GRIDSEED G-BLADE Overclocking 7Mh/s, improvements and repair

[J4bberwock]

Lost them while busy trying to find a short on the board, 3 Ferrites were blown and the FET is shorted !!! Waiting for replacement FET's now.

Thanks J4bberwock

anything else than the ferrites and fet blown?

it can be helpful for some other people.

[Blisk]

Lost them while busy trying to find a short on the board, 3 Ferrites were blown and the FET is shorted !!! Waiting for replacement FET's now.

Thanks J4bberwock

anything else than the ferrites and fet blown?

it can be helpful for some other people.

What he did to blow them up?

[BAUAR]

Hello all,

One of my gridseed blade failes.

I fitted the following a 35v 220uf capacitor, 43k ohm resistor. I put three fans on all sides. Blade very shining.
Power cables began to melt

Can anyone provide me with the schematics so i can repair it?

Thanks!

[J4bberwock]

Hello all,

One of my gridseed blade failes.

I fitted the following a 35v 220uf capacitor, 43k ohm resistor. I put three fans on all sides. Blade very shining.
Power cables began to melt

Can anyone provide me with the schematics so i can repair it?

Thanks!

Hi,
there are no real schematics.
I'm using the 8 chip ones from gridseed github.
https://github.com/gridseed/gc3355-doc

43k resistor is only if you are crazy like me and if you are confident you can fix anything that will break or lose your money
39k is a safe one.

only one extra fan is needed to blow from one side of the powerboard.
plus dissipators under the PCB where the mosfets are, and on the mosfets, you can even add some to the inductance/coil/choke

Your cable started to melt because the wires inside are too thin.
You are asking for 150 watt / 12 amps on each PCB.
I'm using 0.75mm² wires with a screw terminal instead of the original 12v plug connector on the blade.


solder it under the PCB, not on the same side there was the 12v plug. It will be easier to screw/unscrew.
The solder point near the ferrites is the positive
The solder point nearest from the edge of the PCB is negative.

[ktbken]

Hi
I have 3 mildly over clocked blades which have been running fine until today when one of them has blown on both blades the ferrite bead marked (what is the "a" for) afb26 next to the usb connector.
Needles to say they don't work any more  does anyone have any idea of  A/ what the spec is for that component is and B/ The hard one any idea why this may have happened.

The other 2 are hashing fine.

Ken

[J4bberwock]

Hi
I have 3 mildly over clocked blades which have been running fine until today when one of them has blown on both blades the ferrite bead marked (what is the "a" for) afb26 next to the usb connector.
Needles to say they don't work any more  does anyone have any idea of  A/ what the spec is for that component is and B/ The hard one any idea why this may have happened.

The other 2 are hashing fine.

Ken

overheating can cause many components to fail.
the other reason is a short somewhere on the board.
maybe the upper gate mosfet could have died.
check for continuity between pin 4 (gate) and the 4 pins soldered to the same pad (drain).
It's been explained in the 5chips pod thread
https://bitcointalk.org/index.php?topic=519112.msg6722468#msg6722468

on the later blades, gridseed replaced the ferrite beads with zero resistors of the same 1206 size

[Blisk]

OK now I am soure that power plug was problem that my blades stops to work.
I changed them and now works ok

[ktbken]

Hi
I have 3 mildly over clocked blades which have been running fine until today when one of them has blown on both blades the ferrite bead marked (what is the "a" for) afb26 next to the usb connector.
Needles to say they don't work any more  does anyone have any idea of  A/ what the spec is for that component is and B/ The hard one any idea why this may have happened.

The other 2 are hashing fine.

Ken

overheating can cause many components to fail.
the other reason is a short somewhere on the board.
maybe the upper gate mosfet could have died.
check for continuity between pin 4 (gate) and the 4 pins soldered to the same pad (drain).
It's been explained in the 5chips pod thread
https://bitcointalk.org/index.php?topic=519112.msg6722468#msg6722468

on the later blades, gridseed replaced the ferrite beads with zero resistors of the same 1206 size

Well one is back working using one of the fb off of a bad 5 chip and all seems to be fine with it.
Next to bypass the plugs and solder direct to pcb for power  and moooreeee heat sinks )

[Blisk]

post some pictures

[GenTarkin]

update here, I attempted the wire mod from the VRM SMD yellow cap to both the middle(on one blade) and furthest(the rest of my blades) of the 3 ASIC caps(same type of cap) on both +/-

Results arent as great as Id hoped.
The middle one , while working great for the first 2 SMD caps in the ASIC area .. didnt really help the last cap much.
So the last 3 blades, I connected the wires to the last cap, this worked better for all 3.

1.3v for VRM cap - doesnt change of course.
Results:
stock: 1.24-1.235v(first to last) for middle caps
wiremod: 1.255-1.255v(consistent for all 3 now) for middle caps.

So, it helped a bit, HW error dropped a bit at same speed.
What I dont get is why the voltage didnt go up closer to the VRM cap of 1.3v =(

Solder joints look good.
What I used: 20ga copper wire, solid core.

[racebyu]

update here, I attempted the wire mod from the VRM SMD yellow cap to both the middle(on one blade) and furthest(the rest of my blades) of the 3 ASIC caps(same type of cap) on both +/-

Results arent as great as Id hoped.
The middle one , while working great for the first 2 SMD caps in the ASIC area .. didnt really help the last cap much.
So the last 3 blades, I connected the wires to the last cap, this worked better for all 3.

1.3v for VRM cap - doesnt change of course.
Results:
stock: 1.24-1.235v(first to last) for middle caps
wiremod: 1.255-1.255v(consistent for all 3 now) for middle caps.

So, it helped a bit, HW error dropped a bit at same speed.
What I dont get is why the voltage didnt go up closer to the VRM cap of 1.3v =(

Solder joints look good.
What I used: 20ga copper wire, solid core.

Maybe its your wire resistance?
You could go to a thicker gauge wire.

[racebyu]

I have heat sinks glued to the back of my board under the regulator section.
Also noted for cooling if you do not us a additional fan for cooling this area
you can extend the Blade shroud out with cardboard on all side to help keep
the airflow directly over your bottom mounted heat sinks.

I will be adding a higher voltage capacitor, removing FB's and also adding heat
sinks to the 3 main FET's on the top side of the board. How did everyone
secure the heat sinks to the power section, sticky thermal tape or thermal epoxy
on the 3 tops?

I will add a 10 turn 50k pot and adjust to around 39k and see what hash rate
is possible with freq range from 900hz to 1khz, will post up picks soon.

[xcooling]

@GenTarkin please post pics of the wiremod

@racebyu very thin layer of super glue is how i attached all my heatsinks.

[natbyte]

Looking to use CSD16321 as a replacement with dual IR5300

http://uk.mouser.com/Search/Refine.aspx?Ntt=CSD16321&bdprt=1&cartid=50231000

Mouser lists 2 components for this part, I assume the FET I need is CSD16321Q5C can you confirm J4bberwock please, thanks.

[J4bberwock]

Looking to use CSD16321 as a replacement with dual IR5300

http://uk.mouser.com/Search/Refine.aspx?Ntt=CSD16321&bdprt=1&cartid=50231000

Mouser lists 2 components for this part, I assume the FET I need is CSD16321Q5C can you confirm J4bberwock please, thanks.

they are the same, but the Q5C is supposed to have better cooling, so maybe better for us.
they are the ones I have too.

CSD16321 is for upper gate, instead of 5301/5302

my irfb7430 arrived at lunch time, so I'll try it this afternoon.

[GenTarkin]

@GenTarkin please post pics of the wiremod

@racebyu very thin layer of super glue is how i attached all my heatsinks.

Ill try to do that =) , maybe Ill put on some bigger wire... but 18AWG is gonna be harder to solder on these small contacts, specially since all I have my hands on is stranded 18AWG.

[ZiG]

@GenTarkin please post pics of the wiremod

@racebyu very thin layer of super glue is how i attached all my heatsinks.

Ill try to do that =) , maybe Ill put on some bigger wire... but 18AWG is gonna be harder to solder on these small contacts, specially since all I have my hands on is stranded 18AWG.

Stranded is BETTER, General...

[wolfey2014]

@GenTarkin please post pics of the wiremod

@racebyu very thin layer of super glue is how i attached all my heatsinks.

Ill try to do that =) , maybe Ill put on some bigger wire... but 18AWG is gonna be harder to solder on these small contacts, specially since all I have my hands on is stranded 18AWG.

What gauge is your current wire? 18AWG may be overkill anyway.

Wolfey2014

[kingscrown]

dcent thread. im in EU but buy from gaw - check my review of the blade - http://fuk.io/gridseed-g-blade-vs-gawminers-cloud-hosting-comparision-and-review/
cloud hosting from gaw is decent

[GenTarkin]

@GenTarkin please post pics of the wiremod

@racebyu very thin layer of super glue is how i attached all my heatsinks.

Ill try to do that =) , maybe Ill put on some bigger wire... but 18AWG is gonna be harder to solder on these small contacts, specially since all I have my hands on is stranded 18AWG.

Stranded is BETTER, General...

Ha it is? psh well.. I just tried 18 AWG stranded copper wire, was using 20 AWG solid before.
Voltage went from 1.255 to 1.258 =P ROFL!

I guess, at this point, would have to run many upon many wires ... not worth it =(

Here be picture:
http://imgur.com/JXIPbZ1