[DIY] - Reward $100 | Antminer S1/S3 Blade on Raspberry Pi - page 28.

idonothave

full member

Activity: 238

Merit: 100

Quote from: pekatete on October 31, 2014, 02:08:34 PM

I run one board using an el-cheapo CP2102 and WITHOUT a heatsink and got thus:

Obviously, the errors are way out there (due to the lack of a heatsink, I think), but poolside speed was 44 GH/s @ freq 400 (average in an hour) with: -o stratum+tcp://stratum.mypool.mine:3333 -u pekatete -p x --bmsc-freq 0F81
I am currently running with --bmsc-freq 5081 (i.e freq 425)

So I am at a loss as to why some people are only managing a miniscule 12GH/s on a board .... care to share your settings / software / etc ... ?

J4bberwock - way to go! now for the software ..... but seeing what i am getting from a stock cgminer without a heatsink, I am hopeful to be able to squeeze a bit more with one on, which, aside from the fun of it, may be a done deal!

I guess Your blade comes from unit not pensil undervolted, right?

pekatete

hero member

Activity: 518

Merit: 500

I run one board using an el-cheapo CP2102 and WITHOUT a heatsink and got thus:

Obviously, the errors are way out there (due to the lack of a heatsink, I think), but poolside speed was 44 GH/s @ freq 400 (average in an hour) with: -o stratum+tcp://stratum.mypool.mine:3333 -u pekatete -p x --bmsc-freq 0F81
I am currently running with --bmsc-freq 5081 (i.e freq 425)

So I am at a loss as to why some people are only managing a miniscule 12GH/s on a board .... care to share your settings / software / etc ... ?

J4bberwock - way to go! now for the software ..... but seeing what i am getting from a stock cgminer without a heatsink, I am hopeful to be able to squeeze a bit more with one on, which, aside from the fun of it, may be a done deal!

J4bberwock

hero member

Activity: 840

Merit: 1000

Quote from: idonothave on October 31, 2014, 01:46:47 PM

Quote from: J4bberwock on October 31, 2014, 05:30:48 AM

And my best results are with this command line:
cgminer.exe -o stratum+tcp://thepool:9999 -u Jabberwock -p x --bmsc-freq 0F81

giving the following result:

Code:

Block: 146edd3a...  Diff:36G  Started: [11:25:07]  Best share: 12.4K
-------------------------------------------------------------------------------
[P]ool management [S]ettings [D]isplay options [Q]uit
AMU 0:                | 47.37G/42.72Gh/s | A:1248 R:0 HW:744 WU:369.0/m
AMU 1:                | 45.07G/42.26Gh/s | A: 992 R:0 HW:711 WU:372.6/m
-------------------------------------------------------------------------------
[2014-10-31 11:25:20] Recv 22 freq getstatus=801f00010b
[2014-10-31 11:27:38] Accepted 629e10e6 Diff 665/64 AMU 0
[2014-10-31 11:27:49] Accepted 03c799d7 Diff 68/64 AMU 0
[2014-10-31 11:27:53] Accepted 02148e3a Diff 123/64 AMU 0

The pool is getting nearly 50Gh on each panel, 100Gh stable, it's slowly going up.

Edit,
I'll try the chaining this afternoon with a second S1 that is supposed to be dead, just in case it was the control board that failed.
But first, I'll need reflowing the same CP2102 as on the first board since only 1 is detected.

Correct me, but it seems there are ~60% of errors at AMU0 and ~70% at AMU1. It is not 42 at the end but 16 resp 12 GHs at pool stats.

You should be right about the pool stats, and they should be even worst since the accepted are multiplied by the difficulty, and the HW errors, but for some reasons, I'm getting 90-105 GH on multipool.
And it's my only miner on this pool.

idonothave

full member

Activity: 238

Merit: 100

Quote from: J4bberwock on October 31, 2014, 05:30:48 AM

And my best results are with this command line:
cgminer.exe -o stratum+tcp://thepool:9999 -u Jabberwock -p x --bmsc-freq 0F81

giving the following result:

Code:

Block: 146edd3a...  Diff:36G  Started: [11:25:07]  Best share: 12.4K
-------------------------------------------------------------------------------
[P]ool management [S]ettings [D]isplay options [Q]uit
AMU 0:                | 47.37G/42.72Gh/s | A:1248 R:0 HW:744 WU:369.0/m
AMU 1:                | 45.07G/42.26Gh/s | A: 992 R:0 HW:711 WU:372.6/m
-------------------------------------------------------------------------------
[2014-10-31 11:25:20] Recv 22 freq getstatus=801f00010b
[2014-10-31 11:27:38] Accepted 629e10e6 Diff 665/64 AMU 0
[2014-10-31 11:27:49] Accepted 03c799d7 Diff 68/64 AMU 0
[2014-10-31 11:27:53] Accepted 02148e3a Diff 123/64 AMU 0

The pool is getting nearly 50Gh on each panel, 100Gh stable, it's slowly going up.

Edit,
I'll try the chaining this afternoon with a second S1 that is supposed to be dead, just in case it was the control board that failed.
But first, I'll need reflowing the same CP2102 as on the first board since only 1 is detected.

Correct me, but it seems there are ~60% of errors at AMU0 and ~70% at AMU1. It is not 42 at the end but 16 resp 12 GHs at pool stats.

J4bberwock

hero member

Activity: 840

Merit: 1000

Quote from: CHAOSiTEC on October 31, 2014, 10:13:40 AM

a board with connectors for 8 boards would be sweet, since i have 20 blades laying around, and the fewer boards the better :-p

ohh and mounting holes :-p

CP2108 will do it.
and mounting holes should be added for the definitive boards.

I've tested chaining, and it works as expected.
The second board still have some issues with the USB port and the second cp2102 isn't working.
Edit: I simply forgot to add a wire for the 5v supply since the TPS2044 isn't soldered. I now have 3 fully working boards.

CHAOSiTEC

legendary

Activity: 1358

Merit: 1002

Quote from: J4bberwock on October 31, 2014, 05:46:52 AM

It will obviously need a case, mstrongbow will get 1 prototype to make it.

a board with connectors for 8 boards would be sweet, since i have 20 blades laying around, and the fewer boards the better :-p

ohh and mounting holes :-p

J4bberwock

hero member

Activity: 840

Merit: 1000

It will obviously need a case, mstrongbow will get 1 prototype to make it.

J4bberwock

hero member

Activity: 840

Merit: 1000

And my best results are with this command line:
cgminer.exe -o stratum+tcp://thepool:9999 -u Jabberwock -p x --bmsc-freq 0F81

giving the following result:

Code:

Block: 146edd3a...  Diff:36G  Started: [11:25:07]  Best share: 12.4K
-------------------------------------------------------------------------------
[P]ool management [S]ettings [D]isplay options [Q]uit
AMU 0:                | 47.37G/42.72Gh/s | A:1248 R:0 HW:744 WU:369.0/m
AMU 1:                | 45.07G/42.26Gh/s | A: 992 R:0 HW:711 WU:372.6/m
-------------------------------------------------------------------------------
[2014-10-31 11:25:20] Recv 22 freq getstatus=801f00010b
[2014-10-31 11:27:38] Accepted 629e10e6 Diff 665/64 AMU 0
[2014-10-31 11:27:49] Accepted 03c799d7 Diff 68/64 AMU 0
[2014-10-31 11:27:53] Accepted 02148e3a Diff 123/64 AMU 0

The pool is getting nearly 50Gh on each panel, 100Gh stable, it's slowly going up.

Edit,
I'll try the chaining this afternoon with a second S1 that is supposed to be dead, just in case it was the control board that failed.
But first, I'll need reflowing the same CP2102 as on the first board since only 1 is detected.

J4bberwock

hero member

Activity: 840

Merit: 1000

Code:

[2014-10-31 11:02:12] -----------------start freq-------------------
[2014-10-31 11:02:12] Send frequency 820f8100
[2014-10-31 11:02:13] Send freq getstatus 84000400
[2014-10-31 11:02:15] ------recv freq getstatus no data finish------
[2014-10-31 11:02:16] -----------------start nonce------------------
[2014-10-31 11:02:16] Bmsc send golden nonce
[2014-10-31 11:02:16] Bmsc recv golden nonce timeout

It will fail this way uless you start cgminer first and when it's waiting for usb device, you plug the antminer.
then, you have this:

Code:

 Block: 132534a4...  Diff:36G  Started: [11:04:19]  Best share: 953
-----------------------------------------------------------------------------
 [P]ool management [S]ettings [D]isplay options [Q]uit
 AMU 0:                | 1.647G/3.013Gh/s | A:64 R:0 HW:20 WU: 28.7/m
 AMU 1:                | 3.414G/2.346Gh/s | A:64 R:0 HW: 1 WU: 30.5/m
-----------------------------------------------------------------------------
 [2014-10-31 11:04:34] AMU0: invalid nonce - HW error
 [2014-10-31 11:04:34] AMU0: invalid nonce - HW error
 [2014-10-31 11:04:34] AMU0: invalid nonce - HW error
 [2014-10-31 11:04:34] AMU0: invalid nonce - HW error
 [2014-10-31 11:04:34] AMU0: invalid nonce - HW error
 [2014-10-31 11:05:06] Accepted d575eeb9 Diff 307/64 AMU 0
 [2014-10-31 11:05:12] Accepted 44c7041c Diff 953/64 AMU 1
 [2014-10-31 11:05:55] AMU1: invalid nonce - HW error
 [2014-10-31 11:06:34] AMU0: invalid nonce - HW error
 [2014-10-31 11:06:41] AMU0: invalid nonce - HW error
 [2014-10-31 11:07:03] Stratum connection to pool 0 interrupted
 [2014-10-31 11:07:03] Pool 0 difficulty changed to 16
 [2014-10-31 11:07:03] Pool 0 difficulty changed to 64

J4bberwock

hero member

Activity: 840

Merit: 1000

Quote from: CHAOSiTEC on October 31, 2014, 04:16:16 AM

Quote from: sobe-it on October 30, 2014, 09:15:13 AM

My one blade has been running for days with out any heatsink on the back. The chips just have the little stick on ones. 40-50 ghs (solo mining) using ~60 watts. It seems the hashing speed depends on the clock/chip command. The blade is undervolted, still trying ti figure out how my one blade fried.

could you enlighten us on the resistor value you choose for this, and your cmd line?

I'm not sure the solomining can give accurate value unless you have something else than cgminer to read the hashrate

J4bberwock

hero member

Activity: 840

Merit: 1000

Just tested the PWM control, and it works perfectly.
Fan speed can now be adjusted with the trim pot.

Currently replacing the silabs driver with zadig to test the hashing

CHAOSiTEC

legendary

Activity: 1358

Merit: 1002

Quote from: sobe-it on October 30, 2014, 09:15:13 AM

My one blade has been running for days with out any heatsink on the back. The chips just have the little stick on ones. 40-50 ghs (solo mining) using ~60 watts. It seems the hashing speed depends on the clock/chip command. The blade is undervolted, still trying ti figure out how my one blade fried.

could you enlighten us on the resistor value you choose for this, and your cmd line?

powersup

sr. member

Activity: 406

Merit: 250

Quote from: J4bberwock on October 30, 2014, 05:15:48 PM

Quote from: abzalus on October 30, 2014, 03:50:26 PM

WOW what a cool thing!!! When are you starting to taking orders! Hahaha but the real question is how much this board can value?

They should be 15-20$
I'll check the price of the parts, but the fixed prototypes should be something like 10$.

I just tested chaining, and it seems to work.
Tomorrow, I'll test the PWM on the fan.

nice work J4bberwock. I'm keen to get my hands on a prototype.

J4bberwock

hero member

Activity: 840

Merit: 1000

Quote from: abzalus on October 30, 2014, 03:50:26 PM

WOW what a cool thing!!! When are you starting to taking orders! Hahaha but the real question is how much this board can value?

They should be 15-20$
I'll check the price of the parts, but the fixed prototypes should be something like 10$.

I just tested chaining, and it seems to work.
Tomorrow, I'll test the PWM on the fan.

tarmi

legendary

Activity: 1232

Merit: 1011

J4bberwock count me in for one board, thank you.

HerbPean

legendary

Activity: 1638

Merit: 1005

Quote from: J4bberwock on October 30, 2014, 03:48:42 PM

Quote from: zziggi on October 30, 2014, 01:01:12 PM

Use 3.8.5 from bitmaintech. https://github.com/AntMiner/AntGen1/blob/master/cgminer/cgminer-run-windows-20131224.zip?raw=true

I'll try it again with my board instead of the cheap CP2102 from china.
BTW, after reflowing, both CP2102 are detected correctly.
I'll solder a second one tomorrow and test the chaining before sending them to CHAOSITEC and pekatete.
I'll have a few spare boards I can sell for the price of the parts + shipping, if someone is interested.
I ordered 10 PCB, and I currently have the components to make at least 6 boards.

I want one ! Cheesy

abzalus

newbie

Activity: 3

Merit: 0

J4bberwock

hero member

Activity: 840

Merit: 1000

Quote from: zziggi on October 30, 2014, 01:01:12 PM

Use 3.8.5 from bitmaintech. https://github.com/AntMiner/AntGen1/blob/master/cgminer/cgminer-run-windows-20131224.zip?raw=true

I'll try it again with my board instead of the cheap CP2102 from china.
BTW, after reflowing, both CP2102 are detected correctly.
I'll solder a second one tomorrow and test the chaining before sending them to CHAOSITEC and pekatete.
I'll have a few spare boards I can sell for the price of the parts + shipping, if someone is interested.
I ordered 10 PCB, and I currently have the components to make at least 6 boards.

idonothave

full member

Activity: 238

Merit: 100

Quote from: zziggi on October 30, 2014, 01:01:12 PM

Use 3.8.5 from bitmaintech. https://github.com/AntMiner/AntGen1/blob/master/cgminer/cgminer-run-windows-20131224.zip?raw=true

Thank You. I used https://github.com/AntMiner/AntGen1/blob/master/cgminer/cgminer-run-linux and it makes ~42GHs / 72W. I have no idea about temperature, but it is not very hot, I use notebook cooling desk as a cooler. I had a lot of fun with that. It is not the most efficient miner I have ever seen but ugliest for sure

HerbPean

legendary

Activity: 1638

Merit: 1005

Nice ! Can't wait to get one to play with

Topic: [DIY] - Reward $100 | Antminer S1/S3 Blade on Raspberry Pi - page 28. (Read 82179 times)