GridSeed 5-chip USB miner voltage mod

[nemercry]

Actually i would be really carefull and NOT just soldering it. It's actually a frequency filter which acts also like a fuse.
As its possible that your pod is still alive, dont break it with just soldering the broken parts.

From wikipedia:
A ferrite bead is a passive electric component that suppresses high frequency noise in electronic circuits. It is a specific type of electronic choke. Ferrite beads employ the dissipation of high frequency currents in a ferrite ceramic to build high frequency noise suppression devices. Ferrite beads may also be called blocks, cores, rings, EMI filters, or chokes.[1]

@Wolfey: Please stop to give out such advices and dont post your picture in every second post with highest resolution.

...
What the FB is or what it does is irrelevant.
...

You actually made me laugh hard .

So, you decide to bore people with our antics,,,okay- fine!
and I quote....
"which acts also like a fuse" WRONG!
Laugh at that 

Each object which bricks at a certain amount of power is technically a fuse, yes. 
So over 3A the ferrite pearl bricks.

[1714766552]

1714766552

[1714766552]

1714766552

[1714766552]

1714766552

[simon66]

Revert back?

That would be a bit boring, no? Down at 850 it's doing minerside as before the mod, and poolside slightly worse.
The output leads to the same direction: Apr 12 17:41:13 bfgminer[23935]: 20s:373.7 avg:360.5 u:309.2 kh/s
I checked with the multimeter, it's certainly a 47k resistor. That's all really strange...

I get the same but if you leave it mining for a while, the pool side will stabilize. Keep in mind that pool side values are way off and it doesn't really matter. All you should care is "accepted" shares and you get paid per accepted share and not Kh/s value.

What pool you use?

[david123]

I get the same but if you leave it mining for a while, the pool side will stabilize. Keep in mind that pool side values are way off and it doesn't really matter. All you should care is "accepted" shares and you get paid per accepted share and not Kh/s value.

What pool you use?

Yeah, that's right.. Currently I use ipominer.com, where the rate is usually quite accurate and stabilizes after just a couple of minutes. If I let it run @850 for 10 minutes, I see 340 for example, and if I go up to 1150 or 1200, the reported rate stabilizes quickly at a lower level, like 280, where it stays. The accept rate is higher at 1200, but as ipominer is vardiff and bfgminer reports wrong diff "0", I'm not sure if this has any meaning..

[wolfey2014]

Actually i would be really carefull and NOT just soldering it. It's actually a frequency filter which acts also like a fuse.
As its possible that your pod is still alive, dont break it with just soldering the broken parts.

From wikipedia:
A ferrite bead is a passive electric component that suppresses high frequency noise in electronic circuits. It is a specific type of electronic choke. Ferrite beads employ the dissipation of high frequency currents in a ferrite ceramic to build high frequency noise suppression devices. Ferrite beads may also be called blocks, cores, rings, EMI filters, or chokes.[1]

@Wolfey: Please stop to give out such advices and dont post your picture in every second post with highest resolution.

...
What the FB is or what it does is irrelevant.
...

You actually made me laugh hard .

So, you decide to bore people with our antics,,,okay- fine!
and I quote....
"which acts also like a fuse" WRONG!
Laugh at that  

Each object which bricks at a certain amount of power is technically a fuse, yes.  
So over 3A the ferrite pearl bricks.

NO!
It is litterally a BLOWN component!!   Just because a transistor blows, doesn't make it a fuse, active or passive, technically or otherwise! Geesh!
Note: Anyone who refers to a component as an "object" is illiterate! Anyone who calls a blown ferrite bead or any other component a fuse, other than an actual fuse, is incompetent and not to be trusted with electronics work!  

[jamieb81]

I get the same but if you leave it mining for a while, the pool side will stabilize. Keep in mind that pool side values are way off and it doesn't really matter. All you should care is "accepted" shares and you get paid per accepted share and not Kh/s value.

What pool you use?

Yeah, that's right.. Currently I use ipominer.com, where the rate is usually quite accurate and stabilizes after just a couple of minutes. If I let it run @850 for 10 minutes, I see 340 for example, and if I go up to 1150 or 1200, the reported rate stabilizes quickly at a lower level, like 280, where it stays. The accept rate is higher at 1200, but as ipominer is vardiff and bfgminer reports wrong diff "0", I'm not sure if this has any meaning..

I had troubles with BFGminer, miners were shown mining in BFG with 0HW errors but I was getting litteraly no hash speed on pool side,
switched back to CPUminer and hashrate and share rate flew trough the roof.

PS: how's that IPO pool going so far? good results?

[david123]

I had troubles with BFGminer, miners were shown mining in BFG with 0HW errors but I was getting litteraly no hash speed on pool side,
switched back to CPUminer and hashrate and share rate flew trough the roof.

PS: how's that IPO pool going so far? good results?

Ow, thanks, I didn't come up with that (I mean: that it is a software problem). A bit unsatisfying that Hashra will not be useable as intended
then, but anyways..

IPO is really nice, I'll stay with them. But it's difficult to estimate the payout, it depends a lot on when you sell these coins. It's kind of a mixture
of mining and trading

[usao]

Revert back?

That would be a bit boring, no? Down at 850 it's doing minerside as before the mod, and poolside slightly worse.
The output leads to the same direction: Apr 12 17:41:13 bfgminer[23935]: 20s:373.7 avg:360.5 u:309.2 kh/s
I checked with the multimeter, it's certainly a 47k resistor. That's all really strange...

I get the same but if you leave it mining for a while, the pool side will stabilize. Keep in mind that pool side values are way off and it doesn't really matter. All you should care is "accepted" shares and you get paid per accepted share and not Kh/s value.

What pool you use?

Specifically, the pool-side accepted shares are what our pay is calculated from, which may or may not match the miner accepted shares.

[amix]

Has anyone tried just using a sharp nail clipper/scissor to remove the resistor? Simple, quick and sweet!

oh you got luckly. i wouldnt recommend that one as it has a strong possibility of damaging your board, ie breaking a track with it. the tracks are really small and can get pulled up or broke if too much force is applied.

[david123]

So thanks Jamie! cpuminer didn't work at all, a precompiled cgminer underperformed similarly as the hashra-bfgminer,
but a freshly compiled cgminer did the trick. Working fine now

I made a couple of pics!

The workplace (I use an Atten 8502D with a build-in hot air unit and a 0,5 mm soldering tip):

The miner before changing the resistor (I used Kapton tape):

The result:

Happily hashing now :

I'd like to thank all of you who participated in this thread, in particular the one who discovered this trick (I believe I used the
description of Wolfey with regard to which resistor needs to be exchanged for what, but I'm not sure if he was the first to find
out. If so, thanks Wolfey!)

I'll take care of my other units now

[RowanX]

Regarding TP mod..

http://cryptomining-blog.com/wp-content/uploads/2014/03/gridseed-5-chip-asic-new-version.jpg
I presume you throw away the heat pad seen in this photo.. then do the TP stuff afterwards? And has anyone got a photo of the other side?

Secondly.. I don't have vodka.. can I use Listerine to clean the chips prior to TP application?

[jamieb81]

Regarding TP mod..

http://cryptomining-blog.com/wp-content/uploads/2014/03/gridseed-5-chip-asic-new-version.jpg
I presume you throw away the heat pad seen in this photo.. then do the TP stuff afterwards? And has anyone got a photo of the other side?

Secondly.. I don't have vodka.. can I use Listerine to clean the chips prior to TP application?

Don't be that affraid, I'm still using the stock TP's and works fine. Of course you can be a perfectionist.

Mouthwash to clean the pads? 

With the 5v fanmod they are really cold

[Xer0]

so anyone found out, how to mod the unit in a way you can switch voltages with the vid=1 option of cpuminer, like the first post mod, but using the voltages of 400kH and 500kH mods?

so you just can go back to almost stock software side,
the time when efficiency becomes more important

[RowanX]

Regarding TP mod..

http://cryptomining-blog.com/wp-content/uploads/2014/03/gridseed-5-chip-asic-new-version.jpg
I presume you throw away the heat pad seen in this photo.. then do the TP stuff afterwards? And has anyone got a photo of the other side?

Secondly.. I don't have vodka.. can I use Listerine to clean the chips prior to TP application?

Don't be that affraid, I'm still using the stock TP's and works fine. Of course you can be a perfectionist.

Mouthwash to clean the pads? 

With the 5v fanmod they are really cold

My two gridseeds are not hardware modded, but one of them gives 10x more HW errors than the other, even at a lower frequency. It is that one that I am interested in improving the heatsink contact with, to see if that is where the problem lies.

Mouthwash to clean the pads? Well it is a high concentration of alcohol... so? And it would leave the gridseed smelling great!

[jamieb81]

nothing wrong with it, you could aslo use some kind of eau de cologne on a q tip to clean it. I just thought it was funny 

yes you can try to do it to see if the errors go away, but some people run the gridseeds without even a heatsink unmodded,

so my guess is that it wont have much effect

[Xer0]

Isopropanole is the stuff you`re looking for...

[Cablez]

Yes, get some rubbing alcohol form your local pharmacy or grocery store.  Liquor is better suited for drinking and mouthwash is just wrong.

[HashFire]

OK just got my resistors today, I undid the previous 2 solder mods and put on this one.  At 1000mhz i don't seem to be accepting any shares, also using 0 Watts measured with the Kilowatt.  Any ideas?

https://i.imgur.com/EgWRBBH.jpg

https://i.imgur.com/0nWg0uA.jpg

Double check the continuity with your DVM..
The inside solder point looks open...
Solder it to one or the other points next to each other. They are the same connection on the pcb as one was for a capacitor next to R52
Always double check.
It really sounds like your resistor is not in circuit.

I have resoldered this twice now and tested with DMM in OHM mode, I get a drop across the reistor for sure, but still not working.  I think my problem is I have messed up the pad on the outside where the resistor goes.  Anywhere else I can tie into?  Maybe from the bottom if I have to?

[wolfey2014]

OK just got my resistors today, I undid the previous 2 solder mods and put on this one.  At 1000mhz i don't seem to be accepting any shares, also using 0 Watts measured with the Kilowatt.  Any ideas?

Double check the continuity with your DVM..
The inside solder point looks open...
Solder it to one or the other points next to each other. They are the same connection on the pcb as one was for a capacitor next to R52
Always double check.
It really sounds like your resistor is not in circuit.

I have resoldered this twice now and tested with DMM in OHM mode, I get a drop across the reistor for sure, but still not working.  I think my problem is I have messed up the pad on the outside where the resistor goes.  Anywhere else I can tie into?  Maybe from the bottom if I have to?

Well, shit fire!, Hashfire! Looks like you went and messed it up pretty good there matey!
Yep, I see cold solder joints all over the place i.e. OPENS = not a closed circuit.
If that particular resistor is out of circuit, you will see what you're seeing, no share submittals.
Are you using resin core solder? Doesn't look like there's any resin there. This is what cleans and helps the solder stick to the copper conductors. Without resin (flux) you are going to have the problems you've been having with soldering. Also temperature needs to be correct, even a little bit too high, but not much!

The corresponding pad (solder point) to the outer connection you are trying to make is the one just to the back (closest to the outside edge of the pcb) solder point where you soldered the inner point to.
Look at where the old resistor was connected. Those are the two solder points you can solder your resistor to. Trying to solder a 1/4W resistor to those two very close in points is going to be very difficult for a beginner or novice. It's even a tiny bit difficult for me but I can do it via techniques I know.

De-solder the resistor from the card. Use flux/resin core solder or some flux and put some on the solder points, then tin (solder) the points so they are clean and nice and rounded on the tops. Using resin will allow the solder to flow and it will cooperate with you and gravity perfectly.

Then, in the same way, tin the ends of the resistor leads. Then, heat both the solder point and the resistor tips at the same time and let them melt together into one beautiful SHINY solder blob.
The solder will shine brightly when done correctly. It will be dull if not.

Other than this, if you can't get it done right, please send it to someone who is qualified - like me - to do it for you!

Believe me, if you keep trying and fail, you just might do irreversible damage to those connections, then that will be it. Time to buy another miner to replace the nice pretty paper weight you just made  

Here is the result of doing it correctly....*******************************

*************************************************************
Remember, I can do this mod for you or anyone who needs it done by a professional. PM me for details.
And...
as usual, you do this mod at your own risk and expense.
Good luck!
Wolfey2014

[ZiG]

OK just got my resistors today, I undid the previous 2 solder mods and put on this one.  At 1000mhz i don't seem to be accepting any shares, also using 0 Watts measured with the Kilowatt.  Any ideas?

Double check the continuity with your DVM..
The inside solder point looks open...
Solder it to one or the other points next to each other. They are the same connection on the pcb as one was for a capacitor next to R52
Always double check.
It really sounds like your resistor is not in circuit.

I have resoldered this twice now and tested with DMM in OHM mode, I get a drop across the reistor for sure, but still not working.  I think my problem is I have messed up the pad on the outside where the resistor goes.  Anywhere else I can tie into?  Maybe from the bottom if I have to?

Are you sure it is not a software...or pool issue...!

Did you try another MINER...or POOL...?

[24Kilo]

I have been using the girnyau's gridseed fork of cgminer compile for Windows by sang - https://bitcointalk.org/index.php?topic=519112.msg5799813#msg5799813

It has been rock solid, often running 7 - 10 days without a restart.

Today I compiled girnyau's gridseed fork of cgminer into BAMT 1.2 which I have been using for months.

The 'standard' Mhz setting for the sang Windows compile is 888Mhz, which I have been using without a hitch, but in Linux this fork of cgminer is very unstable at 888Mhz. I found that have use a factor of '50', so 800Mhz, 850Mhz, 900Mhz, and etc. I am currently running 850Mhz.

Can anyone else attest to this instability?

And what is the preferred 'cgminer' for running Gridseeds with Linux server?

Thanks!