As of 25 September, the Bitcoin network was running at 926.4 EH/s with difficulty at 132.76 T, and hashprice stood at $41.20 per PH/s per day, according to bitcoinminingstock.io. At the summer low, hashprice had fallen 18.34% in 30 days to $28.68 per PH/s per day (Bitcoin.com News, 26 June).
That is a recovery of roughly 40% in a few months, and it has pushed a number of older machines switched off over the summer back above their shutdown line at some power rates.
A miner that has sat idle for months should not simply be plugged back in. Dust, moisture, aged connectors and compromised firmware can each lead to burnt boards, fire, or a whole batch of machines failing at once. This is the order to check things in before power-on.
1. Physical and environmental checks: preventing shorts
Dust and moisture are the biggest risks after a long shutdown.
- Clean thoroughly. Use an air compressor or high-power blower on the hashboards, fans and power supply interior. Make sure compressed air is dry and oil-free, or you are spraying moisture onto the boards. Hold fan blades still while blowing them out: letting a fan free-spin at high speed damages the bearings. Dust that has absorbed moisture becomes conductive, and powering on without cleaning is an easy way to burn chips.
- Dry out. If the machines were stored somewhere damp, move them to a dry, ventilated area, or treat them with a dehumidifier or drying equipment before applying power.
- Check for foreign objects. Look inside the chassis for insects, rodent droppings or loose screws.
2. Electrical and connection checks: preventing burnt cables
- Look for carbonised connectors. Unplug the PSU 6-pin connectors and ribbon cables and inspect the power supply and hashboard sockets for yellowing, blackening or deformation. Any carbonisation means replacing the cable, the connector, or if necessary the hashboard. Replacing only the cable and leaving a damaged socket means the new cable will burn too.
- Reseat everything. Reseat every control board ribbon cable and hashboard power cable. Connectors loosen in storage, which shows up as boot errors or local overheating.
- Test the power supply on its own. With the hashboards disconnected, power the supply alone, confirm its fan runs normally, and measure the output with a multimeter. Check the reading against that model’s rated range, which differs between supplies. The APW121215, for example, has an adjustable 12–15 V output, and its revisions differ — see APW121215 version letters.
3. Firmware and network checks: preventing hashrate theft
- Reflash firmware. Older machines such as early S19 units and the M20/M30 series may have been scanned or infected with hashrate-stealing malware while idle. Before restarting, reflash them uniformly by SD card or with the manufacturer’s batch tool. If you plan to run third-party firmware, confirm compatibility before flashing the latest official release: some official updates restrict downgrading or installing third-party firmware. See Overclocking vs Undervolting.
- Change default passwords. Change default logins such as root and admin as soon as the machine is up, to stop infection spreading across the local network.
- Verify pool settings. Re-confirm the pool address and worker names. Some pools may have changed settings or shut down while the machines were offline.
4. Restart in stages, and rerun the numbers
- Start with one or two machines. Run them for 20–30 minutes and check:
- whether hashrate reaches the rated figure
- whether every board shows its full chip count, with no missing chips or dropped boards
- whether PCB and chip temperatures are normal and fan speed is not running high
- Recalculate at your real power rate. Older models in the 30–50 J/TH range are extremely sensitive to electricity price.
Example (illustrative, at the 25 September hashprice): an S19 95T at about 34 J/TH draws roughly 3,230 W. That is about $3.91 of revenue and about 77.5 kWh of power per day, which puts break-even at around $0.050/kWh.
| Power price | Power cost / day | Net / day |
|---|---|---|
| $0.03/kWh | $2.33 | about +$1.58 |
| $0.04/kWh | $3.10 | about +$0.81 |
| $0.05/kWh | $3.88 | roughly break-even |
If hashprice fell back to the summer level of $28.68, the break-even power price for the same machine would drop to about $0.035/kWh.
The point: a restart decision should not rest on “it is profitable today”. Leave room for hashprice to fall back, count labour, maintenance and depreciation, and only restart at scale once the cash flow is positive with that margin included.
5. Line up parts and repair before the restart
When machines come back in batches, failures tend to arrive in clusters: the same batch of power supplies, the same cable type, the same generation of fans all show their problems in the first hours.
- Spares first, then power-on. Whatever steps 1 and 2 turn up — power supplies, fans, cables and adapters — get the replacements in hand before restarting. Ordering after the machines are already down makes every day of waiting a day of lost revenue. See Cheap Part, Expensive Downtime.
- Do not let a faulty board hold up the whole machine. If a test run shows a board with missing chips or dropping out, swap in a spare board to get the machine running, then send the faulty board for repair.
- Put it all in one quote request. Send the miner model, part numbers, quantities, shipping destination and how urgent it is. Where both are available, NODE29 quotes China shipping and US stock side by side.