Topic: Setting up my first *ghetto* mining rig!! (Read 1835 times)

Good news. Tonight the rig did not shut down. It's still going. Almost 24 hours now. So hopefully I did solve the problem!! ETH price keeps going up. I must expand soon :p

To be clear what means DIY'ing a server PSU: you have to cut wire into into cords, crimp all terminals (and solder - for better contact), slap on 6/6+2 connectors and solder the cabling to PSU's output, or use breakout with terminal blocks.
Parts can be bought cheap from: China (terminals, connectors, PicoPSU) and the electrical supply store (wire).

Time wise, it takes from 2 to 8 hours to fully DIY one PSU, depending on the options. I understand your friend wants to help you, but modding 17 PSUs - that's rather part-time job, than friendly help  
Doing everything wisely decreases the total cost by a good margin.

PM me if you need specific build details and purchase instructions. Tip required  

modification is not a problem, I have also a friend who is electronics technician and he can do the work for me for free. The problem is that pico psu is as expensive as a separate cheap psu for the motherboard and also the pci-e cables/plugs for the breakout boards also cost a lot. Do you also have any source where I can get just the 6+2 plugs and the wires separately cheaper than the ready made ones from ebay? For 17 rigs this will cut down a lot of the cost.

Server PSUs are the best choice when you are able to DIY and source cheap components. A fully DIY wired PSU w/Pico will cost ~85-100$ in material (DPS-1200FB), labor cost - YMMV.
Frankly DIY seems to be not your case, so better stick with ATX power supplies.

You'll find cheap breakout boards only for DPS-1200FB and the likes. Because they are Chinese.
There exist breakout boards for beefier power supplies - DPS-2000BB, IBM 2980W, etc. They cost >40 USD.
Link: http://www.gekkoscience.com/products/IBM2K_board.html

I will reiterate - if you feel unsure to modify server PSUs, go for decent ATX PSUs.

Yesterday the rig worked for 12 hours straight. New record, but still something happened while I was asleep.

04:55:12:889   1188   ETH: Share accepted (94 ms)!

04:55:12:889   1188   new buf size: 0
04:55:12:920   180c   ADL: Cannot find any AMD adapters
04:55:12:920   180c   em hbt: 0, fm hbt: 47,
04:55:12:920   180c   watchdog - thread 0 (gpu0), hb time 328
04:55:12:936   180c   watchdog - thread 1 (gpu0), hb time 140
04:55:12:936   180c   watchdog - thread 2 (gpu1), hb time 171
04:55:12:936   180c   watchdog - thread 3 (gpu1), hb time 343
04:55:12:936   180c   watchdog - thread 4 (gpu2), hb time 46
04:55:12:952   180c   watchdog - thread 5 (gpu2), hb time 250
04:55:12:952   180c   watchdog - thread 6 (gpu3), hb time 234
04:55:12:952   180c   watchdog - thread 7 (gpu3), hb time 47
04:55:12:952   180c   watchdog - thread 8 (gpu4), hb time 47
04:55:12:967   180c   watchdog - thread 9 (gpu4), hb time 282
04:55:12:967   180c   watchdog - thread 10 (gpu5), hb time 297
04:55:12:967   180c   watchdog - thread 11 (gpu5), hb time 78
04:55:13:749   1188   ETH: checking pool connection...
04:55:13:749   1188   send: {"worker": "", "jsonrpc": "2.0", "params": [], "id": 3, "method": "eth_getWork"}

04:55:13:842   1188   got 248 bytes
04:55:13:842   1188   buf: {"id":3,"jsonrpc":"2.0","result":["0xdebce9c6359ad9a76860c3d61eeacfa9468dec29c19096eb8575a3b49123be79","0x9d2e3fc27256471aefd0ac060863e72fa7ca01ae239f872242f7c9797c1eb775","0x0112e0be826d694b2e62d01511f12a6061fbaec8bc02357593e70e52ba","0x4898b9"]}

04:55:13:842   1188   parse packet: 247
04:55:13:842   1188   ETH: job is the same
04:55:13:842   1188   new buf size: 0
04:55:17:530   180c   ADL: Cannot find any AMD adapters
04:55:18:077   2248   recv: 60
04:55:18:093   2248   srv pck: 59
04:55:19:718   2248   ADL: Cannot find any AMD adapters
04:55:19:718   2248   srv bs: 0
04:55:19:734   2248   sent: 190
04:55:22:343   180c   ADL: Cannot find any AMD adapters
04:55:23:765   1188   ETH: checking pool connection...
04:55:23:765   1188   send: {"worker": "", "jsonrpc": "2.0", "params": [], "id": 3, "method": "eth_getWork"}

04:55:24:578   1188   send: {"id":6,"jsonrpc":"2.0","method":"eth_submitHashrate","params":["0x7ce67c5", "0x000000000000000000000000000000000000000000000000000000006b073ac5"]}

04:55:26:672   180c   ADL: Cannot find any AMD adapters
04:55:32:797   180c   ADL: Cannot find any AMD adapters
04:55:33:782   1188   ETH: checking pool connection...
04:55:34:688   1188   send: {"worker": "", "jsonrpc": "2.0", "params": [], "id": 3, "method": "eth_getWork"}

04:55:41:814   180c   ADL: Cannot find any AMD adapters
04:55:48:362   1188   ETH: Stratum - socket send failed 10054, disconnect
04:55:59:738   180c   ADL: Cannot find any AMD adapters

This is my bat. file:

setx GPU_FORCE_64BIT_PTR 0
setx GPU_MAX_HEAP_SIZE 100
setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_SINGLE_ALLOC_PERCENT 100
EthDcrMiner64.exe -epool eu1.ethermine.org:4444 -ewal 0xD2F39eC8d4161019428395a8136793BC4E3c9C33.Rig1 -epsw x

BTW my meter came. Corsair link wattage indication is pretty accurate!

Also I downloaded the new version of claymore. They say it fixes some stuff. Hope it helps.

- fixed critical issues in remote management feature (attacker could crash miner even in read-only mode).
- now miner supports up to #299 epoch.
- in rare cases ADL API calls can hang, now watchdog checks it as well.

Also, my old HP mouse died somehow. From the cold?

I watched some more tutorials on how to setup claymore. I saw one guy that set the variables on system variables on windows. I did that too now. I downloaded and installed the beta drivers again. And I started it again 20 minutes ago. The "ADL: Cannot find any AMD adapters" thingy that was all over the place in all the logs, now it's gone. Haven't seen it once in these 20 minutes. I have a good feeling about this now.

ok, so I am ready to order all the appropriate parts but as far as I understand, to make each of my rig work I need: a separate cheap 300w standard atx psu (about $20) just for the motherboard, a DPS-1200FB ($69) the cheapest I can find now on ebay (which is actually the DPS-1200FB-1, dont know if its the same) https://www.ebay.com/itm/HP-1200W-Power-Supply-578322-B21-579229-001-570451-001-570451-101-DPS-1200FB-1/232579507890?hash=item3626d1a6b2:g:81IAAOSwqiVaHttv to power 4x r9 290/390 gpus on each rig, a breakout board ($8.19) and the pci-e cables ($29.99) . Or instead of the standard atx psu, a pico psu adapter ($49) to power the motherboard from the server psu. Am I correct?
So, if I sum it all, the cost isn't so low as I  initially thought.  Its roughly  $127. On newegg I can find some 1200W standard atx psus for the same amount of money and I will also have warranty. So whats the point to go for the server psu? https://www.newegg.com/Product/Product.aspx?Item=9SIA2J53KH0075
I need to setup about 16-17 rigs with 4x gpus each. So I am looking to keep the psu cost as low as possible but also to have reliability. Maybe there are some good server psu alternatives that are even cheaper and will do the job just fine or this is the cheapest I can go for my use?

For example I just found this monster on ebay for less cost than the DPS-1200FB. https://www.ebay.com/itm/Dell-PowerEdge-M1000e-Server-2700W-Switching-Power-Supply-E2700P-00-TJJ3M/302565706413?epid=21005587000&hash=item4672524aad:g:sDUAAOSw8UZaMw92
With this one I could feed 2 or 3 rigs at the same time, But are there any breakout boards for this? I cant find any

I installed Sapphire Trixx instead of MSI afterburner because I was thinking maybe that program was the problem. :/
Why don't use use the blockchain ones?

I'm using 16.11.5 drivers for 290 series@W10, they just don't like drivers later than that. Basic DDU + reinstall should help, MSI afterburner also needs some settings to be enabled.

@QuintLeo Yes that plate is annoying, wonder why someone who disassembles ref cooler doesn't get rid of it. I mean, you can always put it back...

That message, is throughout all of my logs. All the time. :/ Right now it's running 8 hours straight after completely disabling any windows updates from registry. Previously I had it, "ask me when to download". But that "ADL: Cannot find any AMD adapters" message is still there all the time. Tomorrow if it keeps running overnight I''m gonna go back to the blockchain drivers. I'm also gonna pull an Ethernet cable to make my connection 100% stable.

 R9 290 is generally going to make you more money mining ETH, where it is a legitimate 29-30 Mhash card but will easily pull 250-300 watts per card to get to that hashrate in most cases.
 I'm not sure what kind of power usage they have mining ZEC as I never had mine doing that for more than a very short period of time during the EARLY ZEC days last year.

 3 of those Bladecenter PS should be plenty, if you can find the right breakout boards for them to make them usable.

 2 would probably be viable, they should be able to handle about 8 cards each PCI-E AND riser - depends on how big an ATX supply you use to run the MB itself and how many of the cards the ATX supply can run.

 About the first "mod" any high-end blower model card with a bloody DVI connector should do is to pull that bloody airflow blocking mounting plate OFF (if it's in a riser rig and doesn't use the mounting holes part).

 Helps a lot on 1080 blower models too, though my ASUS 1080ti blower card doesn't have a DVI connector and doesn't really need the mod.

 Also helped noticeably on the one Sapphire "mining" card I got that had the congested "original" type mounting plate instead of the later models that opened up that mounting plate.


 "Cannot find any AMD adapters" is probably a DRIVER crash that took more than a fractional second to "recover".

Stick with DPS-1200FB, it's the most suitable PSU for your needs, IMO. There is plenty of breakout boards available for this model.

will this one handle 4x r9 290 or 4x r9 390?
https://www.ebay.com/i/122841705413?rt=nc
[/quote]
It's not the best PSU you could find. Go for Delta DPS-1200 (any modification that has only +12V output is good).
PicoPSU - 180/250 W
Breakout - at least with 3 oz copper, look for it to have beefy connectors
I won't advise any cables, many of them are crap, especially the cheap ones.

Is this the same as the standart dps-1200? https://www.ebay.com/itm/Delta-700-1200W-Power-Supply-DPS-1200ABA/331650462356?hash=item4d37e87294:g:Fq8AAOSwZ4dZMEdk
there are many variations dps-1200fb, dps-1200TB etc. Are all the same? also the breakout board fits all of them or there is specific designed breakout board for each variation?

It's not the best PSU you could find. Go for Delta DPS-1200 (any modification that has only +12V output is good).
PicoPSU - 180/250 W
Breakout - at least with 3 oz copper, look for it to have beefy connectors
I won't advise any cables, many of them are crap, especially the cheap ones.

This error has nothing to do with the internet connection. Check for driver issues.

Today it stayed online for 8+ hours. And it didn't crash it it restarted for some reason. This is the logs before restart.

5:48:371   de4   ETH: GPU0 27.028 Mh/s, GPU1 27.086 Mh/s, GPU2 27.721 Mh/s, GPU3 10.925 Mh/s, GPU4 10.912 Mh/s, GPU5 26.094 Mh/s
04:15:49:262   444   ADL: Cannot find any AMD adapters
04:15:52:763   444   ADL: Cannot find any AMD adapters
04:15:53:606   1a34   recv: 60
04:15:53:606   1a34   srv pck: 59
04:15:53:810   1a34   ADL: Cannot find any AMD adapters
04:15:53:810   1a34   srv bs: 0
04:15:53:810   1a34   sent: 188
04:15:56:232   444   ADL: Cannot find any AMD adapters
04:15:56:732   de4   ETH: checking pool connection...
04:15:56:732   de4   send: {"worker": "", "jsonrpc": "2.0", "params": [], "id": 3, "method": "eth_getWork"}

04:15:59:998   444   ADL: Cannot find any AMD adapters
04:16:04:326   de4   send: {"id":6,"jsonrpc":"2.0","method":"eth_submitHashrate","params":["0x0", "0x000000000000000000000000000000000000000000000000000000007cc1fc2c"]}

04:16:04:483   444   ADL: Cannot find any AMD adapters
04:16:07:061   de4   ETH: checking pool connection...
04:16:08:249   de4   send: {"worker": "", "jsonrpc": "2.0", "params": [], "id": 3, "method": "eth_getWork"}

04:16:10:374   444   ADL: Cannot find any AMD adapters
04:16:14:609   444   ADL: Cannot find any AMD adapters
04:16:15:515   f6c   recv: 60
04:16:16:296   f6c   srv pck: 59
04:16:17:843   f6c   ADL: Cannot find any AMD adapters
04:16:19:609   de4   Socket was closed remotely (by pool)
04:16:19:093   444   ADL: Cannot find any AMD adapters
04:16:19:219   f6c   srv bs: 0
04:16:27:704   de4   ETH: Connection lost, retry in 20 sec...
04:16:25:625   f6c   sent: 160
04:16:30:407   444   ADL: Cannot find any AMD adapters
04:16:51:128   444   ADL: Cannot find any AMD adapters
04:16:51:534   444   em hbt: 33204, fm hbt: 0,
04:16:51:941   444   watchdog - thread 0 (gpu0), hb time 60047
04:16:53:503   444   WATCHDOG: GPU 0 hangs in OpenCL call, exit
04:16:54:144   444   watchdog - thread 1 (gpu0), hb time 62031
04:16:54:160   444   WATCHDOG: GPU 0 hangs in OpenCL call, exit
04:16:54:175   444   watchdog - thread 2 (gpu1), hb time 62109
04:16:54:175   444   WATCHDOG: GPU 1 hangs in OpenCL call, exit
04:16:54:191   444   watchdog - thread 3 (gpu1), hb time 61891
04:16:54:191   444   WATCHDOG: GPU 1 hangs in OpenCL call, exit
04:16:54:238   444   watchdog - thread 4 (gpu2), hb time 62188
04:16:54:347   444   WATCHDOG: GPU 2 hangs in OpenCL call, exit
04:16:54:832   444   watchdog - thread 5 (gpu2), hb time 62532
04:16:54:910   444   WATCHDOG: GPU 2 hangs in OpenCL call, exit
04:16:55:363   444   watchdog - thread 6 (gpu3), hb time 63188
04:16:55:426   444   WATCHDOG: GPU 3 hangs in OpenCL call, exit
04:16:55:660   444   watchdog - thread 7 (gpu3), hb time 63672
04:16:55:957   444   WATCHDOG: GPU 3 hangs in OpenCL call, exit
04:16:56:004   444   watchdog - thread 8 (gpu4), hb time 63829
04:16:56:019   444   WATCHDOG: GPU 4 hangs in OpenCL call, exit
04:16:56:176   444   watchdog - thread 9 (gpu4), hb time 64234
04:16:56:316   444   WATCHDOG: GPU 4 hangs in OpenCL call, exit
04:16:56:426   444   watchdog - thread 10 (gpu5), hb time 64562
04:16:56:566   444   WATCHDOG: GPU 5 hangs in OpenCL call, exit
04:16:56:566   444   watchdog - thread 11 (gpu5), hb time 64500
04:16:56:597   444   WATCHDOG: GPU 5 hangs in OpenCL call, exit
04:16:55:301   b3c   recv: 60
04:16:56:676   b3c   srv pck: 59
04:16:56:894   444   ADL: Cannot find any AMD adapters
04:16:57:238   b3c   ADL: Cannot find any AMD adapters
04:16:57:582   b3c   srv bs: 0
04:16:57:785   b3c   sent: 138
04:17:01:551   444   Restarting OK, exit...

This is the next log:

04:17:15:959   128   Check and remove old log files...
04:17:16:615   128   args: -epool eu1.ethermine.org:4444 -ewal 0xD2F39eC8d4161019428395a8136793BC4E3c9C33.Rig1 -epsw x
04:17:16:772   128   
04:17:17:068   128   ÉÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ»
04:17:17:147   128   º     Claymore's Dual ETH + DCR/SC/LBC/PASC GPU Miner v10.1      º
04:17:17:240   128   ÈÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍͼ
04:17:17:334   128   
04:17:17:678   128   ETH: 1 pool is specified
04:17:17:709   128   Main Ethereum pool is eu1.ethermine.org:4444
04:17:17:787   128   DCR: 0 pool is specified
04:17:50:760   948   
04:18:25:154   948   
04:19:07:111   948   
12:38:43:859   948   Miner cannot initialize for 5 minutes, need to restart miner!
12:38:43:874   948   ADL: Cannot find any AMD adapters
12:38:45:975   948   Restarting OK, exit...

Then I found it restarted, working, waiting on the sign in page.
I cannot find out why it restarted. Today I'm gonna connect it with Ethernet. I found a cable long enough.

this particular motherboard I've found has all 4 pci-e functional simultaneously so its ok.
Can you suggest a good and cheap 1200w server psu, a picoPSU and an appropriate breakout board with cables included to order them all together from ebay? There are too many for sale and I am lost.

will this one handle 4x r9 290 or 4x r9 390?
https://www.ebay.com/i/122841705413?rt=nc

It happens the miner to crash the rig if Internet connection is lost. Buy a better quality Wi-Fi receiver if you can't wire the miner to LAN.
Post the last lines of miner logs here.

I can't find out why it keeps crashing after a while. If you have any idea what to check or what to try after let me know. I'm all out of ideas! Maybe a different driver version or something. But which one. I've tried latest and beta drivers, OC or stock speeds, and I've been monitoring it, it's not overheating. It's somewhat random but I've noticed a couple of patterns like both nights crashing at 3:15-20am. But could be a coincidence.
Could it be a problematic PSU? Hope not.

Edit: I was just right infront of a crash. My wifi had an interruption and the reconnecting in 20secs message appeared in claymore a couple of times. As cards were slowing down it froze and crashed. I know wifi for a miner is not ideal but where is is located connecting it with a cable is very hard. I was thinking of a wifi repeater and then ethernet cable. I hope maybe that will fix the interruptions and maybe the crashes?

Even if a 775 MB has >4 PCI-E slots, not all of them may be functional simultaneously. Look for higher-end P45 models (Asus P5Q-E for example). You can slap any CPU, but preferably a dual-core and 4GB RAM.
The server PSU delivers +12V voltage to +12V consumers: GPUs, additional fans. You can power the ATX standard consumers - MB, HDD/SSD, through a DC-DC converter (picoPSU) or using a separate ATX PSU (you'll need to synchronize the server PSU with the ATX one).
You have 3 connection options for server PSU:
1. solder a couple bunch of power cables to PSU's output and split cables to each consumer
2. use a breakout board and connect consumers through detachable cables
3. solder beefy power cords to PSU's output and distribute to each consumer through a terminal block

Maybe I will go for another motherboard because I am searching all over the internet now and I cant find this model in stock.
So I think its better idea to stick with some cheap 775 socket motherboards with at least four pci-e slots each.
I have found some very cheap with ram and cpu included under 40 euros.
So do I need a 1200watt server psu for each of these motherboards? is this ok for 4x r9 290 overclocked or I need more capacity?
also can anyone tell me how these server psus work? I see they attach pci-e cables on them for the gpus (through a custom adapter) but there is no atx plug or cpu plug for the motherboard.