-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
How Do I Responsibly Report a Powerwall 2 Hardware Safety Defect?
- Thread starter mudder
- Start date
-
- Tags
- hazard powerwall 2
They already have Kapton in a few odd places, so yeah I suppose that would work.
Cheaper to replace the aluminum bracket. All they need to do is change this mounting tab from 45 degrees to 90 degrees, just like the other tabs. My guess is the EE who designed the PCB specified all tabs at 90 degrees, and then at some point the ME team changed that one tab to 45 degree to save a few pennies in the folding process.
charlesj
Active Member
I meant to respond to this but forgot, sorry.
I unfortunately do not have access to UL 1741, so I cannot really respond to your questions. I work on products that typically fall under other standards like UL 61010.
By your 0.38mm estimate, this is not a standard 1.6mm/0.64" thick PCB?
0.38mm = .015 inch, so only 100V/mil is needed for 1,500V main transient ignoring the other elements:
Soldermask can be >500V per mil
Conformal coating >1000V/mil
Also, if I'm viewing this correctly, the bracket is attached via a press nut so it is airgapped from the PCB.
1.6mm is a typical PCB thickness, but PCBs often have more than 2 layers. Distance between internal layers are less. I'm less familiar with UL 1741, but other standards I am more familiar with are OK with fairly high voltages between layers on PCB layers because there is solid insulating material (prepreg) between them.
The fact that there is an airgap is not necessarily OK, it depends on the distance as there are minimum clearance numbers you must maintain from line to ground. There are also creepage numbers (across a surface like the top of bottom of a PCB) that you must maintain. Typically creepage numbers are much larger than clearance numbers.
Right, but given the description, it seems to be two layer. Bracket is adjacent to 1 and they show the trace on the opposite side and call it layer 2. If the inner trace was biased to the attachment side, it seems like that would be shown instead.
Like your experience, the interlayer is insulation anyway and 0.38mm seems sufficient.
Yah, I understand creepage and clearance. With sufficiently insulated conductors, there is no creepage nor clearance requirement.
That the bracket does not appear to mount flush to the board seemed relevant when addressing item 2: "Therefore, there is no conformal coating where the aluminum tab intersects the board edge."
Is this assuming a delamination of the PCB then jumping the air gap?
Yes, thinking about this more I'm less certain it's a problem. OP wrote:
How does OP know that the 0.38mm is primarily FR4 and not prepreg? How many mils of prepreg would be required for adequate spacing for observation #1?
I think I don't entirely understand the geometry in question for observation #2. My understanding is that there mains trace comes to the board edge and there is inadequate air clearance from the side of the PCB to the aluminum bracket. Or else the mounting screw / insert on the board has inadequate surface creepage to the mains trace on the bottom.
Prepreg can also be FR-4 material. FR4 type is >20kV/mm so even 0.38mm is sufficient for mains and transients. I think transient is 1,500 working is only 160V peak for US split phase.
It comes near the edge. To conduct via creepage or clearance, there would need to be voids in the layer. Then it's a matter of the stand off distance and path length to the bracket.
Although, it could be that the bracket attaches to the visible side, in which case the access cutouts are there for push nut insertion, not assembly.
If the board is indeed just two layers, then there would be no prepreg, just FR4, Cu, and soldermask layers, and as you say as much as 1.6mm between the traces. But then I don't understand why OP says 0.38mm if the issue is between traces on the top and bottom. Maybe a through via?
I guess we'll have to wait for him to respond. Without an actual board and bracket, or the CAD files, it's a bit hard to clearly grasp.
As an end user with two PW2’s in garage attached to interior wall, what is recommend heat/smoke alarm and install method where ceiling height is a factor; can heat sensor go on the wall just above PW or would it be preferable to build a small shelf from wall so heat sensor is directly above PW casing?
And would 12 inches be appropriate height above Powerwall or something like 24 inches?
And would 12 inches be appropriate height above Powerwall or something like 24 inches?
You should follow the installation guidance on the particular products that you are installing. They are generally pretty specific on where and how the sensors should be installed. Heat(p/smoke/flames all rise...
Yes, I’m aware of heat rising so ceiling would be logical location but the height and size of the room suggests something lower would be more sensitive, although ceiling likely captures and pools the heat so likely best option despite the distance from the source of the heat.
stopcrazypp
Well-Known Member
Read the manual of your specific device. Not all devices are designed to allow being wall mounted and still maintain effectiveness.
For example, if it's a IR/thermal cam based heat sensor, if you wall mount it close to the heat source, it may be out of the FOV of the sensor, so actually that will delay detection (the heat has to reach other areas of the garage first).
Last edited:
No, a new bracket is required. Without that tab (e.g. if you remove it with a Dremel), the 16 pound PCB assembly could flex during transit, which is never good for electronics.
Total PCB thickness is 2.7 mm. IMO this isn't thick enough considering the PCBA weighs 16 pounds.
PCB has QTY4 total layers, with QTY2 layers glued to each side of a ~1.9 mm core.
You are not viewing this correctly. The entire aluminum bracket profile is coplanar to the PCB top surface.
Specifically, the stackup is: PCB||bracket||PEM nut.
Clarification: FR4 is technically a type of prepreg, but I understand your general question. Based on my most generous reading of the code, the absolute bare minimum distance between line and chassis is 0.8 mm, no matter what material is placed between said distance. However, a more strict interpretation (and probably the correct one) is that the minimum distance is not less than 3 mm.
Correct: L2 mains trace is too close to the aluminum bracket, both directly through the prepreg+solder mask+conformal coating (0.38 mm), and the exposed milled edge (trace distance to edge+0.38 mm).
6000 V as used in this application.
Note my previous code interpretation that 0.8 mm spacing is the absolute minimum allowed by the most generous interpretation of the code, regardless of the material placed between said spacing. A more strict intepretation is not less than 3 mm in Category IV mains applications (which grid-connected inverters must adhere to, even though they're technically Category III).
There is no standoff distance; the bracket is exactly coplanar with the PCB top surface. The PEM nut does not contact the PCB; it's on the opposite side of the bracket.
The plane of the PCB is vertical?
Ok, so the photo is showing the L2 vopper two layers down. Gotcha.
Why 6kV?
What code requirement is 0.8mm based on?
Similar threads
