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

[idonothave]

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.

Can You share what vendor id device did You use and what version of cgminer? I have never reached more than 4GHs.

[zziggi]

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

[J4bberwock]

Quick update,
here is the homemade stencil before etching.



And here is the soldered board after I fixed the minor issues. The TPS2044 has been bypassed.




The BOM for this one has been reduced to the following:

ref	value	Qty
C1	680pF	1
C2	0,01µF	1
C3	100µF	1
C4	0,1µF	1
C5	10µF	1
C6,C7	33pF	2
C8	4,7µF	1
C9,C10	1µF	2
C11	0,1µF	1
CON1	USB-MINI-B	1
D1,D2	4148	2
IC1	NE556	1
IC2,IC3	CP2102	2
J1	USB_2	1
P1,P2	CONN_8	2
R1	1k	1
R2	10k	1
R3	1,5k	1
R4..R11	27	8
R12..R15	15k	4
R16	2,2k	1
R17	15k	1
RV1	100k	1
U1,U2	SN75240	2
U3	TUSB2036	1
X1	6MHz	1

The definitive board will use one or two CP2108 to allow up to 4 or 8 hashing units.

[HerbPean]

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

[idonothave]

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

[J4bberwock]

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.

[abzalus]

Quick update,
here is the homemade stencil before etching.

https://i.imgur.com/PZuPrZs.jpg?1

And here is the soldered board after I fixed the minor issues. The TPS2044 has been bypassed.

https://i.imgur.com/795EYCu.jpg?1


The BOM for this one has been reduced to the following:

ref	value	Qty
C1	680pF	1
C2	0,01µF	1
C3	100µF	1
C4	0,1µF	1
C5	10µF	1
C6,C7	33pF	2
C8	4,7µF	1
C9,C10	1µF	2
C11	0,1µF	1
CON1	USB-MINI-B	1
D1,D2	4148	2
IC1	NE556	1
IC2,IC3	CP2102	2
J1	USB_2	1
P1,P2	CONN_8	2
R1	1k	1
R2	10k	1
R3	1,5k	1
R4..R11	27	8
R12..R15	15k	4
R16	2,2k	1
R17	15k	1
RV1	100k	1
U1,U2	SN75240	2
U3	TUSB2036	1
X1	6MHz	1

The definitive board will use one or two CP2108 to allow up to 4 or 8 hashing units.

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

[HerbPean]

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 !

[tarmi]

J4bberwock count me in for one board, thank you.

[J4bberwock]

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.

[powersup]

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.

[CHAOSiTEC]

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?

[J4bberwock]

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

[J4bberwock]

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]

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]

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]

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

[CHAOSiTEC]

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]

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.

[idonothave]

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.