Hacker News Reader

Comment by onethought | Hacker News Reader

I mine on renewables and with 3000 series gpus. I didn’t want to do any of this, I ended up here because my energy retailer decided to tax me for exporting too much energy to the grid. So short of earthing the electricity, my solution was to burn it up in mining, and improve my solar roi.

The 3000 series gpus are the most efficient hashes/watt. So actually by disabling this, you are driving people towards less efficient/worse for the environment ASICS machines.

Yes I batteried my home, and use my solar on everything I can first... there is still excess... especially at peak.

Pretty sure this dismantled most of your argument.

View on HN

You are the exception of all exceptions. What percentage of GPUs being purchased for the purposes of crypto are for reasons like yours vs being bought by already wealthy conglomerates and mining groups to further enrich the new crypto multi-millionaire/billionaire class.

Aren’t wealthy just buying giant TH asic machines?

The gpus are good exactly because they are efficient but if you didn’t care about power consumption, ASICS are more profitable (in terms of cost of setup vs return from mining).

The primary reason I went with gpus, is because I can shut them off based on solar excess in 100w increments. So if I’m 200w in excess, turn 2 cards on. 300w - 3 cards etc.

They're not. Asics are more a function of the coin being mined. Namely, if asics are out and competitive, they generally replace gpus altogether as they're both more powerful and efficient. See btc, where gpu mining is completely worthless. Eth on the other hand has no notable asic out AFAIK, and requires gpu mining. Eth miners, regardless of scale, thus require gpus.

btc mining isn't worthless - just you don't directly mine. You sell your compute to the pools.

This is a bit of a two wrongs don't make a right: the fact that your energy retailer is taxing you doesn't (by itself) make it 'right' to use that energy for mining.

Ofcourse I totally understand you do as you do, on an individual basis, but that just stresses the point that this can only be regulated at the group/population level.

Not really. I’m not sure I even believe the “grids weren’t designed for feed in” bs. Wires don’t care where the electrons come from, and they have no problem with me drawing the same load I feed in... so something doesn’t add up. But I’m no electrician... just cry a little at 7-10kw excess that i either have to pay to feed in or earth.

I don’t see anything I’m doing as a wrong. Literally just wasted energy... why not convert it to $?

https://www.landisgyr.com/ezine-article/making-way-two-way-p...

> In the old days, when power companies were the only generators in town, control devices like regulators, capacitors or relays were designed to assume that power flowed in only one direction. “If they saw power flowing in the other direction, they typically tripped or misbehaved, causing customer outages or power quality issues,” Kuloor says. This tendency has prompted many utilities to update such equipment with reverse-power-flow logic or, in the case of mechanical devices that have no software, replace devices altogether.

Not all local grids are prepared for that.

Additionally, there's the signaling of power supply that generators do between each other.

https://en.wikipedia.org/wiki/Utility_frequency#Frequency_an...

https://solarbuildermag.com/policy/grid-saturation-lessons-f...

The different parts of the grid will use fluctuations in the frequency of the grid to indicate if they've got too much power, or can't supply enough.

This can get more complicated with single house putting power back in that they're matching the frequency, but not controlling it as such. If there is too much supply from single house sources, it can cause the frequency of the grid to vary too much and potentially damage the power plant generators.

When this becomes exaggerated - all the houses start producing more power because the sun came out (or the opposite that it clouds over), this can increase the costs for power.

You've also got the Hawaii problem. The Hawaii grid doesn't have the luxury of shipping its power to the next state over.

https://pv-magazine-usa.com/2020/02/05/how-to-solve-hawaiis-...

https://www.hakaimagazine.com/features/the-hot-mess-of-hawai...

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

https://blogs.scientificamerican.com/plugged-in/3-reasons-ha...

There are times when the rooftop solar in Hawaii exceeded the total demand for power. This meant there was too much power in the wires and that generators needed to shut down (or get into problems with pollution and inefficient combustion).

> The grid can only accept as much power as the island is consuming. Juario must mix and match different sized generators to balance what solar rooftops are producing while ensuring that the generators have enough “spinning reserve”—room to throttle up and down to handle those grid surprises. The Maui Electric chief operator must also keep the generators running hot to prevent inefficient combustion from sending dirtier exhaust up the stacks and violating the air quality rules that protect residents’ health.

https://spectrum.ieee.org/energy/renewables/can-smarter-sola... has a neat map. There were sections of Hawaii where solar was supplying 250% of the demand in that area.

https://www.vox.com/energy-and-environment/2018/11/30/178686... is another useful bit to read. "electrical grid oversupply" is the term to search.

This all feels like a buffering problem. Can power companies not just buy big batteries... you wouldn't even need that many, just enough to signal generators to do the right thing.

(Also no idea all the gadgets in-between house and generators, but assuming all the gadgets can handle load going in and out)

Part of that storage problem is what Tesla and Australia did. https://www.greentechmedia.com/articles/read/australia-picks...

And there are other approaches to power storage. Thermal power storage is popular / useful with large solar installations. Hydro dams associated with a reservoir often have pumped storage where they can pump water back into the reservoir and then use it when its needed.

The signaling and frequency matching is still an issue.

I recall a company I worked at in California had two sets of generators. They had a diesel emergency generator for the data center and also a set of natural gas generators to cut down on power costs (when the price went high). The issue with the natural gas ones is that they needed something else to provide the utility frequency, something about them not being stable/consistent on their own.

Rooftop systems have the possibility / likelihood of also destabilizing the frequency in the grid. https://en.wikipedia.org/wiki/Rooftop_photovoltaic_power_sta...

> An electrical power system containing a 10% contribution from PV stations would require a 2.5% increase in load frequency control (LFC) capacity over a conventional system an issue which may be countered by using synchronverters in the DC/AC-circuit of the PV system. The break-even cost for PV power generation was in 1996 found to be relatively high for contribution levels of less than 10%. While higher proportions of PV power generation give lower break-even costs, economic and LFC considerations impose an upper limit of about 10% on PV contributions to the overall power systems.

There's a reference to that at https://www.worldcat.org/title/ieee-transactions-on-energy-c... - which I don't have access to to quote from.

The key point to this is that if solar power is increased, then the grid as a whole needs to also increase its power to get back in control of the load frequency. That is likely what you're seeing. By itself, this storage isn't sufficient for answering that issue. Ideally, the solar systems would have their own local batteries and respond as part of a smart grid to contribute according to the load frequency. ... But that costs more money for the installation and consumers are hesitant to do that.

https://www.energy-storage.news/news/tesla-powerwalls-hooked...

https://www.provisionsolar.com/general/grid-outages-and-the-...

I think your last challenges the concept of who exactly is the consumer?

Thanks for the thoughtful reply!

Very true. If you're pushing power out onto the grid, you're no longer a consumer but rather a producer and there are other issues that come into play.

Twenty, thirty years ago the amount of power from rooftop installations being pushed out onto the grid was minimal compared to the size of the grid and it wasn't an issue.

If you had an energy producer that was producing 10% of the power and doing significant swings in its production without participating on the wholesale energy market (and not signaling those swings)... the regulators would have shut them down.

But when you've got 10,000 people each contributing 0.001% of the grid power (and all having the same swings)... and not participating on the wholesale market and not signaling their swings, that can be disruptive (and damaging).

https://learn.pjm.com/electricity-basics/market-for-electric... is also a good read and https://www.pjm.com/markets-and-operations.aspx is the live market for that part of the grid.

Out of curiosity, has the mining yet payed for the rig/card?

No where near. But if you factor the tax I was paying (I.e assuming I didn’t earth it). It will reach cost neutral in a year and be returning margin after that.

They are increasing this tax as well shortly... so it will get better.

Reply on news.ycombinator.com