GridSeed 5-chip USB miner voltage mod

[jamieb81]

YoYo

Everything working fine.

Was just that the solder was not making contact properly.

And as I was launched also removed the SMD's and soldered the 47K's in place, took me a while with my cheapo solder iron but it's done and

the GS's are hashing happily right now at 1000mhz will go gradually until 1200 to see what's the sweet spot.

[Zor]

Thanks Wolfey and everyone else whom have contributed to this thread.

I bought a Weller 35watt iron and a pencil tip for it.

Picked my worst GS unit and now the little bugger has been running at 1200 for the past couple errors w/o a HW error (yet).

I did notice that the heatsink wasn't tightened down properly thus one of the GS chips wasn't making full contact with the thermal pad.  That might have something to do with the HW errors as has been mentioned before.

I'm going to begin modding my other GS units starting tomorrow ... not today.

It was a nerve-wracking experience for me.  But I have one under my belt now.  I at least know the general idea now.

[vipgelsi]

Worked fine for me.

[jamieb81]

Thanks Wolfey and everyone else whom have contributed to this thread.

I bought a Weller 35watt iron and a pencil tip for it.

Picked my worst GS unit and now the little bugger has been running at 1200 for the past couple errors w/o a HW error (yet).

I did notice that the heatsink wasn't tightened down properly thus one of the GS chips wasn't making full contact with the thermal pad.  That might have something to do with the HW errors as has been mentioned before.

I'm going to begin modding my other GS units starting tomorrow ... not today.

It was a nerve-wracking experience for me.  But I have one under my belt now.  I at least know the general idea now.

I feel you

my first one was also quite nerve wrecking !

[wolfey2014]

YoYo

Everything working fine.

Was just that the solder was not making contact properly.

And as I was launched also removed the SMD's and soldered the 47K's in place, took me a while with my cheapo solder iron but it's done and

the GS's are hashing happily right now at 1000mhz will go gradually until 1200 to see what's the sweet spot.

great on getting the 5V mod working,, correctly
1200 is gonna be shaky with a 47k resistor.
You should use at least a 50k if not 52k...but at 52, it will start to suck too much juice and not yield maximum results, per my sweet spot test reports...
Lettuce know how you get along, lil doggie!

[jamieb81]

YoYo

Everything working fine.

Was just that the solder was not making contact properly.

And as I was launched also removed the SMD's and soldered the 47K's in place, took me a while with my cheapo solder iron but it's done and

the GS's are hashing happily right now at 1000mhz will go gradually until 1200 to see what's the sweet spot.

great on getting the 5V mod working,, correctly
1200 is gonna be shaky with a 47k resistor.
You should use at least a 50k if not 52k...but at 52, it will start to suck too much juice and not yield maximum results, per my sweet spot test reports...
Lettuce know how you get along, lil doggie!

So what's the best Mhz setting you experienced with the 47K ?

[wolfey2014]

YoYo

Everything working fine.

Was just that the solder was not making contact properly.

And as I was launched also removed the SMD's and soldered the 47K's in place, took me a while with my cheapo solder iron but it's done and

the GS's are hashing happily right now at 1000mhz will go gradually until 1200 to see what's the sweet spot.

great on getting the 5V mod working,, correctly
1200 is gonna be shaky with a 47k resistor.
You should use at least a 50k if not 52k...but at 52, it will start to suck too much juice and not yield maximum results, per my sweet spot test reports...
Lettuce know how you get along, lil doggie!

So what's the best Mhz setting you experienced with the 47K ?

its in the booK!

[Zor]

Thanks Wolfey and everyone else whom have contributed to this thread.

I bought a Weller 35watt iron and a pencil tip for it.

Picked my worst GS unit and now the little bugger has been running at 1200 for the past couple errors w/o a HW error (yet).

I did notice that the heatsink wasn't tightened down properly thus one of the GS chips wasn't making full contact with the thermal pad.  That might have something to do with the HW errors as has been mentioned before.

I'm going to begin modding my other GS units starting tomorrow ... not today.

It was a nerve-wracking experience for me.  But I have one under my belt now.  I at least know the general idea now.

I feel you

my first one was also quite nerve wrecking !

I was freaking out, man.

I couldn't get the original resistor off so I YouTube'd how to get the fuggers off ... lol

Just switched sides heating up the solder and pressed it away with the tip.

Then mounting the axial resistor was entertaining ... solder didn't want to stick to the outer-most pad worth a #%^! so I dropped a ball of solder on there and put the axial over top of it resting on it and melted the ball of solder and pressed the axial down into it.  That worked.

Then I got an eraser and rubbed off the contact of the R52 pad (noticed the clean contact transferred solder beautifully on my throw-away motherboard) and that wasn't much of a pain in the but, just seeing stuff without a magnifying glass kinda sucked donkey.  Fortunately I'm near-sighted ... never thought I'd be happy to be near-sighted -- ever.

So I had it all good to go and noticed the heatsink was gonna make contact with the axial so guess what ... got to re-do R52 because I had to cut the axial resistor wire more.  Didn't really want to be electrifying the heatsink, of all things.

It was a BIG learning experience.

[BCata]

Any significant chance of bricking devices with this mod? I sent 30 units to be moded and 6 of them came back "dead" That is 20%! Basically they appear to be hashing in cgminer but no shares are submitted at all. All 6 units were working fine at 850 Mhz before the mod. If anybody experienced something like this before, please let me know what you did to solve it.

[volder]

Any significant chance of bricking devices with this mod? I sent 30 units to be moded and 6 of them came back "dead" That is 20%! Basically they appear to be hashing in cgminer but no shares are submitted at all. All 6 units were working fine at 850 Mhz before the mod. If anybody experienced something like this before, please let me know what you did to solve it.

Yes, I had the same problem. Turns out i didn't have enough power to the units. When modded they suck more juice! Take 10 of them and move them to another power supply, I bet they start working then.

[sang]

Any significant chance of bricking devices with this mod? I sent 30 units to be moded and 6 of them came back "dead" That is 20%! Basically they appear to be hashing in cgminer but no shares are submitted at all. All 6 units were working fine at 850 Mhz before the mod. If anybody experienced something like this before, please let me know what you did to solve it.

Either power issue or someone sucked at soldering and borked the connections.

[BCata]

Any significant chance of bricking devices with this mod? I sent 30 units to be moded and 6 of them came back "dead" That is 20%! Basically they appear to be hashing in cgminer but no shares are submitted at all. All 6 units were working fine at 850 Mhz before the mod. If anybody experienced something like this before, please let me know what you did to solve it.

Yes, I had the same problem. Turns out i didn't have enough power to the units. When modded they suck more juice! Take 10 of them and move them to another power supply, I bet they start working then.

It is not a power issue. I am using an ATX PSU with 4 units per PCI-e line. Also tried them individually with a dedicated 6A 12V power supply. I am also thinking it must be either the soldering or maybe bad SMDs or something. One unit is showing short-circuit signs with "spark" noise and the ref light flashing really fast. My question is, even if it is bad soldering, are there any changes of getting them to work by redoing the mod? Or are they completely dead now?

One thing I did notice, the damaged units only eat about half watts as the good ones (8-9 watts at the wall while the moded ones are a bit over 15 watts)

[sang]

Any significant chance of bricking devices with this mod? I sent 30 units to be moded and 6 of them came back "dead" That is 20%! Basically they appear to be hashing in cgminer but no shares are submitted at all. All 6 units were working fine at 850 Mhz before the mod. If anybody experienced something like this before, please let me know what you did to solve it.

Yes, I had the same problem. Turns out i didn't have enough power to the units. When modded they suck more juice! Take 10 of them and move them to another power supply, I bet they start working then.

It is not a power issue. I am using an ATX PSU with 4 units per PCI-e line. Also tried them individually with a dedicated 6A 12V power supply. I am also thinking it must be either the soldering or maybe bad SMDs or something. One unit is showing short-circuit signs with "spark" noise and the ref light flashing really fast. My question is, even if it is bad soldering, are there any changes of getting them to work by redoing the mod? Or are they completely dead now?

One thing I did notice, the damaged units only eat about half watts as the good ones (8-9 watts at the wall while the moded ones are a bit over 15 watts)

Try running the "damaged" units at a much lower freq, like 600. Most likely either the solder joints are crap and can't carry the current or the person really messed up and just flat out made it a bridge by using too much solder.

In each of the few cases I have fixed a botched solder job (either by myself or someone else) on a gridseed, the unit was back to operating as expected after the fix. Really depends on the degree of failure but if the person just messed up soldering only that SMD then chances are the gridseed is just fine.

[jamieb81]

Thanks Wolfey and everyone else whom have contributed to this thread.

I bought a Weller 35watt iron and a pencil tip for it.

Picked my worst GS unit and now the little bugger has been running at 1200 for the past couple errors w/o a HW error (yet).

I did notice that the heatsink wasn't tightened down properly thus one of the GS chips wasn't making full contact with the thermal pad.  That might have something to do with the HW errors as has been mentioned before.

I'm going to begin modding my other GS units starting tomorrow ... not today.

It was a nerve-wracking experience for me.  But I have one under my belt now.  I at least know the general idea now.

I feel you

my first one was also quite nerve wrecking !

I was freaking out, man.

I couldn't get the original resistor off so I YouTube'd how to get the fuggers off ... lol

Just switched sides heating up the solder and pressed it away with the tip.

Then mounting the axial resistor was entertaining ... solder didn't want to stick to the outer-most pad worth a #%^! so I dropped a ball of solder on there and put the axial over top of it resting on it and melted the ball of solder and pressed the axial down into it.  That worked.

Then I got an eraser and rubbed off the contact of the R52 pad (noticed the clean contact transferred solder beautifully on my throw-away motherboard) and that wasn't much of a pain in the but, just seeing stuff without a magnifying glass kinda sucked donkey.  Fortunately I'm near-sighted ... never thought I'd be happy to be near-sighted -- ever.

So I had it all good to go and noticed the heatsink was gonna make contact with the axial so guess what ... got to re-do R52 because I had to cut the axial resistor wire more.  Didn't really want to be electrifying the heatsink, of all things.

It was a BIG learning experience.

I see you had exactly the same problems as me.

solder was very hard to stick on the outer pad, so I picked up little balls of solder from out of my sponge put it on the pad then heated it with
the iron to make it stick and then take the resistor and melt the tin again so that the resistor goes into the little solder ball.

I also had a hard time getting the smd's off somehow they just did not want to desolder but found a way.

[jamieb81]

A little Note about my experience using BFgminer and CPUminer.

BFGminer runs shows the grids but is somehow painfully slow, in the BFGminer console I get close to 980Kh/s that shows

but on the pool side only getting very low, like 450/500, also noticed that the pods red light were almost not blinking.


Switched over to CPUminer again and I'm getting roughly 500kh/s per Gridseed again


So BFG is a no go for me.

[CryptKeeper]

Anyone tried to mine BTC with a gridseed after doing the 47K mod?

Do you need more cooling or is it going up in flames?

[bon4ire]

A little Note about my experience using BFgminer and CPUminer.

BFGminer runs shows the grids but is somehow painfully slow, in the BFGminer console I get close to 980Kh/s that shows

but on the pool side only getting very low, like 450/500, also noticed that the pods red light were almost not blinking.


Switched over to CPUminer again and I'm getting roughly 500kh/s per Gridseed again


So BFG is a no go for me.

I did the same.  Was using cgminer and bfgminer. With my 14 seeds I gained 1000khs pool side by switching back to using cpuminer.  It opened a lot of command windows but when stacked, acts like a status indicator.  It's pretty cool.  I even wrote a script to handle all of them at the same time or set a range of seeds to mine particular coins with very little effort.

[smracer]

It is not a power issue. I am using an ATX PSU with 4 units per PCI-e line. Also tried them individually with a dedicated 6A 12V power supply. I am also thinking it must be either the soldering or maybe bad SMDs or something. One unit is showing short-circuit signs with "spark" noise and the ref light flashing really fast. My question is, even if it is bad soldering, are there any changes of getting them to work by redoing the mod? Or are they completely dead now?

One thing I did notice, the damaged units only eat about half watts as the good ones (8-9 watts at the wall while the moded ones are a bit over 15 watts)

Fast red blinking light means the solder is not touching both sides of the resistor.  I had one out of the 100 I modded flash red like that.  I put the hot air gun on the resistor for 5 sec and it fixed it.

[wolfey2014]

A little Note about my experience using BFgminer and CPUminer.

BFGminer runs shows the grids but is somehow painfully slow, in the BFGminer console I get close to 980Kh/s that shows

but on the pool side only getting very low, like 450/500, also noticed that the pods red light were almost not blinking.


Switched over to CPUminer again and I'm getting roughly 500kh/s per Gridseed again


So BFG is a no go for me.

I did the same.  Was using cgminer and bfgminer. With my 14 seeds I gained 1000khs pool side by switching back to using cpuminer.  It opened a lot of command windows but when stacked, acts like a status indicator.  It's pretty cool.  I even wrote a script to handle all of them at the same time or set a range of seeds to mine particular coins with very little effort.

I've always said, cpuminer RULES! 
Care to share your script with other Windows users?
I'm using Win 7 32 bit...
Oh, and make sure you turn OFF FIFO buffers for each comm port.
They just get in the way and eventually crash your mining.
I've had mine turned off for a couple weeks now and my miners have been mining away happily and stably 24/7, without interruption!  It has also allowed maximum hash rates at the pool!

[amix]

I couldn't get the original resistor off so I YouTube'd how to get the fuggers off ... lol

Just switched sides heating up the solder and pressed it away with the tip.

Little tip fro removing thoose SMD resistors. Apply solder to both points and drag it across to form a solder bridge on top of the resistor. This allows you to heat both sides together. then heat the solder and swipe it away wiht the tip. Youll lose the orignal resistor in the blob of solder though. Works quite nicely. Or on theese as there are so small use a wider tip to contacvt both edges together - this works best if you have a spare iron, then you dont need to wait for it to cool so you can change tips