Pages:
Author

Topic: GekkoScience BM1384 Project Development Discussion - page 28. (Read 146665 times)

legendary
Activity: 1498
Merit: 1030

Does anyone have a suggestion as to a link comparing electric heat to miners?

I have electric heat, and I would have to think that since 99.9% of the energy used in a miner is expressed as heat, they would be nearly as efficient as baseboard heaters. Minus the fan, commo, et cetera.


 Ballpark 99% of the electric used by a miner would end up as heat, with pretty good efficiency at spreading it due to the heatsink and fan(s). Light from the LEDs and whatever goes OUT of the miner as communications signal(s) would be about the only part that doesn't end up as heat.


Quote

Chili miners


 I thought chili were grown, not mined?

 9-)

hero member
Activity: 767
Merit: 500
Novak heated part of his house last winter with an AM Tube. I kept my entire apartment comfortably above freezing with a Dragon, S3-supplemented when necessary. Previous winter the house and shop were warmed substantially by AM Cubes and S1. Pretty handy.

yup, why run a space heater when you can run gear and make BTC aside from keeping warm.

but sobe-it, where can you get a miner for $20 and spit out 700W of heat?!
wait.. ebay, time to get the ye'ol Block Eruptor Cube..
sr. member
Activity: 453
Merit: 250
Novak heated part of his house last winter with an AM Tube. I kept my entire apartment comfortably above freezing with a Dragon, S3-supplemented when necessary. Previous winter the house and shop were warmed substantially by AM Cubes and S1. Pretty handy.

yup, why run a space heater when you can run gear and make BTC aside from keeping warm.
legendary
Activity: 3374
Merit: 1859
Curmudgeonly hardware guy
Novak heated part of his house last winter with an AM Tube. I kept my entire apartment comfortably above freezing with a Dragon, S3-supplemented when necessary. Previous winter the house and shop were warmed substantially by AM Cubes and S1. Pretty handy.
sr. member
Activity: 322
Merit: 250
PlanetCrypto, thank you so very much for the detailed answer. It is as I suspected, and I appreciate your thoughts. Nice set up, if you come across any pictures, I'd love to see one installation. I'll check out the links later.

You're right, I would MUCH prefer a C1 instead of the S3+, but I don't think it makes any sense to chase down a cheap one now and throw MORE money into BTC mining. If one came my way, fine, but I don't want to chase for it.

That said, I'm enjoying this little project, and would consider buying whatever comes from the mind of sidehack and Novak to upgrade the S3+. Sidehack and I don't agree on economics and what constitutes profit, but I like how he works and want to support the effort.

sr. member
Activity: 462
Merit: 250

Does anyone have a suggestion as to a link comparing electric heat to miners?

To the best of my knowledge there isn't one.

I have electric heat, and I would have to think that since 99.9% of the energy used in a miner is expressed as heat, they would be nearly as efficient as baseboard heaters. Minus the fan, commo, et cetera.

This isn't to say that they can't use more energy as a minimum than I would otherwise use from baseboard heaters, I'm trying to decide whether to keep my last S3+ and use it for heat, come October or so.

Per the physics laws of the conservation of energy, the only energy that does not get converted to heat in a miner is the energy used to light the LED's.
Everything else gets converted to heat, at a rate of ~3.15 BTU per watt.
So an overclocked S3 ends up being a ~1200 - ~1300 BTU heater.
A C1 that uses 4 S3 hash boards dissipates about ~2500 BTU per unit.

We have 5 installations that use BM C1's for comfort heat (my own home included) typically the coolant flow from the C1's is manifolded and pumped (w/ Taco hydronic infloor circulators) through PEX tubing as in floor heat.
Last winter all the installations heated the respective buildings at a profit. Most ROI'ed the hardware cost.

A couple of the installations use "off peak" or "Dual Fuel" priced electricity ($0.0465 KWh and $0.08 KWh respectively) this would make an S3 or a C1 profitable given the current network difficulty and price of BTC.
Worst case scenario given the above, is it might cost you a few penny's per day to run them. But certainly cheaper than resistive baseboard heat where there is no compensation to offset the electrical cost.

We are actually working on building a replacement 2 phase immersion cooled solid state hot water boiler. Targeted for the residential comfort heating market/industry. A prototype may be viewed here: https://www.youtube.com/watch?v=fStHV_Q-x7s Which BTW is using 2 S3 hash boards. Our YouTube channel is located here: https://www.youtube.com/user/PlanetCrypto

As a standalone unit, we like the C1's versus the S3's as they make less noise during operation.

Stated simply, I live in Minnesota and heated my home at a profit last winter.
I anticipate doing this again this autumn/winter/spring.
I currently have 2 C1's running in the living room.

If I were you (and I'm not) I'd keep the S3's, strategically place them around the house as needed, and run them in lieu of the baseboard heat. I would probably electrically disconnect the baseboard heaters, wire in outlets, and plug the miner(s) into those outlets.

Hope this helps.
sr. member
Activity: 322
Merit: 250

Does anyone have a suggestion as to a link comparing electric heat to miners?

I have electric heat, and I would have to think that since 99.9% of the energy used in a miner is expressed as heat, they would be nearly as efficient as baseboard heaters. Minus the fan, commo, et cetera.

This isn't to say that they can't use more energy as a minimum than I would otherwise use from baseboard heaters, I'm trying to decide whether to keep my last S3+ and use it for heat, come October or so.

Damn it, here I am, getting sucked into this damn hell ass bitcoin dealio again....
legendary
Activity: 1302
Merit: 1068
Just curious what you guys envision doing with the old cards from S1's S3's and C1's if sidehack does end up producing an upgrade kit?

Are the old boards pretty much scrap at that point? Do they have resale value?


S1 are probably done for, they bring little profit only to people with dirt cheap electricity. S3 are still good imo. I would run all the S3 i could get my hand on.
I guess the S1 extras could be resold to people that want to start little just to learn mining and maybe people who want to run them on free electricity.

if you can profit at 1-1.3w/gh, you can pencil mod the boards low enough to cool them with only a fan, no heatsink. connect a bunch of boards with spacers and a pair of 120mm fans and you can still use them to get 40-60GH/board

however, the modding takes at least 10min/board (closer to 20 all things considered)

I have about 1.1w~/GH, depending on the fan's drain, i run the whole thing, they run at 70gh per board. They're pretty quiet, i like them, they just take a fair bit of space for the hash, but i'll probably bring a few inside my room during winter since at 1krpm they are near silent and they will warm up my room a bit. Smiley

Everything is suddenly more profitable when you need the heat anyways.
legendary
Activity: 3374
Merit: 1859
Curmudgeonly hardware guy
When I did an undervolt study on the S1, my test unit got down around 0.8W/GH board level.
legendary
Activity: 2128
Merit: 1005
ASIC Wannabe
Just curious what you guys envision doing with the old cards from S1's S3's and C1's if sidehack does end up producing an upgrade kit?

Are the old boards pretty much scrap at that point? Do they have resale value?


S1 are probably done for, they bring little profit only to people with dirt cheap electricity. S3 are still good imo. I would run all the S3 i could get my hand on.
I guess the S1 extras could be resold to people that want to start little just to learn mining and maybe people who want to run them on free electricity.

if you can profit at 1-1.3w/gh, you can pencil mod the boards low enough to cool them with only a fan, no heatsink. connect a bunch of boards with spacers and a pair of 120mm fans and you can still use them to get 40-60GH/board

however, the modding takes at least 10min/board (closer to 20 all things considered)
legendary
Activity: 4354
Merit: 3614
what is this "brake pedal" you speak of?
yup MX4

used it on all my folding@home GPU/CPU rigs as well as *cough* my bfl *cough* stuff. as well as my regular daily driver PCs.

the ICD diamond stuff has better heat transfer but is a major PITA to apply. MX4 is very easy to apply, non conductive/non capacitive, no break in period and works great.. much better than the artic silver people still love to use. 

sr. member
Activity: 462
Merit: 250
(I'm kinda leaning on PlanetCrypto to pull that off), probably after the S7 is in the wild, it shouldn't take long to accomplish something.

I have money/BTC in hand "he who shall not be named" won't take it.

Sandwiches sound good, Jimmy Johns it is then.
sr. member
Activity: 322
Merit: 250
Regarding burger funds - maybe buy Compacs instead? I've never really felt great about taking in unspecified donations, so buying a stickminer you at least get something in exchange and I'm figuring on funneling profits into dev anyway.
buy a stick or two and I will try to make a compac solo mining club.

https://bitcointalksearch.org/topic/i-am-floating-an-idea-here-a-sidehack-usb-stick-solo-pool-1157503


admission to the club would be point 1 or more sticks to a proxy that points to ck's solo pool

Since a stick costs about .11 btc  if we point 100 sticks to a proxy and hit a block we split 24 btc well maybe 24.5 btc 100 ways which is .245 btc for the stick = profit 

all ideas welcome on that thread.

That's brilliant!

I've got two, count 'em ~TWO~ coming my way (possibly a third which I am not anticipating).

I might just direct the awesome power of these two Powersauce bars, that unleash the awesome power of apples...wait, I mean, these two Compacs towards such a worthy, and likely wasteful, endeavour.
legendary
Activity: 1484
Merit: 1004
Great Idea phil I will for sure join the pool if you start it. If you need any help let me know.

legendary
Activity: 1274
Merit: 1000

buy a stick or two and I will try to make a compac solo mining club.

https://bitcointalksearch.org/topic/i-am-floating-an-idea-here-a-sidehack-usb-stick-solo-pool-1157503


admission to the club would be point 1 or more sticks to a proxy that points to ck's solo pool

Since a stick costs about .11 btc  if we point 100 sticks to a proxy and hit a block we split 24 btc well maybe 24.5 btc 100 ways which is .245 btc for the stick = profit 

all ideas welcome on that thread. 

That's a great idea, I love it!  Had some surprise bills show up so I'm yet to order my sticks, but when I do count me in.
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
Actual S1 boards from Bitmain, it is not easy to make them run from a USB bus. S3 and S5 boards are probably less difficult but still annoying and require special interface hardware. The only thing I'm saying about S1 boards is they will be replaced with something better.

Our boards would be USB native. This means they could tie to any USB-enabled computer running cgminer with driver-gekko compiled. I'm fairly certain everyone will agree this is better for a continuously upgradeable solution. When I say I'd like to consider S1 as a standard, I mean a mechanical standard - physical board dimension, screw hole position, power dissipation, etc. I do not want to use Bitmain controllers to define any sort of standard, as the controllers are hardware-designed to talk to their own chips and this leaves out the ability to easily make boards from other chips with different protocols or which connect to anything that isn't a Bitmain controller.

If we sold "upgrade kits", we'd probably have an option to buy a Pi or something similar (probably a different board which handles simultaneous ethernet and USB better) alongside various numbers of boards.

Regarding burger funds - maybe buy Compacs instead? I've never really felt great about taking in unspecified donations, so buying a stickminer you at least get something in exchange and I'm figuring on funneling profits into dev anyway.

buy a stick or two and I will try to make a compac solo mining club.

https://bitcointalksearch.org/topic/i-am-floating-an-idea-here-a-sidehack-usb-stick-solo-pool-1157503


admission to the club would be point 1 or more sticks to a proxy that points to ck's solo pool

Since a stick costs about .11 btc  if we point 100 sticks to a proxy and hit a block we split 24 btc well maybe 24.5 btc 100 ways which is .245 btc for the stick = profit 

all ideas welcome on that thread. 
legendary
Activity: 3374
Merit: 1859
Curmudgeonly hardware guy
Actual S1 boards from Bitmain, it is not easy to make them run from a USB bus. S3 and S5 boards are probably less difficult but still annoying and require special interface hardware. The only thing I'm saying about S1 boards is they will be replaced with something better.

Our boards would be USB native. This means they could tie to any USB-enabled computer running cgminer with driver-gekko compiled. I'm fairly certain everyone will agree this is better for a continuously upgradeable solution. When I say I'd like to consider S1 as a standard, I mean a mechanical standard - physical board dimension, screw hole position, power dissipation, etc. I do not want to use Bitmain controllers to define any sort of standard, as the controllers are hardware-designed to talk to their own chips and this leaves out the ability to easily make boards from other chips with different protocols or which connect to anything that isn't a Bitmain controller.

If we sold "upgrade kits", we'd probably have an option to buy a Pi or something similar (probably a different board which handles simultaneous ethernet and USB better) alongside various numbers of boards.

Regarding burger funds - maybe buy Compacs instead? I've never really felt great about taking in unspecified donations, so buying a stickminer you at least get something in exchange and I'm figuring on funneling profits into dev anyway.
hero member
Activity: 868
Merit: 1000
So the S1 extras would be the blades and the controller?  IE the new upgrade is pretty much just using the chassis, fan and heatsink, with boards and controller being included in the upgrade? Or is the "upgrade controller" up to the end user?

Regarding the S1 blades, are you saying those are USB tethered, and can be run from a hub via CGminer?  I have been working through the thread were they were trying to run the S1 from an Rpi and it seemed non trivial from what ive read so far.  Is that inaccurate? Is it really just a usb connection and a matter of compiling CGminer correctly? 

Keep the metal bits and fans.

Bin the blades (Ebay?)

Get ready to buy some good, non-metalic heatsink paste.

Then get ready to send some "Burger" funds to GekkoScience. ....When the "chips" go on sale we can all try for a "Barbecue"
legendary
Activity: 1008
Merit: 1001
I'd say keep the metal parts. Since boards are USB-tethered and run directly off cgminer (S1 boards would have fan headers as well) you could hub several together off a single controller or toss a Pi on each. Using USB spec and a good cgminer driver removes the hardware-specificity of the controller setup on S1/3/5, but unfortunately it also sorta removes the ability to use the exiting controller from S1/3/5 unless someone can come up with a ready solution for breaking out USB on those. I doubt the S1 controller can do it at all. You'll want heatsinks, fans and the frame pieces but most likely nothing electronics would carry over.It's unfortunate, but we think that's for the best in the long run. We want to treat the S1 form factor as a standard for midsize miner (and if you're following discussion for our rack standard proposal, the board size and shape might enter that formfactor as well), which means that using a fairly flexible standardized protocol is better than trying to back every new product into a particular hardware-specific connectivity implementation.

In related news...

I'll probably focus on that board in the short term, and use some of our BM1384 one- and two-chip boards for testing the microcontroller integration. All of the digitals would transfer directly from the 4-chip pod to a 30-chip board; I've already got a good handle on the power systems and layout for the 30-chip board so if we can ever get some BM1385 samples to play with (I'm kinda leaning on PlanetCrypto to pull that off), probably after the S7 is in the wild, it shouldn't take long to accomplish something.

Thanks for the Info, i personally am absoloutly fine with that. I will build me my own custom frame and cooling with some small heat sinks on every BM1385. I am thinking of something like building a giant chimney with one big fan and a board carrousel beneath it, give it a nice colour and my girlfriend is happy to have a new fitment yeeeaaaahhhhhh  Grin Shocked Cool
Pages:
Jump to: