[Guide] Dogie's Comprehensive ASICMiner Blade Setup - page 10.

alwaysthirsty

newbie

Activity: 13

Merit: 0

Quote from: matt4054 on January 21, 2014, 01:45:55 PM

Quote from: alwaysthirsty on January 21, 2014, 01:18:37 PM

This is whats heppening right now. Strangly the psu fan doesnt move at all. The blade seemed to be powered, because i can access the configuration page. Cant figure out whats wront.

It looks like your chips aren't getting enough power. Status should be like:

Code:

Chip: OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

And you have

Code:

Chip: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

It seems that your PSU is providing enough power for the Ethernet / control logic to work, but not the chips. The fact that the PSU fan is not spinning is probably because your PSU is at very low load (since the Blades aren't hashing).

How did you power the Blades, i.e. how many +12V and GND lines for each Blades, how many Blades and wattage of your PSU?

I tried to wire it as in the OP, molex splitter, two yellow cables going to the + of the blade and four black to the -. I always considered myself somewhat sloppy when it comes to dealing with hardware, it is possible i have messed it up with screwing? there are 4 of these screws on the connecter, what should be their position?

Im using 550W PSU, i have two of those, decided to go with them because they have two 12v lines each. Heres the exact model http://www.super-flower.com.tw/products_detail.php?class=2&sn=12&ID=110&lang=en

Btw status CHIP:OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO is displayed even if not hashing?

matt4054

legendary

Activity: 1946

Merit: 1035

Quote from: alwaysthirsty on January 21, 2014, 01:18:37 PM

This is whats heppening right now. Strangly the psu fan doesnt move at all. The blade seemed to be powered, because i can access the configuration page. Cant figure out whats wront.

It looks like your chips aren't getting enough power. Status should be like:

Code:

Chip: OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

And you have

Code:

Chip: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

It seems that your PSU is providing enough power for the Ethernet / control logic to work, but not the chips. The fact that the PSU fan is not spinning is probably because your PSU is at very low load (since the Blades aren't hashing).

How did you power the Blades, i.e. how many +12V and GND lines for each Blades, how many Blades and wattage of your PSU?

alwaysthirsty

newbie

Activity: 13

Merit: 0

http://i.imgur.com/56Ec95A.png

This is whats heppening right now. Strangly the psu fan doesnt move at all. The blade seemed to be powered, because i can access the configuration page. Cant figure out whats wront.

@matt4054 thanks for the input, much appreciated

matt4054

legendary

Activity: 1946

Merit: 1035

Quote from: broketech on January 21, 2014, 05:02:37 AM

Code:

./mining_proxy.py -o stratum.bitcoin.cz -p 3333 -gp PORT_TO_SET_IN_BLADE_CONF

I think this is the minimum for what you would need on the proxy.

Actually the minimum would just be:

Code:

./mining_proxy.py -o

It will default to accepting getwork on port 8332 and connecting to the specified stratum pool on port 3333. If that's fine for your pools and Blade configs, then you can make it really that short Wink

Quote from: alwaysthirsty on January 20, 2014, 12:11:43 PM

Hey guys. Is it a good idea to cool my blades with standart fan like this one http://www.mh-china.com/UploadFile/2006112010464134010.jpg ? i got 5 blades on hand

The desk fan works very well for me with 4 OC'ed Blades, with an aluminium dead* Macbook Pro as additional heatsink and stand Wink

*dead because I mined with it for over 3 month 24/7, so I would say don't do it!

broketech

member

Activity: 104

Merit: 10

Code:

./mining_proxy.py -o stratum.bitcoin.cz -p 3333 -gp PORT_TO_SET_IN_BLADE_CONF

I think this is the minimum for what you would need on the proxy. Make sure you are pointing at your local ip with the proxy and not slushes pool. You probably know that.

To re state, you should use your worker details from your slush account: user_workerid,pass:user_workerid:pass

alwaysthirsty

newbie

Activity: 13

Merit: 0

Quote from: broketech on January 20, 2014, 02:11:56 PM

Quote

http://www.mh-china.com/UploadFile/2006112010464134010.jpg

Could be better, but I think the general advice is to blow along the direction of the heatsink fins. You might need something with denser airflow.

Quote

in the field "user:pass" you should enter your login details to slush's pool? Like that "sampleuser:password123,sampleuser:password123" ? Not workers' details?

That is where you would put your worker details. sample_1:pass,sample_1:pass, use your worker info. You may need to set up a stratum proxy though, not sure if slush has getwork ports open still.

Its a high rpm fan i hope the airflow would be sufficient in density.

The stratum proxy is running in the background in cmd window. The blade doesnt mine, just configured one as in the OP.Could it be firewall?

Thanks for the advice.

broketech

member

Activity: 104

Merit: 10

Quote

http://www.mh-china.com/UploadFile/2006112010464134010.jpg

Could be better, but I think the general advice is to blow along the direction of the heatsink fins. You might need something with denser airflow.

Quote

in the field "user:pass" you should enter your login details to slush's pool? Like that "sampleuser:password123,sampleuser:password123" ? Not workers' details?

That is where you would put your worker details. sample_1:pass,sample_1:pass, use your worker info. You may need to set up a stratum proxy though, not sure if slush has getwork ports open still.

alwaysthirsty

newbie

Activity: 13

Merit: 0

Hey guys. Is it a good idea to cool my blades with standart fan like this one http://www.mh-china.com/UploadFile/2006112010464134010.jpg ? i got 5 blades on hand

Another newbie question:

in the field "user:pass" you should enter your login details to slush's pool? Like that "sampleuser:password123,sampleuser:password123" ? Not workers' details?

broketech

member

Activity: 104

Merit: 10

Quote

He will likely replace it with a V2 as V1s haven't been made in months.

It is a v2, and I bought it second hand. Followed most of this thread, and was fishing for some kind of technical diagram or schematic. Far as I can tell the logic is off in the reporting/work queuing end of the circuit.

My next step when I get time was to see if the logic chips are playing ball, and to see what kind of negative voltage they can handle.

dogie

legendary

Activity: 1666

Merit: 1185

dogiecoin.com

Quote from: matt4054 on January 18, 2014, 04:44:08 PM

Yep, so I'm afraid your hardware is at fault. With V1 blades you could just replace the logic board, but I guess you'll have to RMA the whole board through friedcat if it still works.

He will likely replace it with a V2 as V1s haven't been made in months.

matt4054

legendary

Activity: 1946

Merit: 1035

Yep, so I'm afraid your hardware is at fault. With V1 blades you could just replace the logic board, but I guess you'll have to RMA the whole board through friedcat if it still works.

broketech

member

Activity: 104

Merit: 10

Code:

3:10:39,660 INFO proxy client_service.handle_event # Setting new difficulty: 7.63793220111
2014-01-18 13:10:39,660 INFO proxy client_service.handle_event # New job 71341695752773876472819158801180146042 for prevhash 4efb41c5, clean_jobs=True
2014-01-18 13:11:02,700 INFO proxy client_service.handle_event # Setting new difficulty: 7.63793220111
2014-01-18 13:11:02,700 INFO proxy client_service.handle_event # New job 46988138074816531316494145925908141405 for prevhash 4957db65, clean_jobs=True

./mining_proxy.py -o 10.0.0.2 -p 8982 -gp 42222 (Miner is 10.0.0.250, machine with proxy/wallet is 10.0.0.2)

This is some sample output from the proxy. When hashed against a wallet is shows a small number of received, but when the proxy is used there is no activity.

When tested with a known good miner it looked similar to your output, with found shares.

matt4054

legendary

Activity: 1946

Merit: 1035

What is the output from mining_proxy?

Normal output while proxying for Blades look like this:

Code:

2014-01-18 22:01:27,772 WARNING proxy getwork_listener._on_submit # [0ms] Share from 'matt4054_blade2' accepted, diff 32
2014-01-18 22:01:27,786 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade4' asks for new work
2014-01-18 22:01:27,789 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade1' asks for new work
2014-01-18 22:01:27,796 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade4' asks for new work
2014-01-18 22:01:27,799 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade1' asks for new work
2014-01-18 22:01:27,815 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade4' asks for new work
2014-01-18 22:01:27,823 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade4' asks for new work
2014-01-18 22:01:27,845 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade3' asks for new work
2014-01-18 22:01:27,854 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade3' asks for new work
2014-01-18 22:01:27,877 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade3' asks for new work
2014-01-18 22:01:27,884 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade3' asks for new work
2014-01-18 22:01:27,891 INFO proxy jobs.submit # Submitting 606fcb89
2014-01-18 22:01:27,891 WARNING proxy getwork_listener._on_submit # [0ms] Share from 'matt4054_blade3' accepted, diff 32
2014-01-18 22:01:27,969 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade3' asks for new work
2014-01-18 22:01:27,990 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade3' asks for new work
2014-01-18 22:01:28,010 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade4' asks for new work
2014-01-18 22:01:28,019 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade4' asks for new work
2014-01-18 22:01:28,035 INFO proxy jobs.submit # Submitting ab7c5c0d
2014-01-18 22:01:28,036 WARNING proxy getwork_listener._on_submit # [0ms] Share from 'matt4054_blade3' accepted, diff 32
2014-01-18 22:01:28,043 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade4' asks for new work
2014-01-18 22:01:28,046 WARNING proxy getwork_listener._on_submit # [473ms] Share from 'matt4054_blade3' accepted, diff 32
2014-01-18 22:01:28,062 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade1' asks for new work
2014-01-18 22:01:28,092 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade1' asks for new work
2014-01-18 22:01:28,125 INFO proxy getwork_listener._on_authorized # Worker 'matt4054_blade1' asks for new work

broketech

member

Activity: 104

Merit: 10

Quote

You should configure slush's mining_proxy locally, make sure it works, and configure your Blade to use it. Then post a screenshot of your configuration page (port 8000) here, just in case it helps.
It is normal behavior for the Blade to reset every 2 min or so when it can't get work from the getwork proxy.

Proxy confirmed to work with cgminer on a separate machine, calling on the getwork port. I knew about the resetting every two minutes.

I've even let it run on the off chance it would just start hashing (fix itself lel).

Too bad there isn't a schematic out there, or the pic c source for the controller.

matt4054

legendary

Activity: 1946

Merit: 1035

Quote from: broketech on January 18, 2014, 03:28:14 PM

Held off on posting until I read the entire thread. I bought my blade second hand as defective, for the challenge of tracing it. It is AM BE REV 2.01 HK, green board.

I have tried mining against a coin wallet and against slush's proxy. I'm no slouch when it comes to networking. It displays all OOOOOOOO for the chips, and resets every two mins or so. I have put tcpdump on it, and the blade requests work (and sometimes even shows in the received area), but does not hash on anything. I have checked the voltages across the board, using values from this thread. The voltages across all the hash lanes are very consistent.
Some of the pins on the pic are sightly off the pads, but it is on the ethernet side of the chip and I have no problem contacting the config page.

For my next step I was going to check the logic chips in the hope that something is keeping the rails from reporting back. It does get warm enough to be hashing, and the led only blinks at boot.

Any further input would be appreciated.

You should configure slush's mining_proxy locally, make sure it works, and configure your Blade to use it. Then post a screenshot of your configuration page (port 8000) here, just in case it helps.

It is normal behavior for the Blade to reset every 2 min or so when it can't get work from the getwork proxy.

broketech

member

Activity: 104

Merit: 10

Held off on posting until I read the entire thread. I bought my blade second hand as defective, for the challenge of tracing it. It is AM BE REV 2.01 HK, green board.

I have tried mining against a coin wallet and against slush's proxy. I'm no slouch when it comes to networking. It displays all OOOOOOOO for the chips, and resets every two mins or so. I have put tcpdump on it, and the blade requests work (and sometimes even shows in the received area), but does not hash on anything. I have checked the voltages across the board, using values from this thread. The voltages across all the hash lanes are very consistent.
Some of the pins on the pic are sightly off the pads, but it is on the ethernet side of the chip and I have no problem contacting the config page.

For my next step I was going to check the logic chips in the hope that something is keeping the rails from reporting back. It does get warm enough to be hashing, and the led only blinks at boot.

Any further input would be appreciated.

hephaist0s

hero member

Activity: 711

Merit: 532

Where can I get more of the little rectangular fuses that go in the later-generation blades?

defcon23

legendary

Activity: 1120

Merit: 1002

no, dont do it.

J_Dubbs

sr. member

Activity: 322

Merit: 250

Quick question about cooling...

I have 9 blades on a backplane, all are running fine and cooling seems good enough from what I can see. I have six 120mm CoolerMaster "Sickleflow" fans pushing air across in one direction (3 push and 3 pull).

I have a 10th blade ordered and I'm not sure the air will push across as well on the end, lighting sucks but the pic below might help illustrate.

my question: should I buy some of those aluminum adhesive heatsinks for the blade on order? Or can I get by without them? Obviously, it will occupy the slot on the end of the backplane.

louis1584

newbie

Activity: 11

Merit: 0

ok I have tried connecting to slushes pool and the bitminter pool using both stratum proxy and bfgminer I have tried turning my firewall off. Any help would be great Thanks

this is what I get the the cmd window
014-01-11 05:24:50,530 ERROR protocol protocol.dataReceived # Processing of mes
sage failed
Traceback (most recent call last):
File "stratum\protocol.pyo", line 185, in dataReceived
File "stratum\protocol.pyo", line 216, in lineReceived
'rotocolException: Cannot decode message 'POST / HTTP/1.1
2014-01-11 05:24:50,535 INFO stats stats.print_stats # 2 peers connected, state
changed 1 times
2014-01-11 05:24:50,540 INFO stats stats.print_stats # 3 peers connected, state
changed 1 times
2014-01-11 05:24:50,552 ERROR protocol protocol.dataReceived # Processing of mes
sage failed
Traceback (most recent call last):
File "stratum\protocol.pyo", line 185, in dataReceived
File "stratum\protocol.pyo", line 216, in lineReceived
'rotocolException: Cannot decode message 'POST / HTTP/1.1
2014-01-11 05:24:50,559 INFO stats stats.print_stats # 2 peers connected, state
changed 1 times
2014-01-11 05:24:50,562 INFO stats stats.print_stats # 3 peers connected, state
changed 1 times
2014-01-11 05:24:50,572 ERROR protocol protocol.dataReceived # Processing of mes
sage failed
Traceback (most recent call last):
File "stratum\protocol.pyo", line 185, in dataReceived
File "stratum\protocol.pyo", line 216, in lineReceived
'rotocolException: Cannot decode message 'POST / HTTP/1.1
over and over and over

Topic: [Guide] Dogie's Comprehensive ASICMiner Blade Setup - page 10. (Read 580771 times)