[ANN] Spondoolies-Tech - carrier grade, data center ready mining rigs

[GoRDoN-ChoU]

https://i.imgur.com/3NeiVUI.jpg
modules and controller is fine, will it still cover in warranty if I replace the sockets my self?

got the PSU and miner today,PSU is still working, I found that the cables is came out from the PSU, so now I'm confused, why original cables can't take this amp?!
my another SP20 with same company's PSU and same cables just working fine!
https://i.imgur.com/VuBflrK.jpg
one pin is broke!!
but I think it is not the main cause
https://i.imgur.com/2D0HsQf.jpg
https://i.imgur.com/D23ihSQ.jpg
shit....
https://i.imgur.com/HQKoHiu.jpg

[1714063211]

1714063211

[1714063211]

1714063211

[MrTeal]

one pin is broke!!
but I think it is not the main cause

I would say that's very likely the main cause. While I can't comment with certainty without knowing more details on the cables and the pin arrangement, if you have a wire break on the circular connector the unit is then going to be trying to pull all the current through two (or one) pin on the Molex end. That's a surefire recipe for a burned cable.

[GoRDoN-ChoU]

one pin is broke!!
but I think it is not the main cause

I would say that's very likely the main cause. While I can't comment with certainty without knowing more details on the cables and the pin arrangement, if you have a wire break on the circular connector the unit is then going to be trying to pull all the current through two (or one) pin on the Molex end. That's a surefire recipe for a burned cable.

but how to explain other three cables?

[promojo]

shit....

What PSU is this?  Manufacturer?

[GoRDoN-ChoU]

shit....
https://i.imgur.com/HQKoHiu.jpg

What PSU is this?  Manufacturer?

Super flower Leadex Platinum 1200W

[klondike_bar]

one pin is broke!!
but I think it is not the main cause

I would say that's very likely the main cause. While I can't comment with certainty without knowing more details on the cables and the pin arrangement, if you have a wire break on the circular connector the unit is then going to be trying to pull all the current through two (or one) pin on the Molex end. That's a surefire recipe for a burned cable.

but how to explain other three cables?

The 6pin PCIe cable connects to the PSU with 2x12V and 2xGND pins. If one pin was broken, it means that twice the current was going through the other singular pin of the same type. this wire clearly swelled up and began melting at the PSU connection.

w/ regards to the SP20 end:  its likely that the pcie cable may drive different pins with different pins at the PSU connector. Thus if one pin was broken, at the PCIe connector 1-2 of the pins may not have been active, thus forcing large amperage through only 1 or two pins, causing a fire and risking shorts inside the SP20 as well.

[MrTeal]

one pin is broke!!
but I think it is not the main cause

I would say that's very likely the main cause. While I can't comment with certainty without knowing more details on the cables and the pin arrangement, if you have a wire break on the circular connector the unit is then going to be trying to pull all the current through two (or one) pin on the Molex end. That's a surefire recipe for a burned cable.

but how to explain other three cables?

Cascading failure, at least on the one board. You can see that it's the grounds on that plug that are burnt, but the 12V lines look good. Assuming a common ground on the two PCIe plugs, if the one cables has a failure that overloads the ground and melts the plug, not only is the wire and probably the PCB now really hot, but you're now trying to pull the full current from two plugs through only 3 wires and contacts. The already hot wires overheat due to dissipating 4x as much power (I^2R), and the other cable goes poof. Explaining why the other board fried and the control board didn't fry (again assuming both boards share a common ground through the control board) is a little tougher, but it's not unlikely that the other wires bring hot enough to melt off all the insulation wouldn't cause the other wires (that are common at the PSU) to heat to the point that the connection becomes compromised.

Again, this is all just speculation. That's not necessarily what happened but I would say it's a likely scenario. I'm a bit confused though. The red connector that's still plugged in looks like one of the Super Flower ones, but what are the burned flat cables with the circular connectors and the black sleeved cables from?
Edit: I see that it's a SF Arctic Platinum, not a Leadex. That makes more sense.

[tvasbn]

Arctic Platinum isnt SF best, Leadex on the other hand are one of the best PSU(EVGA 1300 is base i think on SF Leadex Gold)

[Duce]

I have a Rosewell 1300 watt gold rated power supply http://www.rosewill.com/products/1799/ProductDetail_Overview.htm that uses the same connectors for the "Modular" design termination to the power supply. I can tell you that at stock settings for an SP20, they were getting very hot in an 25C ambient environment. I clocked the unit down and have already ordered a replacement power supply for it (and took a little "heat" for doing so). I use an EVGA 1300G2 on the other unit with no issues. I used the Rosewell with my Gridseeds before but due to the low current draw had no issues. I haven't looked up the spec on that particular connector but apparently due to the closed environment of the connector it is not conducive for heat dissipation. This just justifies my action about replacing the power supply when I saw the picture this morning (in case my wife asks).

[faetos]

My engineer put the SP20 through a tough quality test.  In the end he seemed happy.

http://i100.photobucket.com/albums/m25/notlist3d/SP20/IMG_3290_zpsef9937a0.jpg

I like your plan.

"But honey its the perfect present for junior, and all the family"

an SP20 with a bag of Beaver Nuggets in every stocking! Buc-ee's rules!

[GoRDoN-ChoU]

one pin is broke!!
but I think it is not the main cause

I would say that's very likely the main cause. While I can't comment with certainty without knowing more details on the cables and the pin arrangement, if you have a wire break on the circular connector the unit is then going to be trying to pull all the current through two (or one) pin on the Molex end. That's a surefire recipe for a burned cable.

but how to explain other three cables?

Cascading failure, at least on the one board. You can see that it's the grounds on that plug that are burnt, but the 12V lines look good. Assuming a common ground on the two PCIe plugs, if the one cables has a failure that overloads the ground and melts the plug, not only is the wire and probably the PCB now really hot, but you're now trying to pull the full current from two plugs through only 3 wires and contacts. The already hot wires overheat due to dissipating 4x as much power (I^2R), and the other cable goes poof. Explaining why the other board fried and the control board didn't fry (again assuming both boards share a common ground through the control board) is a little tougher, but it's not unlikely that the other wires bring hot enough to melt off all the insulation wouldn't cause the other wires (that are common at the PSU) to heat to the point that the connection becomes compromised.

Again, this is all just speculation. That's not necessarily what happened but I would say it's a likely scenario. I'm a bit confused though. The red connector that's still plugged in looks like one of the Super Flower ones, but what are the burned flat cables with the circular connectors and the black sleeved cables from?
Edit: I see that it's a SF Arctic Platinum, not a Leadex. That makes more sense.

Controller is still working, as you can see of first picture the miner is mining, if they share a common ground by controller, it should be burnd too, because the controller can't take this amp.

[wh00per]

Controller is still working, as you can see of first picture the miner is mining, if they share a common ground by controller, it should be burnd too, because the controller can't take this amp.

The weakest element in the chain fails first .. then sharing a ground does not mean that everything connected to that ground fails.

if you have 2 ground wires (paralell) and 2 live wires (paralell) if one ground or live wire is broken, the whole demand will try to squeeze through the only other one good wire available, exceeding its rating .. and burning out. You see, the current tries to flow to the load and then back.. and it does not have enough copper to go through ..

Now.. the controller has the same ground but there is no demand of high amperage to run it .. that is the cables GET burnt bringing power for the whole unit, while the controller is still running. Most of the power is routed to the hashing chips, and not through the controller board. That's expected.

Your car has the same ground .. it might get on fire due to bad cabling or contacts in the high amperage lines, however, that does not fry your radio in the dashboard

Merry Christmas!

[GoRDoN-ChoU]

Controller is still working, as you can see of first picture the miner is mining, if they share a common ground by controller, it should be burnd too, because the controller can't take this amp.

The weakest element in the chain fails first .. then sharing a ground does not mean that everything connected to that ground fails.

if you have 2 ground wires (paralell) and 2 live wires (paralell) if one ground or live wire is broken, the whole demand will try to squeeze through the only other one good wire available, exceeding its rating .. and burning out. You see, the current tries to flow to the load and then back.. and it does not have enough copper to go through ..

Now.. the controller has the same ground but there is no demand of high amperage to run it .. that is the cables GET burnt bringing power for the whole unit, while the controller is still running. Most of the power is routed to the hashing chips, and not through the controller board. That's expected.

Your car has the same ground .. it might get on fire due to bad cabling or contacts in the high amperage lines, however, that does not fry your radio in the dashboard

Merry Christmas!

My understand of MrTeal said:
one connector (ground) burned
another connector (ground) of the same module burned because 2times amp
so now current go through the ground line of the controller to the other module and they burned too

[MrTeal]

Controller is still working, as you can see of first picture the miner is mining, if they share a common ground by controller, it should be burnd too, because the controller can't take this amp.

The weakest element in the chain fails first .. then sharing a ground does not mean that everything connected to that ground fails.

if you have 2 ground wires (paralell) and 2 live wires (paralell) if one ground or live wire is broken, the whole demand will try to squeeze through the only other one good wire available, exceeding its rating .. and burning out. You see, the current tries to flow to the load and then back.. and it does not have enough copper to go through ..

Now.. the controller has the same ground but there is no demand of high amperage to run it .. that is the cables GET burnt bringing power for the whole unit, while the controller is still running. Most of the power is routed to the hashing chips, and not through the controller board. That's expected.

Your car has the same ground .. it might get on fire due to bad cabling or contacts in the high amperage lines, however, that does not fry your radio in the dashboard

Merry Christmas!

My understand of MrTeal said:
one connector (ground) burned
another connector (ground) of the same module burned because 2times amp
so now current go through the ground line of the controller to the other module and they burned too

Not exactly. The bolded is what I suggested (and I've seen happen before, on Minifit Jr and other connectors).

The next part (current flows through the controller, burns it up) is what you would expect on first inspection if you had good connection between the PSU and board on the 12V side, the board was still drawing full power, and the full current ground return was trying to flow back to the PSU through the controller. That assumes that the board keeps drawing full power though. There's a bunch of reasons why that assumption might not hold true though. For instance, the ground lift when the last cable to the board burned out could easily have tripped a protection on the VRMs and shut them down.

I was suggesting that maybe the cable to the second board was compromised by the over temperature on the first cable.

[luki19994]

Sp31 run 32 days , 26 Hours offline yesterday , then run again , no answer whats the Problem from minersource , but after they need 13 days !! to get it mining for me!(sp31)
I'm not really supriseded.
Hashing with 4,83

[notlist3d]

Sp31 run 32 days , 26 Hours offline yesterday , then run again , no answer whats the Problem from minersource , but after they need 13 days !! to get it mining for me!(sp31)
I'm not really supriseded.
Hashing with 4,83

I would contact minersource.  It is not the SP team hosting it.   I'm not sure on minersource but it is Christmas so they might be on skeleton crew.

[luki19994]

I know !!
Sp31 runs really good!
Minersource Hmm??
I make 1,9 btc last 33 days.
Ghas.io
With a sp31
Wrong Pool?

[Dans]

I ordered some SP35's last night at 4:30pm PST and received tracking numbers at 1:03am PST!   

Great service from the Spondoolies-Tech team, as usual!!

I'm a long time customer of Spondoolies-Tech and i would highly recommend them over any other manufacturer.   They stand by their product and their customers, its hard to find a company like this these days!     The quality of their product is top notch, I've run every model of miner they have produced and the quality cannot be beat -- and their very easy to use UI is excellent!

Keep up the good work Spondoolies-Tech team, i cannot wait for the next product release!

[thomas_s]

I know !!
Sp31 runs really good!
Minersource Hmm??
I make 1,9 btc last 33 days.
Ghas.io
With a sp31
Wrong Pool?

There was one SP31 that was put offline last night, however that customer has not sent us any emails asking why it went down. That machine was taken offline for non payment of colocation fees.

[Dabs]

I like big fat wires. But they tend to be more expensive. (In response to wires melting.)