Topic: GRIDSEED G-BLADE Overclocking 7Mh/s, improvements and repair - page 2. (Read 74150 times)
December 03, 2015, 10:23:46 AM
This machine down, saving how to modify the circuit
Hey there,
i have an Gridseed miner and today there was the big bang:
One of the blades gets a short Circuit on 12V...
When i plug in the 12V Power Supply is fast go on and off (short cut protetction or so).
The USB part is working fine, the red light goes on and off.
I haven't overclocked the blade or have changed any resistors. The only thing is I
used an other fan (12V 0.26A)... I read in this thread that when you overclock the blade the mosfets can burn out due overhat, so
could it be that the Mosfets gets too hot or something and burned out because i changed the fan?
On the Other blade side the fan and so on is working.
Optically tere is no burning mark or bad capacitors.
Are there any measure points to test the chips or something?
Please help me
i have an Gridseed miner and today there was the big bang:
One of the blades gets a short Circuit on 12V...
When i plug in the 12V Power Supply is fast go on and off (short cut protetction or so).
The USB part is working fine, the red light goes on and off.
I haven't overclocked the blade or have changed any resistors. The only thing is I
used an other fan (12V 0.26A)... I read in this thread that when you overclock the blade the mosfets can burn out due overhat, so
could it be that the Mosfets gets too hot or something and burned out because i changed the fan?
On the Other blade side the fan and so on is working.
Optically tere is no burning mark or bad capacitors.
Are there any measure points to test the chips or something?
Please help me
Actually, i just realized i can use thos tiny step-down converter modules, as a replacement for the filtering cap, since those can configured to not step down voltage at all, you still get 12 volts, and a good replacement of the stock filtering capaticor.
one issue though, is that i don't know about it's current capabilities so this is something i will eventually test.
so after some tests it appears that those step-down converters aren't good for this because one can only supply up to 3 amps, with a voltage drop (supplied 12 v 18a output 9v 3a) and with this output the converter got so hot it left a burn mark of my wooden desk. so i guess surfacemounting that cap is just not avoidable..... even though that step-down converter would still be a good addition to filtering.. one issue though, is that i don't know about it's current capabilities so this is something i will eventually test.
Actually, i just realized i can use thos tiny step-down converter modules, as a replacement for the filtering cap, since those can configured to not step down voltage at all, you still get 12 volts, and a good replacement of the stock filtering capaticor.
one issue though, is that i don't know about it's current capabilities so this is something i will eventually test.
in the mean time, here is a picture of the step down converter module:
https://i.imgur.com/z5FdYHT.jpg
one issue though, is that i don't know about it's current capabilities so this is something i will eventually test.
in the mean time, here is a picture of the step down converter module:
https://i.imgur.com/z5FdYHT.jpg
Ok here some answers from LSE again :-)
if you have a normal C2 solder iron tip you can do most of the work just fine :-)
Look at my shop on http://www.LSE.solar and just over the SILVER kit :
Gridseed Blade tuning kit (silver)
This tuning kit is for a Gridseed Blade 80 Chip system. It includes as follow :
(2) 2 Terminal
(2) 39Kohm resistor
(2) Capacitor Aluminum 220UF 35V 20% SMD
(1) 1mm drillbit
(4) 10*10*5mm Aluminium Heatsink with Thermal Pad
(2) 8*8*5mm Aluminium Heatsink with Thermal Pad
Those parts will get you to where you wana be :-) if you put your solder iron to work and heat up the SMD itself it will melt both solder ends and you can lift it easy of the Circuit board, same way you can apply the new SMD resistor :-). Sounds crappy but does work with those simple SMD resistors just fine.
The CAP is also easy to do. Just pull it twisting of the board and than apply with the Solder iron heat to the Pin's to remove those.
If you need some nice heat resistant solder I can put that into the envelop too :-) . Once the CAP is removed you put the new one in place and just apply to the end tips some solder while holding it down to the board. And done it that job.
Last remove the cheap Terminal...... Once removed you use the Terminal and hold it to the 12V+ which is the closest PIN to the 3 ferrite beads in black. in the middle of that you drill with the drillbit the first hole. than you mark up the 2nd. And drill there.
IMPORTANT : DO NOT drill the first one else where than right in the middle of the 12V slit. Otherwise you can shorten the circuits because you touch ground which is just 2mm away from the 12V+ ....... my 2 sence .. POOR DESIGN :-(
After that you just apply those 2 heat sinks on the MOSSFETS and done :-)
Also nice tip apply a strip of normal black electrotape to reduce the Airflow towards the back exit so it's forced down more to the PCB Board to cool it on the backside.
Does the error rate will be lower ... NOT IF YOU DO NOT USE THE CORRECT FREQUENCY :-)
Do only jump in 25khz jumps. So like 800,825,850,875,900..... and so on.
900 should be best option, maybe 925. Without with the MOSSFET you can do 1025 stable.
I hope it helps.
Lumanet
https://i.imgur.com/OqH5ORh.jpg
Hello everyone, i have been reading this thread from page 1 - 25 until my brains exploded, it's definetly awesome to read all of your progress but the stuff you do is way over my limit, i know basic electronics and i just got my G-Blade a few days ago.
So at stock 600 MHz it mines at a puny 3.2 MH/s (WTF?? why??)
When overcloking to 838 MHz one blade works at a fairly low amount of H/W errors but the second one gets quite a lot of them, so i underclocked it a little to 825 MHz to make it happy, but anyway back to topic, i wanted to get the most i possible could out of my blade by replacing that lame power jack with a screw terminal (or just fuckin' soldering wires directly to the blades) and i also want to add heatsinks to all of the overheating part, my question is, that if i do all of that, will that make my miner a little more stable with less hardware errors? and also, i might just do the simple version of J4bberwock's voltage mod (just the 39k resistor) but the problem is that i definetly don't have the right equiptment or the money for the right equipment to do SMD soldering, so do you think a thru-hole 39k 5% tolorance resistor will work?
Also, at some point i overread that @J4bberwock connects his miners to the 6 pin PCIe connectors of a high quality ATX PSU, and after a simple google search it seems like EVERYONE connects their miners to the PCIe connector, is there a specific reason why?
Sorry for bombarding with questions, i'm really anxious to know what i can do!!
(The amount of times i edited this post is too damn high!!!)
So at stock 600 MHz it mines at a puny 3.2 MH/s (WTF?? why??)
When overcloking to 838 MHz one blade works at a fairly low amount of H/W errors but the second one gets quite a lot of them, so i underclocked it a little to 825 MHz to make it happy, but anyway back to topic, i wanted to get the most i possible could out of my blade by replacing that lame power jack with a screw terminal (or just fuckin' soldering wires directly to the blades) and i also want to add heatsinks to all of the overheating part, my question is, that if i do all of that, will that make my miner a little more stable with less hardware errors? and also, i might just do the simple version of J4bberwock's voltage mod (just the 39k resistor) but the problem is that i definetly don't have the right equiptment or the money for the right equipment to do SMD soldering, so do you think a thru-hole 39k 5% tolorance resistor will work?
Also, at some point i overread that @J4bberwock connects his miners to the 6 pin PCIe connectors of a high quality ATX PSU, and after a simple google search it seems like EVERYONE connects their miners to the PCIe connector, is there a specific reason why?
Sorry for bombarding with questions, i'm really anxious to know what i can do!!
(The amount of times i edited this post is too damn high!!!)
Ok here some answers from LSE again :-)
if you have a normal C2 solder iron tip you can do most of the work just fine :-)
Look at my shop on http://www.LSE.solar and just over the SILVER kit :
Gridseed Blade tuning kit (silver)
This tuning kit is for a Gridseed Blade 80 Chip system. It includes as follow :
(2) 2 Terminal
(2) 39Kohm resistor
(2) Capacitor Aluminum 220UF 35V 20% SMD
(1) 1mm drillbit
(4) 10*10*5mm Aluminium Heatsink with Thermal Pad
(2) 8*8*5mm Aluminium Heatsink with Thermal Pad
Those parts will get you to where you wana be :-) if you put your solder iron to work and heat up the SMD itself it will melt both solder ends and you can lift it easy of the Circuit board, same way you can apply the new SMD resistor :-). Sounds crappy but does work with those simple SMD resistors just fine.
The CAP is also easy to do. Just pull it twisting of the board and than apply with the Solder iron heat to the Pin's to remove those.
If you need some nice heat resistant solder I can put that into the envelop too :-) . Once the CAP is removed you put the new one in place and just apply to the end tips some solder while holding it down to the board. And done it that job.
Last remove the cheap Terminal...... Once removed you use the Terminal and hold it to the 12V+ which is the closest PIN to the 3 ferrite beads in black. in the middle of that you drill with the drillbit the first hole. than you mark up the 2nd. And drill there.
IMPORTANT : DO NOT drill the first one else where than right in the middle of the 12V slit. Otherwise you can shorten the circuits because you touch ground which is just 2mm away from the 12V+ ....... my 2 sence .. POOR DESIGN :-(
After that you just apply those 2 heat sinks on the MOSSFETS and done :-)
Also nice tip apply a strip of normal black electrotape to reduce the Airflow towards the back exit so it's forced down more to the PCB Board to cool it on the backside.
Does the error rate will be lower ... NOT IF YOU DO NOT USE THE CORRECT FREQUENCY :-)
Do only jump in 25khz jumps. So like 800,825,850,875,900..... and so on.
900 should be best option, maybe 925. Without with the MOSSFET you can do 1025 stable.
I hope it helps.
Lumanet
Hello everyone, i have been reading this thread from page 1 - 25 until my brains exploded, it's definetly awesome to read all of your progress but the stuff you do is way over my limit, i know basic electronics and i just got my G-Blade a few days ago.
So at stock 600 MHz it mines at a puny 3.2 MH/s (WTF?? why??)
When overcloking to 838 MHz one blade works at a fairly low amount of H/W errors but the second one gets quite a lot of them, so i underclocked it a little to 825 MHz to make it happy, but anyway back to topic, i wanted to get the most i possible could out of my blade by replacing that lame power jack with a screw terminal (or just fuckin' soldering wires directly to the blades) and i also want to add heatsinks to all of the overheating part, my question is, that if i do all of that, will that make my miner a little more stable with less hardware errors? and also, i might just do the simple version of J4bberwock's voltage mod (just the 39k resistor) but the problem is that i definetly don't have the right equiptment or the money for the right equipment to do SMD soldering, so do you think a thru-hole 39k 5% tolorance resistor will work?
Also, at some point i overread that @J4bberwock connects his miners to the 6 pin PCIe connectors of a high quality ATX PSU, and after a simple google search it seems like EVERYONE connects their miners to the PCIe connector, is there a specific reason why?
Sorry for bombarding with questions, i'm really anxious to know what i can do!!
(The amount of times i edited this post is too damn high!!!)
anyway, can we get back to my original questions? So at stock 600 MHz it mines at a puny 3.2 MH/s (WTF?? why??)
When overcloking to 838 MHz one blade works at a fairly low amount of H/W errors but the second one gets quite a lot of them, so i underclocked it a little to 825 MHz to make it happy, but anyway back to topic, i wanted to get the most i possible could out of my blade by replacing that lame power jack with a screw terminal (or just fuckin' soldering wires directly to the blades) and i also want to add heatsinks to all of the overheating part, my question is, that if i do all of that, will that make my miner a little more stable with less hardware errors? and also, i might just do the simple version of J4bberwock's voltage mod (just the 39k resistor) but the problem is that i definetly don't have the right equiptment or the money for the right equipment to do SMD soldering, so do you think a thru-hole 39k 5% tolorance resistor will work?
Also, at some point i overread that @J4bberwock connects his miners to the 6 pin PCIe connectors of a high quality ATX PSU, and after a simple google search it seems like EVERYONE connects their miners to the PCIe connector, is there a specific reason why?
Sorry for bombarding with questions, i'm really anxious to know what i can do!!
(The amount of times i edited this post is too damn high!!!)
sure would be nice if you could share specs of the unit and price so others could benefit from it too.
THX
Lumanet
http://imgur.com/a/Tf2KrTHX
Lumanet
I'm not quite sure about the price though, i got it for free
Sorry I had burned your party. But finding a PSU which has 2 real 12V Rails is like winning in Lotto.
http://www.use.com/images/s_2/Hantol_Silent_ATX_500W_Power_Supply_7522d24941723e028d9f_1.jpg
This 500Watt PSU for example only has a 16Amp rail. So one normal Gridseed miner already maxes it out :-( .
This feed focus on the heavy overclocking of those blades and that's why I call out to be very careful and do your math twice before you cause an fire. This is why I just recommend it to check twice. Also if you do your math just for LAB giggles 500Watt on pure 12V can only do 41AMP MAX but this is LAB condition. Most PSU's like the one above doesn't even get close to is. Also most PSU have a cheap cables which can't take a lot of load even the CE specs tell you differently, but that's a reason why MADE in CHINA is not MADE in USA or GERMANY. As I said earlier we did overclock with hardware mod's here which cause a much higher draw from the PSU so that's why I called it out just to be safe. Because the last thing you want is a house on Fire.
If you followed a bit my hardware mod's which do ramp Gridseeds for less than 10$ USD up from normal 600-800 up to close to 1100Hz. so you get a very nice and safe 3.3-3.6 MHz per Blade out without busting it's limit's all the way to the MAX.
Lumanet
Well then i guess i won the lottery, doing a simple route trace on that psu show 2 12 volt rails going to 2 seprate overcurrent protection circuits. Also worth mentioning this PSU isn't one of those scary PSUs that burn your motherboard after a month of use, its a mid-range PSU with very good component and PCB quality for its price.http://www.use.com/images/s_2/Hantol_Silent_ATX_500W_Power_Supply_7522d24941723e028d9f_1.jpg
This 500Watt PSU for example only has a 16Amp rail. So one normal Gridseed miner already maxes it out :-( .
This feed focus on the heavy overclocking of those blades and that's why I call out to be very careful and do your math twice before you cause an fire. This is why I just recommend it to check twice. Also if you do your math just for LAB giggles 500Watt on pure 12V can only do 41AMP MAX but this is LAB condition. Most PSU's like the one above doesn't even get close to is. Also most PSU have a cheap cables which can't take a lot of load even the CE specs tell you differently, but that's a reason why MADE in CHINA is not MADE in USA or GERMANY. As I said earlier we did overclock with hardware mod's here which cause a much higher draw from the PSU so that's why I called it out just to be safe. Because the last thing you want is a house on Fire.
If you followed a bit my hardware mod's which do ramp Gridseeds for less than 10$ USD up from normal 600-800 up to close to 1100Hz. so you get a very nice and safe 3.3-3.6 MHz per Blade out without busting it's limit's all the way to the MAX.
Lumanet
sure would be nice if you could share specs of the unit and price so others could benefit from it too.
THX
Lumanet
Sorry I had burned your party. But finding a PSU which has 2 real 12V Rails is like winning in Lotto.
http://www.use.com/images/s_2/Hantol_Silent_ATX_500W_Power_Supply_7522d24941723e028d9f_1.jpg
This 500Watt PSU for example only has a 16Amp rail. So one normal Gridseed miner already maxes it out :-( .
This feed focus on the heavy overclocking of those blades and that's why I call out to be very careful and do your math twice before you cause an fire. This is why I just recommend it to check twice. Also if you do your math just for LAB giggles 500Watt on pure 12V can only do 41AMP MAX but this is LAB condition. Most PSU's like the one above doesn't even get close to is. Also most PSU have a cheap cables which can't take a lot of load even the CE specs tell you differently, but that's a reason why MADE in CHINA is not MADE in USA or GERMANY. As I said earlier we did overclock with hardware mod's here which cause a much higher draw from the PSU so that's why I called it out just to be safe. Because the last thing you want is a house on Fire.
If you followed a bit my hardware mod's which do ramp Gridseeds for less than 10$ USD up from normal 600-800 up to close to 1100Hz. so you get a very nice and safe 3.3-3.6 MHz per Blade out without busting it's limit's all the way to the MAX.
Lumanet
Well then i guess i won the lottery, doing a simple route trace on that psu show 2 12 volt rails going to 2 seprate overcurrent protection circuits. Also worth mentioning this PSU isn't one of those scary PSUs that burn your motherboard after a month of use, its a mid-range PSU with very good component and PCB quality for its price. http://www.use.com/images/s_2/Hantol_Silent_ATX_500W_Power_Supply_7522d24941723e028d9f_1.jpg
This 500Watt PSU for example only has a 16Amp rail. So one normal Gridseed miner already maxes it out :-( .
This feed focus on the heavy overclocking of those blades and that's why I call out to be very careful and do your math twice before you cause an fire. This is why I just recommend it to check twice. Also if you do your math just for LAB giggles 500Watt on pure 12V can only do 41AMP MAX but this is LAB condition. Most PSU's like the one above doesn't even get close to is. Also most PSU have a cheap cables which can't take a lot of load even the CE specs tell you differently, but that's a reason why MADE in CHINA is not MADE in USA or GERMANY. As I said earlier we did overclock with hardware mod's here which cause a much higher draw from the PSU so that's why I called it out just to be safe. Because the last thing you want is a house on Fire.
If you followed a bit my hardware mod's which do ramp Gridseeds for less than 10$ USD up from normal 600-800 up to close to 1100Hz. so you get a very nice and safe 3.3-3.6 MHz per Blade out without busting it's limit's all the way to the MAX.
Lumanet
Sorry I had burned your party. But finding a PSU which has 2 real 12V Rails is like winning in Lotto.
This 500Watt PSU for example only has a 16Amp rail. So one normal Gridseed miner already maxes it out :-( .
This feed focus on the heavy overclocking of those blades and that's why I call out to be very careful and do your math twice before you cause an fire. This is why I just recommend it to check twice. Also if you do your math just for LAB giggles 500Watt on pure 12V can only do 41AMP MAX but this is LAB condition. Most PSU's like the one above doesn't even get close to is. Also most PSU have a cheap cables which can't take a lot of load even the CE specs tell you differently, but that's a reason why MADE in CHINA is not MADE in USA or GERMANY. As I said earlier we did overclock with hardware mod's here which cause a much higher draw from the PSU so that's why I called it out just to be safe. Because the last thing you want is a house on Fire.
If you followed a bit my hardware mod's which do ramp Gridseeds for less than 10$ USD up from normal 600-800 up to close to 1100Hz. so you get a very nice and safe 3.3-3.6 MHz per Blade out without busting it's limit's all the way to the MAX.
Lumanet
This 500Watt PSU for example only has a 16Amp rail. So one normal Gridseed miner already maxes it out :-( .
This feed focus on the heavy overclocking of those blades and that's why I call out to be very careful and do your math twice before you cause an fire. This is why I just recommend it to check twice. Also if you do your math just for LAB giggles 500Watt on pure 12V can only do 41AMP MAX but this is LAB condition. Most PSU's like the one above doesn't even get close to is. Also most PSU have a cheap cables which can't take a lot of load even the CE specs tell you differently, but that's a reason why MADE in CHINA is not MADE in USA or GERMANY. As I said earlier we did overclock with hardware mod's here which cause a much higher draw from the PSU so that's why I called it out just to be safe. Because the last thing you want is a house on Fire.
If you followed a bit my hardware mod's which do ramp Gridseeds for less than 10$ USD up from normal 600-800 up to close to 1100Hz. so you get a very nice and safe 3.3-3.6 MHz per Blade out without busting it's limit's all the way to the MAX.
Lumanet
The main thing you look for the on an PSU is the 12V RAIL. In Picture above I red marked it for you. As more AMP you have on your Rail as more security / safty you build in that your PSU won't burn up and cause a potential fire.
So a GridSeed Blade only uses 100Watt maybe 120-140Watt if extreme overclocked. The other thing Jabberwork can also confirm is the AMP draw over those thinner ATX Cables cause heat on the Cable so as more you use in parallel as less heat you have on the cable. This is one reason why I highly recommend to go full on 14G-12G 120V Cable with solid copper core like shown in the picture. Those cables can take a BIG Load without producing the heat because they are solid copper. Also an very important factor which nobody tells you here is HEAT means you loose energy and efficiency. Not much but it ends up if you do run it like we do in Data Center settings.
So do your math :
12V and 100Watt = 8.33 AMP
Yes you see it right one Gridseed Blade alone draws on DC Volt over 8 Amp. So we normally do have up to 4 units on 1 PSU so 32AMP is been drawn. And 70Amp can be handled. If you want to operate like a Data Center meaning 24h a day on your mining rigs than go this route so you don't burn down your house. Always leave 20-30% room on the 12V Rail otherwise you burn out those PSU's and you have to buy another one.
So a GridSeed Blade only uses 100Watt maybe 120-140Watt if extreme overclocked. The other thing Jabberwork can also confirm is the AMP draw over those thinner ATX Cables cause heat on the Cable so as more you use in parallel as less heat you have on the cable. This is one reason why I highly recommend to go full on 14G-12G 120V Cable with solid copper core like shown in the picture. Those cables can take a BIG Load without producing the heat because they are solid copper. Also an very important factor which nobody tells you here is HEAT means you loose energy and efficiency. Not much but it ends up if you do run it like we do in Data Center settings.
So do your math :
12V and 100Watt = 8.33 AMP
Yes you see it right one Gridseed Blade alone draws on DC Volt over 8 Amp. So we normally do have up to 4 units on 1 PSU so 32AMP is been drawn. And 70Amp can be handled. If you want to operate like a Data Center meaning 24h a day on your mining rigs than go this route so you don't burn down your house. Always leave 20-30% room on the 12V Rail otherwise you burn out those PSU's and you have to buy another one.
dunno what you talking about buddy, i'm powering my G-Blade with an ATX PSU with the blades clocked at 825 MHz and using 18 AWG wire (i removed the power jacks and soldered 18 AWG wires dircetly to the blades) and it doesnt even get warm. to the touch, the wires feel like any other wire around the house. except that, the maximum amp rating of the 12v rails on that PSU is 18 amps, so i got 9.67 amps of headroom, (and even if the miner takes more then 8.33 amps, which i'm 100% sure it does, at that clock speed, i got 2 seperate 12v rails, so that's another 18 free amps) which is more then enough, i could probably even power a few 5chip miners that i have laying around, all from the same 500W supply!
ATX are CHEAP !!! So that's the main reason why. But if you do have a few of those like we do. We use our own produced PSU and breaker boards which allow you to hook up to 4units to 1 PSU + Breakerboard. If you want we can post photos too.
The 12V RAIL is the key. most PSU's have a low 12v RAIL if you want to run it 24/7 go and get a Server PSU those units are build for 24/7 operation.
We sell those also at http://www.LSE.Solar
Olaf
and yes please do show me some pics
Inside of the PSU so for example the ATX PSU the 12V Rail is accessed best over the 6pin or 8 pin connectors.
This is an example how we do run our Antminer S3 powered.
The main thing you look for the on an PSU is the 12V RAIL. In Picture above I red marked it for you. As more AMP you have on your Rail as more security / safty you build in that your PSU won't burn up and cause a potential fire.
So a GridSeed Blade only uses 100Watt maybe 120-140Watt if extreme overclocked. The other thing Jabberwork can also confirm is the AMP draw over those thinner ATX Cables cause heat on the Cable so as more you use in parallel as less heat you have on the cable. This is one reason why I highly recommend to go full on 14G-12G 120V Cable with solid copper core like shown in the picture. Those cables can take a BIG Load without producing the heat because they are solid copper. Also an very important factor which nobody tells you here is HEAT means you loose energy and efficiency. Not much but it ends up if you do run it like we do in Data Center settings.
So do your math :
12V and 100Watt = 8.33 AMP
Yes you see it right one Gridseed Blade alone draws on DC Volt over 8 Amp. So we normally do have up to 4 units on 1 PSU so 32AMP is been drawn. And 70Amp can be handled. If you want to operate like a Data Center meaning 24h a day on your mining rigs than go this route so you don't burn down your house. Always leave 20-30% room on the 12V Rail otherwise you burn out those PSU's and you have to buy another one.
Here is a Setup I do have for long term testing in my Garage (original unmodded units):
ATX are CHEAP !!! So that's the main reason why. But if you do have a few of those like we do. We use our own produced PSU and breaker boards which allow you to hook up to 4units to 1 PSU + Breakerboard. If you want we can post photos too.
The 12V RAIL is the key. most PSU's have a low 12v RAIL if you want to run it 24/7 go and get a Server PSU those units are build for 24/7 operation.
We sell those also at http://www.LSE.Solar
Olaf
and yes please do show me some pics
Hello everyone, i have been reading this thread from page 1 - 25 until my head exploded, it's definetly awesome to read all of your progress but the stuff you do is way over my limit, i know basic electronics and i just got my G-Blade a few days ago.
At stock 600 MHz it mines at a puny 3.2 MH/s (WTF??)
When overcloking to 838 MHz one blade works at a fairly low amount of H/W errors but the second one gets quite a lot of them, so i underclocked it a little to 800 MHz to make it happy, but anyway back to topic, i wanted to get the most i possible could out of my blade by replacing that lame power jack with a screw terminal (or just fuckin' soldering wires directly to the blade) and i also want to add heatsinks to all of the overheating part, my question is, that if i do all of that, will that make my miner a little more stable with less hardware errors? and also, i might just do the simple version of J4bberwock's voltage mod (just the 39k resistor) but the problem is that i definetly have the right equiptment or the money for the right equipment to do SMD soldering, so do you think a thru-hole 39k resistor will work?
Also, at some point i overread that @J4bberwock connects his miners to the 6 pin PCIe connectors on a high quality ATX PSU, and after a simple google search it seems like EVERYONE does that, is there a specific reason why?
Sorry for bombarding with questions, i'm really anxious to know what i can do!!
At stock 600 MHz it mines at a puny 3.2 MH/s (WTF??)
When overcloking to 838 MHz one blade works at a fairly low amount of H/W errors but the second one gets quite a lot of them, so i underclocked it a little to 800 MHz to make it happy, but anyway back to topic, i wanted to get the most i possible could out of my blade by replacing that lame power jack with a screw terminal (or just fuckin' soldering wires directly to the blade) and i also want to add heatsinks to all of the overheating part, my question is, that if i do all of that, will that make my miner a little more stable with less hardware errors? and also, i might just do the simple version of J4bberwock's voltage mod (just the 39k resistor) but the problem is that i definetly have the right equiptment or the money for the right equipment to do SMD soldering, so do you think a thru-hole 39k resistor will work?
Also, at some point i overread that @J4bberwock connects his miners to the 6 pin PCIe connectors on a high quality ATX PSU, and after a simple google search it seems like EVERYONE does that, is there a specific reason why?
Sorry for bombarding with questions, i'm really anxious to know what i can do!!
ATX are CHEAP !!! So that's the main reason why. But if you do have a few of those like we do. We use our own produced PSU and breaker boards which allow you to hook up to 4units to 1 PSU + Breakerboard. If you want we can post photos too.
The 12V RAIL is the key. most PSU's have a low 12v RAIL if you want to run it 24/7 go and get a Server PSU those units are build for 24/7 operation.
We sell those also at http://www.LSE.Solar
Olaf
Hello everyone, i have been reading this thread from page 1 - 25 until my head exploded, it's definetly awesome to read all of your progress but the stuff you do is way over my limit, i know basic electronics and i just got my G-Blade a few days ago.
At stock 600 MHz it mines at a puny 3.2 MH/s (WTF??)
When overcloking to 838 MHz one blade works at a fairly low amount of H/W errors but the second one gets quite a lot of them, so i underclocked it a little to 800 MHz to make it happy, but anyway back to topic, i wanted to get the most i possible could out of my blade by replacing that lame power jack with a screw terminal (or just fuckin' soldering wires directly to the blade) and i also want to add heatsinks to all of the overheating part, my question is, that if i do all of that, will that make my miner a little more stable with less hardware errors? and also, i might just do the simple version of J4bberwock's voltage mod (just the 39k resistor) but the problem is that i definetly have the right equiptment or the money for the right equipment to do SMD soldering, so do you think a thru-hole 39k resistor will work?
Also, at some point i overread that @J4bberwock connects his miners to the 6 pin PCIe connectors on a high quality ATX PSU, and after a simple google search it seems like EVERYONE does that, is there a specific reason why?
Sorry for bombarding with questions, i'm really anxious to know what i can do!!
At stock 600 MHz it mines at a puny 3.2 MH/s (WTF??)
When overcloking to 838 MHz one blade works at a fairly low amount of H/W errors but the second one gets quite a lot of them, so i underclocked it a little to 800 MHz to make it happy, but anyway back to topic, i wanted to get the most i possible could out of my blade by replacing that lame power jack with a screw terminal (or just fuckin' soldering wires directly to the blade) and i also want to add heatsinks to all of the overheating part, my question is, that if i do all of that, will that make my miner a little more stable with less hardware errors? and also, i might just do the simple version of J4bberwock's voltage mod (just the 39k resistor) but the problem is that i definetly have the right equiptment or the money for the right equipment to do SMD soldering, so do you think a thru-hole 39k resistor will work?
Also, at some point i overread that @J4bberwock connects his miners to the 6 pin PCIe connectors on a high quality ATX PSU, and after a simple google search it seems like EVERYONE does that, is there a specific reason why?
Sorry for bombarding with questions, i'm really anxious to know what i can do!!
Morning ,
my name is Olaf Becker and I am the owner of LSE.Solar. So first of all if you do have a thin soldering iron you can do this work all by yourself.
2nd of all if you want all the parts please reach out to us. We do sell those in set's I will post those parts you need onto our Shop and you can buy those from us without missing a single peace :-)
Also if you do have a Dremel we can show you with a think drillbit how to replace this cheap jack with srew down Terminal like I shown above in pictures. Importance there is you drill exact otherwise you can shorten the circuits.
http://www.LSE.Solar
regards
Olaf
Hello everyone, i have been reading this thread from page 1 - 25 until my brains exploded, it's definetly awesome to read all of your progress but the stuff you do is way over my limit, i know basic electronics and i just got my G-Blade a few days ago.
So at stock 600 MHz it mines at a puny 3.2 MH/s (WTF?? why??)
When overcloking to 838 MHz one blade works at a fairly low amount of H/W errors but the second one gets quite a lot of them, so i underclocked it a little to 825 MHz to make it happy, but anyway back to topic, i wanted to get the most i possible could out of my blade by replacing that lame power jack with a screw terminal (or just fuckin' soldering wires directly to the blades) and i also want to add heatsinks to all of the overheating part, my question is, that if i do all of that, will that make my miner a little more stable with less hardware errors? and also, i might just do the simple version of J4bberwock's voltage mod (just the 39k resistor) but the problem is that i definetly don't have the right equiptment or the money for the right equipment to do SMD soldering, so do you think a thru-hole 39k 5% tolorance resistor will work?
Also, at some point i overread that @J4bberwock connects his miners to the 6 pin PCIe connectors of a high quality ATX PSU, and after a simple google search it seems like EVERYONE connects their miners to the PCIe connector, is there a specific reason why?
Sorry for bombarding with questions, i'm really anxious to know what i can do!!
(The amount of times i edited this post is too damn high!!!)
So at stock 600 MHz it mines at a puny 3.2 MH/s (WTF?? why??)
When overcloking to 838 MHz one blade works at a fairly low amount of H/W errors but the second one gets quite a lot of them, so i underclocked it a little to 825 MHz to make it happy, but anyway back to topic, i wanted to get the most i possible could out of my blade by replacing that lame power jack with a screw terminal (or just fuckin' soldering wires directly to the blades) and i also want to add heatsinks to all of the overheating part, my question is, that if i do all of that, will that make my miner a little more stable with less hardware errors? and also, i might just do the simple version of J4bberwock's voltage mod (just the 39k resistor) but the problem is that i definetly don't have the right equiptment or the money for the right equipment to do SMD soldering, so do you think a thru-hole 39k 5% tolorance resistor will work?
Also, at some point i overread that @J4bberwock connects his miners to the 6 pin PCIe connectors of a high quality ATX PSU, and after a simple google search it seems like EVERYONE connects their miners to the PCIe connector, is there a specific reason why?
Sorry for bombarding with questions, i'm really anxious to know what i can do!!
(The amount of times i edited this post is too damn high!!!)
Maybe this is why I do own and operate a Data Center with my own mining rigs.
We pay 25 USD per Gridseed Blade and making per unit around $1 currently per day in profit after electric costs. Try to beat that.
One thing you have to learn is upfront Investment with ROI if you don't do this right you pay out of your own pocket.
We pay 25 USD per Gridseed Blade and making per unit around $1 currently per day in profit after electric costs. Try to beat that.
One thing you have to learn is upfront Investment with ROI if you don't do this right you pay out of your own pocket.
Again, you're just trying to be right! My client's Zues miners hit ROI month$ ago! It'$ ALL PROFIT BABY! SORRY BUT YOU CANNOT BEAT THAT!
Oh, and there is no way that 5.2MHS will earn $1 per day! Don't be ridiculous! You mean more like 40c to 50c per day if that! And that is IF You can keep them from overheating TRYING to keep up with ANY ZEUS miner running at STOCK SPEED!
Maybe this is why I do own and operate a Data Center with my own mining rigs.
We pay 25 USD per Gridseed Blade and making per unit around $1 currently per day in profit after electric costs. Try to beat that.
One thing you have to learn is upfront Investment with ROI if you don't do this right you pay out of your own pocket.
We pay 25 USD per Gridseed Blade and making per unit around $1 currently per day in profit after electric costs. Try to beat that.
One thing you have to learn is upfront Investment with ROI if you don't do this right you pay out of your own pocket.
You still don't have a clue LOL. Waste your money and time on Zeus while we go with Gridseeds which now are SFarts. Cant' wait for the new chip to come into our labs.
You need to learn how to READ, fool! The comparison is between GS Blade 5.2MH vs 5MHS per card of which there are 2 per heat sink = 10MHS WITH ZERO MODS! I LOVE the 5 chip pods! Started out with them. Anyway....NO ARGUMENT POSSIBLE THERE! KEEP TRYING TO BE RIGHT! GOOD LUCK!
Jump to: