,but using that miner pretty much requires 'free' electricity nowadays.
Topic: Bitcrane T-110 miners - using damaged miners to build into working miners (Read 474 times)
April 23, 2018, 10:13:37 AM
That's a cool story hardware -wise ,
but using that miner pretty much requires 'free' electricity nowadays.
,but using that miner pretty much requires 'free' electricity nowadays.
April 23, 2018, 10:10:50 AM
Hello, last month I bought to electric-tech on eBay for $190 a BitCrane T110 which had both watercooling dead. I replace them with "Cooler Master Nepton 120XL" which are compatible with "Coolit". I replaced the thermal compound with "Thermal grizzly Conductonaut" after having insulated the components around the chip with Neutral Silicon compound. With the frequency set to 725 on /usr/local/bin/page/config/pools.conf.json the T110 run above 1,1 TH/s with the temperature arount 73°C. Then I replaced the noisy fan with the Cooler Master low noise ones and reduced the frequency to 500 Mhz.
Now it run by not making more noise than a PC and gives between 900 GH/s to 1010 GH/s at 77°C which is not bad.
OK now there are Avalon 841 which gives 13 TH/s and some other which gives up to 18 TH/s but there are also as noisy as Fireman Siren.
I think you can get some money by putting them on eBay. There are certainly still a few people who would be delighted to buy them.
Best Regards and good mining !
machproo
Now it run by not making more noise than a PC and gives between 900 GH/s to 1010 GH/s at 77°C which is not bad.
OK now there are Avalon 841 which gives 13 TH/s and some other which gives up to 18 TH/s but there are also as noisy as Fireman Siren.
I think you can get some money by putting them on eBay. There are certainly still a few people who would be delighted to buy them.
Best Regards and good mining !
machproo
December 05, 2017, 09:52:47 AM
Glad to see this (and Dogie's) thread so I don't have to necrobump anything.
I'm in a very similar situation, as I wanna get a bunch of old T-110 units I have lying around working.
Kinda silly, sure, but if people are selling 333mhz block erupters for $100 as "Collectables" then I can try bringing life into these old boards.
The problem I had was the web interface wouldn't initialise cards, if they would it wouldn't be for very long.
I have (had) 4 T-110's, between them two of 8 UltraHoist boards work fine.
I've torn them all apart, and have the two working boards churning away currently (Hooked up to a Pi manually running cgminer)
As for the other 6 boards, two have corrosion damage so I'll save them for last, but the remaining four all have the same problem: I'll run cgminer, they'll WORK, but rapidly hit the 96c thermal throttle, then the 100+ thermal shutdown (Maybe withing 5 seconds of starting?), then reboot and start again.
At first I thought it was the thermal paste, but after re-applying it there was no difference, messing around with HCM told me that all the die's are at the same temp (couldn't get the original tool at https://github.com/BitCrane/cgminer to compile, I'll mention HCM later), so it's not uneven paste, finally I removed the watercooler unit and found the pump clogged up with white gunk, something I usually find in AIO coolers after three or so years of heavy use, quite normal.
(Dogie mentioned dying PSU's in his thread, I've been testing the boards between PSU's, and a bad board will be bad no matter the PSU, same goes for good, so I doubt they're any problem)
I could test by putting one of the good coolers onto a "Bad" board, or vise versa, but I've ordered a few fans and an AIO kit just to see if they help, not wanting to do anything at all to a board if it's working as is.
The other issue I've ran into with all this is that the boards are set to run at 0.82v, even one that looked like it was factory sealed to an extent.
HCM compiled for my Pi isn't able to set anything, and trying to set a default voltage or clock speed just returns the same output as reading the existing settings from the board without modifying anything.
Trying to run it on the cubieboards is futile, as they try running cgminer no matter what, and will respawn them if you try closing the processes.
I've even removed the cgminer binary, no difference, "Device busy"
Next I'll try flashing the boards using a config file like the manual describes, but first I'm downloading a new image for the cubieboard, as the home folder of the one I was using was littered with programs and files (possibly from a past user, possibly from bitcrane. The downloaded image should tell us).
I'll also try using the x86 version of HCM they provide once I can grab a x86 linux box, and see if that helps at all.
Not sure if this'll help anyone, but this is just the troubleshooting I've done with my boards so far.
CHECK THEM WATER COOLERS AND TEMPS IN CGMINER
I'm in a very similar situation, as I wanna get a bunch of old T-110 units I have lying around working.
Kinda silly, sure, but if people are selling 333mhz block erupters for $100 as "Collectables" then I can try bringing life into these old boards.
The problem I had was the web interface wouldn't initialise cards, if they would it wouldn't be for very long.
I have (had) 4 T-110's, between them two of 8 UltraHoist boards work fine.
I've torn them all apart, and have the two working boards churning away currently (Hooked up to a Pi manually running cgminer)
As for the other 6 boards, two have corrosion damage so I'll save them for last, but the remaining four all have the same problem: I'll run cgminer, they'll WORK, but rapidly hit the 96c thermal throttle, then the 100+ thermal shutdown (Maybe withing 5 seconds of starting?), then reboot and start again.
At first I thought it was the thermal paste, but after re-applying it there was no difference, messing around with HCM told me that all the die's are at the same temp (couldn't get the original tool at https://github.com/BitCrane/cgminer to compile, I'll mention HCM later), so it's not uneven paste, finally I removed the watercooler unit and found the pump clogged up with white gunk, something I usually find in AIO coolers after three or so years of heavy use, quite normal.
(Dogie mentioned dying PSU's in his thread, I've been testing the boards between PSU's, and a bad board will be bad no matter the PSU, same goes for good, so I doubt they're any problem)
I could test by putting one of the good coolers onto a "Bad" board, or vise versa, but I've ordered a few fans and an AIO kit just to see if they help, not wanting to do anything at all to a board if it's working as is.
The other issue I've ran into with all this is that the boards are set to run at 0.82v, even one that looked like it was factory sealed to an extent.
HCM compiled for my Pi isn't able to set anything, and trying to set a default voltage or clock speed just returns the same output as reading the existing settings from the board without modifying anything.
Trying to run it on the cubieboards is futile, as they try running cgminer no matter what, and will respawn them if you try closing the processes.
I've even removed the cgminer binary, no difference, "Device busy"
Next I'll try flashing the boards using a config file like the manual describes, but first I'm downloading a new image for the cubieboard, as the home folder of the one I was using was littered with programs and files (possibly from a past user, possibly from bitcrane. The downloaded image should tell us).
I'll also try using the x86 version of HCM they provide once I can grab a x86 linux box, and see if that helps at all.
Not sure if this'll help anyone, but this is just the troubleshooting I've done with my boards so far.
CHECK THEM WATER COOLERS AND TEMPS IN CGMINER
November 16, 2017, 07:28:46 AM
These are ancient asic technology by today's standards and woefully inefficient. They'll cost you 10x more to run than they mine. Salvage the cubieboards, power supplies and cables and toss the rest.
Thanks -ck.
I actually want to get them mining again if possible. I know they are high watt devices for low hashing power (approximately 1GH/s per watt) by today's standards, but aside from my S7, they are what I have. My budget is these or nothing. I'm currently on an electricity plan with my energy provider, so electricity prices are not an issue at the moment. I would like to use some of them alongside the S7 for an extra boost in my hash rates. They also make great room heaters in winter...
November 13, 2017, 07:45:22 PM
These are ancient asic technology by today's standards and woefully inefficient. They'll cost you 10x more to run than they mine. Salvage the cubieboards, power supplies and cables and toss the rest.
November 13, 2017, 09:39:41 AM
Just wondering if anyone has had any luck with refurbishing / rebuilding broken or sub functional Bitcrane T-110 and T-110s miners.
I have 5x T-110 and 2x T-110s - all of which are either running off a single Hashfast module or not functioning at all. I hope to get the 2x T-110s miners working again by using parts from the 5x T-110 miners. If I can also get one or two T-110 miners also working that would be a bonus. I know I have at least 6 working power supplies, and also 6 cubieboards working, and from the hashing I see when I can get some of them to power up, I think I have at least 10 Hashfast modules that should work.
Also if anyone has had any success recovering or underclocking the Hashfast boards, using the Hashfast Characterization Monitor (HCM) on the cubian OS installed with these miners, I'd be happy to learn from their experience.
Any pointers or suggestions would be appreciated.
I have 5x T-110 and 2x T-110s - all of which are either running off a single Hashfast module or not functioning at all. I hope to get the 2x T-110s miners working again by using parts from the 5x T-110 miners. If I can also get one or two T-110 miners also working that would be a bonus. I know I have at least 6 working power supplies, and also 6 cubieboards working, and from the hashing I see when I can get some of them to power up, I think I have at least 10 Hashfast modules that should work.
Also if anyone has had any success recovering or underclocking the Hashfast boards, using the Hashfast Characterization Monitor (HCM) on the cubian OS installed with these miners, I'd be happy to learn from their experience.
Any pointers or suggestions would be appreciated.
Jump to: