• We will be performing upgrades on the forums and server over the weekend. The forums may be unavailable multiple times for up to an hour each. Thank you for your patience and understanding as we work to make the forums even better.

Bitcoin Miner Shelving bond or not to bond

sparkmatic

Member
Location
California
Occupation
Electrician/Instrumentation
Is it more of an issue to bond the metal shelving to the system. The equipment is cord and plug connected from the individual miner's to the PDU, the PDU s feed from 150A breaker. The miners are supplied with 240V all cords run from PDU's along shelving . I see a potential (small as it may be) for a rack or more to become energized. I would hope the equipment ground would clear a fault but wanted to get some input.

Again do not want to create a hazard by bonding if it will not be effective

Thanks
 

suemarkp

Senior Member
Location
Kent, WA
Occupation
Retired Engineer
I fought this enough with regular equipment racks. I thought most of the equipment would bond the rack if bolted to the rack rail through its rack ears. But we used non-marring rack mount screws which are effectively insulated. And many rack ears and rails are painted. I found some rack mounting bolts made by Panduit that have notches under the head to penetrate paint, and thread forming threads to cut any paint in the rack threads. When I measured it with a bonding meter, it wasn't always as good as what I wanted (depended on how many cord and plug items were mounted to the rack, and whether the rack rails were painted insulating them from the frames). I finally got to the point where I installed a ground bar to the rack rails with these panduit screws and put a jumper to the frame. Most of these rack frames had a bonding screw lug, so at least the basic structure got bonded if you ran a bonding conductor to that screw.

If these shelves are just metal shelves and not tested/listed for grounding/bonding, things may be even more challenging because of painted structure. Toothed lock washers can help to bond interconnect pieces of metal bolted together. But it still may not be great.

I'm not sure how much of a hazard cords running across a painted metal shelf are. On equipment racks, there are places where power enter the racks, and power cords that didn't get tided down could get pinched in doors. I remember moving a rack once not knowing it had power coming into it from under the floor, and the rack had no wheels. I saw sparks when I ran over the power cord. It didn't trip a breaker or energize the rack, but made me aware of some of the problems.
 

Fred B

Senior Member
Location
Upstate, NY
Occupation
Electrician
Couple of questions.
Is the PDU or other equipment on the shelfing unit provided a EGC?
Is the Equipment electrically isolated from the shelfing units? (i.e. rubber feet)

No requirement in code to specifically bond the shelfing unit but if the mfg asks for it to be bonded then it would then by default it would require bonding. If the equipment is isolated or not provided with EGC then it would seem bonding would be prudent to clear a potential fault onto the shelfing units.
 

suemarkp

Senior Member
Location
Kent, WA
Occupation
Retired Engineer
what's a bonding meter?
A meter that measures very low resistance. They typically have 4 probes to zero out the lead resistance. It has two parts (mine was gun like in one hand and a round probe on a cord for the other, and each part had 2 very sharp metal probes which could poke through paint or surface coatings. It measures in milliohms and can probably down down to 10 milliohms accurately. Typical UL limits for bonding metal cases is 100 milliohms.

This one was what I used most often: https://bcdelectronics.com/wp-content/uploads/2020/09/product_hero_M1Milliohm-1-1200x1200.png
 

sparkmatic

Member
Location
California
Occupation
Electrician/Instrumentation
If you're talking standard equipment racks, the shelves bolt to the rack.... don't know if the paint on the rack and shelf are an effective insulator.
Why would bonding create a hazard?
I doubt it would create a hazard in most cases but I always like to look at safety from more than one perspective. If it is not effectively bonded with a path back to the source a fault may not clear by tripping the OCD. Just covering all bases
 

sparkmatic

Member
Location
California
Occupation
Electrician/Instrumentation
Couple of questions.
Is the PDU or other equipment on the shelfing unit provided a EGC?
Is the Equipment electrically isolated from the shelfing units? (i.e. rubber feet)

No requirement in code to specifically bond the shelfing unit but if the mfg asks for it to be bonded then it would then by default it would require bonding. If the equipment is isolated or not provided with EGC then it would seem bonding would be prudent to clear a potential fault onto the shelfing units.
Yeah and the shelves are not likely to become energized but the miners are on 240V to ground not 120 so it is less than the 250V I found in the NEC, but again just getting input to make the best informed decision possible. And, the one manufacture does say to bond the racking others do not and we have a bit of both.

Thanks
 

sparkmatic

Member
Location
California
Occupation
Electrician/Instrumentation
The PDU's are feed vial a cable tray and when I check continuity from its grounding conductor to various bolts on the shelves it show they are fairly well bonded but that is only on a few test we have thousands.

And as for the equipment I did not thing of the feet I will check that but the cords are basically zipped tied to the shelving for cable management and the cords power the units via 240 V not 120 which is why management is looking into a best path forward for safety of workers.
Yeah and the shelves are not likely to become energized but the miners are on 240V to ground not 120 so it is less than the 250V I found in the NEC, but again just getting input to make the best informed decision possible. And, the one manufacture does say to bond the racking others do not and we have a bit of both.

Thanks
 
Top