Author

Topic: My residential Solar + Mining farm (Read 2283 times)

full member
Activity: 325
Merit: 110
August 16, 2018, 12:47:27 AM
#60
aren't you afraid of fire hazard with the rigs so close to the ceiling that seems a little "weak"?

Yes definitely a risk to be mitigated.


I have resorted to turning off rigs and fans when the sun goes down Smiley
Z9 mini's run always.

It's the end of winter here - we get about 9.5 hrs full generation growing to 11.5 hrs in summer.
sr. member
Activity: 854
Merit: 277
liife threw a tempest at you? be a coconut !
July 19, 2018, 07:49:59 AM
#59
Most sublime post of the year !!! Impressive, thank you very much... so much panels, so few rigs... much impressed !

aren't you afraid of fire hazard with the rigs so close to the ceiling that seems a little "weak"?
full member
Activity: 1148
Merit: 132
July 19, 2018, 01:49:08 AM
#58
hey im in the process  of specing out a 100 panel 300 watt solar array and also installing
the panels myself, I am debating on doing it using microinverters for each panel to supply full ac power rather than dc and then connect that to my sub panel. 

any thoughts on this ?


From what I understand; Microinverters cost more and greatly improve performance in shading / partial shading conditions. They also allow for panel level diagnostics. They have more points of failure which is good and bad. Good because you know exactly what to fix and it doesn't knock out a whole string.

If you have shading issues, then absolutely consider them. Otherwise, String inverters have lower capital cost and higher rate of return.



I do have shading issues on my garage which i plan to put 40 of the panels up on , my neigbor has a large tree, sometimes during the day it shadows parts of my mining garage roof.  The main house which has no shading issues is very high , and can handle at least 8 panels since its so big and unshaded , maybe go with string connectors there but again its hard to access the roof, a bad string of panels would be a nightmare to fix.

my plan is to use 20 or 30 of the 80 main house panels for the house and run the rest back to the mining
garage via conduit and berried cable.
full member
Activity: 1148
Merit: 132
July 19, 2018, 01:43:31 AM
#57
String inverters have lower capital cost and higher rate of return.

For this sole quoted reason, I avoid microinverters. Theres lots of products out there that claim to optimize your panels along with many other features, but when it comes down to it, just step back and assess your purpose for building your solar array.

My opinion is that if you are in it to offset mining electricity costs and turn a decent profit then go with string inverters. If you are a efficiency hobbyist or a green peace person then go with microinverters.

Yea i was told by supplier the micro inverters add at least $120 to the cost of each panel

the fear of string connectors is if one panel or connector has an issue , breaks dies or loses connection the entire array becomes compromised.  Thats a big issue for me becuase my roof is really high, it could take days to get someone up there to fix it
sr. member
Activity: 784
Merit: 282
July 19, 2018, 01:12:53 AM
#56
String inverters have lower capital cost and higher rate of return.

For this sole quoted reason, I avoid microinverters. Theres lots of products out there that claim to optimize your panels along with many other features, but when it comes down to it, just step back and assess your purpose for building your solar array.

My opinion is that if you are in it to offset mining electricity costs and turn a decent profit then go with string inverters. If you are a efficiency hobbyist or a green peace person then go with microinverters.
full member
Activity: 325
Merit: 110
July 19, 2018, 01:03:53 AM
#55
hey im in the process  of specing out a 100 panel 300 watt solar array and also installing
the panels myself, I am debating on doing it using microinverters for each panel to supply full ac power rather than dc and then connect that to my sub panel. 

any thoughts on this ?


From what I understand; Microinverters cost more and greatly improve performance in shading / partial shading conditions. They also allow for panel level diagnostics. They have more points of failure which is good and bad. Good because you know exactly what to fix and it doesn't knock out a whole string.

If you have shading issues, then absolutely consider them. Otherwise, String inverters have lower capital cost and higher rate of return.

full member
Activity: 1148
Merit: 132
July 18, 2018, 11:18:27 PM
#54
Very cool.  I also have a residential solar system to assist with power cost for mining.  I pay a flat rate of $0.12 per kWh year round.  Currently have 84 260W panels and am in the process of adding an additional 48 280W panels.  The 84 panels produced the following numbers for me this summer:

March 2,391 kWh
April 2,121 kWh
May 2,268 kWh
June 2,578 kWh
July 2,846 kWh
August 2,475 kWh
September 2,172 kWh
October 2,152 kWh

My usage last month was 6,644 kWh, so the 48 additional panels will help, but I'll need another 100 panels at least to "break even".  And since I also constantly add miners, I don't think I'll ever "catch up".

I do all the work myself, so the cost is pretty close to $1 per watt and I get 30% Federal tax credit, so that makes it closer to $0.70 per watt.

hey im in the process  of specing out a 100 panel 300 watt solar array and also installing
the panels myself, I am debating on doing it using microinverters for each panel to supply full ac power rather than dc and then connect that to my sub panel. 

any thoughts on this ? I got a qoute from alibaba for 1000 290 watt mono panels with mounting stuff and the micorinverts already drilled and connected to the solar panel mounts underside for $10k or so( that
exculdes the ac connectors since i have to send measurements of my garage roof first) any

thoughts ?
sr. member
Activity: 784
Merit: 282
June 20, 2018, 07:44:00 PM
#53

And to those who say don't get solar just for mining, I disagree.  Those mining should be the FIRST to look into solar because they will ROI faster from the cost savings.  Overages that would normally be fed back into the grid (with no credit in my situation) are more likely to be consumed in right-sized solar grids supporting mining ops. 


I completely agree. Everyone uses electricity, whether to heat your shower, watch TV, etc. Solar is a good investment whether or not you are into crypto mining.


It's worth the investment only in places where the government gives you an incentive, something in return for the green energy initiative and is a viable place for maximum sunlight, for example tropical countries, also the khw must be greater than lets say $0.20. Australia fits perfectly well the requirements.


Our country gives no incentive at all for using solar energy. Whats worse is that if you want to sell solar energy back to the grid they will only buy it from you at less than half the price they charge per kwh. But despite this, solar is still a profitable investment.

I have the same figures as posted by OP above, 2-3 years time to ROI.
newbie
Activity: 53
Merit: 0
June 20, 2018, 07:22:19 PM
#52
Congratulation, nice and great project sir. I want do that too someday, until the price of solar panel more cheaper at my country. Because i think that alternative source power can make our mining activity more profitable for every month. Thank for info.
sr. member
Activity: 2142
Merit: 353
Xtreme Monster
June 20, 2018, 06:16:19 PM
#51
Install cost was $14,000 USD

Normal consumption would be 4,650 kWh per month. At 18.3c this is $874 per month.

Subsidised cost is $14 daily = $434

Saving per month $480 (54.9% discount on 18.3c = 8.25c)

Solar pays for itself in 29 months (2y 5m 5d)

It's worth the investment only in places where the government gives you an incentive, something in return for the green energy initiative and is a viable place for maximum sunlight, for example tropical countries, also the khw must be greater than lets say $0.20. Australia fits perfectly well the requirements.
legendary
Activity: 2296
Merit: 1031
June 20, 2018, 12:53:08 PM
#50
So I'm thinking more and more about adding solar and yes it has to do with mining.  Now, before I would do this I thought about the condition of my roof.  It's about 15 years old so I figured I better re-roof before adding tons of solar panels.  Also, I operate as a small business S-corporation so I would account for such things as roof renovation and solar installation as business expenses thus reducing my tax obligation.

Any opinions on accounting for such home improvements as business expenses?  My opinion is that the driving factor of roof renovation and solar would be to reduce mining costs in an environmentally friendly way.  Surely this seems to be a legitimate business expense.

And to those who say don't get solar just for mining, I disagree.  Those mining should be the FIRST to look into solar because they will ROI faster from the cost savings.  Overages that would normally be fed back into the grid (with no credit in my situation) are more likely to be consumed in right-sized solar grids supporting mining ops. 









Sorry if this is considered a necro post but a lot of the info here is still relevant and I don't really like searching between so many threads for the same info and I really like this thread because of how relevant everything still is.
full member
Activity: 325
Merit: 110
January 18, 2018, 06:17:27 PM
#49
Can you give me the tl;dr.

How much was the solar panel set-up?
How much would your consumption cost on average monthly from your utility?
How much are you saving on average each month?


Install cost was $14,000 USD

Normal consumption would be 4,650 kWh per month. At 18.3c this is $874 per month.

Subsidised cost is $14 daily = $434

Saving per month $480 (54.9% discount on 18.3c = 8.25c)

Solar pays for itself in 29 months (2y 5m 5d)
full member
Activity: 462
Merit: 115
January 18, 2018, 04:47:47 PM
#48
Can you give me the tl;dr.

How much was the solar panel set-up?
How much would your consumption cost on average monthly from your utility?
How much are you saving on average each month?
full member
Activity: 325
Merit: 110
January 18, 2018, 04:08:22 PM
#47
OK, I finally got around to installing;
- Another shelf
- 4 * exhaust fans (combined; 152W, 1,440m² (15,500 ft²) / min)
- 2 * 20A breakers
- 8 * outlets
- extra rig




The two 20A breakers, with 20A wiring, are wired directly to 2 * double outlets and allow for a maximum of 8A on a socket, up to 16A total (per breaker)

So that's 3,840W * 2 = 7,680W total.
My 4 smart plugs are rated to 2,400W so no troubles there in putting 1920W on each.

Thanks to my smart meter, I know exactly what I use currently. Right now I'm pulling a constant draw of 5,230W

You can see in this graph, grid-consumption is zero from about 6:30am to 4:30pm. This is when I'm self-consuming. Production is so high that I still manage to feed back to the grid around 75kWh. Seeing as I get credited for feeding in at a rate of 50% the grid rate, this effectively offsets another 37.5kWh of usage. So my 67kWh of grid-use is offset down to 29.5kWh, which is roughly $7 per day. Not bad for 5,230W constant draw! I wish they were all sunny days like this example.



I have 5 more cards at the post office, ready for pick-up. They will go straight into spare slots on these motheboards, bringing me to 5800W constant draw. Remembering that my new breakers are good for 7,680W, there is still room to expand! The limiting factor now is shelf space. But I think clearly there is room for 1 more rig on the left shelf.

More soon!
full member
Activity: 350
Merit: 100
December 13, 2017, 05:40:27 AM
#46

No wonder you have such a pessimistic view of solar. $40k for 10KW of panels that make 20KWh in summer.

I had a record the other day - 141KWh off my 21.6KW of panels. 12th day of summer too.

40 summer 20 winter.
sr. member
Activity: 257
Merit: 250
December 13, 2017, 04:55:09 AM
#45
A terrific project, if I had money, I would certainly have embodied something like that. Unfortunately in our climatic zone there are not many sunny days ...
full member
Activity: 325
Merit: 110
December 13, 2017, 04:22:24 AM
#44
Anyone thinking about going solely strictly for mining, just stop it already lol.

I do have a 10kw panel on my house, we got it before I started mining, it is currently producing 20kw/day in the summer it was doing doing about 40kw/day.

Basically it barely offsets 1 rig lol. So if you're thinking you're going to plop down 40k to offset the power cost of 1 or 2 mining rigs.... you're a bad dude lol

You must be retarded if you going to pay 40k for 10kW system lol

It was what it was. Solar City is probably the most expensive, that's who we went with last year.

Done and paid for.

No fucks given.

No wonder you have such a pessimistic view of solar. $40k for 10KW of panels that make 20KWh in summer.

I had a record the other day - 141KWh off my 21.6KW of panels. 12th day of summer too.
full member
Activity: 350
Merit: 100
December 13, 2017, 04:04:33 AM
#43
Anyone thinking about going solely strictly for mining, just stop it already lol.

I do have a 10kw panel on my house, we got it before I started mining, it is currently producing 20kw/day in the summer it was doing doing about 40kw/day.

Basically it barely offsets 1 rig lol. So if you're thinking you're going to plop down 40k to offset the power cost of 1 or 2 mining rigs.... you're a bad dude lol

You must be retarded if you going to pay 40k for 10kW system lol

It was what it was. Solar City is probably the most expensive, that's who we went with last year.

Done and paid for.

No fucks given.
full member
Activity: 236
Merit: 250
December 13, 2017, 03:37:10 AM
#42
You must be retarded if you going to pay 40k for 10kW system lol
$1/W is the going rate for a few panels, you should be able to get them for even less in bulk.
newbie
Activity: 25
Merit: 0
December 11, 2017, 07:53:28 AM
#41
Excellent topic: using renewable energies for mining. This is a great Project. Thanks for sharing
legendary
Activity: 1274
Merit: 1000
December 09, 2017, 08:40:42 PM
#40
were hoping to be solar by next summer and were not renting it because the GOV perks you should get even if you rent or part of it  you don't get but if you Buy it you get it all, to me the company's that rent out solar are F*** you over, all you get out of it is maybe cheap Electricity no increase in poverty value etc ... and they get 1000's in cash you should get even if you rent for going solar.. but that's how the GOV set it up ripe off the little guy and give it to the rich ... an that tax cut they want to do to us in the US is bad ..... which is not law yet.



Don't pay 40 k for any solar system on any house/home the Price is dropping that fast .. at most you should only pay 10k if that Ive seen some system were i live go for 7k for a 10kW system or less  ... i would shop around before buying a solar system for a house/home and don't rent.

Where is this? United states of Trump?


I wonder to , don't worry that a-hole won't be around much longer and i hope the democrats learned from this once they get control back an they use it right ...after all every thing we have in the US was done for the people by the people or by the Democrats with the Republicans fighting it all the way an once the Republicans got control they can't get shit done  an what they want to do will hurt us an make things worse, i always thought it was the Democrats causing problem's like all most everyone was saying, was i wrong  ..I learned a lot with the want a be dictator we have for a President in the US. I'm 61 and never really read the  Constitution for the United States of America in detail with trump I had to.


cost of build solar?

Avg cost in the US around 5 to 10k or less for a full solar system you can use like you do normal Electricity .. but buy it don't rent it .

You can buy a Portable system but I'm not to sure how well they work and are they worth it ? ..just for mining if that's what you want .

i want full solar so i don't have to pay the Electric Company like i do now i know i may have to for night time use but selling to Electric Company is a nice feeling knowing they have to buy it back by law in my state ... an I'll max run it because i can ...

And i was told don't know how true it is, you can buy it over sea or out of the US even cheaper the reason the cost is so high in the US is greed and what GOV lets them do to us and there not held accountable for most of it, so they keep doing it .
full member
Activity: 336
Merit: 100
December 09, 2017, 08:40:35 PM
#39
Anyone thinking about going solely strictly for mining, just stop it already lol.

I do have a 10kw panel on my house, we got it before I started mining, it is currently producing 20kw/day in the summer it was doing doing about 40kw/day.

Basically it barely offsets 1 rig lol. So if you're thinking you're going to plop down 40k to offset the power cost of 1 or 2 mining rigs.... you're a bad dude lol

You must be retarded if you going to pay 40k for 10kW system lol
full member
Activity: 336
Merit: 100
December 09, 2017, 08:38:59 PM
#38
I paid $6200 for 5kW system (Perth WA), it is producing 10MW ($2500 worth) per year, and I sell absolutely nothing back to the grid, so the ROI on the solar is 2,5 years, which is bloody awesome.
full member
Activity: 350
Merit: 100
December 09, 2017, 08:19:09 PM
#37
Anyone thinking about going solely strictly for mining, just stop it already lol.

I do have a 10kw panel on my house, we got it before I started mining, it is currently producing 20kw/day in the summer it was doing doing about 40kw/day.

Basically it barely offsets 1 rig lol. So if you're thinking you're going to plop down 40k to offset the power cost of 1 or 2 mining rigs.... you're a bad dude lol
full member
Activity: 325
Merit: 110
December 09, 2017, 08:13:54 PM
#36
Where is this? United states of Trump?

Hi all. I live in Australia...

Cheesy

I've added the second shelf and some cooling. I'm currently waiting on 4 * Vega64 and a few 1070ti's in post... more pics next week!
newbie
Activity: 42
Merit: 0
December 03, 2017, 08:18:20 AM
#35
Where is this? United states of Trump?
legendary
Activity: 1764
Merit: 1002
December 03, 2017, 08:09:22 AM
#34
try use regulator direct to mobo&GPU rather than using inverter to psu, you are losing 30%-40% efficiency
full member
Activity: 224
Merit: 100
CryptoLearner
December 03, 2017, 05:25:32 AM
#33
I calculate the solar installation to ROI in 4.5 years at absolute worst case scenario (where 100% is export). So we will be far in front of that. I justified the solar on it's own before I added miners into the mix.
Oh damn, this is a nice ROI, here if you can ROI in 15 years you are lucky. Not enough sun  Angry

Keep up the good work, it's awesome  Cool
full member
Activity: 236
Merit: 250
December 02, 2017, 11:49:32 PM
#32
Batteries would need to be half their current cost for me to consider it. The problem I have is that I can store 60KWh per day and batteries are $10,000AUD per 12KWh...
Nissan Leaf modules are going for about $200/kWh, probably even less in quantity. There are also LiFePO4 cells which are even cheaper (like the ones I used in my setup) but building up a sizable pack out of those is a lot of work...
full member
Activity: 325
Merit: 110
December 02, 2017, 10:29:46 PM
#31
Have you thought about finding a way to feed the DC from the solar panels directly into the 12V consumers and bypass the two transformation stages?

My Inverter is quite efficient at max 98.1% efficiency. Typical is 97%. Platinum PSU's are 93%. It's important to note that at this stage, my mining consumption is just 15% of my peak solar output. So the vast majority of generation is feeding back into the grid and earning me money.

What do u say is more econmoical adding more rigs and increase revenue or add solar power and decrease power costs

I basically filled my roof with solar panels. I have one orientation that is 30% less effective due to being south-west. It aslo has shading issues from a neighbouring 2 storey home. So I put as many up there that I could and that's that. I encourage all to do the same but obviously different countries have different economics for payback.

What's coming down from his panels aren't 12V, they're likely strings somewhere in the 150-300V range.  Even individually, larger panels like this individually are above 25V themselves.  Not to mention the solar power isn't constant enough; you'd be varying input voltage constantly.  You need the inverter to smooth it out and make it a clean and stable power.

Correct. Each 'string' is 360V. The North orientation is 360V at 18A and the East orientation is 360V at 36A. Each panel is 30V.

Look nice mate, nice solution, but how much time do you expect to ROI, must have cost quite a bit  Grin

PS : please clean up this CPU fan on your rig  Cry  Grin

I calculate the solar installation to ROI in 4.5 years at absolute worst case scenario (where 100% is export). So we will be far in front of that. I justified the solar on it's own before I added miners into the mix.

It's not going to work...

You haven't factored in the draw when you're charging your Tesla and I know you want one!

They can't build them quick enough for me Smiley

cost of build solar?

$18,500 AUD ($14,000 USD) With a better than 4.5 year payback. From crypto earnings and tradings, I have paid for all computer equipment and 25% of solar install already Wink

As for batteries the Lithion/Ion or polymer packs are still way too expensive even if the solutions such as the Tesla one are quite interesting, on the other hand the old acid one are pretty cheap I think but dunno what are the price in Australia especially considering import tariffs and what's not, the battery should help with excess power but more importantly making you graph look linear.

Batteries would need to be half their current cost for me to consider it. The problem I have is that I can store 60KWh per day and batteries are $10,000AUD per 12KWh...
jr. member
Activity: 58
Merit: 4
December 01, 2017, 04:29:48 PM
#30
48 V DC power supplies
https://www.powerstream.com/DC-PC-48V.htm

Efficiency seems to be around 75%, which is less than impressive.


Supermicro also has some:
http://www.supermicro.com.tw/products/nfo/power_supply.cfm?pg=PWS&type=-48Vdc#type
Don't find any info on efficiency but likely much better.
member
Activity: 210
Merit: 10
November 30, 2017, 11:57:42 PM
#29
Correct which somewhat defeats the purpose of Maximum Power Point Tracking.  The battery will store the energy that otherwise just turns into heat.  If you're not utilizing the potential of the panel at all times, you're wrecking your ROI versus what you get off the grid.

Also if you're doing this off a 12 volt panel and using it at a 12v power supply, you're artificially dragging the voltage down using the MPPT (unless your input tolerance is way higher).  This makes cabling costs and voltage drop an even bigger factor and a full stop for anything larger than 1000W.
If you make sure the load is always at least what the solar panels are capable of producing, then the full power generated by the panels can be used. The remainder will be pulled from the mains. You can even have just some of the panels supplying the hacked PSUs with the remainder supplying an inverter as is usually done. In that configuration, the required inverter will be quite a bit smaller, since those aren't cheap. (The idea of hacking a hybrid car inverter into a solar inverter/rectifier is an interesting one since those are dirt cheap for their power capacity, but requires a very expert level of electronics knowledge.)

Man if you have the time, knowledge, and parts to make a hybrid inverter into a solar one, more power to you!  Yea that's totally a pun, but truly a very fun idea.
full member
Activity: 236
Merit: 250
November 30, 2017, 10:41:23 PM
#28
Excellent project, feeling jealous of you. It is really a great way to use solar power in mining.
I have to admit that I'm jealous of my best friend Naomi Wu! Not only is she smart, she's very beautiful and artistic. I kind of wish I could be as good looking as her so that I could host a nice TV show of my own.

I also have to say that another one of my friends, Allie Moore, is really jealous of me now. For a bit of background info, she and I have been trying to "one up" each other in going green. That kind of slowed down once I got a Prius since she couldn't afford a new car, then she needed to replace the HVAC system in her house so I suggested one of the Gree smart systems. She didn't care very much about my cryptocurrency mining (not even how I largely focus on energy efficient coins), but add the solar and the fame and I managed to rise to a level that would be very difficult for her to one up.

To get back on topic, the best way we could promote solar power is to share information on how to build effective DIY solar setups. From beginners who need a bit of guidance to put together their first solar power system to experts who design open hardware for use in solar power. Solar power and cryptocurrency mining unfortunately don't tend to go together as a good investment. For those who have built large solar mining setups, what are they planning to do with the solar panels once mining is no longer profitable?
member
Activity: 154
Merit: 10
DEm1CKDTViM1y9YmEcBaktNLWVx8rwuQUm
November 30, 2017, 10:18:24 PM
#27
Excellent project, feeling jealous of you. It is really a great way to use solar power in mining.
full member
Activity: 236
Merit: 250
November 30, 2017, 10:17:18 PM
#26
Correct which somewhat defeats the purpose of Maximum Power Point Tracking.  The battery will store the energy that otherwise just turns into heat.  If you're not utilizing the potential of the panel at all times, you're wrecking your ROI versus what you get off the grid.

Also if you're doing this off a 12 volt panel and using it at a 12v power supply, you're artificially dragging the voltage down using the MPPT (unless your input tolerance is way higher).  This makes cabling costs and voltage drop an even bigger factor and a full stop for anything larger than 1000W.
If you make sure the load is always at least what the solar panels are capable of producing, then the full power generated by the panels can be used. The remainder will be pulled from the mains. You can even have just some of the panels supplying the hacked PSUs with the remainder supplying an inverter as is usually done. In that configuration, the required inverter will be quite a bit smaller, since those aren't cheap. (The idea of hacking a hybrid car inverter into a solar inverter/rectifier is an interesting one since those are dirt cheap for their power capacity, but requires a very expert level of electronics knowledge.)
hero member
Activity: 588
Merit: 501
November 30, 2017, 08:39:22 PM
#25

We have night-time tariff but it is on a separate meter called controlled load (off-peak). It also needs approval to connect to things like hot water and pool pump. It must be hard-wired and connecting anything else to this is illegal.

A residential customer can at any time change from a 'flat-rate' plan to a 'time of use plan', which has lower night-time rates, but you are killed during the double hump morning/afternoon peaks - up to 40c per KWh at those times. They are just outside my generation window. But I'll admit that I haven't looked too deep into this and will do so after your reminder.

Australia commenced solar incentives in 2009 and initially the feed-in tariff was 60c. This was heavily exploited and nerfed to 6c within 2 years. So today at 13c we are quite happy.
The other type of incentive is a rebate of about -30% install costs.
Typical usage-based solar systems in Australia have a payback period of 4-5 years.

No, I do not have a battery. When they are cheaper, I will definitely get one.


First of all thanks for the reply.

I see so it's quite different and you have alot of constraint with the off-peak power, unlike here, but it might be because here in France we get most of the electricity from nuclear so it's pretty cheap and abundant.
I think one, should do a simulation comparing both, there must be someway to simulate your powerusage vs time vs the variable rate.

4-5 Years sounds really reasonsable for a payback I think it's pretty interesting to do such a project especially when you couple it to mining.

As for batteries the Lithion/Ion or polymer packs are still way too expensive even if the solutions such as the Tesla one are quite interesting, on the other hand the old acid one are pretty cheap I think but dunno what are the price in Australia especially considering import tariffs and what's not, the battery should help with excess power but more importantly making you graph look linear.
member
Activity: 182
Merit: 10
November 30, 2017, 07:56:18 PM
#24
cost of build solar?
full member
Activity: 322
Merit: 101
November 30, 2017, 06:38:19 PM
#23
It's not going to work...

You haven't factored in the draw when you're charging your Tesla and I know you want one!
member
Activity: 210
Merit: 10
November 30, 2017, 01:24:03 PM
#22
Unless you live in the desert, even with MPPT, the voltage won't be stable enough.  The moment a cloud hits a panel, you're going to see voltage swings without a buffer.  If the 5-8% loss is of that great concern, having a battery as a voltage buffer is the safest way of doing it.  Not to mention the PSU and GPU don't keep a stable draw, it varies constantly.  The MPPT needs something to power track against.  If both the power source and load are constantly varying, that's not the best.  

I think the OP is on the best route.  Operating large solar systems at 12V is NOT cost effective.  The cable sizing required to bring the amount of amps from any length of distance at 12V is outright impossible and certainly not cost conscious.  Keep in mind there are MANY factors in the overall efficiency of a solar system, cabling and inverter are certainly big ones.  Having the solar system bring in the high voltage (the whole point of strings) and invert it (that inverter is likely 98% efficient) to clean 120VAC is worth the few % lost.

I was a solar designer and installer for almost a decade and the question of can I do this right off the panel came up very often.  I have a very similar mining/solar set up myself.
The normal PSUs running from the mains hold the voltage constant. The hacked PSUs (that accept on the order of 300V DC and perform MPPT) just take load off the normal PSUs when solar is available. Could also just tap into the primary side DC bus with one set of PSUs but then the solar setup will have a connection to mains, probably causing problems when connecting many PSUs together.

For my setup, I can (and have) run it without a battery if I add a PSU with a series diode, connecting it where the battery would normally go. It just wouldn't be efficient at using the solar panel since the miners only use 16W or so and the 100W panel almost always makes more than that during the day.

Correct which somewhat defeats the purpose of Maximum Power Point Tracking.  The battery will store the energy that otherwise just turns into heat.  If you're not utilizing the potential of the panel at all times, you're wrecking your ROI versus what you get off the grid.

Also if you're doing this off a 12 volt panel and using it at a 12v power supply, you're artificially dragging the voltage down using the MPPT (unless your input tolerance is way higher).  This makes cabling costs and voltage drop an even bigger factor and a full stop for anything larger than 1000W.
full member
Activity: 236
Merit: 250
November 30, 2017, 11:11:27 AM
#21
Unless you live in the desert, even with MPPT, the voltage won't be stable enough.  The moment a cloud hits a panel, you're going to see voltage swings without a buffer.  If the 5-8% loss is of that great concern, having a battery as a voltage buffer is the safest way of doing it.  Not to mention the PSU and GPU don't keep a stable draw, it varies constantly.  The MPPT needs something to power track against.  If both the power source and load are constantly varying, that's not the best.  

I think the OP is on the best route.  Operating large solar systems at 12V is NOT cost effective.  The cable sizing required to bring the amount of amps from any length of distance at 12V is outright impossible and certainly not cost conscious.  Keep in mind there are MANY factors in the overall efficiency of a solar system, cabling and inverter are certainly big ones.  Having the solar system bring in the high voltage (the whole point of strings) and invert it (that inverter is likely 98% efficient) to clean 120VAC is worth the few % lost.

I was a solar designer and installer for almost a decade and the question of can I do this right off the panel came up very often.  I have a very similar mining/solar set up myself.
The normal PSUs running from the mains hold the voltage constant. The hacked PSUs (that accept on the order of 300V DC and perform MPPT) just take load off the normal PSUs when solar is available. Could also just tap into the primary side DC bus with one set of PSUs but then the solar setup will have a connection to mains, probably causing problems when connecting many PSUs together.

For my setup, I can (and have) run it without a battery if I add a PSU with a series diode, connecting it where the battery would normally go. It just wouldn't be efficient at using the solar panel since the miners only use 16W or so and the 100W panel almost always makes more than that during the day.
member
Activity: 210
Merit: 10
November 30, 2017, 08:32:17 AM
#20
What's coming down from his panels aren't 12V, they're likely strings somewhere in the 150-300V range.  Even individually, larger panels like this individually are above 25V themselves.  Not to mention the solar power isn't constant enough; you'd be varying input voltage constantly.  You need the inverter to smooth it out and make it a clean and stable power.
Hack some cheap server PSUs to perform MPPT, then parallel the output with unmodified server PSUs to get a constant voltage. Hack the load share to give preference to the solar PSUs, with the normal PSUs supplying the remainder. It will take a lot of electronics knowledge to do it.

Unless you live in the desert, even with MPPT, the voltage won't be stable enough.  The moment a cloud hits a panel, you're going to see voltage swings without a buffer.  If the 5-8% loss is of that great concern, having a battery as a voltage buffer is the safest way of doing it.  Not to mention the PSU and GPU don't keep a stable draw, it varies constantly.  The MPPT needs something to power track against.  If both the power source and load are constantly varying, that's not the best.  

I think the OP is on the best route.  Operating large solar systems at 12V is NOT cost effective.  The cable sizing required to bring the amount of amps from any length of distance at 12V is outright impossible and certainly not cost conscious.  Keep in mind there are MANY factors in the overall efficiency of a solar system, cabling and inverter are certainly big ones.  Having the solar system bring in the high voltage (the whole point of strings) and invert it (that inverter is likely 98% efficient) to clean 120VAC is worth the few % lost.

I was a solar designer and installer for almost a decade and the question of can I do this right off the panel came up very often.  I have a very similar mining/solar set up myself.
full member
Activity: 224
Merit: 100
CryptoLearner
November 30, 2017, 05:03:06 AM
#19
Look nice mate, nice solution, but how much time do you expect to ROI, must have cost quite a bit  Grin

PS : please clean up this CPU fan on your rig  Cry  Grin
full member
Activity: 236
Merit: 250
November 30, 2017, 02:47:34 AM
#18
What's coming down from his panels aren't 12V, they're likely strings somewhere in the 150-300V range.  Even individually, larger panels like this individually are above 25V themselves.  Not to mention the solar power isn't constant enough; you'd be varying input voltage constantly.  You need the inverter to smooth it out and make it a clean and stable power.
Hack some cheap server PSUs to perform MPPT, then parallel the output with unmodified server PSUs to get a constant voltage. Hack the load share to give preference to the solar PSUs, with the normal PSUs supplying the remainder. It will take a lot of electronics knowledge to do it.
sr. member
Activity: 588
Merit: 335
Steady State Finance
November 30, 2017, 02:40:56 AM
#17
Amazing U'r mining farm, Congrat. I'm speechless  Lips sealed, Thank's for sharing

member
Activity: 210
Merit: 10
November 30, 2017, 01:45:52 AM
#16
Have you thought about finding a way to feed the DC from the solar panels directly into the 12V consumers and bypass the two transformation stages?

That could potentially give you another 10 - 15% more net power by reducing losses. I know they do things like that in large datacenters sometimes. Unfortunately, 12V PSUs are still not mass produced and therefore expensive. However, compared to high end consumer PSUs and solar panels, it may be very profitable. You could perhaps implement it by using some properly dimensioned battery bank like an UPS and then add some custom wiring to the GPUs.


What's coming down from his panels aren't 12V, they're likely strings somewhere in the 150-300V range.  Even individually, larger panels like this individually are above 25V themselves.  Not to mention the solar power isn't constant enough; you'd be varying input voltage constantly.  You need the inverter to smooth it out and make it a clean and stable power.
full member
Activity: 151
Merit: 100
November 30, 2017, 01:24:57 AM
#15
Wow, really cool project. It seems that with the sky-high electric, solar is a good option for you. Thanks for sharing.
full member
Activity: 236
Merit: 250
November 29, 2017, 10:53:56 PM
#14
Have you thought about finding a way to feed the DC from the solar panels directly into the 12V consumers and bypass the two transformation stages?

That could potentially give you another 10 - 15% more net power by reducing losses. I know they do things like that in large datacenters sometimes. Unfortunately, 12V PSUs are still not mass produced and therefore expensive. However, compared to high end consumer PSUs and solar panels, it may be very profitable. You could perhaps implement it by using some properly dimensioned battery bank like an UPS and then add some custom wiring to the GPUs.

I have done that, but on a much smaller scale with one 100W solar panel. The miners (cheap smartphones and small ASIC board) run from a TI DC/DC module. At 89% efficient, it's a bit dated, but works well enough for now. (I plan to replace it with a homemade 96% converter based on a LT chip just for the fun of it.) There's a LiFePO4 battery pack to provide power at night (although the main purpose is as a UPS/Ohmconnect battery for my IT setup) and since the solar panel easily generates more than what the miners use, I have the DC/DC converter (custom design) in my PC to use whatever is left over.

BTW, I would suggest not having a DC bus voltage that can be both higher and lower than the load voltage, since buck boost converters are less efficient than plain buck converters. The reason that was used in my PC was because back then, lead acid was the only economical option for a UPS battery. (And going 24V would require two batteries which doubles the cost.) When that needed replacement, I replaced it with LiFePO4 which is very close to a drop in replacement. If I were to redesign it from scratch, I would use LiMn batteries which is 14.4V nominal for a 4S configuration, so the DC/DC for the PC would only need to be a buck converter.

What would be interesting for larger setups would be to get a hybrid car inverter and hack it to act as a grid tie inverter/rectifier. With continuous power ratings well into the 10s of kW, it would be plenty for a very sizable mining setup. It would also require a very expert level of electronics knowledge.
full member
Activity: 205
Merit: 100
November 29, 2017, 10:49:41 PM
#13
Nice Man ! I was thinking of adding solar power to my system soon as compared to adding more rigs
What do u say is more econmoical adding more rigs and increase revenue or add solar power and decrease power costs
I was thinking that I should add more rigs till I hit the 5kW mark before adding the solar panels for now only using 1800 watt on 2 rigs will expand soon
Anyways good luck ahead and keep us posted of all the troubles etc Smiley
full member
Activity: 325
Merit: 110
November 29, 2017, 10:32:47 PM
#12
I'm thinking I want to try a vega rig but I need to go and do some more research I think!

Cheers for your info!

Even if I could buy a Vega for as cheap as a 1070ti, I would be reluctant to build another Vega system. I'm sure with new drivers and optimised Linux settings, they would be great. But on windows they are a major headache.
Nvidia on Linux I could get running set and forget in about 15 minutes from the boxes arriving.
newbie
Activity: 32
Merit: 0
November 29, 2017, 10:22:10 PM
#11
WOW, very cool dude!!
Hope it will turn out to be a good choice for you Smiley

Go Green Energy!!  Cheesy
member
Activity: 151
Merit: 11
Website link - Bitcoin Unicorn
November 29, 2017, 09:59:42 PM
#10
Gah -I want to do this here in NZ - our power costs are about the same. Also I'm sick of being at the mercy of our unreliable power suppliers. I looked into solar for a new house build but at the time my main concern was cost savings (no mining impact then). And at the time - there was no cost benefit.

I need to readdress and factor in clean energy, mining and cost on my next evaluation.

So jelly of your rigs too - I'm thinking I want to try a vega rig but I need to go and do some more research I think!

Cheers for your info!
jr. member
Activity: 34
Merit: 10
November 29, 2017, 09:44:21 PM
#9
This is a great idea to save energy for mining. Currently, in my country, there are no investors using solar energy for mining. I am very supportive of your idea and will certainly study more about it.
full member
Activity: 558
Merit: 194
November 29, 2017, 08:56:17 PM
#8
Have you thought about finding a way to feed the DC from the solar panels directly into the 12V consumers and bypass the two transformation stages?

That could potentially give you another 10 - 15% more net power by reducing losses. I know they do things like that in large datacenters sometimes. Unfortunately, 12V PSUs are still not mass produced and therefore expensive. However, compared to high end consumer PSUs and solar panels, it may be very profitable. You could perhaps implement it by using some properly dimensioned battery bank like an UPS and then add some custom wiring to the GPUs.

Interesting idea!  One would need some sort of DC to DC converter to get to a constant 12V output.  And also some way to switch back to AC power at night when the sun doesn't shine.  Here's what the DC output of a typical solar panel looks like during the day:



Solar inverters are quite efficient at 97% typically, and good Platinum PSUs are around 94% efficient.  So there would be a little gain to be had for sure.
full member
Activity: 558
Merit: 194
November 29, 2017, 08:51:25 PM
#7
Very cool.  I also have a residential solar system to assist with power cost for mining.  I pay a flat rate of $0.12 per kWh year round.  Currently have 84 260W panels and am in the process of adding an additional 48 280W panels.  The 84 panels produced the following numbers for me this summer:

March 2,391 kWh
April 2,121 kWh
May 2,268 kWh
June 2,578 kWh
July 2,846 kWh
August 2,475 kWh
September 2,172 kWh
October 2,152 kWh

My usage last month was 6,644 kWh, so the 48 additional panels will help, but I'll need another 100 panels at least to "break even".  And since I also constantly add miners, I don't think I'll ever "catch up".

I do all the work myself, so the cost is pretty close to $1 per watt and I get 30% Federal tax credit, so that makes it closer to $0.70 per watt.
jr. member
Activity: 58
Merit: 4
November 29, 2017, 08:46:41 PM
#6
Have you thought about finding a way to feed the DC from the solar panels directly into the 12V consumers and bypass the two transformation stages?

That could potentially give you another 10 - 15% more net power by reducing losses. I know they do things like that in large datacenters sometimes. Unfortunately, 12V PSUs are still not mass produced and therefore expensive. However, compared to high end consumer PSUs and solar panels, it may be very profitable. You could perhaps implement it by using some properly dimensioned battery bank like an UPS and then add some custom wiring to the GPUs.
member
Activity: 98
Merit: 10
November 29, 2017, 08:31:23 PM
#5
Congrats mate! I am surprised that electricity is ho high in Australia. Hopefully mining will help you pay off those panels

A good friend of mine who lives in Arizona (not mining) makes so much power with solar that he is pushing it back to the grid and gets a check every month.
full member
Activity: 325
Merit: 110
November 29, 2017, 08:23:11 PM
#4
First of all congrats for your project and thanks for sharing, From the data you are sharing I can assume you don't have any power storage solution or at least of scale am I right ? Also I have a second question regarding Australia do you guys have a day and night electricity rates, typically here in Europe we get a day tarif which is more expensive since there a full load on the grid than the night price and this can be beneficial for such a project.

Another thing I find weird there in Australia is that the feed in tariff is less than the price of electricity, where as most countries in western Europe do the opposite putting an incentive for people to go green

We have night-time tariff but it is on a separate meter called controlled load (off-peak). It also needs approval to connect to things like hot water and pool pump. It must be hard-wired and connecting anything else to this is illegal.

A residential customer can at any time change from a 'flat-rate' plan to a 'time of use plan', which has lower night-time rates, but you are killed during the double hump morning/afternoon peaks - up to 40c per KWh at those times. They are just outside my generation window. But I'll admit that I haven't looked too deep into this and will do so after your reminder.

Australia commenced solar incentives in 2009 and initially the feed-in tariff was 60c. This was heavily exploited and nerfed to 6c within 2 years. So today at 13c we are quite happy.
The other type of incentive is a rebate of about -30% install costs.
Typical usage-based solar systems in Australia have a payback period of 4-5 years.

No, I do not have a battery. When they are cheaper, I will definitely get one.
member
Activity: 460
Merit: 12
November 29, 2017, 08:14:51 PM
#3
Wow, impressive! I'm jealous Grin
hero member
Activity: 588
Merit: 501
November 29, 2017, 08:00:55 PM
#2
First of all congrats for your project and thanks for sharing, From the data you are sharing I can assume you don't have any power storage solution or at least of scale am I right ? Also I have a second question regarding Australia do you guys have a day and night electricity rates, typically here in Europe we get a day tarif which is more expensive since there a full load on the grid than the night price and this can be beneficial for such a project.

Another thing I find weird there in Australia is that the feed in tariff is less than the price of electricity, where as most countries in western Europe do the opposite putting an incentive for people to go green
full member
Activity: 325
Merit: 110
November 29, 2017, 07:41:38 PM
#1
Hi all. I live in Australia and pay $0.24.2AUD per KWh (US 18c) of electricity. I have always wanted to install solar, even prior to mining crypto. So when the feed-in tariff was recently increased to 13c (US 10c), I pulled the trigger and had a large system installed.

Here is an aerial shot of my 72 * 300W panels -


The 20KW inverter -


All-time generation data showing daily average production of 93.4 KWh -



A typical sunny day parabola showing >5KW of production from 7am - 5pm (10 hours) -


Due to the parabolic nature of generation, I calculated the 'green % time' in 200W increments of constant draw (mining), all the way up to 3000W, on average throughout the year. As you can see, the first 200W is a massive 49% green while the 2800W-3000W are 38% green -
Code:
--- 	  200W	400W	600W	800W	1000W	1200W	1400W	1600W	1800W	2000W	2200W	2400W	2600W	2800W	3000W
January  56% 54% 53% 53% 52% 52% 51% 50% 49% 48% 47% 47% 46% 45% 45%
February  54% 53% 53% 52% 50% 49% 47% 47% 46% 45% 45% 45% 45% 44% 43%
March  49% 47% 45% 45% 44% 43% 43% 42% 41% 40% 39% 39% 38% 37% 37%
April  46% 45% 44% 43% 42% 40% 40% 39% 38% 38% 37% 37% 36% 36% 36%
May  42% 39% 38% 37% 37% 37% 37% 36% 36% 35% 34% 33% 32% 31% 31%
June  38% 38% 37% 37% 36% 35% 34% 33% 32% 31% 30% 30% 29% 29% 28%
July  39% 38% 38% 37% 37% 37% 37% 36% 36% 36% 34% 33% 32% 32% 32%
August  46% 44% 42% 40% 39% 38% 38% 38% 37% 37% 37% 36% 36% 35% 34%
September 49% 46% 45% 45% 45% 45% 44% 44% 43% 42% 41% 40% 40% 40% 39%
October  52% 51% 49% 49% 48% 48% 48% 48% 47% 46% 45% 44% 43% 43% 42%
November  58% 56% 54% 52% 50% 50% 50% 49% 48% 48% 47% 47% 46% 44% 44%
December  60% 55% 53% 53% 53% 52% 51% 51% 50% 48% 47% 46% 46% 46% 45%
---
Total  49% 47% 46% 45% 44% 44% 43% 43% 42% 41% 40% 40% 39% 39% 38%

Finally, and what was originally the point of this exercise, I will calculate the amount (in terms of W drawn at the wall) of mining equipment I can have running 24/7 while maintaining a zero sum electricity bill.
---THESE NUMBERS ARE ALL IN AUD---

My historical usage is roughly 5,300 kWh per year, of which, 2,000 kWh is expected to be during generation hours. Lets say 800 kWh of generation hours is baseload (200W), which does effect the constant draw and 1200 kWh is on-demand during peak generation (dishwasher / washing etc) which doesn't effect constant draw due to excessively high peak generation.

Combined total generation: 31,844 kWh
Non-Mining self-consumption: 2,000kWh ($0)
Ordinary Export: 29,844 kWh @ 12.87c ($3,841)
Non-Mining Grid consumption: 3,300 kWh ($799)
Service Charge: 365days @ 90.2c ($329)

Typically, I would see a yearly credit of $2,713 on my bill. So lets attempt to use that by way of mining;

Our household baseload is at 200W, so a mining baseload starts there. We have $2713 to 'spend'. We have to factor in that for every 200W of mining equipment, we must also reduce our FIT;

Code:
Draw    Cost    Norm_Cost
200W $327 $424
400W $331 $424
600W $333 $424
800W $334 $424
1000W $336 $424
1200W $337 $424
1400W $338 $424
1600W $339 $424
1800W $341 $424
2000W $342 $424
2200W $344 $424
2400W $345 $424
2600W $346 $424
2800W $348 $424
3000W $349 $424
So we begin to reduce our $2,713 by each increment (starting with $331) and we get to;
1800W constant draw with $24 left over.
So we can have mining equipment with a constant draw of 1600W at the wall and maintain a $0 bill.

Over the last 3 months I have acquired gear and built rigs around the 1.6KW figure.
Here is the frame I use which is made out of $10 worth of aluminium angle and two pieces of 4*2. You just need some tin snips and screws -


You can use angle off cuts for PSU/HDD guide channels. Just put felt buttons under the motherboard -


I have a 5 * 1080ti mining rig -


And a 4 * Vega56 mining rig -


Which together draw 1730W (who's counting Smiley )-


Time will tell just how my rigs balance out. Regardless, I am keen to expand anyway - the summer months will be better for mining due to extended daylight. Janurary is 45% green right up at 3000W. And in winter I can move rigs into bedrooms for some heating capacity.
I am keen to get another rig going for the 3 months of summer. Here is my garage set up so far showing two rigs, the inverter and my IT cabinet. Yes I will add some decent cable management or even run new outlets Smiley As you can see, I have room for another shelf and perhaps 2 more rigs -


Current ventillation Wink . I will be adding exhaust fans above rigs soon -


Version 1 next to Version 2 (I'm improving). Oh boy how sexy are Vegas? Smiley -


Here is a screen management of the miners. The 1080ti rig is mining equihash on nvOC. The Vega rig is managed through Chrome Desktop, running windows 10 and GPU0 hashes slightly slower due to no HDMI dummy. It still it manages 7000 H/s Cryptonight. The other window is the gaming PC that I manage them from, which also has 3 nvidia cards mining equihash.



------------------------------------------------------------------------------

UPDATE 19/01/18

OK, I finally got around to installing;
- Another shelf
- 4 * exhaust fans (combined; 152W, 1,440m² (15,500 ft²) / min)
- 2 * 20A breakers
- 8 * outlets
- extra rig





The two 20A breakers, with 20A wiring, are wired directly to 2 * double outlets and allow for a maximum of 8A on a socket, up to 16A total (per breaker)

So that's 3,840W * 2 = 7,680W total.
My 4 smart plugs are rated to 2,400W so no troubles there in putting 1920W on each.

Thanks to my smart meter, I know exactly what I use currently. Right now I'm pulling a constant draw of 5,230W

You can see in this graph, grid-consumption is zero from about 6:30am to 4:30pm. This is when I'm self-consuming. Production is so high that I still manage to feed back to the grid around 75kWh. Seeing as I get credited for feeding in at a rate of 50% the grid rate, this effectively offsets another 37.5kWh of usage. So my 67kWh of grid-use is offset down to 29.5kWh, which is roughly $7 per day. Not bad for 5,230W constant draw! I wish they were all sunny days like this example.



I have 5 more cards at the post office, ready for pick-up. They will go straight into spare slots on these motheboards, bringing me to 5800W constant draw. Remembering that my new breakers are good for 7,680W, there is still room to expand! The limiting factor now is shelf space. But I think clearly there is room for 1 more rig on the left shelf.

More soon!

------------------------------------------------------------------------------

Jump to: