Bitmain just announced the AntMiner A3 - 815GH/s Siacoin miner - Blake(2b)

[greyday]

Ummm wtf this video is going around saying the Antminer a3 emits HCHO/Formaldehyde

https://www.youtube.com/watch?v=YjjRClk1se0

this can cause cancer and irritation.. is there any truth to this video?

Possibly, but all that video says is that his phone or whatever meter he's using is finding it in the room he's in. It's not very scientific.

[glosed]

Are HW errors at all related to the pool you are mining from? My hunch is no... but I was sitting in the other room and my miner shot up to 100% fan speed for no reason (when I checked the miner status PCB/Chip temps were like 30/40 so that wasn't the cause). But my error rate was nearly 3% which is way above what it has been to date (read: the 22 hours or so it's been running).

HW = Hardware Faults. Not at all related to where the machine is pointed.

Are you overclocked?

Default 600 with temps on the lower end. I Save+Applied my miner config to reset the Miner Status page and everything seems to be back to normal. Now I'm just wondering what the hiccup was... oh well.

[greyday]

Default 600 with temps on the lower end. I Save+Applied my miner config to reset the Miner Status page and everything seems to be back to normal. Now I'm just wondering what the hiccup was... oh well.

It's just Chinatown, Jake.

[ylif]

• Mining pool selection - I set up SiaMining at first and was getting around 80SC/hour. In the spirit of decentralizing and supporting good pools, I swapped over to Luxor, and my SC/hour was much smaller. After over 90 minutes it was estimating about 30SC, so I reverted back to SiaMining. However, I just refreshed the Luxor page and it is showing 362SC for just under 2 hour's work. Why does the number fluctuate so much, and why was that figure so delayed? Is that typical with Luxor, or is it due to the nature of PPLNS? Also, it seemed to take MUCH longer for my hashrate to reach ~800 on Luxor (30 mins or so) vs. SiaMining where I had full hash within a couple minutes. Or should I go with F2Pool instead? Haven't tried that one out yet.
• HW error rate - People have been mentioning HW errors as "a lot of errors" or "an acceptable rate", but I haven't seen hard stats on what those rates actually are.
   Is 0.01% too high? 0.1%? 1%? I'm currently at 0.006%...

I am jealous of your temps.

1) Mining pool--Luxor is having some sort of problem where their luck is currently at about 40%, so they switched over to PPLNS from PPS. Which means you get paid out in an average over time, but with their luck that low you're still going to make less right now (in theory). Siamining is PPS and currently paying out well. I'm with you though, I prefer Luxor and will switch back when whatever problem they're having gets sorted.

2) .01% is fine. .1% is ok. 1% is too much (opinions vary, but that's my standard).

3) Your cold air is your advantage, these things run hot. But I have mine overclocked (650 currently but usually 631). Overclocking voids your warranty, fair warning there. Your at-the-wall is your call; you can calculate profit with usage deducted at whattomine.com, it's usually pretty close. On a 20a 110 circuit, I'd suggest limiting wattage to 1700W for safety, you likely won't get anywhere near that though. HW isn't necessarily higher at higher frequencies, it depends. 631 drops more than 650 for me. It's more machine dependent.

Are HW errors at all related to the pool you are mining from? My hunch is no... but I was sitting in the other room and my miner shot up to 100% fan speed for no reason (when I checked the miner status PCB/Chip temps were like 30/40 so that wasn't the cause). But my error rate was nearly 3% which is way above what it has been to date (read: the 22 hours or so it's been running).

Also, it looks like there's something up with SiaMining, wondering if that could be the cause of the fan speed? Pool hashrate dropped like a rock. I figured it could be due to some miners going to AntPool, since https://siastats.info/mining_pools.html shows about 300TH/s that I don't think was there before. But then I noticed my own hashrate had gone down. To rule out that being the result of me getting a spike in HW errors or something else happening like it resetting and spiking the fanspeed, all/many of the addresses here: https://siamining.com/addresses are also getting cut in half...as if I needed a reason to switch back to Luxor, it does seem like their luck issues are getting better.
 
https://i.imgur.com/jJFgYJN.png

I found out that once you have overclocked, the hardware become less stable. Even you reture to the default 600, the hw error will suddenly  jump in certain amount of time.(chips temp are <70) So I guess overclock is one way ticket, and there won’t be anyway you can return to normal!

[glosed]

• Mining pool selection - I set up SiaMining at first and was getting around 80SC/hour. In the spirit of decentralizing and supporting good pools, I swapped over to Luxor, and my SC/hour was much smaller. After over 90 minutes it was estimating about 30SC, so I reverted back to SiaMining. However, I just refreshed the Luxor page and it is showing 362SC for just under 2 hour's work. Why does the number fluctuate so much, and why was that figure so delayed? Is that typical with Luxor, or is it due to the nature of PPLNS? Also, it seemed to take MUCH longer for my hashrate to reach ~800 on Luxor (30 mins or so) vs. SiaMining where I had full hash within a couple minutes. Or should I go with F2Pool instead? Haven't tried that one out yet.
• HW error rate - People have been mentioning HW errors as "a lot of errors" or "an acceptable rate", but I haven't seen hard stats on what those rates actually are.
   Is 0.01% too high? 0.1%? 1%? I'm currently at 0.006%...

I am jealous of your temps.

1) Mining pool--Luxor is having some sort of problem where their luck is currently at about 40%, so they switched over to PPLNS from PPS. Which means you get paid out in an average over time, but with their luck that low you're still going to make less right now (in theory). Siamining is PPS and currently paying out well. I'm with you though, I prefer Luxor and will switch back when whatever problem they're having gets sorted.

2) .01% is fine. .1% is ok. 1% is too much (opinions vary, but that's my standard).

3) Your cold air is your advantage, these things run hot. But I have mine overclocked (650 currently but usually 631). Overclocking voids your warranty, fair warning there. Your at-the-wall is your call; you can calculate profit with usage deducted at whattomine.com, it's usually pretty close. On a 20a 110 circuit, I'd suggest limiting wattage to 1700W for safety, you likely won't get anywhere near that though. HW isn't necessarily higher at higher frequencies, it depends. 631 drops more than 650 for me. It's more machine dependent.

Are HW errors at all related to the pool you are mining from? My hunch is no... but I was sitting in the other room and my miner shot up to 100% fan speed for no reason (when I checked the miner status PCB/Chip temps were like 30/40 so that wasn't the cause). But my error rate was nearly 3% which is way above what it has been to date (read: the 22 hours or so it's been running).

Also, it looks like there's something up with SiaMining, wondering if that could be the cause of the fan speed? Pool hashrate dropped like a rock. I figured it could be due to some miners going to AntPool, since https://siastats.info/mining_pools.html shows about 300TH/s that I don't think was there before. But then I noticed my own hashrate had gone down. To rule out that being the result of me getting a spike in HW errors or something else happening like it resetting and spiking the fanspeed, all/many of the addresses here: https://siamining.com/addresses are also getting cut in half...as if I needed a reason to switch back to Luxor, it does seem like their luck issues are getting better.
 
https://i.imgur.com/jJFgYJN.png

I found out that once you have overclocked, the hardware become less stable. Even you reture to the default 600, the hw error will suddenly  jump in certain amount of time.(chips temp are <70) So I guess overclock is one way ticket, and there won’t be anyway you can return to normal!

I was not overclocked. My HW rate typically fluctuates between 0.01% and 0.1%, but 3% is quite large. I feel like something may have happened to a lot of Antminers... my File System and Kernel versions are both from December, but it seems a bit odd that a big spike in my fanspeed and HW rate would occur at the precise moment there is a very noticeable dip in hash rates across SiaMining and F2Pool. Maybe it is coincidence... as greyday advised, I'm going to forget it, because it's Chinatown 

[ylif]

• Mining pool selection - I set up SiaMining at first and was getting around 80SC/hour. In the spirit of decentralizing and supporting good pools, I swapped over to Luxor, and my SC/hour was much smaller. After over 90 minutes it was estimating about 30SC, so I reverted back to SiaMining. However, I just refreshed the Luxor page and it is showing 362SC for just under 2 hour's work. Why does the number fluctuate so much, and why was that figure so delayed? Is that typical with Luxor, or is it due to the nature of PPLNS? Also, it seemed to take MUCH longer for my hashrate to reach ~800 on Luxor (30 mins or so) vs. SiaMining where I had full hash within a couple minutes. Or should I go with F2Pool instead? Haven't tried that one out yet.
• HW error rate - People have been mentioning HW errors as "a lot of errors" or "an acceptable rate", but I haven't seen hard stats on what those rates actually are.
   Is 0.01% too high? 0.1%? 1%? I'm currently at 0.006%...

I am jealous of your temps.

1) Mining pool--Luxor is having some sort of problem where their luck is currently at about 40%, so they switched over to PPLNS from PPS. Which means you get paid out in an average over time, but with their luck that low you're still going to make less right now (in theory). Siamining is PPS and currently paying out well. I'm with you though, I prefer Luxor and will switch back when whatever problem they're having gets sorted.

2) .01% is fine. .1% is ok. 1% is too much (opinions vary, but that's my standard).

3) Your cold air is your advantage, these things run hot. But I have mine overclocked (650 currently but usually 631). Overclocking voids your warranty, fair warning there. Your at-the-wall is your call; you can calculate profit with usage deducted at whattomine.com, it's usually pretty close. On a 20a 110 circuit, I'd suggest limiting wattage to 1700W for safety, you likely won't get anywhere near that though. HW isn't necessarily higher at higher frequencies, it depends. 631 drops more than 650 for me. It's more machine dependent.

Are HW errors at all related to the pool you are mining from? My hunch is no... but I was sitting in the other room and my miner shot up to 100% fan speed for no reason (when I checked the miner status PCB/Chip temps were like 30/40 so that wasn't the cause). But my error rate was nearly 3% which is way above what it has been to date (read: the 22 hours or so it's been running).

Also, it looks like there's something up with SiaMining, wondering if that could be the cause of the fan speed? Pool hashrate dropped like a rock. I figured it could be due to some miners going to AntPool, since https://siastats.info/mining_pools.html shows about 300TH/s that I don't think was there before. But then I noticed my own hashrate had gone down. To rule out that being the result of me getting a spike in HW errors or something else happening like it resetting and spiking the fanspeed, all/many of the addresses here: https://siamining.com/addresses are also getting cut in half...as if I needed a reason to switch back to Luxor, it does seem like their luck issues are getting better.
 
https://i.imgur.com/jJFgYJN.png

I found out that once you have overclocked, the hardware become less stable. Even you reture to the default 600, the hw error will suddenly  jump in certain amount of time.(chips temp are <70) So I guess overclock is one way ticket, and there won’t be anyway you can return to normal!

I was not overclocked. My HW rate typically fluctuates between 0.01% and 0.1%, but 3% is quite large. I feel like something may have happened to a lot of Antminers... my File System and Kernel versions are both from December, but it seems a bit odd that a big spike in my fanspeed and HW rate would occur at the precise moment there is a very noticeable dip in hash rates across SiaMining and F2Pool. Maybe it is coincidence... as greyday advised, I'm going to forget it, because it's Chinatown 

Actually quite similar situation here other than overclocking. 0.1% is already too high according to the customer service. This  may even because this A3 1st batch has not even finish its development with fully  testing. It’s all about win the time to conquer the sc1’s market. Otherthan rebooting, i don’t think there is anything we can do at this time as every minute now  is so valuable in the future. And do not hesitate to control the hw, high hw will damage the hashboard no matter what your temp is。 i have already did some research on this hw error thing.

[glosed]

• Mining pool selection - I set up SiaMining at first and was getting around 80SC/hour. In the spirit of decentralizing and supporting good pools, I swapped over to Luxor, and my SC/hour was much smaller. After over 90 minutes it was estimating about 30SC, so I reverted back to SiaMining. However, I just refreshed the Luxor page and it is showing 362SC for just under 2 hour's work. Why does the number fluctuate so much, and why was that figure so delayed? Is that typical with Luxor, or is it due to the nature of PPLNS? Also, it seemed to take MUCH longer for my hashrate to reach ~800 on Luxor (30 mins or so) vs. SiaMining where I had full hash within a couple minutes. Or should I go with F2Pool instead? Haven't tried that one out yet.
• HW error rate - People have been mentioning HW errors as "a lot of errors" or "an acceptable rate", but I haven't seen hard stats on what those rates actually are.
   Is 0.01% too high? 0.1%? 1%? I'm currently at 0.006%...

I am jealous of your temps.

1) Mining pool--Luxor is having some sort of problem where their luck is currently at about 40%, so they switched over to PPLNS from PPS. Which means you get paid out in an average over time, but with their luck that low you're still going to make less right now (in theory). Siamining is PPS and currently paying out well. I'm with you though, I prefer Luxor and will switch back when whatever problem they're having gets sorted.

2) .01% is fine. .1% is ok. 1% is too much (opinions vary, but that's my standard).

3) Your cold air is your advantage, these things run hot. But I have mine overclocked (650 currently but usually 631). Overclocking voids your warranty, fair warning there. Your at-the-wall is your call; you can calculate profit with usage deducted at whattomine.com, it's usually pretty close. On a 20a 110 circuit, I'd suggest limiting wattage to 1700W for safety, you likely won't get anywhere near that though. HW isn't necessarily higher at higher frequencies, it depends. 631 drops more than 650 for me. It's more machine dependent.

Are HW errors at all related to the pool you are mining from? My hunch is no... but I was sitting in the other room and my miner shot up to 100% fan speed for no reason (when I checked the miner status PCB/Chip temps were like 30/40 so that wasn't the cause). But my error rate was nearly 3% which is way above what it has been to date (read: the 22 hours or so it's been running).

Also, it looks like there's something up with SiaMining, wondering if that could be the cause of the fan speed? Pool hashrate dropped like a rock. I figured it could be due to some miners going to AntPool, since https://siastats.info/mining_pools.html shows about 300TH/s that I don't think was there before. But then I noticed my own hashrate had gone down. To rule out that being the result of me getting a spike in HW errors or something else happening like it resetting and spiking the fanspeed, all/many of the addresses here: https://siamining.com/addresses are also getting cut in half...as if I needed a reason to switch back to Luxor, it does seem like their luck issues are getting better.
 
https://i.imgur.com/jJFgYJN.png

I found out that once you have overclocked, the hardware become less stable. Even you reture to the default 600, the hw error will suddenly  jump in certain amount of time.(chips temp are <70) So I guess overclock is one way ticket, and there won’t be anyway you can return to normal!

I was not overclocked. My HW rate typically fluctuates between 0.01% and 0.1%, but 3% is quite large. I feel like something may have happened to a lot of Antminers... my File System and Kernel versions are both from December, but it seems a bit odd that a big spike in my fanspeed and HW rate would occur at the precise moment there is a very noticeable dip in hash rates across SiaMining and F2Pool. Maybe it is coincidence... as greyday advised, I'm going to forget it, because it's Chinatown 

Actually quite similar situation here other than overclocking. 0.1% is already too high according to the customer service. This  may even because this A3 1st batch has not even finish its development with fully  testing. It’s all about win the time to conquer the sc1’s market. Otherthan rebooting, i don’t think there is anything we can do at this time as every minute now  is so valuable in the future. And do not hesitate to control the hw, high hw will damage the hashboard no matter what your temp is。 i have already did some research on this hw error thing.

Thanks for that intel. Happened again to me this morning, right around 8am I was woken up by the fan speeds ramping to 100%. I just Save&Applied miner config to reset the machine, and when I did I was getting HW out the wazoo - over 100% on board 2, which to date has been the board that typically reports the least HW. Lots of x among the o's, and board 1 wasn't even hashing at 100GH/s. I tried another time and the same thing happened, which was strange because usually it's easy to fix. So I turned my unit off to go back to sleep for a bit and deal with it later. Woke up a couple hours later, fired it up, and she;'s cruising along just fine.

I'm glad that the issue is temporarily fixable, what I'm more concerned about is, what if this happens when I'm not at my apartment to make sure it's not running on big HW for hours on end?

[RuslanBozhy]

Can someone explain me why difficulty is falling down after it's soared up? http://skrinshoter.ru/i/040218/MwgR4Rsz.png

[glosed]

Can someone explain me why difficulty is falling down after it's soared up? http://skrinshoter.ru/i/040218/MwgR4Rsz.png

Here's my best explanation (which might not be accurate so someone please let me know).

There are three main differentiators between Bicoin's difficulty algorithm and Sia's that are at play here (there are more differences, but these seem to be the most important):

1. Difficulty adjusts every block, rather than every 2016 blocks (~2 weeks).
2. Difficulty adjustments are limited to 0.4% per block, while Bitcoin's adjustment is unbounded.
3. Rather than adjusting difficulty to meet the targeted blocktime on a go-forward basis, Sia looks at time since genesis and adjusts blocktimes to get back on track.

https://www.reddit.com/r/siacoin/comments/7rz606/update_on_the_block_139000_hardfork/

The above post explains it much better than I could, but in the context of the A3's hitting the network:

1. A3's start going online and the hashrate increases much faster than the difficulty can adjust for.
2. Difficulty increases, and continues to increase, even after hashrate flattens out to slow down block generation until we are in line with where we should be relative to genesis.
3. Once we are closer to aligning actual block solves to when they should be happening, difficulty starts to adjust back down so we don't overshoot and fall behind schedule.

[Searing]

• Mining pool selection - I set up SiaMining at first and was getting around 80SC/hour. In the spirit of decentralizing and supporting good pools, I swapped over to Luxor, and my SC/hour was much smaller. After over 90 minutes it was estimating about 30SC, so I reverted back to SiaMining. However, I just refreshed the Luxor page and it is showing 362SC for just under 2 hour's work. Why does the number fluctuate so much, and why was that figure so delayed? Is that typical with Luxor, or is it due to the nature of PPLNS? Also, it seemed to take MUCH longer for my hashrate to reach ~800 on Luxor (30 mins or so) vs. SiaMining where I had full hash within a couple minutes. Or should I go with F2Pool instead? Haven't tried that one out yet.
• HW error rate - People have been mentioning HW errors as "a lot of errors" or "an acceptable rate", but I haven't seen hard stats on what those rates actually are.
   Is 0.01% too high? 0.1%? 1%? I'm currently at 0.006%...

I am jealous of your temps.

1) Mining pool--Luxor is having some sort of problem where their luck is currently at about 40%, so they switched over to PPLNS from PPS. Which means you get paid out in an average over time, but with their luck that low you're still going to make less right now (in theory). Siamining is PPS and currently paying out well. I'm with you though, I prefer Luxor and will switch back when whatever problem they're having gets sorted.

2) .01% is fine. .1% is ok. 1% is too much (opinions vary, but that's my standard).

3) Your cold air is your advantage, these things run hot. But I have mine overclocked (650 currently but usually 631). Overclocking voids your warranty, fair warning there. Your at-the-wall is your call; you can calculate profit with usage deducted at whattomine.com, it's usually pretty close. On a 20a 110 circuit, I'd suggest limiting wattage to 1700W for safety, you likely won't get anywhere near that though. HW isn't necessarily higher at higher frequencies, it depends. 631 drops more than 650 for me. It's more machine dependent.

Are HW errors at all related to the pool you are mining from? My hunch is no... but I was sitting in the other room and my miner shot up to 100% fan speed for no reason (when I checked the miner status PCB/Chip temps were like 30/40 so that wasn't the cause). But my error rate was nearly 3% which is way above what it has been to date (read: the 22 hours or so it's been running).

Also, it looks like there's something up with SiaMining, wondering if that could be the cause of the fan speed? Pool hashrate dropped like a rock. I figured it could be due to some miners going to AntPool, since https://siastats.info/mining_pools.html shows about 300TH/s that I don't think was there before. But then I noticed my own hashrate had gone down. To rule out that being the result of me getting a spike in HW errors or something else happening like it resetting and spiking the fanspeed, all/many of the addresses here: https://siamining.com/addresses are also getting cut in half...as if I needed a reason to switch back to Luxor, it does seem like their luck issues are getting better.
 

I found out that once you have overclocked, the hardware become less stable. Even you reture to the default 600, the hw error will suddenly  jump in certain amount of time.(chips temp are <70) So I guess overclock is one way ticket, and there won’t be anyway you can return to normal!

That seems to happen anyways.  I have not overclocked my unit and there are still weird spikes in error rates, fan speed, and so forth.  Yes, I have paid close attention to my chip temperatures, keeping them in the 60's/lower 70's.

Sometimes the ASIC maker of any FLAVOR, messes up and is too optimistic on speed....for the NEW ASIC product.....I'd underclock till maybe a firmware update or somethihng to be safe.

(don't have an A3 but I scare easy, what i'd do )

[Tekno2020]

Can someone explain me why difficulty is falling down after it's soared up?

Let me put it simple for you.  SIAs difficulty algorithm is adjusting difficulty to target a 10 minute block discovery cycle.  The addition of ASICs to the network pushed block discovery time to 30minutes.

[DragosD1]

Can someone explain me why difficulty is falling down after it's soared up? http://skrinshoter.ru/i/040218/MwgR4Rsz.png

Also, any explanation why hashrate is falling?Was at peak around 11Ph, then falled back to 7.5 Ph. As we speak is about 7.9 Ph...

[RuslanBozhy]

Can someone explain me why difficulty is falling down after it's soared up?

Let me put it simple for you.  SIAs difficulty algorithm is adjusting difficulty to target a 10 minute block discovery cycle.  The addition of ASICs to the network pushed block discovery time to 30minutes.

so whattomine dot com is lying about mining profit on SIA?

[Tekno2020]

Can someone explain me why difficulty is falling down after it's soared up?

Let me put it simple for you.  SIAs difficulty algorithm is adjusting difficulty to target a 10 minute block discovery cycle.  The addition of ASICs to the network pushed block discovery time to 30minutes.

so whattomine dot com is lying about mining profit on SIA?

Seems to be pretty accurate to me at the moment even with price of SC way down.

http://whattomine.com/coins/161-sc-blake-2b?utf8=%E2%9C%93&hr=815000&p=1275&fee=3&cost=0.1&hcost=2500&commit=Calculate

Hope to get another 43 days at this level before batch 2 lands,,,, then we will see.

[dlezama]

For those interested in A3s (all of them) dropping hashrate at the same time, in all pools, there is a channel discussing that in the unofficial sia discord (https://discord.gg/EVn3HPK). So far the theory is that is caused by some blocks, probably a bug in the A3s. The issue lasts until a block is found and they go back to normal. Boards restart when mining these problematic blocks and fans stay at 100% after that... Very weird.

[Christian Pezza]

What you think will happen after all the A3 orders of March Batch will be delivered?

How much will effect the difficulty?

How much the price of SC will change?
 

[johnny_blaze]

What you think will happen after all the A3 orders of March Batch will be delivered?

How much will effect the difficulty?

How much the price of SC will change?
 

we can only guess...

[glosed]

For those interested in A3s (all of them) dropping hashrate at the same time, in all pools, there is a channel discussing that in the unofficial sia discord (https://discord.gg/EVn3HPK). So far the theory is that is caused by some blocks, probably a bug in the A3s. The issue lasts until a block is found and they go back to normal. Boards restart when mining these problematic blocks and fans stay at 100% after that... Very weird.

Perfect! Thank you, will check it out. Glad I'm not going crazy.

[glosed]

my room temperature is about 25 C, why my chips temp always above 82 C? I already set the fan to 95%.
I rebooted it and it's still the same.
Is there a way to reduce the chips temp?

Improve your cooling on it or underclock it. Not sure if underclocking will void your warranty, or if only overclocking it will... Proceed with caution. When I underclocked mine to 575 I saw a drop in temps of over 4 degrees: http://bitcointalk.org/index.php?topic=2881452

[ylif]

• Mining pool selection - I set up SiaMining at first and was getting around 80SC/hour. In the spirit of decentralizing and supporting good pools, I swapped over to Luxor, and my SC/hour was much smaller. After over 90 minutes it was estimating about 30SC, so I reverted back to SiaMining. However, I just refreshed the Luxor page and it is showing 362SC for just under 2 hour's work. Why does the number fluctuate so much, and why was that figure so delayed? Is that typical with Luxor, or is it due to the nature of PPLNS? Also, it seemed to take MUCH longer for my hashrate to reach ~800 on Luxor (30 mins or so) vs. SiaMining where I had full hash within a couple minutes. Or should I go with F2Pool instead? Haven't tried that one out yet.
• HW error rate - People have been mentioning HW errors as "a lot of errors" or "an acceptable rate", but I haven't seen hard stats on what those rates actually are.
   Is 0.01% too high? 0.1%? 1%? I'm currently at 0.006%...

I am jealous of your temps.

1) Mining pool--Luxor is having some sort of problem where their luck is currently at about 40%, so they switched over to PPLNS from PPS. Which means you get paid out in an average over time, but with their luck that low you're still going to make less right now (in theory). Siamining is PPS and currently paying out well. I'm with you though, I prefer Luxor and will switch back when whatever problem they're having gets sorted.

2) .01% is fine. .1% is ok. 1% is too much (opinions vary, but that's my standard).

3) Your cold air is your advantage, these things run hot. But I have mine overclocked (650 currently but usually 631). Overclocking voids your warranty, fair warning there. Your at-the-wall is your call; you can calculate profit with usage deducted at whattomine.com, it's usually pretty close. On a 20a 110 circuit, I'd suggest limiting wattage to 1700W for safety, you likely won't get anywhere near that though. HW isn't necessarily higher at higher frequencies, it depends. 631 drops more than 650 for me. It's more machine dependent.

Are HW errors at all related to the pool you are mining from? My hunch is no... but I was sitting in the other room and my miner shot up to 100% fan speed for no reason (when I checked the miner status PCB/Chip temps were like 30/40 so that wasn't the cause). But my error rate was nearly 3% which is way above what it has been to date (read: the 22 hours or so it's been running).

Also, it looks like there's something up with SiaMining, wondering if that could be the cause of the fan speed? Pool hashrate dropped like a rock. I figured it could be due to some miners going to AntPool, since https://siastats.info/mining_pools.html shows about 300TH/s that I don't think was there before. But then I noticed my own hashrate had gone down. To rule out that being the result of me getting a spike in HW errors or something else happening like it resetting and spiking the fanspeed, all/many of the addresses here: https://siamining.com/addresses are also getting cut in half...as if I needed a reason to switch back to Luxor, it does seem like their luck issues are getting better.
 
https://i.imgur.com/jJFgYJN.png

I found out that once you have overclocked, the hardware become less stable. Even you reture to the default 600, the hw error will suddenly  jump in certain amount of time.(chips temp are <70) So I guess overclock is one way ticket, and there won’t be anyway you can return to normal!

I was not overclocked. My HW rate typically fluctuates between 0.01% and 0.1%, but 3% is quite large. I feel like something may have happened to a lot of Antminers... my File System and Kernel versions are both from December, but it seems a bit odd that a big spike in my fanspeed and HW rate would occur at the precise moment there is a very noticeable dip in hash rates across SiaMining and F2Pool. Maybe it is coincidence... as greyday advised, I'm going to forget it, because it's Chinatown 

Actually quite similar situation here other than overclocking. 0.1% is already too high according to the customer service. This  may even because this A3 1st batch has not even finish its development with fully  testing. It’s all about win the time to conquer the sc1’s market. Otherthan rebooting, i don’t think there is anything we can do at this time as every minute now  is so valuable in the future. And do not hesitate to control the hw, high hw will damage the hashboard no matter what your temp is。 i have already did some research on this hw error thing.

Thanks for that intel. Happened again to me this morning, right around 8am I was woken up by the fan speeds ramping to 100%. I just Save&Applied miner config to reset the machine, and when I did I was getting HW out the wazoo - over 100% on board 2, which to date has been the board that typically reports the least HW. Lots of x among the o's, and board 1 wasn't even hashing at 100GH/s. I tried another time and the same thing happened, which was strange because usually it's easy to fix. So I turned my unit off to go back to sleep for a bit and deal with it later. Woke up a couple hours later, fired it up, and she;'s cruising along just fine.

I'm glad that the issue is temporarily fixable, what I'm more concerned about is, what if this happens when I'm not at my apartment to make sure it's not running on big HW for hours on end?

I have come acrossed  that start up issue too. What I realized is that the temp was too low （30-49）when i reboot it. I did some talk with people from bitmain，and they suggested me to raise the temp. So I turned off the 100%fan speed and it then started up just fine with higher temperature. （still less than 74）although the hw issue  still happens, but it only happen like once in a few hours. The overall hashrate was not really affected by those sudden drops，as it returned normal quick. Now I just leave my a3 in this way，as anything else are fine. I will keep monitoring it until roi. I may try overclocking again in some days but I‘m just not sure.