Topic: Innosilicon t3 50 arrived. - page 2. (Read 746 times)

Heh, I have no skin in the game with any manufacturer, I've had good deliveries and slow deliveries, I've had cancelled orders because of over subscription, suppliers who I lost coin on because of exchange rate fluctuations, and vice versa.

I think things for you have changed, now you are stressed about mining because its a bigger concern and you have partners, so 8 days of delay, which in the past would have been nothing, is now a big thing for you.

For me, it feels the same now as it did when I had 100TH. The "business" of mining feels the same, the suppliers feel the same, the experience is the same. Maybe thats why it doesn't stress me out, but then I don't have business partners to appease

For now, the T3-50T appears to be mining at spec. But then so do the M21S's that I got at the same time, and people have had problems with them too. Go figure.

Although I am pissed with Inno that these T50's I've had to take as replacements for the more efficient T39 and T43's that I actually ordered. That's an annoyance but I'll live with it now that gear is getting harder to get.

Bitmain pissed me off too with their tiny window of purchase opportunity and their limit of 2 from a batch for the T17-38, but a deep breath and all is good again.

yep  my inno's run in  a 90f space they have filtered air,  but the room is always 80f to 95f.

At factory+  it has been doing 51th.  

I have no issues for it at factory+

But it is so far off spec on low setting I find it annoying to say the least.

40t  and  2460 watts is simply not close at all to 33t and 1650 watts.

  To put it simply  when it roi's down the road and the ½ ing comes around  the option to drop it to  33t and 1650 watts does not exist. They advertised  that it could do that.
So come may of 2020  it will be an issue.

They also advertised it would ship in August it shipped on the 9th of sept.  

I think  one problem with mining is the now mentally   So who cares if it  won't downclock in May of 2020  that is far off.  Okay  you could be right may not be a problem in may of 2020.  but shipping 9 days late  was now and that was 100usd in coins.  They did not offer a refund of that or even a coupon.

https://bitcointalksearch.org/topic/the-best-in-class-btc-miner-innosilicon-t357t-t3-50t-shipping-in-september-5185022  I asked for a coupon  for make up for the 2 advertising wrongs  and offered to buy more gear.

which means I don't hate the gear. (I don't like their advertising)  2 fails  one fail for now and one for later.

Look mining is a business  we all try to do what is best for our ownself  and maybe for others.

I would buy one from them if they gave me a coupon to make up for late shipping and for bad downclocking.  

The gear to me is worth  about 1600 not the 2100 they are now asking.  It they had shipping on time  I would ask to pay 1700 or 1800 for one in 7 days.  not the 2100 they are asking.

My T3-50's are up and after a couple of hours are running on default settings @ 50T with +/- fluctuations of only 1%

They are in a datacenter though so probably have better cooling available to them than where Phil has his.

I think if your really read everything in that post you wouldn't be making such a statement, I clearly mentioned


Which is the proper way of doing things, I also don't understand how is it logical to compare a gear which you can get in 4-7 days to a gear that will only ship in late November and will probably only be received in December.

if you can get a T17 now for 1145$ then there is no math needed, go for it, but if you have to wait till December, a gear like the one in review will be at your door step in just a couple of days, and you will have at least 65 days lead , that is almost 750$ of profit ( before the power bill) , also all bitmain gears are sold out at the moment, so the earliest batch will probably be next year January.  

The math is really simple here, and it all depends on how much do you have to pay for the gear, I personally don't like per-orders of any kind, I base my business plans on the current situation and don't like to have my BTC sitting in a manufacturer wallet for 3-5 months to get the gear, I either pay more now or just hold BTC, but it's different for everyone.

The T17 38T costs in the last Bitmain nov. batch $1145. So this is $30 per TH and not $45.

Late product earns less. Philip's comment on returns is incorrect, technically we do not allow returns after an order is submitted to manufacturer, though we treat any issues case by case.

Hopefully Philip can find time to double the FW version is latest and we can hope for more efficiency gain.

These numbers are not bad to be honest, the T3 50th goes for 2000$ in china from resellers, the T17-40th on the other hand goes for about 1800$ , the latter does about 58 watts a TH.

cost per th

T3  = 40$
T17= 45$

40th*5$ = T17 is 200$ more expensive

assuming power cost is 6 cents per kw:

40TH on the T3 cost  =   3.91 $ / day
40TH on the T17 cost=   3.34 $ /day

a difference of  0.57$ a day , so to ROI that 200$ price difference in gears price we do the following

200/0.57 =  350 days almost a whole year

let's see how it stands against S17-53 th which goes for 3500$ in china, it does 45w per th

cost per th

T3= 40$
S17 =66$

difference in 50th = 1300$
assuming power cost is 6 cents per kw:

50TH on the T3 cost  =  4.89 $ / day
50TH on the S17 cost=   3.23 $ /day

a difference of  1.66$ a day , so to ROI that 1300$ price difference in gears price we do the following

1300/1.66 = 783 days or 26 months !!

Honestly, with the numbers ,it makes every sense to get the T3 50th for 2000$ than getting any other gear from bitmain.

These numbers are based on the average price of "buy now" from Chinese resllers, buying the gears directly from the manufacturer will be a lot of different, but the waiting time will be different too.

It was not set to factory when I started it it was set to performance.

I did not run it long enough to give a fair number. But it was 52th

I am pretty happy with efficiency I will run it till I go to sleep. It is loud enough that I won’t sleep well.

I will test factory in the morning and performance in the afternoon.

I was annoyed it came late but if I can get 45 or 46 th on factory - -

48:on factory and 52 on performance .

Okay  just fired it up and set to Factory

---

admin
admin

signs you in

the gui looks like  my t3 39 and like my t2 24

I use an hourly kwatt meter  more accurate then the instant versions but  you need to wait for 2 hours to get a decent measurement

I estimate  it will be about 5.4 kwatts maybe more for the 2 hours.

I will photo the  meter and post once  it starts  to mine and finishes tuning.
These take about an hour to fireup fully.

reads                   1165.76
1hr should read     1168.46  actual =?  1169  which is

1169.00
1165.76
      
3.24 or 3240 watts  doing 48 th  this is 67.5 watts a th on factory  not that good.

the gear is cycling which is a no no  I will see if it stops
I suspect that factory is not a good setting for it.



the meter is at 1169 kwatts at  11 am  we are  doing 48th on factory

I did have a lot of recycling for about 30 minutes which has stopped.

so I will test again   11am 1169 kwatts   48 th New York time on factory

I will let it run for a few hours as I want to be accurate.

---

It is doing very close to 49th at factory setting
In 18 minutes it will be 2pm it  be 3 hours since 11 am reading of 1169.00

so 1169 + 9.3 = 1178.3   that would be  3100 watts and hour  or 3100/49 = 63 watts I do not think that will happen



actual 3 hour reading = 1178.71
                                    1169
                                         9.71/3 =  3,237/49 = 66 watt a th at the wall  it is rated 3100/50 = 62 watts a th.  not sure it that is wall or dc rating.

so factory setting is okay  :

49 vs 50
 3237 watts at wall

webpage info below:


so 49 vs 50 is 2%    1th lower which is spec
3237w vs 3100wx is 4.4% or  237w higher which is spec

of course it is 66watts a th  vs 61 watts a th  which is very typical of inno to do exactly this

I let it go for 5 hours to 4pm New York time

so
1185
1169
    16 kwatts spent which is 3.2 or 3200 watts  for 49th  that is   3200/49 = 65.3 watts on factory setting

---

The next test will be lowest setting   as 33th at 1650 watts is promised if exactly on spec it would be 50 watts a th

translation of that is  33th x .95 = 31.35th  and 1650 x 1.1 = 1815 watts  that is 1815/31.35 = 57.9 watts a th

if I can get it to do 52watts a th rather then my 57.9 watt guess I would like that.

I just set it to lowest most efficient speed.  I will wait till 5pm and get readings

oh temps read 69 to 70 c on my 3 boards at factory settings.

1187.51 is the 5pm reading.

34.33  is the hashrate  I will check it later   to see what we do with this setup.

if this gear does 50 watts a th  the 1 hour  meter read would be

1189.23.

Actually if it can do 50 watts a th it has some staying power down the road. If you can buy it at a lower price in a few months.

---

I tested at lowest setting efficient and I am really disappointed in the gear.

it does 40th at 2460 watts which =  61.5 watts a th so it completely missed  the promised low setting numbers of   33.5th and 1650 watts.



I won't order anymore of these.  I want to thank minefarmbuy  for  handling this in a professional manner.
I took the unit to buysolar's house and put in in the barn to us that solar  array
I cranked unit up to factory +   which seems to do 53th at 3600 watts that is 68 watts a th  a lot more than the advertised  62 watts a th.

for now with current diff and coin price  it is okay gear but  it will lose profit at a diff of 23.8   as it will earn 5.62 usd and cost 5.62  if coins are 10200 and your power is 6.5 cents

if the down clocked worked as advertised.  it would earn $3.53 and cost $2.57  to run it when diff is 23.8 and coins are 10200

for the record  I would return this  to inno if I could but I used minefarmbuy  which  does allow a return and minefarmbuy  takes the hit.

inno deserves a hit on this gear.  They completely missed the down clock specs  even if you add in the -+ ranges they blew it .

I would love to see if they put out a firmware that gives the promised 50 watt a th on low speed.

Till then I do not want any more of their gear.

If my testing is wrong  and I had a meter issue someone show the unit at low speed doing close to 33th at 1650 watts.
I am only human and would love to be proven wrong and that  the downclock goes a lot closer to the 50 watts a th promised.

Be advised  that 61watts a th is very far from 50watts a th in the case of the t3 t50  low speed setting.

What's the hashrate on factory and performance? Is it on spec?

Little over 2.7kw sounds pretty good. Keep me posted.

Looks like it settled in at 43 th and does 63.3 watts a th

This is on balanced++ setting.

Of course Inno firmware set fans to 100% and gear is around 59c a board.

Since these go up to Clifton NJ the loud noise is okay but I would prefer 90% not 100%

This are pretty loud I do not think they will please a wife or girlfriend with noise factor.

I set it to balanced ++. It seems to do 43th edit 39th. Watts are at kwatts are at 148.44. I will check in an hour to get a reading.                 
   7:34 pm to 8:34 pm. In theory I will use 2.3 to 2.4kwatts so reading should be 150.84 at 8:34pm