Bitcoin Forum
May 14, 2024, 02:12:12 PM *
News: Latest Bitcoin Core release: 27.0 [Torrent]
 
  Home Help Search Login Register More  
  Show Posts
Pages: « 1 2 3 [4] 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 »
61  Bitcoin / Hardware / Re: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH on: January 01, 2016, 10:13:01 PM
Well the batch eight went into the other direction about .3 watts vs .25 watts.

Just to make sure we are on the same page:

Batch 6 was factory set at 600 MHz for 4.06 TH/s with 135 Chips
Batch 8 was factory set at 700 MHz for 4.73 TH/s with 135 Chips

Hence the .25 watts versus .3 watts because of Batch 8 clocked at a higher frequency.

My question is, can I under clock Batch 8 down to 600 MHz and get the same .25 watts efficiency as the Batch 6 at 600 MHz?

If the efficency is really different, it's very likely that the buck converter is set to another voltage.
I'm back home. Will answer PM, quote requests, arrange a few deliveries first.
Then, I'll look more in depth at what can be done on this buck converter.
Better efficiency should be easy by setting the output voltage slightly lower.
Overvolting might be tricky because the MOSFETs could already be close to their limits at 800MHz. I didn't check the data sheets.
(my batch 8 ran without any issue 10 days at 800, I set it back to 700 so it's not too loud now that we are back.
62  Bitcoin / Legal / Re: Payment in bitcoin law reducing taxes on: January 01, 2016, 04:44:32 PM
we do not speak of bank transfer, we are talking about payment in BTC, which is not any payment for customs, then the seller may also declare a lower value

for example S7 costs $ 1,200 and pay at bitcoin
bitmain could also declare $ 700
This permeterebbe users to pay less money to customs.

but maybe you're financially well, and you do not need to save.
thank you anyway for your time even if you have not voted.

the world becomes a shit for people who do not rebel, and only make us slaves. Angry

It's not as  simple as putting a number on a paper. There is much more things involved when you are a company.

Even if I'd like to pay less taxes, big companies exporting from China are taking huge risks if they are making false declaration on the exported value. I believe you can easily check on Google what the risks are for the company making false declaration.
This value should be the same as the one declared in the results of the company. Large companies are subject to more control than small ones too, at least, here in France.

I'm not taking the risk when selling through my company.
When selling as an individual, it's a different story.

Considering Bitmain is also getting 17% of the declared exported value back from Chinese government.
Would you risk getting penalties for false declaration and at the same time lowering the amount you get back from the government if you were owning Bitmain?
63  Bitcoin / Mining support / Re: Need new s7 hashboard. on: December 31, 2015, 07:05:34 PM
What i did was i powerd off everything and the miner was hot like it had overheated waited for it to cooldown and then powerd on the whole thing.
i meen the hashboards where running for about 30 sec whitout the controlboard powerd on and the whole thing like went into overheat mode.. smoke every where you could smell the burned plastic for days. I gusess if i let it runn for like 1 min or more there could prop. be a fire , Whats more intressting is that the PSU was running while the miner was frying itself or what ever, but something was def. Mellting inside it..  i will open up the miner later today and post some pictuers of its hashbords and the miner so you guys can also take a look  maybe , someone have a clue what to do nexst or how to fix this. ? would app. It

Board 1 and 2 was OK and hashing. But board 3 is showing temp 0 and --- --- sometimes 000 000...




I may be interested in your dead board shipped to France. Can you send me a picture of the board?
64  Bitcoin / Hardware / Re: Large review of Finksy/J4bbrwock server psus compared to ATX psu- photos up on: December 31, 2015, 08:47:57 AM
I'll break this up to make it easier to follow. I'll leave the discussion alone after this post...

Hi Finksy,

Thank you very much for your constructive criticism  Your frankness is much appreciated.  I also appreciate your willingness to contribute HW in Prelude's efforts to determine power efficiency difference between PSU's.  You made very good points about what would probably be the proper way to test PSU efficiency for bitcoin mining applications.  Consistency in HOW the test is performed is extremely important... Making measurements with same meter at same location with same rigs over same period of time.

I hope I did not open up a can of worms.  That was not my intention at all.  I've made too many mistakes in the beginning of my bitcoin mining venture.  I've learned [The hard way] how important it is to make an educated decision about hardware used for mining.  I don't want to continue those mistakes anymore.  I'm hoping this venture will last for many years to come.  So, I'm wanting to make the wisest and most economical decision I can make while growing into the future.

All of the time and energy invested with engineering and design with breakout boards for various PSU's and the commitment to produce these on a relatively large scale for many people by you and J4bberwock is appreciated greatly, Sir!  Thank you both for all you have done and continue to do for the community!

As far as I can tell, the easiest way we'll have to test efficiency will be to compare against a known efficiency PSU with either of the following options:
-Same voltage for both PSUs if the miner is using a string design
-Close voltage for both PSUs if the miner is using a buck converter (The S7 B8 and other with 135chips will work since it is an hybrid with the string fed by a buck converter similar to the design I tried for SFARDS chips)

Ideally, DPS2000 BB vs DPS2500BB, and 2880 vs 2980 with the same miners as load so we can cancel the effect of the built in fans.
I'll test the 2500 vs 2000 next week once I'm back home.

I can estimate, 87% for the 2880 including the fans will be between 90% and 92% without them depending on the load, so not too far away from the platinum efficiency measured on the 2980.
The 4kw may have a better global efficiency including the fans since we are using ultrakaze or similar fans, and only 2 of them for 4kw compared to 3x turbojet fans on the 2880/2980w
65  Bitcoin / Hardware / Re: Bitmain give false? on: December 30, 2015, 09:52:03 PM
for my English I help with google translator, I am sorry,

I'm pissed off because I do not understand why he said more than they cost,

Do you have the proforma invoice from Bitmain that Fedex used for the calculations?
When there is no separate amount declared for Shipping, Fedex and DHL (probably UPS too) will add an arbitrary shipping charge amount to the invoice before they calculate VAT.
It's also possible that at the time Bitmain exported the items, the value increased because BTC exchange rate changed.
You paid a BTC value, not a US$ value, so it's possible that they edited the US$ value at the time of shipping.

I also saw somewhere that Bitmain is getting money back from Chinese governement based on the exported value.
If it's true and proven, overvaluating exported items (chosing the best exchange rate between sale time and shipping time) may be a way to make more income.
66  Bitcoin / Hardware / Re: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH on: December 26, 2015, 07:33:46 PM
It does not seem right to me sincerely lose $ 600 in 4 days Huh Huh

LOL. Forget it. We're talking about bitmain here.

Your only hope is to have the bank cancel the transfer if you want your $600 back.

And in this particular case, if the bank can cancel, they'll charge you something, but it'll be cheaper than $600
67  Economy / Computer hardware / Re: [WTS] Custom server PSU breakout boards. 1200, 1300, 2000, 2880w. PCIE adapter on: December 24, 2015, 10:55:24 AM
Please check PM, I sent message about new order. Tnx.

Yes, I saw the message. I'll be away until the end of the year.
Shipping will resume on the 4th of january for new orders.
68  Bitcoin / Hardware / Re: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH on: December 24, 2015, 09:38:25 AM
Batch 8 on DPS-2000BB @ 743Mhz:



This is exclusive of the fans. System efficiency will depend on your cooling situation.

This works out to just over 0.27W/Gh/s - and the DPS-2000BB can do better with higher input voltage

Vac       207.56
Vsupply 12.032
Vchain   10.281

I'll screenshot the miner interface once it's been running for a few hours but this is essentially 4.95TH; 0.016%HW; 60C board temps

pump up the volt on the dps, it should workout better & can go either higher or lowers the HW errors for sure.

ah i just saw it's a B8 hmmm but good to see some diff #'s.

i saw on phillipmas post & j4b's post they went higher with higher volts but of course it's the earlier batches.

i'm still happy at stock clocks but after the 90 days, he he he he !

Mine is a Batch 8 too, no overvolt for now, just lucky with the chips, I guess.
69  Bitcoin / Hardware / Re: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH on: December 23, 2015, 10:23:28 PM
Yes definitely Luck.  My Batch 8 using the same 220V IBM 2KW won't change in hash rate by moving to 775 or 800M.  Have to increase the voltage to see some gain like what Phillip did.

There is no point in changing the voltage on a Batch 6, 8 or 9. They have a Buck Converter built onto the Hash board that controls the voltage sent to the chips. Good news is that at some point I am sure a mod will be available that will enable the Core Voltage to be adjusted.

I'm interested in changing Vin to the buck converter to find the maximum efficiency, not in changing Vout to overclock the chips; the LM27402 in this configuration should be able to hold Vout of 10.1 with Vin in lower 11s and it's conceivable to see +/- 1% efficiency over the Vin range 11.6 to 13.3. Numbers / photos soon.

The closest Vin is from Vout, the better your efficiency.
I haven't checked if there is lm27402 design tool/spreadsheet available, but I used the one for LM27403 in a design, and I can confirm that changing Vin can give some efficiency improvement.
Unfortunately, we are already quite close, so I don't believe you can improve efficiency by 1%.

Edit:
The design calculator is available, I ran some numbers assuming 200kHz switching frequency (most likely value with 2 mosfest in parallel), and 11.6 to 13.0v won't really change anything in efficiency.
And since the switching frequency is already at 200kHz, there is no gain to expect by changing it neither.

Edit2: I checked the frequency adjust resistor, and it's indeed set for 204kHz, so no gain to expect by lowering the frequency, it's already at its lowest value.

The only option left to get a better efficiency is by changing Vout
70  Bitcoin / Hardware / Re: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH on: December 23, 2015, 08:11:09 PM
Hi Everyone.


So normally I wouldn't rant about support topics, But I have 2 major issues I have been trying to resolve over the last few days with no joy.

Now I will start by saying bitmain have normally been really good and solve issues quickly, but I don't seem to be getting anything  resolved at the moment so any advice would be much appreciated.

The first issue I have is that my latest S7 that is currently sat in customs has an invoice attached saying the value of the 1x S7 is $30,000 and they want £5000 for the VAT.

Now as far as im aware the S7 I ordered should be valued at $1650 but it seems that there has been some sort of admin error on bitmains side, I did receive an email from bitmain stating that we needed to pay our VAT or the machine would be sent back to them, I advised them of the issue and they sent over the correct invoice to me, but Fedex have request an explanation from bitmain and a copy of the invoice directly from them, so far fedex has not received a response and neither have I.

The next issue im having is with one of our current S7 machine from batch 5 has an issue,  I checked all my machines the other day and noticed a very strong electrical burning smell and noticed 2x hashing boards had gone down.

I assume they have died but cannot check as there is a warranty sticker over the machine that I do not want to touch without the permission from bitmain, So far I have emailed info@ sent a message through zendesk and also messaged zendesk when live support was available and have so far received no response.

I feel a little ignored and for the first time in 2 years im annoyed at the way Bitmain is dealing with my issues, we have easily spent well over $30,000 with them, not much compared to some bigger miners but enough to think they would as least help.

Like I said I have been more than happy with Bitmain products and there support in the past, but feel very ignored at the moment, and any advice would be greatly appreciated.



The thing that really bothers me about the S7 is the new Bitmain warranty policy stating that if anything burns on the board, there is no warranty.
It's likely that they wanted to avoid replacing boards overclocked/modded, but it also prevent honest users like you, who either got bad luck, or falling heatsinks (engineering failure) for example, to get warranty.
I hope they will reconsider it, at least for the defects caused by falling heatsinks.

If I'm not mistaken, on a batch 5, there shouldn't be any buck converters, unlike the 135 chips units.
So the most likely dead components are an asic, a capacitor, or one of the small LDOs creating the communication voltages (usually 1.8v).

You can start by shaking the unit to check if you hear a something like a heatsink that went loose.
If that's the case, you can remove one fan without losing the warranty to confirm.

You can try to negociate with them if it's a heatsink issue because there is no way they can take you as responsible for it.

Other burnt parts can be caused either by a user overclocking the unit to death, or simply by one bad component in a batch, so it's a bit harder.

If they are totally out of warranty, I'd be interested to grab one board shipped to France for experimenting.
71  Economy / Computer hardware / Re: (WTB) hashing boards for an S7 Batch1 on: December 23, 2015, 04:57:46 PM
I'm looking for 3 working hash boards from either Batch 1,3, 5 or 6 (600 Mhz).

If you have 1 to 3 boards you could sell,  PM me with cost per board.

Do you mean you have dead boards with 162 chips? I'm interested in at least one shipped to France.
Offering 250$ for the board.



I think he might want to add some extras, kinda interested too in adding more boards to a controller  Smiley .

I have 3 dead boards. I've decided Im going send them to Bitmain. If they can be repaired and if the cost is not too high, I'll get it running in about 3 weeks.

If not, I'll resume my search.

@J4bberwock I'll also ask them to send me back the boards if they can't fix them. We can talk then.

Thanks, keep me udated. For one dead board, I'll offer roughly 1/5th of the price Bitmain will sell a full S7 at the time of the deal.

If they are still under warranty, go for it, it's your best bet.
If they want to charge you the repair, I hope it won't be too high.

One of the things I didn't like about their new warranty policy is that if some heatsinks happen to fall (their fault) and you end up with burnt components on the board, the warranty doesn't apply.
I'm totally ok for no warranty when I play with the board, disassemble things and end up with some magical blue smoke, but not when it's an engineering failure that causes the miner to die.
72  Bitcoin / Hardware / Re: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH on: December 23, 2015, 04:45:15 PM
The earlier batches probably have more, undervolted chips. While the later batches have less, overvolted chips. So i'm guessing the earlier batches with the most chips would be able to get the most out of overvolting.

Assuming nothing else fry when you try it.


There is probably not much in it, except of course for the greater number of chips favouring the 162 chip systems.
The 162 chip miner with 12v across the 18 Chip string gives a core voltage of 0.667V
The 162 chip miner with 12.2v across the 18 Chip string gives a core voltage of 0.678V
The 135 chip miner with 10.12v (measured by J4bberwock) across the 15 Chip string gives a core voltage of 0.675V

A bigger factor may be if they have been selecting BM1385 chips, which of course we will never know?

Rich

The label I've seen on the Batch 8 boards states BM1385+ Anyone can confirm anything regarding older batches?
So maybe hand picked chips, or a minor revision on the second batch from the foundry.
3 month after first release seems correct to have chips manufactured with a minor update on them.
73  Economy / Computer hardware / Re: KNC boards, Jupiter,Saturn. on: December 23, 2015, 11:53:16 AM
You can buy less, but the price will not change. 5 euro piece+ shipping costs.

On each plate is 8 x VRM, the same with Neptune and Titan VRM.s.
VRM (voltage Regulator Module) is just a different name for DC/DC converter.

 Price of the new VRM is 35 USD, if quantity is 500 pieces.

You can see how much is the Jupiter Boards  eBay price .  US $39.00  .    9 available   /  11 sold

http://www.ebay.com/itm/KNC-Jupiter-ASIC-November-Mother-Board-International-OK-/111759749380?hash=item1a05667104:g:uQ4AAOSwgQ9V5TPQ



I never looked for the VRMS modules alone, and wasn't even aware you could buy the things in assembled modules like that.
Where did you find them?
But 5€ is clearly cheap, even if it was just to get the spare parts for Neptune/Titan.
74  Economy / Computer hardware / Re: (WTB) hashing boards for an S7 Batch1 on: December 23, 2015, 09:25:24 AM
I'm looking for 3 working hash boards from either Batch 1,3, 5 or 6 (600 Mhz).

If you have 1 to 3 boards you could sell,  PM me with cost per board.

Do you mean you have dead boards with 162 chips? I'm interested in at least one shipped to France.
Offering 250$ for the board.
75  Bitcoin / Hardware / Re: Antminer s7 issue on: December 23, 2015, 09:15:38 AM
It should have a warranty no matter where you bought it unless it was from someone who bought from them and sold it as is (basically 2nd hand) but why not try anyways. If all else fails you could try to diagnose and try to repair it yourself if your tech savvy. Its not hard to put a chip back on if that's the issue. If not then pay someone who can or just write it off as a loss and move on (maybe sell the board to at least get something back).
If nothing else works, I'll be happy to buy the dead board for a fair price.
Actual miner price is 1400$. It will be 1100 in January with batch 9.
I'll pay 250$+ shipping to France for the dead board.
76  Bitcoin / Hardware / Re: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH on: December 23, 2015, 09:06:38 AM
Yes definitely Luck.  My Batch 8 using the same 220V IBM 2KW won't change in hash rate by moving to 775 or 800M.  Have to increase the voltage to see some gain like what Phillip did.

There is no point in changing the voltage on a Batch 6, 8 or 9. They have a Buck Converter built onto the Hash board that controls the voltage sent to the chips. Good news is that at some point I am sure a mod will be available that will enable the Core Voltage to be adjusted.


Rich
I won't have access to the S7 for the next 8-9 days, but I'll definitely change the 6k8 resistor as soon as I'm back.
800MHz, 5,3Th is still too low. I'm sure we can do 900.
And I'll probably try to get an earlier batch with 162 chips later to see how it can be overclocked.
77  Bitcoin / Hardware / Re: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH on: December 20, 2015, 10:55:34 PM

Any chance the Batch8 is software controllable? digging around vi SSH (root/admin), found these references:
etc/config/cgminer.conf:    "bitmain-voltage" : "0706"
etc/cgminer.conf.factory:   "bitmain-voltage" : "0725"

seems strange that they are different, yet still necessary to reference. Anyone know what voltage the B8 actually runs at?

currenty playing with the values to see if they do something

It would be great if they do something.... but my gut feeling is that those numbers look very familiar from all sorts of conf files in other miners that do nothing?

They don't seem to do anything, but why put a digital pot adjustable over I2C with a PIC just to control it if you just set a fixed value?
Some people with better Antminer firmware knowledge may be able to unlock it if it's locked somehow.
78  Bitcoin / Hardware / Re: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH on: December 20, 2015, 10:45:10 PM
ok, so R15 = Rc1 in the datasheet, and C73, C75 would be Cc1,Cc2.
U2 should be a digital pot made by onsemi, then. In the cat51xx family
And that means that we could have some software control over it.
I spotted the 6k8 as being tied to ground.
I'll do a quick test on one board replacing the 6k8.
I noticed that without the controller plugged, the voltage output at the coil was closer to 8v, can't remember exact value.
And, it went higher to 10.2 with the control board plugged

8V would make sense for the sort of voltage you would expect without the shunting effect of the Digital Pot. So this implies that the PIC is being controlled or at least reset by the controller as opposed to it being completely stand alone as I had assumed?
A trick to feed the chips with a voltage so low they won't start (burn the entire board) if the controller doesn't send  a signal to tell he is ready with ethernet connected?
Yes I guess it makes it safe. CAT51xx looks spot on against the pinout, not sure how that ties to AG30/ AG6K but guessing it's a 10K pot?
I do not have sufficient info with just the pictures to suss out all the connections to the PIC, in particular any interface to the control board?
Any chance the Batch8 is software controllable? digging around vi SSH (root/admin), found these references:
etc/config/cgminer.conf:    "bitmain-voltage" : "0706"
etc/cgminer.conf.factory:   "bitmain-voltage" : "0725"

seems strange that they are different, yet still necessary to reference. Anyone know what voltage the B8 actually runs at?


currenty playing with the values to see if they do something
79  Bitcoin / Hardware / Re: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH on: December 20, 2015, 10:16:48 PM

ok, so R15 = Rc1 in the datasheet, and C73, C75 would be Cc1,Cc2.


U2 should be a digital pot made by onsemi, then. In the cat51xx family

And that means that we could have some software control over it.

I spotted the 6k8 as being tied to ground.
I'll do a quick test on one board replacing the 6k8.

I noticed that without the controller plugged, the voltage output at the coil was closer to 8v, can't remember exact value.
And, it went higher to 10.2 with the control board plugged



8V would make sense for the sort of voltage you would expect without the shunting effect of the Digital Pot. So this implies that the PIC is being controlled or at least reset by the controller as opposed to it being completely stand alone as I had assumed?




A trick to feed the chips with a voltage so low they won't start (burn the entire board) if the controller doesn't send  a signal to tell he is ready with ethernet connected?

Yes I guess it makes it safe. CAT51xx looks spot on against the pinout, not sure how that ties to AG30/ AG6K but guessing it's a 10K pot?

I do not have sufficient info with just the pictures to suss out all the connections to the PIC, in particular any interface to the control board? However several of the connections come to the pads on the board edge probably including programming the PIC. If I had a board I would have a go at reading it in case they have not Read Protected :-)

I have done some maths on the Digital Pot, which has 32 positions, in combination with the 6K8 to ground & series 12K. So assuming it is a 50K Pot we get.

        Supply V   Core V
Min       8.63        0.58
Mid       9.17        0.61
Max     11.87        0.79


Which would be a very sensible range of core voltages, so may be correct?

Rich



Found it:

MCP4017T-503E/L

http://ww1.microchip.com/downloads/en/DeviceDoc/22147a.pdf
80  Bitcoin / Hardware / Re: ANTMINER S7 is available at bitmaintech.com with 4.86TH/s, 0.25J/GH on: December 20, 2015, 05:05:34 PM

ok, so R15 = Rc1 in the datasheet, and C73, C75 would be Cc1,Cc2.


U2 should be a digital pot made by onsemi, then. In the cat51xx family

And that means that we could have some software control over it.

I spotted the 6k8 as being tied to ground.
I'll do a quick test on one board replacing the 6k8.

I noticed that without the controller plugged, the voltage output at the coil was closer to 8v, can't remember exact value.
And, it went higher to 10.2 with the control board plugged



8V would make sense for the sort of voltage you would expect without the shunting effect of the Digital Pot. So this implies that the PIC is being controlled or at least reset by the controller as opposed to it being completely stand alone as I had assumed?




A trick to feed the chips with a voltage so low they won't start (burn the entire board) if the controller doesn't send  a signal to tell he is ready with ethernet connected?
Pages: « 1 2 3 [4] 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 »
Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!