The Porsche SPASM suspension comes with electronically adjustable dampers. From the factory, you get Normal and Sport modes with Sport mode being much stiffer, of course. But why settle for the two presets from Porsche when you have infinitely adjustable dampers? The DSC Sport V3 Controller enables full control of the stock suspension allowing the user to tune the suspension to best suit the roads or tracks the car is being used on.
After reading around the internet a bit, I had learned the housing for the PCBA was completely plastic. I also had seen some rare instances of the unit running very hot and melting the plastic backplate. So, because I can’t leave things alone, I used Autodesk Fusion to design my own backplate to be made from aluminum. It’s amazing the tools available for free these days.
I 3D printed the design out of plastic for a test fit before committing to metal.
The PCBA is offset roughly 3.5mm from the edge of its housing. I made the pad on my backplate 2.5mm to close the gap to ~1.0mm.
Why do you make a 3D print before making the real thing? In order to catch little mistakes like this large gap between my backplate wall and the connector.
Another lesson is to check the accuracy of the 3D print. There’s a tolerance on that too. In this case, for whatever reason, one side was a little stretched which is why the two left screw holes didn’t line up.
12 comments
Really loving this, thank you
Kheim, nice article! Very detailed in your work. I am always here to support your projects. Johnny @ Seal Methods Inc.
Can you share the dimensions of the plate or .STL file?
Any chance you could email the guys over at dsc ? It would seem you have fixed a product flaw. Are these for sale ? Guys on the 991 rennlist would do a group buy I bet
It’s not a hard design problem. From the time I broke out the screw driver and calipers to have the 3d model done was around 30-45min.
Doing another iteration, I would add some fins to the back surface. I guess I could have some more made. Without doing a thorough pricing analysis, I’d guesstimate $250 for the plate with gap pad.
@Khiem,
There is a lot of interest for this right now on a Rennlist thread. Though, several are also hesitant about your first price. Would you be open to posting on Rennlist about a possible group buy or atleast future goals? Or if you have no intention of making more of these, sharing the file? Or even licensing the file?
My one-off unit from Axis was $250; I did ask for a quote at qty 4 and that dropped the unit price down to $160 if I recall correctly. Seal Methods hooked me up with the thermal gap pad which I believe was Saint Gobain TC2006 silicone gap pad which would cost roughly $15 based on some quick online pricing I did. Packaging material, time to package and ship, my time doesn’t come cheap. The fitment isn’t as good as I’d like because the PCB housing is still plastic and it’s not very dimensionally good; it’s got some warp to it. The warp of the base plastic plate was quite bad. Screwing it to the PCB housing helps to straighten it out, but left on its own, it’s not flat at all.
I actually just took out the DSC temporarily because I have to take the car to the dealership to get the ECU reflashed for an updated emissions related thing. There is a bigger gap on one side than the other between my aluminum plate and the plastic housing which I’m not happy about; but again, the plastic housing is a bit warped. To make sure the gap pad actually touches both the PCB and the aluminum plate, I had to go thicker than nominal. Of course, thicker gap pad means higher resistance to conductivity. But the other situation would be zero contact between the gap pad and PCB which would be basically zero conductivity. So, it doesn’t really meet my design standard and why I’m not really angling to make more. But, you know, trying to make lemonade out of lemons.
Anyway, if people are interested, I’ll talk to Raggdoll Motorsport and see if they’d be interested in making this. Just understand it’s not a perfect assembly due to reusing the plastic housing over the PCB which has some warp to it.
My hobby background used to be custom cooling for overclocked gaming computers and GPUs. So, the sorts of things that will melt if they’re not cooled properly. First, I think your solution of putting the thermal pad on the BACK of the PCB is not perfect, but certainly clever and a good first start. What you’re doing is using the PCB, itself, as the primary medium to extract heat into the aluminum plate. For a prototype, i think this is a good first start, because again, it is the PCB boards that we’re all worried about. However, traditionally you want to cool the individual chiplets, MOSFETs, and voltage regulators instead, so that that heat does defaultly not go into the PCB in the first place. This is why if you take apart a motherboard on a computer, or a GPU, or anything PC related, you will find the heat sink is attached to the chiplets/processor/ram/etc via thermal pads or grease, and not the back of a PCB. In your defense, I understand why you’d first try to cool the PCB itself: It’s flat (ideally) or atleast semi flat, and makes the design attempt very simple and easy. So, I would suggest that to improve this design, the next iteration attempt to cool the chiplets directly. So you need to worry about the other side, too. The problem with this proposal is that we don’t know if all the chiplets sit at the same height. If they’ve even 1-2mm off from eachother, then you run into the risk of applying too much pressure to certain components, and zero contact with others, and hence uneven cooling. That can translate to undue physical and heat stress, too. Im afraid there is no easy way to tackle this but by experimenting with thicker thermal pads (e.g. 3mm+) and then looking at the depth of the imprint on thicker pads to see where the contact is actually happening. Previous posts on Rennlist have attempted to isolate which specific chiplet is getting cooked. Even focusing on that one area might be enough. When the time comes, I’d be happy to purchase a brand new board and then loan it out so that you have an unwarped/idealized board to play with. But i don’t want to push you toward a project you dont have time for. Thank you again for trying this.
I could not get my board to separate from the housing easily and I wasn’t going to risk damaging the $1400 I spent. The 4 pairs of fat pins in the middle of the board should be the power output to the four dampers. I’d expect the machining of a custom aluminum housing for the PCB would also be significantly more expensive than the flat backplate I made.
This post has good pictures of the board, all the components are at different heights:
https://rennlist.com/forums/991/1423509-dsc-v3-controller-failed-4.html#post19588975
I did a little reading and typical 12V to 5V converters can operate up to 80C-85C temps which I assume to be ambient. Voltage regulator chips can operate up to 125C component temp and fail over that. ABS plastic has a softening temp around 105C-115C. So it does seem the internal temp is getting much over 100C.
I’d say the ideal solution is a small stir fan in the housing and pumping in some fresh air. Especially in the 911 application where the module is buried, so it won’t get much heat transfer by natural convection. At least on my 718, I can conduct the heat to the aluminum mounting bracket which acts as a heat sink. The DSC is also mounted vertically and in some open space for some natural convection. While I had the DSC out, I did paint my aluminum plate flat black to help promote some radiation heat transfer.
Absent adding an aluminum plate with gap pad like I did, I’d drill holes in the housing holding the PCB on all sides to help get air to circulate through. One would have to be able to remove the PCB from the housing first though which was something I couldn’t do easily. Considering the DSC is pretty buried in the either platform, I don’t see much risk for dust contamination.
It stands – we are interested in the backplate.
The regulators and chips on PCB can be addressed with micro heatsinks, they can get down to 4mmx4mm for 25 cents each off amazon et al. Will be doing this myself.
Have anyone interested shoot me an email at s2kmanifold@gmail.com. I went ahead and updated the design with flanges for the 911 application and I added fins of 1.5mm depth to the backplate to try to get just that little bit more heat transfer. I’ll have to get requoted due to the extra machining operations; hopefully it’s not a large cost impact. If there are at least four people interested, I think we can hit that target $250 price, not including shipping. Figure the plate price will have some incremental increase from the previous quote of $160 at qty 4 due to the machining in of the fins. I do need to get an official quote for the gap pad. I will be on international work travel for a couple weeks this month, so maybe sometime in February we can get the ball rolling if there are at least four people interested and I have the actual quotes for the machined plates and the thermal gap pad. I figure $250 price is roughly 10% margin. I’ve already burned more in what my typical hourly work rate is compared to the profit I’ll make on selling a handful of these, but I do want to help out some folks that are interested. If only one or two people are interested, I can still get them made, just understand the price will be higher. Like I mentioned, my one-off unit was $250 not including the gap pad.
I would also note that while this will definitely help reduce the running temperature of the board, it is not a 100% guaranteed against board failure as there could be other quality issues like poor solder joints, etc.
Quick update. I got an updated quote for the new design with fins machined into the back surface. Qty 1/2/4 piece prices: $400/$300/$200. I’m waiting on the quote for the gap pad.