PCB backplane (~65A) - Experience?
« Reply #12 on: September 02, 2018, 06:27:15 pm »
You'd also need to look at the ambient temp requirement, and the convective cooling.  I did a pcb recently for a project where the tracks on the pcb were the "fuse" and i was fairly surprised at how little copper is actually needed to carry big amps when there is good convective cooling (ie a high surface area and a vertically oriented pcb).

And also, do you need 65amps continuous? Or is this a transient load etc
 

Offline Smokey

  • Super Contributor
  • ***
  • Posts: 2760
  • Country: us
  • Not An Expert
Re: High current on PCB backplane (~65A) - Experience?
« Reply #13 on: September 03, 2018, 03:23:32 am »
Basic questions:
How many high current power rails?  What's the current on each?  What voltage?
What's the total signal count on the backplane connector for each card (is it like 4 power + 200 low voltage IO or something?)
What form factor (2U, 6U, etc?)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf