New miners, and everyone who is going to join the ‘crypto movement’, are interested in the question: how to bring cryptocurrency mining to a safe level. There are several ways to maintain the maximum profitability of mining, and one of the main is to keep the farm in good condition.
Mining farm: step-by-step maintenance instructions
Today we will tell you how to perform scheduled and extended maintenance, how to monitor and take care of your GPUs, and also we`ll give you some life hacks for warranty repairs.
1. Current maintenance: 1-2 months
It is good to carry out regular quick maintenance, once a month, or two months. Check for software updates, check your CPU and GRU for dust, damage, and proper fan operation.
Software
Make sure that your mining software is updated. In any latest version, developers remove vulnerabilities and improve the performance of the software (and bring more money). Updating software for Windows can be difficult due to problems with video drivers, so it is recommended to look for Linux distributions.
As you may face issues with some updates, install them in two steps. First, update a few mining installations and see if everything is in order. If so, install updates for the rest of the farm. Some time ago, Claymore miners stopped working after the HiveOS update, so they had to switch to other miners with a lower hashrate.
Visual inspection
This may seem trivial, but it is important to visually inspect your farm from time to time, especially its remote parts.
First, check the power supply. Turn off the computer to shut the power supply down. Inspect the fan: does it rotate freely? Is there any dust? Does it make strange sounds, and is the rotation speed normal? Unresolved problems with the fan can lead to damage to the power supply and downtime of the mining installation.
Now here’s what to do in case of problems. Unscrew the screws to open the power supply cover, and do the same with the fan housing. If possible, remove the cable – it can be quick-detachable or soldered. Check if there is any dust or other parts that can be removed. If the fan looks damaged, you can buy any regular PC fan to replace it – they usually work fine. Keep in mind that Chieftec fans may require more frequent replacement.
Check the cables and connectors. The connectors of the GPU power cable and the adapter board may wear out when used at maximum power. If the cables are removable, check their sockets on the power supply housing.
If the slots are burned out, you can do two things. First, try to repair the cable yourself: solder it to the connector from another cable. In this case, you will need cables that you did not throw away after receiving the power supply. Secondly, if the connector is badly burned, the PSU should be replaced. Miners report that Chinese-made power supplies can fail less often than Corsair or Chieftec power supplies.
Check the CPU fan. If the processor overheats, it usually shuts down before the damage occurs. However, without a properly functioning of CPU, you will not be able to start the installation. If there is a problem, the message “CPU FAN ERROR” will be displayed, and you will be prompted to press F1 to start the setup. To avoid this, monitor the CPU fan remotely: check if the cables or the adapter board bases are stuck inside it.
2. Scheduled extended maintenance – once every few months
Some of the things listed above need to be done in a more accurate way every few months. They include visual inspection and dust removal.
To get rid of dust, do not use a vacuum cleaner – it will generate static electricity. Instead, take an air compressor. It will spread the dust everywhere, without harming you and the farm. Don’t forget to turn off the settings first!
For extended maintenance, reconnect the cables and connectors, adapter cards, and RAM. Over time, they can become oxidized, which can cause problems with their functioning. At the same time, check the serviceability of the power connectors.
When you turn on your installation again, see how the GPU fan works. If it takes a long time to start the rotation, inspect it more carefully, or consider replacing it. For a while, you can push the blades manually so that they work normally, but remember about the detected fault signal!
3. Annual maintenance
Replacement of the CR2032 CMOS battery. This is a motherboard battery that needs to be replaced every 1-2 years. A low battery charge can cause a BIOS reset, and you can only detect this after a power outage. In addition, the battery can cause corrosion and damage the motherboard. Since it is very expensive to repair the motherboard, change the battery every 1-2 years, and you will be safe.
Take care of your GPU
Graphics cards are the most important equipment for mining most coins, so they need to be given special attention. The critical parameter is the temperature of the GPU: it must be below 70 ° C, otherwise the processor will fail or additional fan power will be required.
- Fan issues
If one of your GPUs is hotter than the others, and the fan speed exceeds the allowed speed, it will probably have problems with rotation. When you do your research, keep in mind that the order of the GPUs displayed in your interface may differ from the actual order of the graphics cards on your motherboard.
Unfortunately, many fan manufacturers use copper bushings instead of bearings. If there is something wrong with the fan, it makes no sense to lubricate it, but it is quite difficult to find a suitable bushing. If you are in such a situation, just buy a new fan on AliExpress. Their wholesale price can be $7.5 per pair, plus cables. It is not difficult to replace the fan: in some cases, you will need to replace a couple of fans at once. You can use your warranty to replace the fans – we will return to this topic soon.
- Thermal paste or thermal plates
Between the graphics chips and the cooling system radiator in your GPU, a thin layer of material prevents the graphics card from overheating – this is thermal paste or thermal plate. If your GPU often overheats, but replacing or cleaning the fan does not help, apply the paste again. Do this once every 3-5 years, not more often – there is a risk that the layer will be too thin, and overheating may occur.
- GPU memory errors.
If you have them, this is a sign that the memory chips of the GPU are worn out. Running damaged memory chips can harm your mining, and lead to a pool lock. To avoid this, reduce overclocking.
4. What parameters can you monitor to ensure the security of your farm?
Constant monitoring will help you prevent unexpected breakdowns, farm downtime and money losses. Here’s what you can control remotely:
- The temperature of the GPU. Control this using the graphical web interface.
- Processor temperature and fan speed. Use the web interface or the command line (lm-sensors tool).
- GPU errors.
Remember that the farm can heat both itself and the surrounding air – the room can quickly become too hot. Use temperature sensors and smart home tools, they will help fans of graphics processors to do their job.
5. Keep a spare hard drive to avoid downtime
Remote monitoring can help prevent many types of breakdowns, but, unfortunately, not all of them. As a precautionary measure, keep additional adapter cards, power supplies and a backup disk on hand to quickly solve the problem and avoid downtime. The configured OS and miner must be pre-installed on your hard disk, so when the hard disk fails, you can replace it as soon as possible.
Here are two tips: add your FARM_HASH to the rig.conf file in Hive OS, and your installation will be displayed immediately after connecting. Secondly, use the web interface to easily configure a new installation.
6. Warranty repair or replacement
If you have a problem with the equipment, and you are sure that it is a defect of the manufacturer, use the warranty. You can apply for a replacement or refund. Unfortunately, miners report that sellers and service centers usually do not like miners, so you may need some perseverance.
If your BIOS has been changed, or you have 4 GPUs connected to the same power supply, the service may reject your refund order, even if there is a manufacturing defect. To avoid this scenario, restore the original hardware configuration before diagnostics.
Each manufacturer and seller has different warranty periods. For mining versions of GPUs, the warranty can be only 3 months, so we recommend buying a regular graphics card for games. Make sure that the seals are not removed when going to diagnostics.
As for the fans, they break most often, so think about replacing them. However, the service center does not compensate you for the downtime of the farm. Order several additional fans on AliExpress, keep them on hand and replace them manually if necessary. The seals will remain safe, and you will still be able to apply for warranty service.
Last thoughts
There are quite a lot of things to do on this list. However, once your farm is working, you will not need much time to track its parameters, order spare equipment on AliExpress, as well as carry out regular cleaning and replacement.
Remember that these things maximize your mining profits, as they reduce the likelihood of downtime.
