DataTank Mining: 1.2MW 3M Novec Immersion Cooled 2PH Mining Container

[2112]

but being proud owners of some pricey storage containers and a chunk of land to keep them on.

I don't think that's the case. They way I understand this DataTank Mining is purposefully set up in Panama so that investment in them doesn't create any mortgage, secured debenture, collateralized loan, etc type of relationship. In other words in case of DataTank Mining going bankrupt Allied Control takes back the possession of the equipment leased to DataTank Mining.

It is tried-an-tested method of diluting the liability, the international version of "HashFast,CA vs. HashFast,DE" that is right now playing in a neighboring subforum.

[jimmothy]

If everyone is moving towards 28nm and lower in not too distant future, then many will realize very quickly that physical limits of W/m²K heat transfer in air and single phase liquid can't be overcome.

Another example of lack of understanding of physics and intentionally confusing analysis. The secondary loop in Allied Control's system is single phase. Obviously this poster is neither semi-literate nor semi-numerate. As of now I see two possibilities:

I'm afraid it might be you who lacks the understanding of physics. There is a very hard limit to the max heat density you can cool via heatsink/fans. Notice how heatsink on high powered chips are always much larger than the chips? With immersion cooling you only need a tiny amount of fluid covering the chip.

DataTank Mining also does not (to my knowledge) have much in the way of IP here, and no chance for patents.  There is nothing to stop copycats once the hard work of developing demand is done.  If large miner manufacturers are genuinely interested, I see no reason to need fundraising through an unlicensed Panamanian exchange.  None.  If large manufacturers are not interested, then investing in this will be no different from investing in all the other Havelock offerings--a sure way to lose money.

You are completely forgetting that they are not selling tanks. They are selling an investment opportunity for mining/hosting.

You are forgetting the very thing you seem to advocate--standard form factor.  There are hundreds of motherboards, using different Intel chips, but all could use the same CPU cooler.  Standard form factor has nothing to do with 2-phase immersion cooling, and nothing is stopping manufacturers from choosing one and adhering to it.  They are doing it as we speak--using commercially available CPU coolers.  Common as dirt and only a bit more expensive.  No need for custom engineering, and readily available at amazing prices.

Please do find the price of the cheapest ~300+W heatsink and 100CFM fan. I think you will find that they are not nearly as cheap as you imagine.

It's also a technology that's been around a bit, and is dirt cheap and simple to repair/replace by unskilled help. 

It might be simple but it is not cheap and it is very time consuming. Try plugging in 4000 fans on to asics inside cases/racks. That would be ~40,000 screws at least.

Smaller node size doesn't mean higher W/m2.  KNC is cooling their 20nm chippery with air. 

No, that is exactly what it means. Surprised a self proclaimed expert on everything like yourself doesn't know this.

In US, we invented air conditioning, putting the whole 2-phase thing into a discrete box.  You'd be surprised how well it copes with those hot summer days.  In the winter, we use another paragon of Yankee ingenuity--the Off Switch.

Air conditioning is the least efficient way to cool hardware. You are looking at 1.3-1.5 PUE minimum and I'm sure with a bit of research you can figure out why it's not comparable even though yes, they both involve a phase change.

[IPO Magic]

@Jimmothy:  Please delete your posts, they contribute nothing.
And stop trying to science.  Your brain just sputters and fumes like Novec in a hot fryer, it's unseemly.

300w heatsink   Bro, download moar ram!

[jimmothy]

@Jimmothy:  Please delete your posts, they contribute nothing.
And stop trying to science.  Your brain just sputters and fumes like Novec in a hot fryer, it's unseemly.

I'd ask you to do the same but ignoring you is much more effective.

300w heatsink   Bro, download moar ram!

Heatsinks transfer heat aka watts. Sorry you don't understand elementary school level science.

[2112]

I'm afraid it might be you who lacks the understanding of physics. There is a very hard limit to the max heat density you can cool via heatsink/fans. Notice how heatsink on high powered chips are always much larger than the chips? With immersion cooling you only need a tiny amount of fluid covering the chip.

Another fine example of somebody with high-school education giving engineering advice. The limit on two-phase immersive/evaporative cooling is created by vapor bubbles that have significantly lower heat conductivity than the liquid phase. For Bitcoin mining this shouldn't be the case because the devices work in a steady state. It really matters only with devices working in the impulse modes, where the time between the power impulses is higher than the time to release the bubble from the surface and replace it with a fresh, colder liquid.

Here is a fun Google Images search:

http://www.google.com/imghp?q=vapotron

Vapotrons were designed for immersive/evaporative cooling with distilled/deionized water, and be capable to operate in the impulse modes of "Class C" analog power amplifiers. Note the carefully designed evaporation grooves. What you can't see is the special processing of the surface to produce many steam nucleation sites.

Again, this doesn't really matter for Bitcoin mining. But for a really high-power application you really would've designed bubbling groves (thick & shallow heatsinks, really) and care about producing many vapor nucleation sites on the surface. This means that mirror-smooth backs of silicon chips are actually rather inefficient for immersive-evaporative cooling.

I remember that there is a really nice high-school level demonstration of this effect. In a really smooth and clean glassware one can heat water in a microwave to more than the boiling point. Then you can induce explosive boiling by dropping a rusty metal pin into the super-heated liquid.

The good thing in Bitcoin mining is that the coarse black plastic IC packages naturally have a lot of nucleation sites.

[jimmothy]

The limit on two-phase immersive/evaporative cooling is created by vapor bubbles that have significantly lower heat conductivity than the liquid phase. For Bitcoin mining this shouldn't be the case because the devices work in a steady state. It really matters only with devices working in the impulse modes, where the time between the power impulses is higher than the time to release the bubble from the surface and replace it with a fresh, colder liquid.

Thanks for the info. But do you not admit that immersion cooling is capable of cooling much higher power densities than air cooling?

I didn't even know that was up for debate.

[Ozziecoin]

Of course the 3 months _initial_ build time is long, but the DataTank plan is long term. After that it only takes a few hours and everything is re-used. Again and again. And don't forget that after the initial phase, the production pipeline is setup and we can produce 1 or a hand full or even more per week.

Re: build time. My understanding is that DTMA will purchase chips at the time it is ready to deploy? Is this correct?

Antirack - can you confirm chip purchase timeline?

Bump. Would like an answer from the horse's mouth on chip purchase timing for DTMA.

[Collider]

Of course the 3 months _initial_ build time is long, but the DataTank plan is long term. After that it only takes a few hours and everything is re-used. Again and again. And don't forget that after the initial phase, the production pipeline is setup and we can produce 1 or a hand full or even more per week.

Re: build time. My understanding is that DTMA will purchase chips at the time it is ready to deploy? Is this correct?

Antirack - can you confirm chip purchase timeline?

Bump. Would like an answer from the horse's mouth on chip purchase timing for DTMA.

That question was already answered. (And might also be in the prospectus).

DTM will purchase chips/harware in time for deployment.

Todays´ chip prices were used as a reference point only.

If more hardware can be deployed for the same price, you will get awarded additional units.

[Ozziecoin]

Of course the 3 months _initial_ build time is long, but the DataTank plan is long term. After that it only takes a few hours and everything is re-used. Again and again. And don't forget that after the initial phase, the production pipeline is setup and we can produce 1 or a hand full or even more per week.

Re: build time. My understanding is that DTMA will purchase chips at the time it is ready to deploy? Is this correct?

Antirack - can you confirm chip purchase timeline?

Bump. Would like an answer from the horse's mouth on chip purchase timing for DTMA.

That question was already answered. (And might also be in the prospectus).

DTM will purchase chips/harware in time for deployment.

Todays´ chip prices were used as a reference point only.

If more hardware can be deployed for the same price, you will get awarded additional units.

Sorry, you are not the horse  . I'm interested in knowing when they will actually buy the chips. If they have bought now then there will be no extra money for extra chips. "In time for deployment" is different to "at time of deployment". If they buy chips now but take two months to install and "test the chps", the ROI will be very different. This affects my decision on whether to invest in dtma or dtmb.

[Collider]

In time for deployment means early enough that finished miners will be ready when the tanks are ready.

If you want to test something with the chips, you obviously don´t buy a bulk amount of chips and have them laying around, you buy some chip samples.

[Ozziecoin]

In time for deployment means early enough that finished miners will be ready when the tanks are ready.

If you want to test something with the chips, you obviously don´t buy a bulk amount of chips and have them laying around, you buy some chip samples.

You don't really think "testing" means testing? 

[Collider]

In time for deployment means early enough that finished miners will be ready when the tanks are ready.

If you want to test something with the chips, you obviously don´t buy a bulk amount of chips and have them laying around, you buy some chip samples.

You don't really think "testing" means testing?  

I am not quite sure what you are implying.

You get some some chips and build your first blade designs.

If it works out fine you go for the bulk order at bulk prices, nobody buys a Ton of chips just to have them lying around and loosing value. (Well anyone but hashlast )

[Ozziecoin]

In time for deployment means early enough that finished miners will be ready when the tanks are ready.

If you want to test something with the chips, you obviously don´t buy a bulk amount of chips and have them laying around, you buy some chip samples.

You don't really think "testing" means testing?  

I am not quite sure what you are implying.

You get some some chips and build your first blade designs.

If it works out fine you go for the bulk order at bulk prices, nobody buys a Ton of chips just to have them lying around and loosing value. (Well anyone but hashlast )

One would hope this is what actually happens, and if this is confirmed by Antirack I'll be happy to invest.

[RoadStress]

In time for deployment means early enough that finished miners will be ready when the tanks are ready.

If you want to test something with the chips, you obviously don´t buy a bulk amount of chips and have them laying around, you buy some chip samples.

You don't really think "testing" means testing?  

I am not quite sure what you are implying.

You get some some chips and build your first blade designs.

If it works out fine you go for the bulk order at bulk prices, nobody buys a Ton of chips just to have them lying around and loosing value. (Well anyone but hashlast )

I think that he is implying that "testing" will be "hashing for ourselves".

[antirack]

They way I understand this DataTank Mining is purposefully set up in Panama so that investment in them doesn't create any mortgage, secured debenture, collateralized loan, etc type of relationship. In other words in case of DataTank Mining going bankrupt Allied Control takes back the possession of the equipment leased to DataTank Mining.

This is incorrect. You speculate information that is openly available on the site or in the prospectus.

- DataTank Mining is not setup in Panama
- DataTank Mining is a real company with real people, not a shelf company
- It is our location with production line, office, CNC machines, etc.
- The address of DataTank Mining is on the web site and in the prospectus, openly
- It is the same address we have used for years

We have also received visits from forum members before, including as early as 2012 or just last week. Including some of the hardware manufacturers, including AM, Spondoolies, etc

This is not a one man show or shelf company.

DataTank Mining does not lease systems from Allied Control.

http://www.datatank-mining.com/faqs/partnership-with-allied-control/

The agreement encompasses to purchase DataTank container mining units at production cost value for internal use at DataTank Mining.

And just to be clear, we have free access to all of Allied Control's resources including production line and so on.

You can contact us any time if you have any questions, also directly. I had previously sent you a PM.

It's a honest business plan with solid technology. It would be extremely nice if you stop to speculate or make wrong claims. Thank you.

[antirack]

My understanding is that DTMA will purchase chips at the time it is ready to deploy? Is this correct?

Antirack - can you confirm chip purchase timeline?

Bump. Would like an answer from the horse's mouth on chip purchase timing for DTMA.

We are working with a number of chip manufacturers and keep working on it. The purchase (= when money changes hands) happens when everything is in place for manufacturing of boards and deployment of hardware. This has many advantages.

After chips are purchased they will go straight to our SMD line and then into our immersion tanks and be online without delay. This process can be made transparent, in order to avoid speculation of mining for self-profit (the "testing" referred to in the other posts). However, there will be shipping involved to the final destination. With previous installations, we have literally produced boards in the morning and deployed them in the afternoon. QC testing is usually done in the production line and takes less than a minute per board (visual inspection and power up/probing with pogo pins in test fixtures + real functional test). The first time the boards start mining will be when they create revenue for the unit holders.

Some parts (PCBs etc) need to be ordered in advance (weeks) for larger quantities. At that time the choice of chip/manufacturer needs to be clear. We have a good solution for that.

Let me know if that helps.

[Ozziecoin]

My understanding is that DTMA will purchase chips at the time it is ready to deploy? Is this correct?

Antirack - can you confirm chip purchase timeline?

Bump. Would like an answer from the horse's mouth on chip purchase timing for DTMA.

We are working with a number of chip manufacturers and keep working on it. The purchase (= when money changes hands) happens when everything is in place for manufacturing of boards and deployment of hardware. This has many advantages.

After chips are purchased they will go straight to our SMD line and then into our immersion tanks and be online without delay. This process can be made transparent, in order to avoid speculation of mining for self-profit (the "testing" referred to in the other posts). However, there will be shipping involved to the final destination. With previous installations, we have literally produced boards in the morning and deployed them in the afternoon. QC testing is usually done in the production line and takes less than a minute per board (visual inspection and power up/probing with pogo pins in test fixtures + real functional test). The first time the boards start mining will be when they create revenue for the unit holders.

Some parts (PCBs etc) need to be ordered in advance (weeks) for larger quantities. At that time the choice of chip/manufacturer needs to be clear. We have a good solution for that.

Let me know if that helps.

Yes good info. Assuming manufacturing is ready to go, what is the estimated timeline from chip purchase to shipping to immersion tanks? Reasonable approximation is fine.

[Collider]

We have literally produced boards in the morning and deployed them in the afternoon

That is for pick and place, aswell as reflow soldering. (Which in moderate volume can obviously be done in one day.)

Larger PCB orders take weeks to do cheaply (the earlier you order your PCBs and the higher the amount, the cheaper per PCB).


If your PCB prototype works well, you finish it and go into mass PCB production.
This happens in parallel to the manufacturing of the containers, so that everything is ready once the containers are ready.

The actual miner production isn´t what takes the most time. Ordering of mass quantities of parts has to be done beforehand in order to avoid shortages.

[NotLambchop]

...
Please do find the price of the cheapest ~300+W heatsink and 100CFM fan. I think you will find that they are not nearly as cheap as you imagine.
...

...

300w heatsink   Bro, download moar ram!

Heatsinks transfer heat aka watts. Sorry you don't understand elementary school level science.

Once again, i'm forced to educate the great unwashed

When I was a little kid, I wondered why resistors weren't simply labeled in volts, so if i wanted to drop the voltage by 2V, all I'd have to do is use a 2V resistor.
I was playing with scraps of wire I stripped with my little teeth.  I was about five years old.
Somewhere, across the globe, Jimmothy was already discovering the thrills of huffing the colorful liquids found under the kitchen sink.

Asking for a 300W heatsink is the equivalent of asking for a 2V resistor, or "a pipe that flows 2 gallons of water per minute."
That's why CPU heatsinks are not rated in Watts.
That's why you need a bigger heatsink to keep your CPU @45 degrees Celsius in a room @30 degrees Celsius than you would need to keep it @50 degrees Celsius in a room @15 degrees Celsius.

I know I promised today would be spent learning about shapes and colors, but ur simply not ready

[Collider]

Please spare this thread with offtopic discussions.

Go build your magic heatsink miners and stop spamming this thread.