Fuse Box PCB

[Wetmelon]

Has anyone tried mounting all of their fuses and relays to a circuit board? How thick did you go on the copper? Which company did you decide to go with? Was it reliable?

Edit: Where did you find terminal blocks capable of high amperage?

Thanks!

Wetmelon.

[murpia]

Originally posted by Wetmelon:
Has anyone tried mounting all of their fuses and relays to a circuit board? How thick did you go on the copper? Which company did you decide to go with? Was it reliable?

Edit: Where did you find terminal blocks capable of high amperage?

Thanks!

Wetmelon.

Like this?

w w w.bgsoflex.com/mspower/mspowercomplete.jpg

w w w.bgsoflex.com/mspower/mspower.html

Regards, Ian

[Kirby]

Originally posted by murpia:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Wetmelon:
Has anyone tried mounting all of their fuses and relays to a circuit board? How thick did you go on the copper? Which company did you decide to go with? Was it reliable?

Edit: Where did you find terminal blocks capable of high amperage?

Thanks!

Wetmelon.

Like this?

w w w.bgsoflex.com/mspower/mspowercomplete.jpg

w w w.bgsoflex.com/mspower/mspower.html

Regards, Ian </div></BLOCKQUOTE>

Frank from UQ on QUT's use of DB9 connectors in the 06 car:
"what the hell is with your connecors, a race car is not a pinball machine or a printer"

[Ockham]

Frank from UQ on QUT's use of DB9 connectors in the 06 car:
"what the hell is with your connecors, a race car is not a pinball machine or a printer"

QFT.

I never really thought about using a PCB for power distribution. It would save a lot of time when wiring the car's power distribution, but if anything ever went wrong, you'd have to replace the whole board. Of course, a PCB is probably more rugged than dozens of individual wires, and I wouldn't miss hand soldering every contact. Clever.

[murpia]

Originally posted by Ockham:
I never really thought about using a PCB for power distribution. It would save a lot of time when wiring the car's power distribution, but if anything ever went wrong, you'd have to replace the whole board.
If anything goes wrong that's what the fuses are for, to protect the loom (and PCB if you use one). Even then, I'd make the PCB the weak link compared to the loom. Replacing a burned-out PCB has to be 100x easier than sorting out a melted loom. Nasty job...

In case it wasn't obvious, that Megasquirt PCB has bases for plug-in fuses and plug-in relays. So fuse blows or relay failures don't need any soldering to fix.

I do wonder if it's worth the effort to do your own PCB though, unless you integrated it into the dash display or something like that where you needed to use a PCB anyway. There are plenty of integrated fuse holder / relay bases already on sale, and there's another thread around about those.

Regards, Ian

[Nicky]

We reuse connectors from our previous wiring harnesses and since this was done many times over re-splicing of wires became an issue. So in a bid to reuse the connectors with their entire length of wire we built a central unit where all power goes through. The board uses screw-in type solderable connectors so if sensors need to be added or removed the entire length of wire can be accomodated. While the connection from the board to the ECU was a fixed standard.

We used Goodsky relays(Sugar cube relays) with a rating of 20amps on each. These are directly solderable. We then used blade fuses on the same board.

The pcb we made had the thickest copper layer available here at the indian market i.e. 75um. While track widths were generally kept at 12-18mm at high loads (20amps0 and 7-12mm at medium loads(5amps).

Again, answering questions about a burnt relay, well we did our own test for the relays at 40amps and they survived an entire minute before burning off. We thought the board wouldn't handle it, but it did quite efficiently and there was hardly any rise in temperature on the PCB (so we're thinking of reducing the track width in the following design).

Again since we had relays for most of the loads,we had separate fuses which were at tops 10amps. So the only region which was any bit warm after a full 15minute run was the region where current entered the board since it was centralised, but we had bolts on the board which provided the ground to the chassis.

Our earlier design included a provision for a CPU fan which could be mounted on the board, but the need never arised.

We separated the starter relay at kept it off the board.

regards,
Nikhil

[Ockham]

Posted by murpia:
There are plenty of integrated fuse holder / relay bases already on sale, and there's another thread around about those.

I'm still watching that thread because I don't like stringing the relay and fuse center together with slightly janky snap-together relay sockets, and since I'm not a huge fan of our current fuse center. However, I've never seen a relay and fuse center with the right number of slots, and with easy-to-find contacts.

A PCB would also remove a lot of parts-hunting pain, and there are lots of companies doing small-batch PCB work. It shouldn't be too difficult to find sponsorship from one, and even if you can't, the boards usually cost ~$5-$20 apiece, with a setup charge ranging from ~$30-$90. Like tube notching, it's a serious labor saver, but you might need to plunk down some cash.

[Nicky]

We designed the board on Eagle (PCB layout software) but since our manufacture needed an autocad model of the pcb he had to redesign it which costed us extra.

Well, the board costed us INR 7000($ 108) including manufacturing, but subsequent boards would cost INR 300($ 6.52) for 2.

Regards,
Nikhil

[Wetmelon]

Originally posted by Nicky:
We designed the board on Eagle (PCB layout software) but since our manufacture needed an autocad model of the pcb he had to redesign it which costed us extra.

Well, the board costed us INR 7000($ 108) including manufacturing, but subsequent boards would cost INR 300($ 6.52) for 2.

Regards,
Nikhil

:O!

I have a completed design in Eagle, but the guy wants a minimum run of 5 boards, for $58 each, plus a tooling fee of $200 (for a total ~$485 Canadian...). By the way, we're not using Relays at all... Power MOSFETs, baby!!

And the idea is two identical, complete "distribution boxes". One of them breaks, you yank it out, stick a new one in, keep driving while you fix the first.

Oh, I should mention that because we're driving 40 Amps (the entire fuse box, meaning everything coming from the battery) and the starter (fused at 50A), we had to go with a 6oz/ft^2 copper board, not a 1oz (which is typical).

[PeterK]

My team is just about ready to place an order for our fuse box and dashboard PCBs (this is the first year we are doing them) and after talking with a few profs and the tech dept here on campus they have suggested http:// w w w . goldphoenixpcb. biz/

I haven't used them myself but our mechatronics prof has used them for the past 2 years and they've done a good job at a very reasonable price. They also offer up to 6oz Cu.

Key thing with any PCB company is to follow their suggestions for hole sizes to minimize tool changes which keeps costs down.

[Wetmelon]

Key thing with any PCB company is to follow their suggestions for hole sizes to minimize tool changes which keeps costs down.

Any idea how to do that in Eagle? Aren't most of the parts already using standard drill sizes?

[PeterK]

I can't speak about Eagle as I've never used it, however I would expect there is a way to play with hole sizes. If you contact your PCB supplier they will be able to tell you what sizes they use as standard (each company is a bit different, but they usually have about 10-20 diameters that you don't have to pay extra for). The most common hole sizes are almost always included but it can pay to check. I've been using UtilBoard (part of MultiSim) and it has been going pretty good so far. I would love to use Cadence OrCAD or Allegro but, sadly, I'm not made of money.

[murpia]

Originally posted by Wetmelon:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Key thing with any PCB company is to follow their suggestions for hole sizes to minimize tool changes which keeps costs down.

Any idea how to do that in Eagle? Aren't most of the parts already using standard drill sizes? </div></BLOCKQUOTE>
Board suppliers will either expect you to produce 'Gerber' files from EAGLE, or accept an EAGLE .brd file direct.

For the former you should be able to specify a drill diameter list (which should match your suppliers need). For the latter the supplier may or may not do that for you, or may charge extra to do it. Also, they may supply a .drc file so you can check your board against their design rules. I use a company that provides a .drc and the results have been good.

I have always found it best when using a pre-supplied component from a library to take a copy of it and modify it as required. I mainly need to do this for SMD pads to make them easier to hand solder. But you can alter drill sizes this way if you need to.

Regards, Ian

[Nicky]

i checked with our manufacturer, different types of gerber files can be generated from Eagle. You'll need to check with your manufacture for gerber files in the true sense vary with the type of machines used in making the positive film and drilling. Again drill sizes for IC pins and most standard packages are available. However for hole size different from these(excluding standard drill sizes), you need to provide the tool else use one from their arsenal. Hope this helped

@Wetmelon
We had MOSFETs in plan when we designed the first board, but got mixed up with power darlingtons and the idea right now is suspended. I do believe that you use atleast 10 times a MOSFET with 10 times the rated load. Gimme an example, say for a 10amp load, what rating do you use?

@PeterK
I tried finding on the website you mentioned the minimum order. I believe that OEM designs are done only in high volume. You don't wanna end up with a million boards http://fsae.com/groupee_common/emoticons/icon_wink.gif
Yours truly,
Nikhil

[PeterK]

Nikhil,

If you mouse over/click on Price Info|PCB Fabrication|PCB Fab Price 1-2 layers you will get to their prototype costing (you have to submit information for a quote for 4 or more layers). They offer a 1000cm^2 or 650cm^2 package. You might end up with a couple of designs, but that's probably a good thing anyway. http://fsae.com/groupee_common/emoticons/icon_wink.gif

[Wetmelon]

@Wetmelon
We had MOSFETs in plan when we designed the first board, but got mixed up with power darlingtons and the idea right now is suspended. I do believe that you use atleast 10 times a MOSFET with 10 times the rated load. Gimme an example, say for a 10amp load, what rating do you use?

Yours truly,
Nikhil

Well I'm using MOSFETs rated for peak 74A for everything. w w w.irf. com/product-info/datasheets/data/irf4905pbf .pdf

For a 10A load, I would say use at least a 15A rated MOSFET. The important thing really is the Drain Current @ Temperature, which has a derating curve that is usually supplied on the datasheet. For example, I need to push 40A for the board. I'm using a 74A MOSFET, but it can only handle 40A drain current AT or BELOW 150 degrees Centigrade, according to the derating graph, AT SATURATION.

The next most important thing to look at is the Rds-On (Which is the Resistance across the Drain-Source junction, at saturation.) This also has a derating curve based on drain current, so be a little bit careful here.

This Rds-On is going to affect how much power is wasted in the MOSFET, and how much heat is generated - P = R*I^2, so if you have 10A, it's 10*10*Rds-On (which for me is .02Ohm), giving me 2W. You'll see on the datasheet "Rjc, Rcs, and Rja", which each have values of degrees/Watt. Note that this is the wattage we just calculated, not the wattage powering the load. If when you multiply it out the total temperature rise (plus ambient) is greater than the max junction temperature, Tjc? (usually 150 or 175C), you'll need a heatsink and can ignore Rja for the most part.

Hmm. I didn't mean to write that much, and you probably know a good chunk of it already, but never hurts to have the information out there :P

We haven't got our boards built yet, so I'm pretty excited/anxious to see how it performs. Somewhat afraid that my Starter MOSFET is going to explode. I have used Yenka from Crocodile Clips as a simulation program to verify the design, but it doesn't have models for everything...

[Nicky]

The first board that we made was huge, i mean the size of an A4 sheet and had all screw-in terminals. Now if i want to cut that down to 15cm x 15 cm. Increase the current load on the board.

Do you provide separate cooling for the board or is ambient cooling enough? Heat sinks for the MOSFETs anyway.

Do you use 4 layered boards for such high currents, using the sandwiched layers for power/gnd?

What sort of connectors do you use? As Wetmelon said before, if one of these burns out, plug it out, plug a new one in. How do you get about that?

@Wetmelon: thanks fer the last post.

Ashwa Racing
R V College of Engineering

[PeterK]

Things to think about with multi-layers... the inside layers can't handle nearly as much current as the outside layers because they can't dissipate heat directly to air. It will likely be cheaper if you can keep it down to two layers. I did in our design. If your traces are designed properly there should be no need for active cooling.

Something you can also do is run two FETs in parallel to split the current (decreasing the power loss, and the heat dissipation, by a factor of four). This only works because the FETs have such a low Rds(on) and a fast switching time. Don't try this with IGBTs http://fsae.com/groupee_common/emoticons/icon_wink.gif

Wetmelon, I'm assuming you're using p-channel so you can do high side switching for the starter. Did you use n-channel for your other loads?

[murpia]

Originally posted by PeterK:
Wetmelon, I'm assuming you're using p-channel so you can do high side switching for the starter. Did you use n-channel for your other loads?
Reading the datasheets, commercial Power Distribution Units (PDUs) from MoTeC, Pi, GEMS etc. use high-side drivers (p-channel) exclusively. There must be good reasons why.

There are a whole bunch of 'smart' high-side drivers around these days with impressive specs that suggest they are very failure-proof. They handle current limits, short circuits, large inductive energies, over temp, etc.

Has anyone done any testing with any of these these?

Regards, Ian

[Wetmelon]

No, I used P-channels to drive all the loads. I use N-channels in some spots to pull the P gate down.

Just to be nice, http:// i55.photobucket. com/albums/g122/wetmelon/Circuits/Automatic%20Car%20Project/sch_2011PWR_002 .png

[Wetmelon]

There must be good reasons why.

Yes, because you want to drive your loads with them connected to ground so you don't have to come BACK to the board. If you drive them with N-Channels, you have to put them on the Drain side if you want them to work properly, meaning you have to bring the GND of the load back to connect to N-Channel. P-Channels allow for GND-Connected loads http://fsae.com/groupee_common/emoticons/icon_smile.gif

I used all P-Channels in my board, with N-Channels as a switch when I had voltage coming IN (I.e. the dash switch, Fan and Fuel Pump ECU signals). If my other post ever goes through, I linked my circuit just to be nice. If it doesn't I can PM you the link.

[Nicky]

@Wetmelon: From your circuit i could straight away infer that all current running to any given load on the car runs through Q1. Wouldn't it worry you to put all that current through just one FET? More than current, woudn't the entire working of the car finally depend on Q1? Doesn't having any redundancy say in the form of another FET working in parallel having the power to source the entire load improve the overall reliability???

Just a suggestion, the rule states that the kill switch can act through a relay and must cut the power to the fuel and ignition systems. And we made use of that to keep the ECU running even when the kill switch is pushed in. This helped us in booting and accessing the data-logging system without priming the fuel system. This prevents frequent booting up of the system. The ECU power control is directly linked to the master switch.

Ashwa Racing
Bangalore