Freezing Your Miners To Life.!

[mikeywith]

We all have heard this old "freeze" your electronic stuff to fix them, some say it works, some say it does not, I have always thought it was a complete B.S, until one of the members here whom I believe was "Favebook" reported some sucssess.

As some of you already know,  I have a lot of dead 17 series hash boards, so I decided to give it a try.

I got myself 10 plastic bags and put 10 hash boards in those bags and buried them in the fridge for about 40 hours or so, I don't recall the exact duration but it was 2 nights and a bit less than 2 days.

I took the plastic bags out of the freezer but I didn't open them because the water condensed on them, I left them to dry overnight and the next day, I tested the 10 boards which were a combination of S17+ and T17 ( the T17s are a pain in the ass to disassemble and reassemble compared to the + and E versions).

The initial test brought 4 boards to life, reporting full asic numbers, I went ahead and flash a custom firmware to see if I get better results, sadly nothing.

After 10 hours when I checked the miners again two out of the four have died again (tears), I think one died after 30 mins or so (judging the hashrate on the pool) the other one lasted for a little under 10 hours.

The other two boards I managed to revive are still hashing after 2 days, it was one T17 hash board and one S17+, all the boards I have tested have no physical damage, I applied a good amount of pressure on all the heatsinks and nothing fell, I also noticed the 2 hash boards that came back to life for a while before resting in piece again show more than 0 asic but not all Asics for an example one of the T17 boards shows 20 Asics.

The T17 board that is still mining on does seem to have an issue with the 4th temp sensor reading, it's showing 0 on both the PCB and Chip but hashing at a modest 12.6TH.

Again, I would like to give credit to Favebook, and encourage you all to try this simple and free trick, I also think my success rate isn't realistic and far from sustainable when the sample size is larger, so don't expect to get 2 out of 10 every time, but really, given how easy and free this is, even with 1 in 20 you are a winner.

[1715460122]

1715460122

[1715460122]

1715460122

[1715460122]

1715460122

[DaveF]

The science behind it as explained to me years ago.
Things shrink when cooled. The boards shrink at one rate, the wire traces in the board shrink at another. The chips shrink at one rate the solder connections at another.

We are talking about fractions of a % of a mm. BUT, something that was just on the edge at room temperature will at times start to work.

After a heating cycle or 2 it goes back to the way it was. But it can work.

Never thought about it when it came to miners.

-Dave

[philipma1957]

So this is like putting it in a 175°F  oven for 45 minutes

[DaveF]

So this is like putting it in a 175°F  oven for 45 minutes

Sort of, that makes the solder softer and can allow it to make better connections that tend to be more permanent then this. You "re-attached" a bad connection.

Cooling / freezing it is more along the lines of whacking something on it's side to get it started. But will eventually lock up again. Kind of like hitting a stuck starter in your car with a hammer so you can get it going to get home and then to the the auto parts store. It might last longer, but you really can't trust it.

-Dave

[HagssFIN]

Has anyone considered putting their bad hash board to oven as some people do for GPUs? In hopes of re-soldering it.

[mikeywith]

Has anyone considered putting their bad hash board to oven as some people do for GPUs? In hopes of re-soldering it.

I have plenty of dead hash boards to test, give me the oven receipie and I'll cook the hell out of them.

The science behind it as explained to me years ago.
Things shrink when cooled. The boards shrink at one rate, the wire traces in the board shrink at another. The chips shrink at one rate the solder connections at another.

We are talking about fractions of a % of a mm. BUT, something that was just on the edge at room temperature will at times start to work.

That explains it very well.


Update:

The 2 hash boards are working just fine still, I also noticed one of the boards that came back to life and then died now shows full Asics (65 asics) but it has a temp sensor error

Code:

2021-02-23 19:04:14 driver-btm-api.c:1134:check_asic_number_with_power_on: Chain[0]: find 65 asic, times 0
2021-02-23 19:04:25 driver-btm-api.c:1134:check_asic_number_with_power_on: Chain[2]: find 65 asic, times 0
2021-02-23 19:04:30 driver-hash-chip.c:266:set_uart_relay: set uart relay to 0x330003
2021-02-23 19:04:30 driver-btm-api.c:435:set_order_clock: chain[0]: set order clock, stragegy 3
2021-02-23 19:04:30 driver-btm-api.c:435:set_order_clock: chain[2]: set order clock, stragegy 3
2021-02-23 19:04:31 driver-hash-chip.c:502:set_clock_delay_control: core_data = 0x34
2021-02-23 19:04:31 driver-btm-api.c:1892:check_clock_counter: freq 50 clock_counter_limit 6
2021-02-23 19:04:31 voltage[0] = 1980
2021-02-23 19:04:31 voltage[2] = 1980
2021-02-23 19:04:31 power_api.c:226:set_working_voltage_raw: working_voltage_raw = 1980
2021-02-23 19:04:32 temperature.c:340:calibrate_temp_sensor_one_chain: chain 0 temp sensor NCT218
2021-02-23 19:04:33 temperature.c:340:calibrate_temp_sensor_one_chain: chain 2 temp sensor NCT218
2021-02-23 19:04:33 uart.c:72:set_baud: set fpga_baud to 12000000
2021-02-23 19:04:34 driver-btm-api.c:293:check_bringup_temp: Bring up temperature is 21
2021-02-23 19:04:34 thread.c:1378:create_check_miner_status_thread: create thread
2021-02-23 19:04:34 thread.c:1368:create_show_miner_status_thread: create thread
2021-02-23 19:04:34 thread.c:1348:create_temperature_monitor_thread: create thread
2021-02-23 19:04:34 frequency.c:514:check_bringup_temp_dec_freq: dec freq = 0 when bringup temp = 21 dec_freq_index=0
2021-02-23 19:04:34 freq_tuning.c:183:freq_tuning_get_max_freq: Max freq of tuning is 650
2021-02-23 19:04:34 driver-btm-api.c:1765:send_null_work: [DEBUG] Send null work.
2021-02-23 19:04:34 thread.c:1338:create_asic_status_monitor_thread: create thread
2021-02-23 19:04:34 frequency.c:1110:inc_freq_with_fixed_vco: chain = 255, freq = 500, is_higher_voltage = true
2021-02-23 19:05:58 power_api.c:352:set_to_voltage_by_steps: Set to voltage raw 2090, step by step.
2021-02-23 19:05:59 power_api.c:85:check_voltage_multi: retry time: 0
2021-02-23 19:06:01 power_api.c:40:_get_avg_voltage: chain = 0, voltage = 20.817602
2021-02-23 19:06:02 power_api.c:40:_get_avg_voltage: chain = 2, voltage = 20.720663
2021-02-23 19:06:02 power_api.c:53:_get_avg_voltage: average_voltage = 20.769132
2021-02-23 19:06:02 power_api.c:71:check_voltage: target_vol = 20.90, actural_vol = 20.77, check voltage passed.
2021-02-23 19:09:27 frequency.c:1152:inc_freq_with_fixed_step: chain = 2, freq_start = 500, freq_end = 520, freq_step = 5, is_higher_voltage = true
2021-02-23 19:09:35 frequency.c:1181:inc_asic_diff_freq_by_steps: chain = 0, start = 500, freq_step = 5
2021-02-23 19:09:41 frequency.c:1181:inc_asic_diff_freq_by_steps: chain = 2, start = 520, freq_step = 5
2021-02-23 19:09:44 driver-btm-api.c:765:set_timeout: freq = 550, percent = 90, hcn = 44236, timeout = 80
2021-02-23 19:09:44 power_api.c:310:set_to_working_voltage_by_steps: Set to voltage raw 1980, step by step.
2021-02-23 19:09:49 power_api.c:85:check_voltage_multi: retry time: 0
2021-02-23 19:09:50 power_api.c:40:_get_avg_voltage: chain = 0, voltage = 19.727040
2021-02-23 19:09:51 power_api.c:40:_get_avg_voltage: chain = 2, voltage = 19.630102
2021-02-23 19:09:51 power_api.c:53:_get_avg_voltage: average_voltage = 19.678571
2021-02-23 19:09:51 power_api.c:71:check_voltage: target_vol = 19.80, actural_vol = 19.68, check voltage passed.
2021-02-23 19:09:51 thread.c:1373:create_check_system_status_thread: create thread
2021-02-23 19:09:52 driver-btm-api.c:2618:bitmain_soc_init: Init done!
2021-02-23 19:09:52 driver-btm-api.c:222:set_miner_status: STATUS_INIT
2021-02-23 19:09:56 driver-btm-api.c:222:set_miner_status: STATUS_OKAY
2021-02-23 19:09:57 frequency.c:205:get_ideal_hash_rate_GH: ideal_hash_rate = 45227
2021-02-23 19:09:57 frequency.c:223:get_sale_hash_rate_GH: sale_hash_rate = 43000
2021-02-23 19:10:00 driver-btm-api.c:1496:dhash_chip_send_job: Version num 4.
2021-02-23 19:10:00 driver-btm-api.c:1644:dhash_chip_send_job: stime.tv_sec 1614107400, block_ntime 1614107373
2021-02-23 19:10:04 thread.c:1273:asic_status_monitor_thread: ERROR: chain 2 get hashrate_reg_counter 63, require 65, failed times 1: xxooo ooooo ooooo ooooo ooooo ooooo ooooo ooooo ooooo ooooo ooooo ooooo ooooo
2021-02-23 19:10:05 thread.c:1293:asic_status_monitor_thread: chain 2 can't get enough hashrate reg val for 0 times.
2021-02-23 19:10:23 thread.c:1273:asic_status_monitor_thread: ERROR: chain 2 get hashrate_reg_counter 5, require 65, failed times 1: ooooo xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx
2021-02-23 19:10:24 temperature.c:843:get_temp_info: read temp sensor failed: chain = 2, sensor = 0, chip = 14, reg = 0
2021-02-23 19:10:24 temperature.c:843:get_temp_info: read temp sensor failed: chain = 2, sensor = 0, chip = 14, reg = 1
2021-02-23 19:10:24 temperature.c:843:get_temp_info: read temp sensor failed: chain = 2, sensor = 1, chip = 10, reg = 0
2021-02-23 19:10:24 thread.c:1293:asic_status_monitor_thread: chain 2 can't get enough hashrate reg val for 0 times.
2021-02-23 19:10:25 temperature.c:843:get_temp_info: read temp sensor failed: chain = 2, sensor = 1, chip = 10, reg = 1
2021-02-23 19:10:25 thread.c:1273:asic_status_monitor_thread: ERROR: chain 2 get hashrate_reg_counter 5, require 65, failed times 1: ooooo xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx
2021-02-23 19:10:25 temperature.c:843:get_temp_info: read temp sensor failed: chain = 2, sensor = 2, chip = 54, reg = 0
2021-02-23 19:10:25 temperature.c:843:get_temp_info: read temp sensor failed: chain = 2, sensor = 2, chip = 54, reg = 1
2021-02-23 19:10:26 temperature.c:843:get_temp_info: read temp sensor failed: chain = 2, sensor = 3, chip = 50, reg = 0
2021-02-23 19:10:26 temperature.c:843:get_temp_info: read temp sensor failed: chain = 2, sensor = 3, chip = 50, reg = 1
2021-02-23 19:10:26 temperature.c:875:get_temp_info: ERROR: chain 2 can get NONE temp info or temp value abnormal, power it off 
2021-02-23 19:10:26 thread.c:1293:asic_status_monitor_thread: chain 2 can't get enough hashrate reg val for 1 times.
2021-02-23 19:10:27 thread.c:1265:asic_status_monitor_thread: ERROR: chain 2 get hashrate_reg_counter 0, require 65, failed times 1
2021-02-23 19:10:28 frequency.c:205:get_ideal_hash_rate_GH: ideal_hash_rate = 22656
2021-02-23 19:10:28 frequency.c:223:get_sale_hash_rate_GH: sale_hash_rate = 22000
2021-02-23 19:40:00 thread.c:259:calc_hashrate_avg: avg rate is 23083.64 in 30 mins
2021-02-23 19:40:00 temperature.c:516:temp_statistics_show:   pcb temp 49~65  chip temp 72~80
2021-02-23 20:10:01 thread.c:259:calc_hashrate_avg: avg rate is 22760.07 in 30 mins
2021-02-23 20:10:01 temperature.c:516:temp_statistics_show:   pcb temp 50~65  chip temp 71~79

Maybe i froze those temp sensors more than I should, or, it could just be a loose heatsink somewhere, but nonetheless, showing full asic count after the freezer visit is a good sign.

[HagssFIN]

@Mikey
Here is one article for gpu: https://www.overclockers.com/the-oven-trick-repairing-your-broken-video-card-with-an-oven/

I haven't tested this so I'M UNABLE TO GIVE MY REVIEW FOR THE PROCESS AND ALL THE RISKS ARE YOUR OWN.

I would recommend using an oven in which you don't cook your food.

[mikeywith]

@Mikey
Here is one article for gpu: https://www.overclockers.com/the-oven-trick-repairing-your-broken-video-card-with-an-oven/

Well, of course, this wouldn't be easy, it says:

While that’s preheating, start removing the heatsink and any other plastic parts on your video card. You essentially want the card naked, with nothing but the PCB and rear bracket

While that is obvious, I seem to have ignored that when I thought about baking those hash boards, many components on the hashboard will melt at 198c, so this isn't a simple operation, removing the plastic 18pin data ribbon cable alone is going to take me a long time (I have shaky hands when I deal with things like that).

With that being said, if someone trusted here tries that and it works, I could hire someone to remove all those components that need removing and try the trick out, but without any reviews from other miners, this would be a waste of time, efforts, probably an oven, and a wife.

[thierry4wd]

Hi! I tried this ... my T17 card showed 24chips, passing through the oven, gradually increasing in temperature, I have absolutely nothing to remove like plastic or other !!! you should know that on these models, bitmain has to use soldering at low temperature ... so no need to go up high temp to recast everything, it is therefore their default (the chips fall) ...

after my test, twice on 2 cards, absolutely no change, I tried at 135c, 140c, 145, and 155c, everything was recast (be careful not to move anything, the chips no longer hold), but unfortunately no results , always 24chips detected ...

watch my video, you have to download it (I don't know how to post a video without having to register somewhere)

https://easyupload.io/h2mmpz