Search content
Sort by
Showing 20 of 93 results by Andareed
Okay, my 'non-pencil-trick' mod is:
Replace C36 10nf with a 100nf capacitor = no more pencil tricks! 'yay'
This includes the other 3 mods i.e. the two jumpers and resistor change out on page 14 (Sandor's mods) of this thread.
I have run over 36 hours with ZERO hardware errors - 1000MHz (stable) at 568KH/s peak - 378KH/s average pool side .07% stale shares out of 587,114 valid shares. Comm port FIFO's were set at 14 Receive and 16 Transmit.
It's still running strong and stable. So I believe this mod is a pass!
I"m looking forward to trying out nemercry's mods when he finally releases the particulars. His results sound fantastic!
As usual, if you do these mods, you do them at your own risk!
Replace C36 10nf with a 100nf capacitor = no more pencil tricks! 'yay'
This includes the other 3 mods i.e. the two jumpers and resistor change out on page 14 (Sandor's mods) of this thread.
I have run over 36 hours with ZERO hardware errors - 1000MHz (stable) at 568KH/s peak - 378KH/s average pool side .07% stale shares out of 587,114 valid shares. Comm port FIFO's were set at 14 Receive and 16 Transmit.
It's still running strong and stable. So I believe this mod is a pass!
I"m looking forward to trying out nemercry's mods when he finally releases the particulars. His results sound fantastic!
As usual, if you do these mods, you do them at your own risk!
378 kh/s is lower than you should expect at 1000 MHz. Reported hardware error counts only tell half the story. A reported hardware represents a nonce that your miner thought was valid, but really isn't. It's way more likely that your miner would be corrupting and missing valid nonces, which are unfortunately not detectable.
I'm pretty skeptical that these are running at 600 kh/s at 1000-1100 MHz. Hash rate should scale roughly linear, so you would need to be running closer to 1400 MHz to achieve those hash rates. It's possible you just go really lucky, or the pool is incorrectly calculating your hashrate.
Re: SCRYPT ASIC miner ready. BTC+LTC ASIC miner; LA1THS, usd4500, stock ready now.
i have a problem 
3 of my miners just died in the last few days
initially i wasn't sure that they were dead , but once i opened them it become obvious
http://imageshack.com/a/img838/8194/9meg.jpg
http://imageshack.com/a/img812/8051/yipu.jpg
3 of my miners just died in the last few days
initially i wasn't sure that they were dead , but once i opened them it become obvious
http://imageshack.com/a/img838/8194/9meg.jpg
http://imageshack.com/a/img812/8051/yipu.jpg
It's possible your PSU went out of spec on the 12V rail, as there's 12V power running in that area. Honestly, I wouldn't trust the PSUs included in the kit, and this is one of the reasons I'm using a spare ATX PSU. The other reason is that my ATX PSU is more efficient and actually has power factor correction. If you have a multimeter, you can try checking the voltage.
The 10 I modded yesterday seem to be rock solid at 938Mhz. I have pushed the non-modded ones as far as I can without too many HW errors. I will solder the rest if these are stable through the weekend.
I also used the --usb :50 command to break them up into 2 groups of 50 miners so I can see them all.
I also used the --usb :50 command to break them up into 2 groups of 50 miners so I can see them all.
Okay, cool! Nice work!
I modded one of mine last night. Solder bridged 2 of the resistor blanks as shown earlier by the original modder. This increases core voltage a tad, I believe. And I think that's all you can safely increase it to.
I haven't tried replacing the resistor with a 38K or the pencil track mod trick yet.
That's for PLL voltage adjustment as I understand it, without delving into the schematic etc. myself.
Is this what you did too?
As for your fine tuning to an odd step of 938Mhz, how did you accomplish this?
I'd like to do the same thing.
Mine is rock steady at 900MHz but at 950Mhz being the next allowed step in cpuminer throws enough HW errors to turn me off to it. So for now. it's 900Mhz which is giving positive results at the pool. $$$
Wolfey2014
If you use https://github.com/dtbartle/cgminer-gc3355, you can specify pll_r, pll_f, and pll_od as part of gridseed-options; you can use https://gist.github.com/dtbartle/5319985b29b7f78c3993 to generate all possible combinations of these, and their resulting frequencies.
GridSeed docs are actually wrong, pll_BS is bit 31, and pll_BP is probably bit 30. If you look at the frequency tables in cgminer, you'll see that for frequencies >= 500 MHz, the highest bit is set.
Actually the gridseed version of cgminer sets bit30 low for 250/400/450mhz, and high for 500mhz and above just as they specified it in the docs.Here's the frequency table from cpuminer (https://github.com/gridseed/cpuminer/blob/master/gc3355.h#L51):
Code:
static const char *cmd_frequency[] = {
"55AAEF0005002001",
"55AAEF000500E001",
"55AAEF0005002002",
"55AAEF0005006082",
"55AAEF000500A082",
"55AAEF000500E082",
"55AAEF0005002083",
"55AAEF0005006083",
"55AAEF000500A083",
"55AAEF000500E083",
"55AAEF0005002084",
"55AAEF0005006084",
NULL
};
"55AAEF0005002001",
"55AAEF000500E001",
"55AAEF0005002002",
"55AAEF0005006082",
"55AAEF000500A082",
"55AAEF000500E082",
"55AAEF0005002083",
"55AAEF0005006083",
"55AAEF000500A083",
"55AAEF000500E083",
"55AAEF0005002084",
"55AAEF0005006084",
NULL
};
Note the 8 in the 2nd-last column, which corresponds to bit 31 (i.e. the high bit).
from the gridseed gc3355 datasheet :
CLUS_ADDR=0xE
REG ADDR
(8 bits)
0x0
[30]pll_BS
0 PLL Band Select
1: High band, 500MHz<=Fvco<=1GHz
0: Low band, 300MHz<=Fvco<=600MHz
the default is 0 so it might requirw a software mod to modify the reg addr to the higher band? or is it already enabled? thought of this as i recall the the PLL voltage being modifed somewhere.. stab in dark maybe.........
CLUS_ADDR=0xE
REG ADDR
(8 bits)
0x0
[30]pll_BS
0 PLL Band Select
1: High band, 500MHz<=Fvco<=1GHz
0: Low band, 300MHz<=Fvco<=600MHz
the default is 0 so it might requirw a software mod to modify the reg addr to the higher band? or is it already enabled? thought of this as i recall the the PLL voltage being modifed somewhere.. stab in dark maybe.........
GridSeed docs are actually wrong, pll_BS is bit 31, and pll_BP is probably bit 30. If you look at the frequency tables in cgminer, you'll see that for frequencies >= 500 MHz, the highest bit is set.
Testing at 1000 MHz now, seems stable with slightly higher voltage.
Did you replace a resistor?
If I perform the mod, then I want to go back to mining at 600-850mhz again for some reason, that still works? (but still using 2W more?)
Of course, the power consumption is directly related to core voltage, so frequency does not matter much in that regard.
If you want to hard mod the voltage to 1.25v, and use different miner software:
Yes, up to 950 MHz would be stable, I am using cpuminer.
Excellent! I knew there had to be a further hardware mod 'jumper' to get stable operation at 1.25V without converting over to a different mining program. I'm using cpuminer and am quite satisfied with it since I have a very small farm anyway.
One thing that still bugs me though is having to reset the USB ports every 24+/- hours.
Anyway, I can make the jumper mods easily as I own a professional Weller soldering station with adjustable temp, very fine tip assortment, de-soldering etc. So no problems making the mods.
I wonder though, have either of you chaps tried seeing if any of the existing header's pins correspond to either of those two jumper points? Or is that header simply a serial port connection. I haven't studied the schematic yet so I realize this may be a stupid question.
Anyway, I'm very interested in this mod for my own miners. Having done years of laboratory level R&D in electronics, this stuff is a piece of cake!
Have the two jumpers stabilized the GS units well enough to overclock above 950 yet?
Man, if we can get these little buggers running at 500KH/s, that would be fantastic!
But, 400KH/s ain't nothin to laugh at either
I have all my units fanless now. If necessary, I can simply use 1 or 2 fans to blow air on the fins indirectly which will of course keep them running cooler.
I might be interested in making this mod for others too if the money is right.
They'd be responsible for all shipping costs and insurance etc. etc.
I'll figure out what is a fair price to charge per mod later if anyone shows interest.
Thanks for coming up with this mod. Sure saved me some time and hassle.
Wolfey2014
The header allows you tap into the nonce reporting/status and switch bootloader mode (so that you can flash the miner firmware - this is what the white button on early miners was connected to). There's a second header spot (consisting of 3 pads) that will give you access to vid0, but nothing for vid1.
Here's what my cgminer looks like after about 10 hours:
Code:
GSD 0: 8D9155794849 850 MHz | 360.7K/360.7Kh/s | A:177152 R: 0 HW: 1 WU: 0.6/m
GSD 1: 8D9111984849 850 MHz | 360.6K/360.7Kh/s | A:194048 R: 0 HW:14 WU: 0.6/m
GSD 2: 8D71428E5654 850 MHz | 360.7K/360.7Kh/s | A:200704 R: 0 HW: 0 WU: 0.7/m
GSD 3: 8D8D36B04849 850 MHz | 360.7K/360.7Kh/s | A:196096 R: 512 HW: 0 WU: 0.6/m
GSD 4: 8D98148C4849 850 MHz | 360.7K/360.7Kh/s | A:195072 R: 0 HW: 0 WU: 0.6/m
GSD 5: 8D833F674849 950 MHz | 403.1K/403.1Kh/s | A:221696 R: 0 HW: 0 WU: 0.7/m
GSD 6: 8D7328515654 850 MHz | 360.7K/360.7Kh/s | A:200192 R: 512 HW: 1 WU: 0.7/m
GSD 7: 8D8E556D4849 850 MHz | 360.7K/360.7Kh/s | A:182784 R: 0 HW: 0 WU: 0.6/m
GSD 8: 8D9916694849 850 MHz | 360.7K/360.7Kh/s | A:200192 R: 0 HW: 0 WU: 0.7/m
GSD 9: 8D9255784849 850 MHz | 360.7K/360.7Kh/s | A:182272 R: 0 HW: 0 WU: 0.6/m
GSD 10: 8D89188D4849 850 MHz | 360.7K/360.7Kh/s | A:202240 R: 0 HW: 0 WU: 0.7/m
GSD 11: 8D86367B4849 850 MHz | 360.7K/360.7Kh/s | A:186880 R: 512 HW: 0 WU: 0.6/m
GSD 1: 8D9111984849 850 MHz | 360.6K/360.7Kh/s | A:194048 R: 0 HW:14 WU: 0.6/m
GSD 2: 8D71428E5654 850 MHz | 360.7K/360.7Kh/s | A:200704 R: 0 HW: 0 WU: 0.7/m
GSD 3: 8D8D36B04849 850 MHz | 360.7K/360.7Kh/s | A:196096 R: 512 HW: 0 WU: 0.6/m
GSD 4: 8D98148C4849 850 MHz | 360.7K/360.7Kh/s | A:195072 R: 0 HW: 0 WU: 0.6/m
GSD 5: 8D833F674849 950 MHz | 403.1K/403.1Kh/s | A:221696 R: 0 HW: 0 WU: 0.7/m
GSD 6: 8D7328515654 850 MHz | 360.7K/360.7Kh/s | A:200192 R: 512 HW: 1 WU: 0.7/m
GSD 7: 8D8E556D4849 850 MHz | 360.7K/360.7Kh/s | A:182784 R: 0 HW: 0 WU: 0.6/m
GSD 8: 8D9916694849 850 MHz | 360.7K/360.7Kh/s | A:200192 R: 0 HW: 0 WU: 0.7/m
GSD 9: 8D9255784849 850 MHz | 360.7K/360.7Kh/s | A:182272 R: 0 HW: 0 WU: 0.6/m
GSD 10: 8D89188D4849 850 MHz | 360.7K/360.7Kh/s | A:202240 R: 0 HW: 0 WU: 0.7/m
GSD 11: 8D86367B4849 850 MHz | 360.7K/360.7Kh/s | A:186880 R: 512 HW: 0 WU: 0.6/m
Have you tried 1000mhz yet?
Yes, it generates hardware errors.
Re: LightningAsic usb miners based Gridseed GC3355 Tech Support Thread
If anyone interested I've modified Andareed's version of cgminer a bit to let you better finetune the gridseed units. There's a big variance in the resistors they used as voltage references so every unit has a different maximum frequency they can run at.
https://github.com/girnyau/cgminer-gc3355
New frequency table: 800, 813, 825, 838, 850, 863, 875, 888, 900, 913, 925, 938, 950, 963, 975, 988, 1000, 1013, 1025, 1038, 1050, 1063, 1075, 1088, 1100, 1113, 1125, 1138, 1150, 1163, 1175, 1188, 1200
You won't need the ghz ones however unless highly overvolted
You can also define a base frequency as usual, then set per device frequencies by the units' serial number if you want:
--gridseed-options freq=888 --gridseed-freq 8D74488A4949=900,8D96227B5449=900,8D7357785355=875
Factory units running since ~48 hours without any hw mods but the fans desoldered:
cgminer version 3.7.2 - Started: [2014-03-16 13:23:19]
--------------------------------------------------------------------------------
(5s):1.891M (avg):1.888Mh/s | A:3357360 R:4832 HW:228 WU:10.5/m
ST: 2 SS: 0 NB: 950 LW: 34253 GF: 2 RF: 0
Connected to litecoinpool.org diff 256 with stratum as user composeit.gridseed
Block: 3a0bf6f9... Diff:5.16K Started: [22:08:37] Best share: 18.8M
--------------------------------------------------------------------------------
*P*ool management *S*ettings *D*isplay options *Q*uit
GSD 0: 8D74488A4949 900 MHz | 382.0K/381.9Kh/s | A:653120 R:2176 HW:204
GSD 1: 8D7135A35355 888 MHz | 376.9K/376.8Kh/s | A:670592 R:1152 HW: 6
GSD 2: 8D96227B5449 900 MHz | 382.0K/381.9Kh/s | A:675936 R: 800 HW: 5
GSD 3: 6D9323805251 888 MHz | 376.9K/376.8Kh/s | A:677392 R: 352 HW: 3
GSD 4: 8D7357785355 875 MHz | 371.4K/371.3Kh/s | A:680320 R: 352 HW: 10
--------------------------------------------------------------------------------
Cheers
https://github.com/girnyau/cgminer-gc3355
New frequency table: 800, 813, 825, 838, 850, 863, 875, 888, 900, 913, 925, 938, 950, 963, 975, 988, 1000, 1013, 1025, 1038, 1050, 1063, 1075, 1088, 1100, 1113, 1125, 1138, 1150, 1163, 1175, 1188, 1200
You won't need the ghz ones however unless highly overvolted
You can also define a base frequency as usual, then set per device frequencies by the units' serial number if you want:
--gridseed-options freq=888 --gridseed-freq 8D74488A4949=900,8D96227B5449=900,8D7357785355=875
Factory units running since ~48 hours without any hw mods but the fans desoldered:
cgminer version 3.7.2 - Started: [2014-03-16 13:23:19]
--------------------------------------------------------------------------------
(5s):1.891M (avg):1.888Mh/s | A:3357360 R:4832 HW:228 WU:10.5/m
ST: 2 SS: 0 NB: 950 LW: 34253 GF: 2 RF: 0
Connected to litecoinpool.org diff 256 with stratum as user composeit.gridseed
Block: 3a0bf6f9... Diff:5.16K Started: [22:08:37] Best share: 18.8M
--------------------------------------------------------------------------------
*P*ool management *S*ettings *D*isplay options *Q*uit
GSD 0: 8D74488A4949 900 MHz | 382.0K/381.9Kh/s | A:653120 R:2176 HW:204
GSD 1: 8D7135A35355 888 MHz | 376.9K/376.8Kh/s | A:670592 R:1152 HW: 6
GSD 2: 8D96227B5449 900 MHz | 382.0K/381.9Kh/s | A:675936 R: 800 HW: 5
GSD 3: 6D9323805251 888 MHz | 376.9K/376.8Kh/s | A:677392 R: 352 HW: 3
GSD 4: 8D7357785355 875 MHz | 371.4K/371.3Kh/s | A:680320 R: 352 HW: 10
--------------------------------------------------------------------------------
Cheers
Cool! Not sure that the resistors are the reason for the variance. The four I measured were within 0.1% of spec. I assume this is just silicon quality differences in the actual ASIC chips.
Here's what my cgminer looks like after about 10 hours:
Code:
GSD 0: 8D9155794849 850 MHz | 360.7K/360.7Kh/s | A:177152 R: 0 HW: 1 WU: 0.6/m
GSD 1: 8D9111984849 850 MHz | 360.6K/360.7Kh/s | A:194048 R: 0 HW:14 WU: 0.6/m
GSD 2: 8D71428E5654 850 MHz | 360.7K/360.7Kh/s | A:200704 R: 0 HW: 0 WU: 0.7/m
GSD 3: 8D8D36B04849 850 MHz | 360.7K/360.7Kh/s | A:196096 R: 512 HW: 0 WU: 0.6/m
GSD 4: 8D98148C4849 850 MHz | 360.7K/360.7Kh/s | A:195072 R: 0 HW: 0 WU: 0.6/m
GSD 5: 8D833F674849 950 MHz | 403.1K/403.1Kh/s | A:221696 R: 0 HW: 0 WU: 0.7/m
GSD 6: 8D7328515654 850 MHz | 360.7K/360.7Kh/s | A:200192 R: 512 HW: 1 WU: 0.7/m
GSD 7: 8D8E556D4849 850 MHz | 360.7K/360.7Kh/s | A:182784 R: 0 HW: 0 WU: 0.6/m
GSD 8: 8D9916694849 850 MHz | 360.7K/360.7Kh/s | A:200192 R: 0 HW: 0 WU: 0.7/m
GSD 9: 8D9255784849 850 MHz | 360.7K/360.7Kh/s | A:182272 R: 0 HW: 0 WU: 0.6/m
GSD 10: 8D89188D4849 850 MHz | 360.7K/360.7Kh/s | A:202240 R: 0 HW: 0 WU: 0.7/m
GSD 11: 8D86367B4849 850 MHz | 360.7K/360.7Kh/s | A:186880 R: 512 HW: 0 WU: 0.6/m
GSD 1: 8D9111984849 850 MHz | 360.6K/360.7Kh/s | A:194048 R: 0 HW:14 WU: 0.6/m
GSD 2: 8D71428E5654 850 MHz | 360.7K/360.7Kh/s | A:200704 R: 0 HW: 0 WU: 0.7/m
GSD 3: 8D8D36B04849 850 MHz | 360.7K/360.7Kh/s | A:196096 R: 512 HW: 0 WU: 0.6/m
GSD 4: 8D98148C4849 850 MHz | 360.7K/360.7Kh/s | A:195072 R: 0 HW: 0 WU: 0.6/m
GSD 5: 8D833F674849 950 MHz | 403.1K/403.1Kh/s | A:221696 R: 0 HW: 0 WU: 0.7/m
GSD 6: 8D7328515654 850 MHz | 360.7K/360.7Kh/s | A:200192 R: 512 HW: 1 WU: 0.7/m
GSD 7: 8D8E556D4849 850 MHz | 360.7K/360.7Kh/s | A:182784 R: 0 HW: 0 WU: 0.6/m
GSD 8: 8D9916694849 850 MHz | 360.7K/360.7Kh/s | A:200192 R: 0 HW: 0 WU: 0.7/m
GSD 9: 8D9255784849 850 MHz | 360.7K/360.7Kh/s | A:182272 R: 0 HW: 0 WU: 0.6/m
GSD 10: 8D89188D4849 850 MHz | 360.7K/360.7Kh/s | A:202240 R: 0 HW: 0 WU: 0.7/m
GSD 11: 8D86367B4849 850 MHz | 360.7K/360.7Kh/s | A:186880 R: 512 HW: 0 WU: 0.6/m
Re: SCRYPT ASIC miner ready. BTC+LTC ASIC miner, LA1THS, usd5000, stock ready now.
Hardware mod that allows stable higher frequencies: https://bitcointalk.org/index.php?topic=519112
Wait! Don't we already have a cpuminer that is supposed to sw that?
http://cryptomining-blog.com/1228-download-cpuminer-oc-edition-for-gridseed-5-chip-gc3355-asics
You can set frequency sky-high, but you'll only hardware errors after some point (usually 850 MHz).
There are SMD jumpers available however I'm not sure how small jumper we would need. One test miner should be with me by thursday. I can't wait to test it out
I can measure later today, but the're almost certainly 0.1"/2.54mm pitch. Something like this would work:
https://www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?storeId=10001&langId=-1&catalogId=10001&pa=2076826&productId=2076826
https://www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?storeId=10001&langId=-1&catalogId=10001&pa=112432&productId=112432
Yeah you are absolutely right. Pencil trick just makes things worse. Even with 850mhz and voltage=1 now I get a lot of HW errors. Time to break out the eraser!
Edit: Erased and back to normal. Zero HW errors @ 850mhz. Damn - now to find someone who can solder a bridge that tiny.
Edit: Erased and back to normal. Zero HW errors @ 850mhz. Damn - now to find someone who can solder a bridge that tiny.
It was really annoying to do. What I did was get a gob onto the general bridge area, and then ran the solder tip (I have a fairly fine tip) along both sides of the bridge to collect any stray solder.
EDIT: Going to try using this to solder the bridge: http://www.ebay.com/itm/151169919397
Nice work
Cant imagine that solder paint will help though. Try a piece of very thin fuse wire a couple of inches long so you can hold it easy. Tack it on to the 2 contacts then break off the excess by bending back and forth or use a small blade to cut it.
Interesting idea, but where do you get fuse wire from? I wonder if folding a few layers of tin foil would work.
Andareed are you running with or without the fan @ 950?
Without. I have two rows of six units lying on their side, top to bottom, with a pair of 120mm fans blowing across them. I've only modded a single unit so far, though 10-15% more heat isn't something I'm going to worry about.
Yeah you are absolutely right. Pencil trick just makes things worse. Even with 850mhz and voltage=1 now I get a lot of HW errors. Time to break out the eraser!
Edit: Erased and back to normal. Zero HW errors @ 850mhz. Damn - now to find someone who can solder a bridge that tiny.
Edit: Erased and back to normal. Zero HW errors @ 850mhz. Damn - now to find someone who can solder a bridge that tiny.
It was really annoying to do. What I did was get a gob onto the general bridge area, and then ran the solder tip (I have a fairly fine tip) along both sides of the bridge to collect any stray solder.
EDIT: Going to try using this to solder the bridge: http://www.ebay.com/itm/151169919397
So just bridge the section you circled? is there a resistor in place you have to remove?
There's no resistor to remove, I just created a solder bridge across the two contacts. Ideally gridseed would have added a header post here, like they did with vid0.
Could this be done via graphite pencil mod?
Possibly, though the pencil mod will decrease resistance, whereas we want the opposite (on the particular resistors I'm looking at). It might be possible to target a different resistor. What this mod is doing is setting vid1 = 1 (solder part) and vid0 = 1 (voltage=1 part). Vid0 and vid1 are connected to a 2-bit VID buck converter, which selects between four output lines (set0, set1, set2, set3). On my miners, the resistance of the resistors connected to these lines are 33, 30, 27, and 36 K (this is different than gridseed's PCB spec!). By default vid0 = 0 and vid1 = 0, so you get 33 K. I experimented with all four vid0/vid1 values and found that power usage (and thus core voltage) is proportional to resistance.
https://github.com/gridseed/usb-miner/blob/master/hardware/GC3355%20USB%20for%20release.pdf?raw=true
| vid0 | vid1 | W | R (K) |
| 0 | 0 | 12.4 | 33 |
| 0 | 1 | 10.4 | 30 |
| 1 | 0 | 11.2 | 27 |
| 1 | 1 | 14.0 | 36 |
wow nice... At 950, what does it hash at? 400 kh/s?
Running stable?
Running stable?
Around 403 kh/s. I let it run for about 8 hours and no hardware errors. Trying 1000 MHz (424 kh/s) now. The solder bridge was a PITA, though I think I've got the technique down and will mod the remaining 11 units I have soon.
EDIT: 1000 MHz produces hardware errors, 975 MHz seems stable.