Pages:
Author

Topic: [Work in progess] Burnins Avalon Chip to mining board service - page 69. (Read 624200 times)

sr. member
Activity: 434
Merit: 250
I didnt know that rpi dont include the sd-card already so i think i have to buy one first since i dont find such card in the moment. But for testing i probably will use a notebook with linux now...

SD cards super cheap thankfully. You'll need a reader on the PC to flash the image to the SD card, and a USB <> MicroUSB cable to power the Pi if you plan to power it off a powered USB hub. Look for them in cell phone accessories, not computer cables area.
legendary
Activity: 2674
Merit: 1083
Legendary Escrow Service - Tip Jar in Profile
Thats where cgminer stops.
I now will try to run it in linux.

I'm running my BitBurners on a Raspberry Pi with cgminer 3.3.2, working perfectly. Running 48 straight without any problems.

I didnt know that rpi dont include the sd-card already so i think i have to buy one first since i dont find such card in the moment. But for testing i probably will use a notebook with linux now...
hero member
Activity: 700
Merit: 500
Thats where cgminer stops.
I now will try to run it in linux.

I'm running my BitBurners on a Raspberry Pi with cgminer 3.3.2, working perfectly. Running 48 straight without any problems.
legendary
Activity: 2674
Merit: 1083
Legendary Escrow Service - Tip Jar in Profile
I had 100 chips ordered in zefir's first batch and today the ready made miners reached. I took fotos of everything.

http://imgur.com/a/nlb09

The first and the last miner in the line needs 2 jumpers to be set. I didnt found them but in the basement i found 2 old harddrives and one motherboard where i could steal 4 jumpers from.

Windows first didnt start but i used zadeg like described for BFL-Miners to install WinUSB. The Miner then start, the lightshow is green, yellow and red (im not sure what the colors mean) and the chips get hot and use wattage. Unfortunately something seems not to be correct with cgminer in windows yet. It seems to start the hashing but no stats show up. And when i close cgminer it still goes on with mining while i dont know what it calculates then.

Code:
[2013-08-12 22:07:03] Started cgminer 3.3.2
 [2013-08-12 22:07:07] BTB0: Reset succeeded
 [2013-08-12 22:07:07] BTB0: Idling 1 miners
 [2013-08-12 22:07:07] BTB0: Core voltage set to 1200 millivolts
 [2013-08-12 22:07:07] Probing for an alive pool
 [2013-08-12 22:07:10] Pool 0 difficulty changed to 8
 [2013-08-12 22:07:11] Disabling extra threads due to dynamic mode.
 [2013-08-12 22:07:11] Tune dynamic intensity with --gpu-dyninterval
 [2013-08-12 22:07:11] Network diff set to 37.4M
 [2013-08-12 22:07:12] Thread 1 being disabled

Thats where cgminer stops.

I now will try to run it in linux.
hero member
Activity: 563
Merit: 500
Thanks for pointing this out. Thats not a good thing. One would need special PSU's that have different 12A-Outputs since a high wattage PSU with only 1 or 2 such channels used with Y-cables would hurt the miners when i understand it correctly.
Its a good thing i didnt buy more psu's yet but others might have bought wrong ones. Or is there a workaround for this?

You don't need special power supplies.  As I understand it we need to do one of two things:

  • EITHER: Power each miner via its PCI-E connector, just like you do with most other miners.  How you do this doesn't matter, as long as you don't overload your PSU or cables
  • OR: Connect your miners together in pairs, using the screw connectors and the solid-core wires that burnin specifies, and then power each pair of miners through the PCI-E connector of one miner in the pair

What you must not do is connect larger groups together using the screw connectors, in an attempt to power three or more miners all from power fed to one of the PCI-E connectors.  Either connect them in pairs, or just ignore the screw connectors and power each miner via its PCI-E connector.

If in any doubt at all, just ignore the screw connectors.

Corrections welcome if I've misunderstood anything.

roy
member
Activity: 80
Merit: 10

Thanks for pointing this out. Thats not a good thing. One would need special PSU's that have different 12A-Outputs since a high wattage PSU with only 1 or 2 such channels used with Y-cables would hurt the miners when i understand it correctly.
Its a good thing i didnt buy more psu's yet but others might have bought wrong ones. Or is there a workaround for this?

You misunderstand.  The information you've been given here tells you to use solid copper wires Awg 12 or 14 to bridge the board terminals and not to exceed 20 amps between the connectors on the boards (240W).

If you stick to this, nothing you do with the PSU cabling can exceed these limits and damage the boards directly.  Of course, with such high loads if you aren't safe and wise its entirely possible to melt upstream connectors and cause a fire risk, or otherwise overload your PSU rails depending on its design.

@Burnin: If the tracks between connectors are rated 20amp does this mean the board can take its own requirements, say 8 amps plus passthrough a further 16 amps between the connectors, subject to not overloading the PCIe connector itself.
legendary
Activity: 2674
Merit: 1083
Legendary Escrow Service - Tip Jar in Profile
I wonder what would be the risk when a psu has for example 400W at one 12A line and i would power one board directly and 3 others with cables? I could use 3mm speaker cables if that helps...

You risk that the weakest part melts.
The power for the 3 others has to go through the PCB from the direct connection to the green terminal then.
burnin already recommended to only connect two boards and use solid wires. I would just follow his recommendations.

I took solid as stable and not static and i still wonder whats the difference. In theory the size of the copper inside tells how easy the electrons can wander inside. So i cant see what difference would there be.

When i think about the setup with the cables... connecting a second board this way would mean that the power would flow completely through the first board until it goes to the second one. In case i overclock the first board would have to take up to 200W and the second 100W. Burnin mentioned that some parts regarding the voltage already is at its limit with such high overclocking/overvolting. So would this setup not be possible for overclocking or are there no parts involved that get at their limits with that overclocking?
As I asked,
https://bitcointalksearch.org/topic/m.2907566
and burnin replied at the top of the page:
https://bitcointalksearch.org/topic/m.2908117

Thanks for pointing this out. Thats not a good thing. One would need special PSU's that have different 12A-Outputs since a high wattage PSU with only 1 or 2 such channels used with Y-cables would hurt the miners when i understand it correctly.
Its a good thing i didnt buy more psu's yet but others might have bought wrong ones. Or is there a workaround for this?
sr. member
Activity: 243
Merit: 250
ALTCOM Ab9upXvD7ChnJxDRZgMmwNNEf1ftCGWrsE
I think it's solid core vs stranded wire. Its just a recommendation by burnin.
I think you are much more likely to get a good connection using solid core wire.
If it were me I would try and use 12 or 14 gauge solid core copper wire if it would fit into the connectors.


Stranded wires without crimped conductor sleeves are a hazard, solid conductors are recommended for fixed installations like this.
And they are big enough for AWG12.

Yes, I have the signed recipe with a epoch time long before the deadline. No response so far.

I will probably drive to Hamburg next week because I’ll be in Hannover for a work related faire anyway.

Yes, i did send it on Friday, just missed to add your tracking number.

Hello Burning¡ I send you a email two weeks ago asking about order 539. I paid with SEPA transfer and i want know if you recibed the money without problems¡ I use my friend partner forum name: jorgeantolini

Thanks¡¡¡
It is marked as processing that means its paid.

In general: anyone with open requests that have not heard from me yet, drop me an email.

The tracks between the connectors can take up to 20Amps.

Btw: quick instructions on how to assemble this thing: http://www.burninmining.com/assembly-instructions-part1/
Sources will be released soon, too.

legendary
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
I wonder what would be the risk when a psu has for example 400W at one 12A line and i would power one board directly and 3 others with cables? I could use 3mm speaker cables if that helps...

You risk that the weakest part melts.
The power for the 3 others has to go through the PCB from the direct connection to the green terminal then.
burnin already recommended to only connect two boards and use solid wires. I would just follow his recommendations.

I took solid as stable and not static and i still wonder whats the difference. In theory the size of the copper inside tells how easy the electrons can wander inside. So i cant see what difference would there be.

When i think about the setup with the cables... connecting a second board this way would mean that the power would flow completely through the first board until it goes to the second one. In case i overclock the first board would have to take up to 200W and the second 100W. Burnin mentioned that some parts regarding the voltage already is at its limit with such high overclocking/overvolting. So would this setup not be possible for overclocking or are there no parts involved that get at their limits with that overclocking?
As I asked,
https://bitcointalksearch.org/topic/m.2907566
and burnin replied at the top of the page:
https://bitcointalksearch.org/topic/m.2908117
legendary
Activity: 2674
Merit: 1083
Legendary Escrow Service - Tip Jar in Profile
When i think about the setup with the cables... connecting a second board this way would mean that the power would flow completely through the first board until it goes to the second one.
No, it doesn't flow through the entire PCB and its power conveter circuits. It will flow just through the PCB traces between the PCI-E connector and the green power connector. As burnin states, these few cm of traces between those 2 connectors are rated for the current of 2 boards (1 non-overclocked board about 40W, so the traces are rated probably for 80-100 W at 12V, which would mean 6-7 A).

So its the same flow of energy like you would power the board through the green connector directly like you have it when pushing 2 times the wattage in per atx and take one half of it out of the green connector again.
Wouldnt it then be possible to not connect the solid cables to the green connector but instead directly to the atx-incoming-power? Then there are no pcb-parts in between that could brake and the attached cables could draw the power to the other boards. I mean that would be the same like using atx-y-cables isnt it? Of course im not sure if such kind of DIY wouldnt involve a portion of risk.

DHL tells me they want to bring my miners to me tomorrow.
member
Activity: 63
Merit: 10
Hello Burning¡ I send you a email two weeks ago asking about order 539. I paid with SEPA transfer and i want know if you recibed the money without problems¡ I use my friend partner forum name: jorgeantolini

Thanks¡¡¡
member
Activity: 98
Merit: 10
My batch #1 order is still processing even though you stated that all batch 1 and some batch 2 orders have shipped.
Could you please look into my order?

Did you fill out Zefir's website with information about you chip order?

Yes, I have the signed recipe with a epoch time long before the deadline. No response so far.

I will probably drive to Hamburg next week because I’ll be in Hannover for a work related faire anyway.
m5
newbie
Activity: 82
Merit: 0
When i think about the setup with the cables... connecting a second board this way would mean that the power would flow completely through the first board until it goes to the second one.
No, it doesn't flow through the entire PCB and its power conveter circuits. It will flow just through the PCB traces between the PCI-E connector and the green power connector. As burnin states, these few cm of traces between those 2 connectors are rated for the current of 2 boards (1 non-overclocked board about 40W, so the traces are rated probably for 80-100 W at 12V, which would mean 6-7 A).
sr. member
Activity: 378
Merit: 250
I think it's solid core vs stranded wire. Its just a recommendation by burnin.
I think you are much more likely to get a good connection using solid core wire.
If it were me I would try and use 12 or 14 gauge solid core copper wire if it would fit into the connectors.
legendary
Activity: 2674
Merit: 1083
Legendary Escrow Service - Tip Jar in Profile
I wonder what would be the risk when a psu has for example 400W at one 12A line and i would power one board directly and 3 others with cables? I could use 3mm speaker cables if that helps...

You risk that the weakest part melts.
The power for the 3 others has to go through the PCB from the direct connection to the green terminal then.
burnin already recommended to only connect two boards and use solid wires. I would just follow his recommendations.

I took solid as stable and not static and i still wonder whats the difference. In theory the size of the copper inside tells how easy the electrons can wander inside. So i cant see what difference would there be.

When i think about the setup with the cables... connecting a second board this way would mean that the power would flow completely through the first board until it goes to the second one. In case i overclock the first board would have to take up to 200W and the second 100W. Burnin mentioned that some parts regarding the voltage already is at its limit with such high overclocking/overvolting. So would this setup not be possible for overclocking or are there no parts involved that get at their limits with that overclocking?
newbie
Activity: 31
Merit: 0
My batch #1 order is still processing even though you stated that all batch 1 and some batch 2 orders have shipped.
Could you please look into my order?

Did you fill out Zefir's website with information about you chip order?
member
Activity: 98
Merit: 10
My batch #1 order is still processing even though you stated that all batch 1 and some batch 2 orders have shipped.
Could you please look into my order?
sr. member
Activity: 243
Merit: 250
ALTCOM Ab9upXvD7ChnJxDRZgMmwNNEf1ftCGWrsE
Now that the boards are for sale is the schematic going to be released? I know it was mentioned previously in the thread but haven't seen them yet. I think having them to look at may answer a lot of questions.

I cant quote burnin directly right now, but his statement towards this question was that he will release them after a reasonable profit was made.
Burnin surely had a lot of work done and needed to invest which should at first returned bevor releasing those data.

I will send out BOM, Stencil-Gerber, drawings and Pick and Place data out to those who have ordered blank PCBs next week, under NDA.
That NDA basically reads: do not redistribute until i open source the project on delivery of the first BitBurners. (5th to 6th of august.)
Unfortunately only a few will get PCBs from the very first batch since i received more chips from Zefir then anticipated.
More PCBs are on their way but will take at least one week longer then the first. (12th to 16th)

Have you send? Nothing in my Inbox.

Did you pay that order?


Question regarding that, that hasn't been mentioned.
If you are powering more than 2 of them through one power source, you are also running that current through the input voltage section of each of the boards.
Thus I imagine that could damage the boards also ... ?
i.e. really, don't do it with more than 2 ... ever ... for that reason also?

The tracks between the 6Pin connector and the screw-terminal are rated for the current of two boards. 
sr. member
Activity: 378
Merit: 250
I would only connect one board with the solid copper wire.
So one of every 2 boards has the main power connected.
Also make sure the connections are tight.
Lose wires could cause the connector to melt.
legendary
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
I wonder what would be the risk when a psu has for example 400W at one 12A line and i would power one board directly and 3 others with cables? I could use 3mm speaker cables if that helps...

You risk that the weakest part melts.
The power for the 3 others has to go through the PCB from the direct connection to the green terminal then.
burnin already recommended to only connect two boards and use solid wires. I would just follow his recommendations.

I wrote it in red and bold, USE SOLID COPPER WIRES
...
Question regarding that, that hasn't been mentioned.
If you are powering more than 2 of them through one power source, you are also running that current through the input voltage section of each of the boards.
Thus I imagine that could damage the boards also ... ?
i.e. really, don't do it with more than 2 ... ever ... for that reason also?
Pages:
Jump to: