Topic: Idea: Creation of "real" air flow table of ASIC models. For soundproof desing. (Read 179 times)
April 22, 2019, 04:39:33 PM
https://www.flickr.com/photos/156748370@N04/38858833380/in/dateposted-public/
https://www.flickr.com/photos/156748370@N04/25798417567/in/dateposted-public/
https://www.flickr.com/photos/156748370@N04/38858844890/in/dateposted-public/
https://www.flickr.com/photos/156748370@N04/39959287004/in/dateposted-public/
https://www.flickr.com/photos/156748370@N04/26797717828/in/dateposted-public/
https://www.flickr.com/photos/156748370@N04/26797731338/in/dateposted-public/
https://www.flickr.com/photos/156748370@N04/26797725718/in/dateposted-public/
This is the box I made for the D3 that I told you about. I made a series of 10 and another of 15 boxes. I just have to sell 5. I did a back filter and another front to take advantage of the heat in winter. Removing the filters is ready to connect 16cm tubes to be able to conduct hot air through a window. I have also sold them for S9 and for T2T. The customers were quite satisfied. Now I'm designing a box that admits two "single tube" asic and double fan (in and out) or one of "double tube" and also double fan by tube (S17, A9 Zmaster, GMO B3 ...). The question I have when I think of how many "single tube" asic and a single fan would be able to introduce in that design (z9 mini, Avalon 741, ...), the same doubt I have with "double tube" and 1 fan for each tube (Z11, E3, Avalon 10, S15, S11, DR5 ...). When you do not physically have the Asic equipment to test, it is a big problem because you want to commit as few faults as possible. The choice of the diameter of the outlet duct is the most important. Soundproofing anyone does ... but getting the equipment is not hot is the most difficult, especially here in Spain 30º centigrade for 5 or 6 months a year. It forces you to oversize ducts and to expel air outside.
That is why I give so much importance to knowing what "real" flow of air has.
Last client instalation in Cordoba. In addition to videos and other things.
https://www.facebook.com/media/set/?set=a.2556806677679815&type=3&__xts__%5B0%5D=68.ARCrlMLIlP9vsGEZazv1xY6Vb664ItFu6KjiLz7xqF6gVStonIWmn9L63OxxU12v0WFckrSbjYfS7moOJPjAaWnRrwpDqVVmY_3tW9kKDvbG9dLmntvpEbT0TvYZzbGZD6in-ZLOjJzBUd7K_ubFEBXeezr8x2l3Kl3uYvwmc7nwbvVUBQ9YOZYftV9n3o2iA7kX2XkCAANb9sGUF4jwtEfeQS1jqdP-5WrXECg7M9NjGsbtAChpQLEohqD16uv023ArI35gcCcqzz_brrV-ZgB5RAs2L6bw4xH-uPiJzyZJYVQkcP9CK-g4r2Q5w9RBg6r0bEcrTD6dVZLurbOfA687XNQ5VdohFaM5KI9qn8ANw0zK63iDjhE6mLhpqMKbdH0XCBY0rX1mk_lA_3m_KKATe2p1cuN2szSoO_CtJ8alytgMGn_GKxDzJG4-RAQRicZDETC4rJJ4GBa0IhJQZ5X8pA&__tn__=-UC-R
April 21, 2019, 04:35:58 PM
It does have some value.
As I know asic builders change fans a lot.
So knowing if the fan you got from them is good enough is info worth knowing.
To remind you you mentioned z11s which are not sha-256.
But your idea applies to all asics including sha-256.
Bitmain s-9 uses different heatsinks on each miner.
I would think you can have large variance with wind resistance if you pick 25 s9s at random .
Avalon a921 have long solid heatsinks so they should all have the same or close to the same.
Whatsminer m10 also have long solid heatsinks they should measure the same.
I am guessing bitmain had to stack two short tubes on the s15 and s17 due to multiple mini heat sinks inside it and of course heavy duty watts
As I know asic builders change fans a lot.
So knowing if the fan you got from them is good enough is info worth knowing.
To remind you you mentioned z11s which are not sha-256.
But your idea applies to all asics including sha-256.
Bitmain s-9 uses different heatsinks on each miner.
I would think you can have large variance with wind resistance if you pick 25 s9s at random .
Avalon a921 have long solid heatsinks so they should all have the same or close to the same.
Whatsminer m10 also have long solid heatsinks they should measure the same.
I am guessing bitmain had to stack two short tubes on the s15 and s17 due to multiple mini heat sinks inside it and of course heavy duty watts
April 21, 2019, 04:29:10 PM
Good afternoon.
I think it is important to make a table of real air flows (m3 / h or cfm) of each model of Asic (air-cooled) that we know. In this way we will be able to manufacture more efficiently our soundproof boxes for Asic.
The "real" flow (at maximum regime) of refrigeration nobody knows, and I tell you why.
The first thing we think is that it would be enough to look at the reference of the fan of that Asic that the manufacturer gives us and see what air evacuation data gives us. But this is not enough, it is a data for nothing reliable ... we must count on the resistance of the heatsinks and internal plates (of different sizes and configuration in each manufacturer and model) that directly reduce the flow. And there are more variables ... equipment with a only fan, there are with double fan, there are fan with 5000 rpm other with 6000 rpm or 7500 rpm, also equipment with fan input different from the output, ... and single-tube Asic and dual-tube Asic.
It would be interesting to know the real value of air movement of each team in its maximum regime (to make sure) of every Asic model. I have a cheap anemometer and because I do not have a measurement area as big as the fans, I do not trust the values.
I have seen that there is a measuring instrument that would work well, it is a balometer.
https://www.tsi.com/getmetafile/7b25020f-3844-4c43-ae5d-663df38b5e20/EBT731_03
Normally used to measure airflow ducts accurately.
Imaginary case:
From my experience in D3, which is a team of "single tube" and double fan I have determined that for a diameter of helix of 10, 5 cm, which is about 86 cm2, it is necessary a tube for the extraction with a section of 200 cm2 about 16cm in diameter (with two meters in length) so that the equipment is almost not affected. If for example I wanted to design a box that included two Z11 (double tube of a single fan per tube) I could only rule by trial and error in the choice of the section of the evacuation tube, come on that first I have no chance to know with something of accuracy the tube section that I will need.
However ... if we had a table with the references of air extraction real and quite accurate if we could know the amount of asics of a given model that an extraction tube would be able to admit. Imagine that I do the measurement with the balometer in the D3 and it tells me that evacuates about 300 m3 / h and that to stabilize the equipment (based on trial and error) I previously determined that I needed a tube of 16cm in diameter (200cm2) so that would associate that volume with that section. Well imagine that the Z11 to be double tube but a single fan per tube extracted about 430 m3 / h (measured with our balometer) ... what tube would use? Applying the "logic" above apply a "rule of three" ... if for 300 m3 / h was needed 200cm2 for 430 m3 / h will be missing X. X = 430x200 / 300 = 286cm2. And as we intend to put two Z11 would be about 286x2 = 572cm2 what standardized soundproof tubes are for that section? ... of diameter of 25cm (490cm2) and the next 31,5cm in diameter (774cm2). In that case I would opt for the 25cm because it happens that increasing the diameter of a tube greatly reduces the contact surface of this with air and it is very likely that this is a sufficient section. But well this is a calculation of "fat brush".
With all this I think it would be very useful to have a list of "real" flows evacuated from the different Asics that are in the market. It is more ... it would not cost anything that was a data more than into the specifications given by the manufacturer on each model.
I know it is a difficult task for us ... a second hand balometer is worth 1100 euros ... but if someone who had one and was in this "world" would be ideal to be providing data.
A greeting.
Any idea or correction that comes to mind is welcome.
(I hope somebody thinks it's a good idea
)
I think it is important to make a table of real air flows (m3 / h or cfm) of each model of Asic (air-cooled) that we know. In this way we will be able to manufacture more efficiently our soundproof boxes for Asic.
The "real" flow (at maximum regime) of refrigeration nobody knows, and I tell you why.
The first thing we think is that it would be enough to look at the reference of the fan of that Asic that the manufacturer gives us and see what air evacuation data gives us. But this is not enough, it is a data for nothing reliable ... we must count on the resistance of the heatsinks and internal plates (of different sizes and configuration in each manufacturer and model) that directly reduce the flow. And there are more variables ... equipment with a only fan, there are with double fan, there are fan with 5000 rpm other with 6000 rpm or 7500 rpm, also equipment with fan input different from the output, ... and single-tube Asic and dual-tube Asic.
It would be interesting to know the real value of air movement of each team in its maximum regime (to make sure) of every Asic model. I have a cheap anemometer and because I do not have a measurement area as big as the fans, I do not trust the values.
I have seen that there is a measuring instrument that would work well, it is a balometer.
https://www.tsi.com/getmetafile/7b25020f-3844-4c43-ae5d-663df38b5e20/EBT731_03
Normally used to measure airflow ducts accurately.
Imaginary case:
From my experience in D3, which is a team of "single tube" and double fan I have determined that for a diameter of helix of 10, 5 cm, which is about 86 cm2, it is necessary a tube for the extraction with a section of 200 cm2 about 16cm in diameter (with two meters in length) so that the equipment is almost not affected. If for example I wanted to design a box that included two Z11 (double tube of a single fan per tube) I could only rule by trial and error in the choice of the section of the evacuation tube, come on that first I have no chance to know with something of accuracy the tube section that I will need.
However ... if we had a table with the references of air extraction real and quite accurate if we could know the amount of asics of a given model that an extraction tube would be able to admit. Imagine that I do the measurement with the balometer in the D3 and it tells me that evacuates about 300 m3 / h and that to stabilize the equipment (based on trial and error) I previously determined that I needed a tube of 16cm in diameter (200cm2) so that would associate that volume with that section. Well imagine that the Z11 to be double tube but a single fan per tube extracted about 430 m3 / h (measured with our balometer) ... what tube would use? Applying the "logic" above apply a "rule of three" ... if for 300 m3 / h was needed 200cm2 for 430 m3 / h will be missing X. X = 430x200 / 300 = 286cm2. And as we intend to put two Z11 would be about 286x2 = 572cm2 what standardized soundproof tubes are for that section? ... of diameter of 25cm (490cm2) and the next 31,5cm in diameter (774cm2). In that case I would opt for the 25cm because it happens that increasing the diameter of a tube greatly reduces the contact surface of this with air and it is very likely that this is a sufficient section. But well this is a calculation of "fat brush".
With all this I think it would be very useful to have a list of "real" flows evacuated from the different Asics that are in the market. It is more ... it would not cost anything that was a data more than into the specifications given by the manufacturer on each model.
I know it is a difficult task for us ... a second hand balometer is worth 1100 euros ... but if someone who had one and was in this "world" would be ideal to be providing data.
A greeting.
Any idea or correction that comes to mind is welcome.
(I hope somebody thinks it's a good idea
) Jump to: