Question about upgrading my Antminer S9s

[carlbutler86]

Hey y'all - been learning a lot on this forum, so thanks for the help!

I currently have a bit over 50 Antminer S9 miners running on 15 Tripp lite 30 amp PDUs. I have 4 S9s running on each PDU and a couple of L3s running to make up the difference in slots. Can't quite get 12 running on a 100 amp box. But I can squeeze in 11.

Each PDU is on a 100 amp box on its' own 20 amp circuit. Total of 500 amps for 15 PDUs. I have an average of about 750TH.

I am planning on upgrading in the next few months with the halving coming up in May and I'm trying to figure out what mining rig would be the best to upgrade to maximize the power and slots that I currently have.

I was thinking of upgrading to the Innosilicon T2T 30th, but I think they are 10 amps each, and that means I can only put 1 or 2 on each PDU correct?

If I can put 2 on each PDU then that means I'd have a total of 900TH, which is a bit more than my 750, so that would be worth it. But if a PDU can only handle 1, then it would only be 450th and that would be a downgrade.

So I guess I'm trying to figure out what mining rig will give me more TH but with less rigs. I plan on selling all 50+ Antminer S9s to try and pay for some of the upgrade.

Any advice or good math skills would really help a potato farmer out haha 

Carl from Idaho!

[1714861339]

1714861339

[1714861339]

1714861339

[1714861339]

1714861339

[1714861339]

1714861339

[1714861339]

1714861339

[Steamtyme]

Man you are in a sweet kind of tough spot lol. Might I suggest moving this topic over to Mining Speculation - Can be done from the lower left of the screen at the bottom of the topic by you -. There will be a more knowledgeable and engaging group there that can give you some decent insights.

Not sure how much research you've done on what's out there but here's a fairly maintained list of Current List of Competitive Hardware - October 2019 It's missing the Canaan 1066 I see has a review done. I also haven't kept pricing that up to date as it has been changing every other week it seems for some.

On to your situation, I'm guessing you have a cheap stable price for power if you are already running 50 S9'S plus extras profitably. Sounds like you want to be in this for the long haul. So my best suggestion as you are in the US, is go with something Bitmain. I read they are starting to use a plant outside China to build gear that ships to the US, this avoids the massive tariff currently being applied. There really just isn't any getting around the savings of 25% tax upfront. More to the point they have also been dropping the prices significantly from time to time. Though I think some people had issues with coupons and reimbursements lately - not really a factor for you atm.

The decision lies in how low your power costs are and what you are willing to spend on the upgrade. That leaves you with 3 choices:

• Does it make sense to pay more for the efficient gear now before the halving?
• Do you wait until after the halving hoping for lower prices on efficient gear being dumped? This comes with a loss of potential earnings on your S9's the same way.
• Do you buy less expensive less efficient gear now and that is your final upgrade.

Not knowing the exact factors a hypothetical buy right now for me looks something like this - You go with the Antminer S17  50 TH @ 2250W . They ship out in 7 days at a cost of 1253 plus shipping duty. I'm assuming you have stable 220 Volts plus but you could put 2 of these per PDU even if they pull hard at say 2400W (just a what if) ~11 amps. For you this means you can with your current footprint double your hashrate to 1500 TH. For the low low price of $37 950 if you grab 30. Depending on your S9 and firmware your nearly twice as efficient, so your monthly utility should be about the same only slightly higher.

Pros of the above are you pull 2X the BTC before the halving that you would have. Your S9's and PSU's still have resale value - maybe 3000$ currently not sure. You create space to grow or at least PSU's and Panels for redundancy or resale.

[carlbutler86]

Thanks for the tip, I just moved it over!

And thanks for the info, to answer your question, yes I have super cheap electric, it might as well be free. So my S9s are profitable for sure. Yup, you're right also, my main factor in making a decision is on budget. I don't have a ton to spend, definitely not that 37k you mentioned. That's why I was thinking of the 30th rigs, since they are around 500 each. And I figured if I waited until March to upgrade, I can maybe get them even cheaper, or even wait til' the halving takes place and buy them even cheaper, I was thinking maybe even 200-300 a piece.

Also how do you figure out how many rigs I can put per PDU? The S9s I figured out by trial and error, and also some math from my electrician. We determined I can put 4 S9s per PDU in most cases.

Yes, I plan to go for the long haul, at least a few years mining with the next upgrade I make. I can probably only afford about 15k (including the profits from the sale of my S9s). So I am trying to figure out what rig I'll be able to buy to fill my slots for that much, as well as what makes sense for the power I have. And of course how many of that rig I should buy.

A friend of mine told me that I could run 3 x Innosilicon T2T 30th on each 30amp PDU. But I'm wondering how that's possible considering they are 2100+ watts (total of 6300) each whereas I run 4 x Antminer S9 on each PDU and they are 1350 watts (total of 5400).  Can I really run 3 x Innosilicon T2T 30th on each 30amp PDU? If so that means I could essentially run 45 of them. For some reason I don't think I can run that many.

I don't really understand the watts / volts / amps etc. - so to me when I see online that the T2T runs around 9-10 amps - that makes me think that I can only run 2 on the PDU since it's connected to a 20amp circuit.

I guess I'm just looking for some advice on what rigs I can run on my PDU's safely, how many of them, and of course which ones are going to be the most affordable, but still increase my hashrate instead of just continuing to run my S9s.

Thanks again for your insight, it's definitely helpful!

[wndsnb]

What voltage is your electrical service? 240V? 208 3 phase?

Each PDU is on a 100 amp box on its' own 20 amp circuit. Total of 500 amps for 15 PDUs. I have an average of about 750TH.

Are you sure each PDU is on a 20A circuit? Doesn't seem like you could run 4 S9s on a single 20A circuit. With new firmware running 1250W per S9 you'd be pulling close to 21A at 240V. Keep in mind that for continuous loads you shouldn't exceed 80% of the circuit capacity, so a 20A circuit maxes out at 16A continuous usage. Did you mean 30A circuit?

I guess it could work if you are running 208 3 phase to the PDUs.

[carlbutler86]

This is what I mean when I say I'm not great with the numbers  

So the easiest way I can explain it, for example, in one room I have a 200 amp box with 6 20amp fuse switches (breaker switches?) and each switch has a PDU hardwired directly to it. And each of those PDUs is running 4 x S9s on "low power mode". Been running these for months with no issues.

I'm running these on standard residential 200amp or 100amp boxes and yes at 240V.

[wndsnb]

In that case, I'd be concerned ... you are sure those are 20A breakers and not 30A? And you're running just normal "low power" and not "enhanced low power"?

Just because it has worked for a few months doesn't mean it's not a fire hazard...

If they are 20A 240 circuits you have 3840W of safe continuous power available, and your current setup isn't really safe even if you are running enhanced low power at 1000W per S9.

If the breaker panels are located close to your PDUs and everything is open (not wired through walls), it may not be very difficult to rewire for 30A circuits.

In that case, you would have 5760W per PDU available.

T2T-30T run 2200W if you believe the manufactures specs, so you'd only be able to run two. Since you say your power is basically free, this would not help. They are only marginally more efficient than your S9s, so it really isn't going to help any way you look at it. If you were to re-wire everything to fully utilize your 500A service with these, you'd only end up with about 20% more hashrate.

You could run two 50TH T17e on a PDU. They are $1007 + shipping + tariff for Jan delivery. So about $1350 each delivered.  Just buy as many of those as you can afford, and fill the remaining capacity with your existing S9s. So with 15K, you could get 11 T17e which would replace 22 of your S9s, giving you about 240TH over what you have now.

Same would work for 64TH S17e, but they are not currently available.


Also another thing to keep in mind, if the service is residential the power company is not expecting you to utilize 100% of the capacity of your service. It may be wise to take a walk outside to see what rating the transformer that your service is coming from is. Many are marked on the outside with the KVA rating. If you are trying to use 500A of 240V, the KVA rating should be higher than 500A x 240V = 120,000VA, or 120KVA. Don't normally see transformers that big in residential areas.

This is an interesting article: https://community.openenergymonitor.org/t/us-200-amp-service-explained/4219

[carlbutler86]

You know what I just double checked and you are right, the breakers are 30amp. I was wondering why my math was off haha  

Thanks for all the extra info and that link, I'm digging and doing more research now, you guys are great!!

[Steamtyme]

Also how do you figure out how many rigs I can put per PDU? I don't really understand the watts / volts / amps etc. - so to me when I see online that the T2T runs around 9-10 amps - that makes me think that I can only run 2 on the PDU since it's connected to a 20amp circuit.

[...]

Can I really run 3 x Innosilicon T2T 30th on each 30amp PDU? If so that means I could essentially run 45 of them. For some reason I don't think I can run that many.

The shortest explanation is a basic electrical calculation - Watts = Volts X Amps , quick switch can get you amps by going Amps = Watts/Volts. Then you look at the circuit and apply an 80% safety factor, so your 30 Amp circuit can hold 24 Amps. This will also mean that your 30 Amp PDU's can do the same but likely distributes 12 Amps on two different circuits internally - mine have 2 sets of sockets.

I don't know what PDU you have but if they are similar to mine, 12 amps per side you won't be able to run 3 T2T models, as you can't split the load of the 3rd across the 2 PDU circuits. Hope that makes sense, it's similar to when people used to use 2 different 110V  15amp circuits to run s9's on 2 power supplies.

Yes, I plan to go for the long haul, at least a few years mining with the next upgrade I make. I can probably only afford about 15k (including the profits from the sale of my S9s). So I am trying to figure out what rig I'll be able to buy to fill my slots for that much, as well as what makes sense for the power I have. And of course how many of that rig I should buy.  

[...]

I guess I'm just looking for some advice on what rigs I can run on my PDU's safely, how many of them, and of course which ones are going to be the most affordable, but still increase my hashrate instead of just continuing to run my S9s.

With your budget it's a tough call where to go for a nice upgrade. You might want to do a blend of the T2T and maybe something like the T17 which is a better bargain due to tariffs if they come into play. With the way the number look now here is pretty much your best case scenario, and becuase it is only an upgrade on fairly efficient still profitable gear for you waiting for a further drop in prices makes sense. It's just not a guarantee what will be available for what price in the future.

When you factor in the US tariffs the T2T is 30 TH at 566$ so 18.86$ if you long play it and order 1 at a time at least a week apart; more like 707.5$ so per TH it's - 23.58$. It can be done though and might actually walk you into the option I think works best for you; blending. Do this steadily for the very cheap hash, but also order T17's, maybe a couple in stock with coupons you can buy from members here on the forum. T17 numbers are as follows - The T17 is 42TH @ 898 in stock so per TH it's 21.38$, Delivery in January is 20.14. Now with a coupon you can save an additional 10%, but you will have to buy at least the first one from a member so more like 6.5%. There are a few of these for sale for 250$.

If I had the budget and the mindset I would order 8 T17's, buy a 689$ coupon from a member here selling 2, your order would be 7184$-689$=6495$ plus shipping. I would double check that these avoid the tariffs being assembled outside China first, but assuming so that is a good first batch to swing into the farm and make space trying to move S9's. With maybe 3 T2T models ordered and delivered over the next month you will be at 416 TH of upgraded gear with 22 S9's cleared out or up for sale. This is a little over half of your upgrade budget to this point, being 8193$.

Now you can see if Bitmain gave you a coupon for this order. If so you can apply it to a less expensive future batch of T17's maybe, depending on if you keep ordering T2T's or re evaluate when you see what pricing looks like at the end of January. I don't know what that leaves you for your old hashrate but we are taking off less than half your S9's so you should still have around 400TH old gear and 416 TH of new gear. New total hashrate with 8443$(added in coupon cost) invested, 22 S9 sets available for sale = 816 TH an increase of 66TH.

Good luck with the upgrades.

[mikeywith]

T2T 30th  VS 2*S9 14th

30TH       VS    28TH
2200w     VS    2500w

So you are saving 150 watts and getting an extra tera hash per replacement , taking Steamtyme words for T2T 30th will cost you 566$ each, and your best bet would be selling an S9 for 150-200$ ? I highly doubt it but say you get 200$ for each, that means you will have to pay 366$ just to save 150w for every 1250w and get 1th worth of hashrate, as far as business is concerned, this is a terrible move.

I would suggest AntminerT17 or even S17, I would go for S17, simply put, I think that dark days are coming rather soon, only those with super-efficient gears or power rate near zero will actually be able to stay in the market.

Halving is coming, prices are falling, no sign for quick recovery, everything seems depressing for us miners, so "pay more now, so you can survive later" is what I think the best step any miner would make.

I would also add a piece of information regarding Bitmain coupons. Usually for most miners, the coupon you get can't be used to purchase the same exact miner, so you will most likely have to sell that coupon you get and buy another. But this all depends on what gear/s do you actually plan on buying, another side note is that 90% of coupons sellers on forum and outside the forum alike are actually scammers, so you want to be careful and always use escrow.

[philipma1957]

So you want to upgrade  50 x 13.5 = 675 th  and maybe 62kwatts of power

I would suggest getting 1 s17 from bitmain

https://shop.bitmain.com/product/detail?pid=00020191212144713309ClUHueAO06A8

cost is 1253 + 117 =  1370  and a possible trump tax of 330 or so

that is  as high as 1700  sell 4 s9    about 600 to 800 back

best case  is small trump tax  of 50 and high sell price of 800
worst case is high trump tax of 330 and low sell price of 600

so best 1420-800 =  a  620 cost
worst    1700-600 = a 1100 cost

about the same  hash rate but a power drop from 5000 to 2500 watts  2.5 x 24 x 30 = 1800 kwatts a month saved in power  at 5 cents  it is a 90 a month savings.

do this now wait until the 1 unit arrives see how trump tax worked   if  trump tax was good repeat .

if trump tax was bad try with a s17 pro next.

[wndsnb]

I'm guessing OPs best course of action is to not upgrade yet.

If "basically free" means <3 cents, don't upgrade yet.  Right now a bunch of S9s are being shut down by these guys with 5 to 6 cent power, and the price of S9s is dropping. OP should wait until the price for 60 W/TH gear (T17, A1066, etc ...) to crash when the 6 cent guys are no longer profitable with it. At that point, upgrading would make sense.

If OP is barely profitable with the S9s (ie 5 cent power), then it would make sense to upgrade now.

[philipma1957]

Yeah that argument could be okay.

or 1 s17 pro
or 1 s17
or 1 t17

but for sure no full upgrade.

His 675 th is about 90 usd a day  minus power and maintenance  if that is really cheap for him  say 30 usd or less  he clears 60 usd.

the t17  is 898 + 117        =  1015    /60 = 17 day payback
the s17 is 1250 + 125       = 1375    /60 =  23 day payback
the s17 pro is 1402 + 128 = 1530   /60 =   26 day payback

getting just 1 of the above   and waiting until  it arrives and is setup  should be okay to do.

[carlbutler86]

Wow you guys are greatttt!! Such amazing math skills you guys have haha - and insight into the different rigs and upsides / downsides.

Right now based on what you are all saying, it does sound like upgrading is a bad idea, especially since my electric is next to nothing - might as well be free.

I do think it wouldn't be a bad idea though to upgrade at least half of my S9s to T2T's mainly because of space / maintenance. I have 12 rigs in one room and 12 in another, but would like to consolidate all 24 into one room. And I could do that if I could swap 24 S9s for 12 T2Ts, which based on your above math it sounds like I can.

Or I could do what some of you are saying and upgrade to S17 pros. But how many could I then fit on 6 PDU's with 200 amps?

Then I can leave my other 26 S9s running as is and have only 12 T2Ts. But I'm going to wait until at least February or March to upgrade, to see where prices are. I originally bought my S9s around $300 each and now look, they are $100 each. So I figure if I just sit tight for a few months I can probably get T2T 30th around $300 or maybe even less if I'm lucky.

Plus the price of BTC might go down according to some traders I talk to, I mean even as I write this it broke 7k support and is now around the 6800 mark.

Guess I can be a little patient, since there is no hurry and my S9s are churning coin in the meantime!

You guys all rock!!  

[mikeywith]

True, sometimes "doing nothing is the best thing to do", however, this depends on many factors, the most important one would be (What does he plan to buy those gears with?)

Using Fiat:
Case 1: Bitcoin price keeps dropping, gear prices too                       > Good to wait.
Case 2: Bitcoin price goes up, so do gear prices                              > Good to upgrade.
Case 3: Price stays flat, difficulty keeps rising, gear prices drop       > Unknown.

Using Bitcoin:
Case 1: Bitcoin price keeps dropping, gear prices too                > Good to buy.
Case 2: Bitcoin price goes up, so does gear' price                     > Unknown.
Case 3: Price stays flat, diff keeps rising, gear prices drop         > Good to buy.

The above ignores that fact that he actually own some gears to sell, which in Fiat case 2 for example will work to his favor as his S9s will be worth more, but if S9 prices goes 10% , more efficient gears will go much higher.

Also, just to let you guys know, you can now buy S9j 14th with Apw3++ power supply + Warranty from Bitmain for 86$ only , which means older S9 models without warranty can probably be sources for 70$ or less, probably even 50$ for large orders ( not sure about the last one yet).

[philipma1957laptop]

I wanted him to add 1s17 and in 2 weeks-3 weeks add another.  It is a hedge this way.

Doing all has one set of risks.
Doing none has another set of risks.

doing one every two or three weeks middles his risk.

[wndsnb]

You could run two 50TH T17e on a PDU. They are $1007 + shipping + tariff for Jan delivery. So about $1350 each delivered.  Just buy as many of those as you can afford, and fill the remaining capacity with your existing S9s. So with 15K, you could get 11 T17e which would replace 22 of your S9s, giving you about 240TH over what you have now.

For the above situation, you'd spend  $14,850. You might get $100 per S9 with PSU, so $2200 back, net cost $12,650 to get an additional 240TH.

With free power, and BTC price of $6900, the 240TH makes about $31.50 per day, so 402 days to make back your investment if the diff to price ratio stays constant. Of course, the halving is way less than 402 days out, so it is actually worse than that.

If you only bought 1, the ROI calc would work out the same.

If you think that this is the bottom, and you don't think your S9s will be profitable with your nearly free power after the halving, then it would make sense to start upgrading. If not, the #s don't look good to me.

If your power cost is lower than most of the industrial big guys, then you'll always be able to just wait for them to start liquidating their unprofitable gear and buy it for very low cost. You take on a lot of extra risk paying a premium to upgrade early. You can buy 56TH of S9s right now for $400, maybe even less. If the break-even power cost for a T17 or A1066 hits 6 cents, you'll probably be able to get them for $400 instead of $1400 now.

Another option is to find a host to send new gear to and keep your S9s running with your super cheap power. When the new gear is no longer profitable at the host you can move them to your location and retire the S9s.

[philipma1957]

Another option is to find a host to send new gear to and keep your S9s running with your super cheap power. When the new gear is no longer profitable at the host you can move them to your location and retire the S9s.

the bold can work if he gets a few s17 and t17 plus of course a decent host.

[carlbutler86]

Man so many choices! You guys are amazing. On one hand I like the idea of waiting for the big guys to liquidate their rigs cheap at some point. And then on the other hand I like the idea of just upgrading a handful of rigs to better ones, but not many, because of the fact that the premium you guys mention is definitely an issue for me.

I have some thinking to do....