Topic: recycle miners exhaust with mini wind turbine (Read 226 times)

Correct. I calculated the absolute best case numbers possible. In the real world, you'll be lucky to get half of that 150w. Instead, you could also lower the heat production of the hashboards by 10-20% with custom firmware/undervolting at the same hashrate. That will conserve energy for both the hashboards and the fans!

There are definitely safer compounds that have a boiling point between 40° C - 80° C. Here's a [good place to start](https://en.wikipedia.org/wiki/List_of_refrigerants). It looks like [Novec](https://www.3m.co.uk/3M/en_GB/novec-uk/applications/fire-suppression/) in particular is very popular for datacenter fire suppression. If there's a leak, it is safe for workers, not flammable, and it leaves no residue. The only problem is it has 1/10th the thermal conductivity of water. Pentane has double the thermal efficiency, but it has a lower boiling point, so it'd only work in colder climates.

Another alternative is to run air through the miners as usual, but build a heat exchanger where the exhaust flows through a big radiator so the heat is transferred to the refrigerant. But this is more complex and adds cost, and efficiency will be lower because the temperature of the exhaust is always cooler than the actual silicon. However, I like this idea the most because it can be retrofitted to existing warehouses as long as they put a manifold behind the miners.

But like I said, it takes a six figure team of technicians to figure this stuff out, let alone bringing it to life, so even Bitmain or the largest farms probably don't want to venture into this idea (except for virtue signalling purposes, which might help because it would stop the greenies from pointing fingers at us miners).

Not only is the return minimal (thanks for providing numbers!), on a conceptual level assuming heatsinks take out 0% of wind energy would mean they would just not cool the chips, meaning if the machinery were to run like this, you would also be able to just run it without fans spinning. And that would be much cheaper than running them and taking out 30% of the energy through an expensive turbine.
I'm putting this in again for anyone who's struggling: https://en.wikipedia.org/wiki/Conservation_of_energy

This honestly doesn't sound like a very bad idea. I'm sure there are non-flammable liquids that could be used for this. Also, water does evaporate below 100°C, and the evaporation is accelerated by any amount of heat (i.e. also <100°C), but I don't think you can run a turbine off of such 'slow evaporation'.

Here's my (informed) opinion on this unique idea:

A replacement fan for an Antminer uses 12v/2.7A, which is 32.4 watts. Let's assume each miner has 4 fans, for a total of 130w of consumption. Assuming the motors are 100% efficient at converting electric energy --> wind energy (which they're not, because otherwise the fans would be silent, not 90 dB), you have 130w of extractable wind power. I'm also assuming the restriction of the airflow passing through the hashboard heatsinks takes out 0% of the wind energy. Finally, I will presume that the exhaust will pass through ducts, in order to bypass Betz's Law.

Assuming the average ASIC farm has a power cost of 5¢/kWh, and your turbine's motor is 100% efficient, the turbine would return $0.15/day. Let's say the turbine has a lifespan of 1000 days. The most you will get back is $150.

Is it worthwhile to build this contraption? If the rate of return on your contraption is less than if you just spent the $$ on more miners, you shouldn't built it. Also, now you have to worry about your wind power project instead of just the mining farm. Not to mention the slowed airflow through the miners, the heat stress and the higher failure rates of your miners. They were not designed with this reduced level of airflow. The more efficient your turbine is, the slower the air will pass through, and the hotter the devices will get.

Is it worth risking your equipment failing in order to save a few bucks on electricity? I think not.

---

Now let me talk about harvesting the heat energy. I actually think it is technically feasible. Of course, water won't work, since your hashboards would need to be 100°C to even generate steam (which will kill them).

But instead of water, you could use a hydrocarbon, like propane, with a lower boiling point. Let's say the hardware is 70° C while the outside air is -10 - 30° C. You can theoretically harvest that energy to spin a turbine, then cool the propane down and it flows back to the miner to get boiled again. There are power plants in the real world which use this process for certain applications. The efficiency is usually 30-40% of the heat energy, so for a 3400w miner, you can get 1200w back.

Now of course, the average person buying ASICs and setting up a farm can't do this. It takes hardware engineers, mechanical engineers and construction workers to design a miner that is propane-cooled. Then there's the problem of the flammability of propane, especially next to a heat source like the ASIC. If the pipe has a leak and the propane meets oxygen, the entire warehouse can go up in flames. The farm owner needs to build an entire piping system, not to mention fire safety regulations. It would take an entire team of engineers to manage the construction. Why not just spend that money (six-seven figures) on more harwdare instead?

So this idea probably won't work even for Bitmain or Innosilicon themselves. Why go through all this trouble just to save 30% on power?? Why hire 10 new engineers just to design a new ASIC that is propane cooled? It only makes sense if the miners are inefficient and power costs are higher, which is at the end of the hardware's lifespan. Of course, the vendors care about selling the miners cheaply, not how long they will last.

What hasn't been discussed is that electronics cooling fans are not terribly efficient. 

Mount them on a plate, fastened to a duct, with the turbine at the opposite end of the duct.  See how much of the power into the fans was even recoverable without the turbulence and Reynolds inefficiency of the miner cooling labyrinth.

Go a step further and drive one turbine directly with another, coupled together. 

The Betz Limit for turbines is 0.593.  0.593**2 = 0.3516, which means the combination of the fans and the turbine, can at most recover 0.3516 of the energy put into the fans.

In air driven turbines. the theoretical maximum efficiency is not achieved. Small air driven turbines achieve about 30% or half the theoretical limit.

Keep in mind with this calculation that this only works if they are burning 320W more than what they actually need, and if there is no way to run the fans slower.
What I'm trying to say: If you could just run the fans slower without effect on mining performance, that will always be the better option than running them at the same speed and harvesting back part of the 'extra energy'.

Imagine that the fans at 100% fanspeed burn 320W, and with the added turbine in the back, still cool the miners effectively. The turbine will take out some of that energy and convert it into electricity.
For simplicity's sake, let's say the amount of drag created by the turbine reduces the airflow so that the miners are effectively only cooled by 160W of 'fan energy' and the rest is taken up by the turbine & converted into 100W of electrical energy.
You've spent 320W, cooled using 160W and got back 100W, thus only really having 'spent' 220W.

If you would've just ran the fans at 50% speed, it would have cost you 160W, so it's 60W cheaper than the turbine setup.

The only situation where this makes sense is if you only need 50% fanspeed but they're fixed to 100% for example. Or if you need 20% fanspeed, but the fans only go down to 40%. Then you could harvest the 50 or 20% difference between 'what the miner needs' and 'what the minimum of the fan is'.

Though in that case it should be possible to just buy slower fans that need less energy and it should be cheaper than buying a turbine.

---

The difference between 'recycling heat' (which is a great idea and does work) and 'recycling airflow' is that in the former case, you're not going to burn more energy due to the fact that you take away heat, because it's taken away (convection) anyway and it's a byproduct.
Instead, in the latter case, by taking out airflow energy, you have to put in the same (or higher) amount of energy in the first place to keep the cooling equal.

It is hard to understand why someone wants to run their electronics hotter.  Because imposing drag on the exhausting air will cause an increase in backpressure, and a decrease in flow.  If you want more flow, to reestablish cooling levels, one will need additional power to turn fans.

If you can find things which need drying, the could be put on the exhaust part of the mine, but I would avoid impeding airflow.

Oh, and the generation of the wind turbine will not likely exceed the power of the fans.  Look at the fan power requirement and multiply by the number of fans, and that will be the highest possible power recovery.  (Except in some highly impractical configurations such as a cooling tower.)

I would consider energy utilization, where gentle heat and airflow could be monetized.  Perhaps drying rugs which have been washed.  Or drying plant matter of some sort.  Perhaps dehydrating food products if the airflow path and cleanliness of the miners can be maintained.

 I suspect that four s17s with 16 fans pushing are burning maybe 320 watts
maybe you recover half of that. So you could run some low power miner maybe a few sidehack usb sticks.

you cant focus enough air to move a big turbine. but there are some small ones for camping and some for rvs .

I used two y ducts stacked as it would not really restrict air flow an would move a small turbine.

stacking more wont work. the air flow will get restricted.

there are small turbines would be fun to see.

if I had to bet on this one, I would probably bet on "no" , but really, as you said, "you can't have this conversation in a "bubble" so to speak ",  there are many factors involved, one key factor would be the cost of energy in the first place, this process of building and running a slightly complicated system comes at a cost, once everything is taken into account, say for an example that total cost of the newly generated energy is 10 cents per kWh while the cost at the wall socket is 20 cents, then maybe it's worth recycling a bit of airflow,  but if it's 14 cents vs 15 cents, then probably not.

Personally, If I had the time and the mood, I would run such an experiment just for sake of it, but I wouldn't invest in turbines and what not just to attempt it, and if I mine at a hot place where the wasted energy of the miner is almost nothing, I would forget about this idea altogether.

And you can't have this conversation in a "bubble" so to speak.

If all we were discussing is theory then yes. But in reality, you have the cost of the turbine, the setting up and and wiring of the turbine into the power of the facility, the maintenance of the turbine and many other things. Will you ever generate enough electric to offset the cost of it or even come close?

-Dave

That's pretty accurate, let me add one more missing point, in some cases, the miner doesn't even need 20%, in very cold places people have to generate heat from somewhere and direct it to the miner, sometimes they will put the miner in a box to keep it warm for a while, and then point the exhaust of this miner to another miner's intake so that it can run, of course, this isn't something you can find in every mining farm, but it exists and has to be manipulated very carefully for it to actually become a net positive.


It's probably not enough with just one miner, I never owned a turbine like that, but I would imagine that you need a lot of miners to spin it without having to intentionally increase the miner's fans, philipma1957 probably knows better, he has a very good experience with solar and wind stuff.

Okay, got you. So the idea is: the miner only really needs 20% fan speed, but the minimum speed is 30% before the fan just turns off, so e.g. if 80% of those 10% 'wasted energy' can be recovered, it's still a net positive. That makes sense.
However I wonder if the energy difference between 20 and 30% fan speed, subtracting losses, is enough to turn a turbine.

The main point is that you do need some 'wasted energy' that can't otherwise be prevented (like running fans slower in the first place) to be able to 'recoup' something without putting more energy into the system.

Perfectly true, now let's stick to this example for a while, imagine the minimum force you can apply to the pedals will put the bicycle at a speed of 5km/h, but you want it to stay at an average of 2.5km/h, and thus you have to hit the break once a while, that will be a loss of energy, which would otherwise be used to run the dynamo.

Now let's reflect that on an s17 miner, we know there is a minimum speed the fans have to spin at and let's say that is 30%, those fans consume about 50w each at 100%, so at 30% we have to burn 15*4 = 60w.

Now if the miner is still cool at this rate, it will "burn" energy by increasing internal voltage to keep the miner warm (miners can't run too cold for reasons not related to this topic) I remember thierry4wd ran the numbers of how much power is wasted when the miner is colder than it wants to be, I will need to find the thread for that, but let's just assume it's an extra 100w burnt to keep the miner warm.


And since as you mentioned, the fans will move less air when having to spin a turbine, it means the miner will run hotter at 30% fan speed and thus it won't have to increase internal voltage and burn energy for nothing, so in theory, we can save 100w per miner which when pointed to a turbine might generate say 20% of that (after all loses that happens in the conversion process)

The key point here is that only the "otherwise wasted energy" should be used to run the turbine, increasing the power consumption on the miner's side just to run the turbine will of course get you a net loss which is what I said in the first post "but your statement is true if OP intentionally runs the miner's fans at higher speeds just to spin the turbine."

I do agree with both of you if this theory is applied on a basic system like that without taking into account anything else, if OP's idea was to plug a fan in the wall socket just to spin the turbine -- then we wouldn't be having this discussion, but there is a slight difference that makes a lot of difference which is saving the energy wasted by the miner.

With that said, there is a thin line between success and failure in this method, if the airflow is reduced by more than it should, the miner will run hotter than it wants to be and it will either have to increase the fan's RPMs (if it can) or drop the frequency it mines at and thus reduces the miner's profit, if it can't do either of that, the miner will RIP.

Dynamo is a good example. When you turn that on, you have to pedal harder to keep the same speed. If there was a way to convert muscle energy into light directly, that would be much better than turning it into more kinetic energy and then using that to power the dynamo.

What's happening with the wind turbine is that you assume the fans run at the same speed and power input and you just slap a turbine in the back, but it will get harder for the fans to run once you add it; 'more harder' in fact than what you get out of the turbine. Hence it's a losing game.

This doesn't take into account the miner at all. It's basically a system with a motor and a generator coupled together. Imagine an electric car: it has regenerative braking, but that, well, slows down the car. You can't drive around and attach a huge dyno to charge the batteries a little without having to put more energy into the motor to keep the same speed. Pretty much the same idea.

You are going to consume more than the [normal driving energy] + [generated energy], in fact it will be: consumed₀ = [normal driving energy] + [generated energy] + [losses]. It means you would get more range with consumed₁ = [normal driving energy]. Without loss, it would be a net zero (nothing won, nothing lost), except more moving parts and having to buy a generator. That's the best case scenario in a world without losses and friction.

It is more like how a dynamo or an altenator works, far from perpetual motion concept, when you place a light on bicycle and a let the wheels induce enough power via the dynami, the lights will turn on, there is no free energy in this concept, it is a simple conversion of energy, similar to OP's idea.

Of course, I would not call this a very effienct way of creating energy, but miners fan have a minumum speed they run at depending on the room temp, if the miner gets too cold the software will increase the voltage to keep the miner hot enough which will increase the overall power consumption, so if you are in a cold place you might very well use all the resistance you get from pushing the exaust through the turbine to help keep the miner warm without having to increase the voltage.

On the other hand, if the turbine creates enough resistance that forces the miner to run at a higher speed that leads to more loss of energy than otherwise the voltage increace would cause -- a loss of energy in the conversion process will happen.

The idea can work at a small scale if the maths are done in a perfect way, a bit of energy can be converted and reused, it also needs to be in a cold place where a large sum of wasted energy is caused by the fact that the miner is raising its voltage to keep the gear running.

I'm pretty sure Loyce is right on this one, though. The system would be creating energy out of nothing if the fans would run at the same speed and same power input with or without turbine.

In other words: the turbine applies force in the opposite direction, obstructing the airflow of the fans to be able to run a generator and convert kinetic to electric energy. This means with the same power going to the fans, they will run slower. Or you increase fan voltage to keep same airflow to the hashboard + turn the turbine.
As you can easily see, let's say you have to add 100W of power to the fans to run the turbine and keep the same fan RPM as without turbine, due to inefficiencies and friction you will get less than 100W of electricity out of the turbine.
Otherwise you'd be creating energy out of thin air; closely related to 'perpetual motion' machines.

If you were not to add any extra power to the fans and the turbine would still turn + the miner wouldn't overheat, you'd be better off just running the fans at that RPM without the turbine in the back since it would be much easier to run the fans and they would consume much less energy. You can't get more energy out of the system than what you put in, simple as that.

https://en.wikipedia.org/wiki/Conservation_of_energy

The fans will have to spin regardless, you can't run an S17 without 4 fans attached, the cfm phill gave about represents 100% fan speed, so he will be getting anywhere from 250 to 450 cfm (S17 will hardly run cool with fans below 50% unless it's placed in a very cold room), but your statement is true if OP intentionally runs the miner's fans at higher speeds just to spin the turbine.

I don't think the idea is all that bad, but the problem would be the hot air you push into the turbine, will the turbine run fine at 70c or anything close to that? also, the pressure will fall drastically the further it goes, so I don't think a few S17s will run a turbine at full speed, I also doubt anyone here or elsewhere has an exact answer to your question, you will have to go through a trial and error process, it's an interesting idea nonetheless.

If you want to use the kinetic energy from the fans: you're wasting energy. It doesn't come for free, you have to put in more power than you take out. It's more efficient to reduce the power to the cooling fans.
If you want to use the thermal energy from the miners: low temperature heat isn't very useful other than for space heating.

so lets say he blends  the miners into a pipe.



          |
 _____|______
 |__________|
   |            |
   |            |              
miner     miner



this pipe could do


https://growershouse.com/y-duct-connectors?

s17 blows about 450 cfm

so 2 with  450 + 450 could exhale maybe 800 cfm

can a 12 inch duct directed at a turbine give decent power?

i think 4 s17 and 2 y vent stacked could run a turbine

So you are planning to recycle the hot air from miners and generate free energy from your mini turbine? Such a good idea.
It is just like the same as reusing the waste heat from ASICminer.

I don't know if one miner could spin a mini wind turbine but if you have many ASIC machines combining them into one exhaust duct maybe it can be without temp issue from your miners.

Why not do some experiment and update here I search on Google it seems no one tried it yet.

so I was thinking can how many mines exhaust would it take to push a wind turbine that is 200 watts. would it be better to combine all of them in to one port?