Bitaxe Longevity: A QHSE Perspective on Underclocking for 24/7 Solo Mining

[AnisEverRise]

Hi,

I'm a Master's student in Quality, Safety, and Environment (QHSE). When I started solo mining with my Bitaxe, my first reflex wasn't "how fast can it go?" but "how long can it last?".

The Setup:
Currently running a Bitaxe Gamma stabilized at 0.9 TH/s. While most users push for the highest hashrate, I’ve deliberately underclocked my unit.

Why Underclocking is the "Smart" Solo Move:

• Thermal Efficiency: Based on risk management principles, reducing heat is the best way to prevent ASIC degradation.
  
• MTBF (Mean Time Between Failures): I am treating my Bitaxe as a 5-year lottery ticket. Running at 0.9 TH/s keeps the chip in a "safe zone" for continuous operation.
  
• Consistency: Higher frequencies often lead to more "HW errors." At 0.9 TH/s, my error rate is near 0%.
  

Conclusion:
In the world of solo mining, you don't need to be the fastest every day; you just need to be "active" when the winning block is found. My Bitaxe is my silent, cool-running ticket to that 3.125 BTC dream.

I'm curious: Have any long-term Bitaxe users here analyzed the degradation of the BM1366/BM1370 chips when kept under 50°C vs 60°C?

[philipma1957]

I have  and somewhere some place I posted that I run at freq 400 at 880gh.

It simply will just run this way.

2 reboots in around  a year

[AnisEverRise]

I have  and somewhere some place I posted that I run at freq 400 at 880gh.

It simply will just run this way.

2 reboots in around  a year

I’m a firm believer in stability over raw power.

Running my Bitaxe at 400MHz (~880GH/s) has proven to be the ultimate "set and forget" strategy. Only 2 reboots in a full year , that’s the kind of reliability I’m looking for.

In solo mining, I'd rather have a steady, cool running ticket in the lottery than a pushed chip that crashes when I'm not looking. Quality uptime is the real winner here

[philipma1957]

here is mine

[AnisEverRise]

here is mine

Hi @philipma1957,

Thank you so much for sharing your data! Seeing those charts is incredibly satisfying from a Quality and Reliability perspective.

Two reboots in an entire year is an outstanding MTBF (Mean Time Between Failures). It proves that the Sweet Spot for these BM chips isn't at the thermal limit, but exactly where you’ve placed it around 400MHz. Your AxeOS dashboard shows a very clean efficiency curve, which is exactly the 'stability over power' model I’m aiming for.

It's interesting to see that even at 880 GH/s, your hash rate history remains remarkably flat. No erratic spikes usually means less thermal throttling and less stress on the voltage regulators.

As a legendary member, have you noticed any significant difference in chip longevity or zombie shares over the years when comparing these underclocked units to those running at stock 500MHz+ frequencies?

[jstefanop]

Hi,

I'm a Master's student in Quality, Safety, and Environment (QHSE). When I started solo mining with my Bitaxe, my first reflex wasn't "how fast can it go?" but "how long can it last?".

The Setup:
Currently running a Bitaxe Gamma stabilized at 0.9 TH/s. While most users push for the highest hashrate, I’ve deliberately underclocked my unit.

Why Underclocking is the "Smart" Solo Move:

• Thermal Efficiency: Based on risk management principles, reducing heat is the best way to prevent ASIC degradation.
  
• MTBF (Mean Time Between Failures): I am treating my Bitaxe as a 5-year lottery ticket. Running at 0.9 TH/s keeps the chip in a "safe zone" for continuous operation.
  
• Consistency: Higher frequencies often lead to more "HW errors." At 0.9 TH/s, my error rate is near 0%.
  

Conclusion:
In the world of solo mining, you don't need to be the fastest every day; you just need to be "active" when the winning block is found. My Bitaxe is my silent, cool-running ticket to that 3.125 BTC dream.

I'm curious: Have any long-term Bitaxe users here analyzed the degradation of the BM1366/BM1370 chips when kept under 50°C vs 60°C?

These chips are designed to run 90c+ junction temps all day. 50-60c is cool in terms of IC temps. You only need to run that low if you are trying to achieve max efficiency.

Major degradation wont occur unless you are running 80-90c 24/7. Keep in mind these chips run from 40c in the front to 80+c in the back due to thermal gradients on the air-cooled systems.

[AnisEverRise]

These chips are designed to run 90c+ junction temps all day. 50-60c is cool in terms of IC temps.

True, the silicon is tough. But as a QHSE student, I look at the System Reliability rather than just the ASIC die.

Running at 0.9 TH/s @ 50°C isn't just about the chip; it's about protecting the VRMs and capacitors on the Bitaxe PCB. High thermal loads accelerate electrolyte depletion in caps and increase thermal fatigue on BGA solder joints over a 5-year horizon.

In solo mining, Uptime > Hashrate. I'd rather have a silent, cool-running unit with 99.9% stability than 5% more hash with a higher risk of hardware failure mid-lottery.

Question: Has anyone measured the efficiency (J/TH) "sweet spot" for the BM1366 at sub-500mV? I'm chasing longevity, not just speed.

[philipma1957]

If air cooled lower temps mean slower fan speeds.

Lower fan speeds mean less vibration

Less vibrations means less mechanical wear and tear on every solder joint.


Thus less micro fractures in the solder joints.

My fans run at 85% .  Not 100%

[NotFuzzyWarm]

The biggest issue with longevity of the Bitaxes is the fact that being an open-source design it is made by scads of mfgrs in China and elsewhere - many (but not all) of them build the miners with the cheapest possible components and seem to have little to no QC. As expected, that leads to many failures. As Sidehack often says, it is a race to the bottom for them.

Given that, ja, under-clocking them is a very good idea.

[AnisEverRise]

If air cooled lower temps mean slower fan speeds.

Lower fan speeds mean less vibration

Less vibrations means less mechanical wear and tear on every solder joint.


Thus less micro fractures in the solder joints.

My fans run at 85% .  Not 100%

Maybe this is the real Golden Ratio for the Bitaxe.From a QHSE (Quality & Health/Safety/Environment) perspective, pushing for 100% fan speed is often a trap. By capping at 85%, we drastically reduce the harmonic vibrations that lead to micro-fractures in solder joints (intergranular cracking) over time.Also, reducing the core voltage below 500mV isn't just about efficiency  it's about the Arrhenius Law. Every 10°C drop effectively doubles the lifespan of the electrolytic capacitors on the board.0.9 TH/s at these temps isn't just underclocking, it's long-term reliability engineering.

The biggest issue with longevity of the Bitaxes is the fact that being an open-source design it is made by scads of mfgrs in China and elsewhere - many (but not all) of them build the miners with the cheapest possible components and seem to have little to no QC. As expected, that leads to many failures. As Sidehack often says, it is a race to the bottom for them.

Given that, ja, under-clocking them is a very good idea.

Maybe you've pinpointed the exact weak link in the chain. If these units are built as a race to the bottom with subpar components, then my approach isn't just a choice it's a necessity. By capping at 0.9 TH/s and 85% fan speed, I am essentially derating the hardware to match its actual build quality. It’s better to have a steady, derated miner that achieves 100% uptime than a pushed unit that risks a catastrophic failure due to thermal fatigue or poor solder quality. I'm prioritizing the MTBF (Mean Time Between Failures) over raw hash power, which is the only logical move for a solo miner in it for the long haul.

[NotFuzzyWarm]

Ja. Sidehack does/did make them to fit into their price range and being Gekko products they are top quality with excellent attention to details and QC. Price matches being at the top end. https://www.gekkoscience.com/shop/ He has often made mention of resellers contacting him about making 'axes for them but all insisted on using sub-par (cheaper) components that they spec. Sidehack of course rightly refuses to do that.

He has gone as far as now making a much improved version with NO ip from OSMU who are the folks behind the BitAxes and own the circuit design and firmware used.

[AnisEverRise]

Ja. Sidehack does/did make them to fit into their price range and being Gekko products they are top quality with excellent attention to details and QC. Price matches being at the top end. https://www.gekkoscience.com/shop/ He has often made mention of resellers contacting him about making 'axes for them but all insisted on using sub-par (cheaper) components that they spec. Sidehack of course rightly refuses to do that.

He has gone as far as now making a much improved version with NO ip from OSMU who are the folks behind the BitAxes and own the circuit design and firmware used.

That’s a solid point. I didn't know the full story behind the pressure from resellers to use cheaper components, but it’s a relief to hear that Sidehack stood his ground. In the long run, saving a few bucks on parts just isn't worth the stability issues, especially for something running 24/7.

I’ve always preferred the Gekko approach because of that focus on build quality. Even if you're underclocking for better efficiency, it’s much safer to do it on a board that’s actually built to last rather than a budget version with sub-par QC.

It'll be interesting to see how the new version without the OSMU IP performs. If the quality stays at that top end, it’s definitely going to be the standard for solo mining.

[NotFuzzyWarm]

The readme in his repo for the GekkoAxe https://github.com/sidehack-gekko/GekkoAxe/ says it all. That miner is the last he is making that are based on any BitAxe IP. His new replacement should go on sale in another month or so.

[AnisEverRise]

The readme in his repo for the GekkoAxe https://github.com/sidehack-gekko/GekkoAxe/ says it all. That miner is the last he is making that are based on any BitAxe IP. His new replacement should go on sale in another month or so.

Thanks for pointing that out! I just checked the repo and you're right, the README is pretty clear about the shift. It’s a bold move to step away from the BitAxe IP, but if it means better optimization and total control over the QC, it’s a win for the community.

[philipma1957]

Ja. Sidehack does/did make them to fit into their price range and being Gekko products they are top quality with excellent attention to details and QC. Price matches being at the top end. https://www.gekkoscience.com/shop/ He has often made mention of resellers contacting him about making 'axes for them but all insisted on using sub-par (cheaper) components that they spec. Sidehack of course rightly refuses to do that.

He has gone as far as now making a much improved version with NO ip from OSMU who are the folks behind the BitAxes and own the circuit design and firmware used.

Yeah I had his  first or second build. And I sent it back for rework when he realized some spec parts suck.

Heatsink was brutal. His rework is flawless so he builds them all that way.