OK just an update. I have done all necessary staff for board production to start. I had big enough order of new boards to think it is safe to go ahead and pay first 50% of the licence. But as I was afraid I can't get them done in Slovenia but will still be done in EU. So tboards will be done(if all go to the plan) on 12th and I should get them day or two later... Then assembly and programming... I think I should be able to have last boards ready for shipping over the weekend.
I also talked with customs today. Some additional chips were processed today the rest will be processed tomorrow and should have the by the end of the week...
Topic: BFL board project COINTAMINATION - EU facility - ORDERS OPENED for Chili - page 26. (Read 73054 times)
November 04, 2013, 03:35:30 PM
November 03, 2013, 04:14:05 PM
I think this is your problem:
https://bitcointalksearch.org/topic/m.3468682
Some chips overheet and slow down...
https://bitcointalksearch.org/topic/m.3468682
Some chips overheet and slow down...
Quote from Spieder on the chili thread:-
"I've been playing around with a chili I bought on ebay with an Evo 212 until my batch 2 boards are done.
I originally booted it up with the Evo 212 and a 120mm fan blowing sideways over the vrms and caps. In bfgminer, cgminer and bfgminer on a raspberry pi the unit runs @ 37-38 ghz @ 72 C - 10% errors.
If I remove the side fan, the unit quickly readjust to 52 C and drops to about 32 ghz. So cooling those vrms is a key piece to achieving higher speeds. I know there's a lot a variability in performance based on chips but I'd tend to think if your cooling option let's you hash at 72 C then you're close to max. If your hashing at 52 C you may have room for improvement.
From the wall with my pi, I'm reading the system is pulling 175 watts at 32 ghz and 209 watts at 37 ghz. So 15% increase in ghz is roughly 25% increase in power.
Now if I could figure out how to get those errors down I'd be happy because my effective is @ 33 ghz. I'd rather run at 33 ghz @ 175 watts with no errors than 38 ghz @ 209 watts with 33 ghz effective."
End quote
I think this covers my situation almost to the letter.
So for first option is to cool the mofsets.
If that has no effect I will look at the heatsink mounting.
Thanks for the lead
November 03, 2013, 03:34:41 PM
I think this is your problem:
https://bitcointalksearch.org/topic/m.3468682
Some chips overheet and slow down...
https://bitcointalksearch.org/topic/m.3468682
Some chips overheet and slow down...
November 03, 2013, 01:46:26 PM
My first observations now that my Chili has been online for 2 days.
The chili starts up with a big power draw, 200w+ and a hash rate of 38/39GH, temp at 75c with a massive error rate 15% (as if in an overclock condition)
If the chips are hot, there is a delay in startup, till the fan drops the temp. (It would possibly be a good idea for power-on after fan-on delay)
Then drops steadily for about 10 mins to 166w 32.5GH 52c, error rate 7.8%
Normal practice would be to start up low and increase to the maximum stable state.
The chili starts at full throttle and then slows down. (you would never do this to a car).
Does any body else have any observations from their chilli?
That is odd behavior. Is that from a cold start when you plug it in, or after it's been sitting awhile and you are just starting the mining software.The chili starts up with a big power draw, 200w+ and a hash rate of 38/39GH, temp at 75c with a massive error rate 15% (as if in an overclock condition)
If the chips are hot, there is a delay in startup, till the fan drops the temp. (It would possibly be a good idea for power-on after fan-on delay)
Then drops steadily for about 10 mins to 166w 32.5GH 52c, error rate 7.8%
Normal practice would be to start up low and increase to the maximum stable state.
The chili starts at full throttle and then slows down. (you would never do this to a car).
Does any body else have any observations from their chilli?
Hi Mr Teal
Cold start and Hot start, from power on.
The only difference is the delay on hot start.That is only for a few seconds, (15 or 20)
After it has been running for about 10 mins it becomes rock steady at 166w 32.5GH 52c, error rate 7.8%.
Stopping and starting mining software causes a small spike in the hash rate but only for about 1 min, most of my miners do this.
I guess the secret is not to power it down to often.
Are these chips like strip lights (needing a power spike to start them?)
November 03, 2013, 10:10:10 AM
Hello Lucko,
If so do you know when the voltage must be switched exactly and which are the values?
Do you know where a documentation of the chips is available about this change? I've only BFL_SHA2_Spec_Rev2.2.pdf here. Maybe there is a newer one?
Maybe a resistor in line of the trimmer can be used to short them before or after initialization, but I need more details.
Cheers...
All is just a guess but if I go from what we see and what MrTeal figure out too I would guess that 0.8 to 0.85V is strange range that hangs chips and they don't respond any more(being busy)... It could be just really unlucky coincident with power module that we used and any other would be just fine. But as I stated at the start this is just a guess what is going on. It explains what is going on but it could be something complicity different. Unfortunately there is just no more time to make any big debugging and changes... About power on... We guess if we use different power supply or a way to cut and then turn on power it would work... But in both cases we need to change the board... Chips just hang and will not work... At some point we did get 16 chips working but that was one time deal at minimal frequency and a lot of luck. In most cases you see 3 to 6 chips working if Atmel comes on line... And all 3 boards acts about the same so I don't think there is a point in guessing that there might be just a assembly problem. Boards were looked over by form in details...
I understood core voltage initialization of the chips must be different from the running core voltage?If so do you know when the voltage must be switched exactly and which are the values?
Do you know where a documentation of the chips is available about this change? I've only BFL_SHA2_Spec_Rev2.2.pdf here. Maybe there is a newer one?
Maybe a resistor in line of the trimmer can be used to short them before or after initialization, but I need more details.
Cheers...
November 03, 2013, 10:05:00 AM
My first observations now that my Chili has been online for 2 days.
The chili starts up with a big power draw, 200w+ and a hash rate of 38/39GH, temp at 75c with a massive error rate 15% (as if in an overclock condition)
If the chips are hot, there is a delay in startup, till the fan drops the temp. (It would possibly be a good idea for power-on after fan-on delay)
Then drops steadily for about 10 mins to 166w 32.5GH 52c, error rate 7.8%
Normal practice would be to start up low and increase to the maximum stable state.
The chili starts at full throttle and then slows down. (you would never do this to a car).
Does any body else have any observations from their chilli?
That is odd behavior. Is that from a cold start when you plug it in, or after it's been sitting awhile and you are just starting the mining software. The chili starts up with a big power draw, 200w+ and a hash rate of 38/39GH, temp at 75c with a massive error rate 15% (as if in an overclock condition)
If the chips are hot, there is a delay in startup, till the fan drops the temp. (It would possibly be a good idea for power-on after fan-on delay)
Then drops steadily for about 10 mins to 166w 32.5GH 52c, error rate 7.8%
Normal practice would be to start up low and increase to the maximum stable state.
The chili starts at full throttle and then slows down. (you would never do this to a car).
Does any body else have any observations from their chilli?
November 03, 2013, 09:10:07 AM
My first observations now that my Chili has been online for 2 days.
The chili starts up with a big power draw, 200w+ and a hash rate of 38/39GH, temp at 75c with a massive error rate 15% (as if in an overclock condition)
If the chips are hot, there is a delay in startup, till the fan drops the temp. (It would possibly be a good idea for power-on after fan-on delay)
Then drops steadily for about 10 mins to 166w 32.5GH 52c, error rate 7.8%
Normal practice would be to start up low and increase to the maximum stable state.
The chili starts at full throttle and then slows down. (you would never do this to a car).
Does any body else have any observations from their chilli?
The chili starts up with a big power draw, 200w+ and a hash rate of 38/39GH, temp at 75c with a massive error rate 15% (as if in an overclock condition)
If the chips are hot, there is a delay in startup, till the fan drops the temp. (It would possibly be a good idea for power-on after fan-on delay)
Then drops steadily for about 10 mins to 166w 32.5GH 52c, error rate 7.8%
Normal practice would be to start up low and increase to the maximum stable state.
The chili starts at full throttle and then slows down. (you would never do this to a car).
Does any body else have any observations from their chilli?
November 03, 2013, 08:13:30 AM
Hello Lucko,
If so do you know when the voltage must be switched exactly and which are the values?
Do you know where a documentation of the chips is available about this change? I've only BFL_SHA2_Spec_Rev2.2.pdf here. Maybe there is a newer one?
Maybe a resistor in line of the trimmer can be used to short them before or after initialization, but I need more details.
Cheers...
About power on... We guess if we use different power supply or a way to cut and then turn on power it would work... But in both cases we need to change the board... Chips just hang and will not work... At some point we did get 16 chips working but that was one time deal at minimal frequency and a lot of luck. In most cases you see 3 to 6 chips working if Atmel comes on line... And all 3 boards acts about the same so I don't think there is a point in guessing that there might be just a assembly problem. Boards were looked over by form in details...
I understood core voltage initialization of the chips must be different from the running core voltage?If so do you know when the voltage must be switched exactly and which are the values?
Do you know where a documentation of the chips is available about this change? I've only BFL_SHA2_Spec_Rev2.2.pdf here. Maybe there is a newer one?
Maybe a resistor in line of the trimmer can be used to short them before or after initialization, but I need more details.
Cheers...
November 03, 2013, 07:09:36 AM
I'd like to keep my orders with you for my 112 chips. PM me the details of the balance to be paid plz.
This is an unfortunate situation for both of us, and I place 100% of the blame on BFL. I mean, the purpose of sample chips is to have a solid reference to design against. The production chips are not supposed to be freakin' *DIFFERENT* in any way, but BFL went and changed them, and now it's costing both of us time and money.
Edit: So I guess there isn't any way to change the low-voltage-at-startup behavior of the existing boards, eh?
Can you please just fill up a form please...This is an unfortunate situation for both of us, and I place 100% of the blame on BFL. I mean, the purpose of sample chips is to have a solid reference to design against. The production chips are not supposed to be freakin' *DIFFERENT* in any way, but BFL went and changed them, and now it's costing both of us time and money.
Edit: So I guess there isn't any way to change the low-voltage-at-startup behavior of the existing boards, eh?
About power on... We guess if we use different power supply or a way to cut and then turn on power it would work... But in both cases we need to change the board... Chips just hang and will not work... At some point we did get 16 chips working but that was one time deal at minimal frequency and a lot of luck. In most cases you see 3 to 6 chips working if Atmel comes on line... And all 3 boards acts about the same so I don't think there is a point in guessing that there might be just a assembly problem. Boards were looked over by form in details...
I have an Accelero Twin Turbo 2 cooler on mine sitting on four pegs or feet to raise it up a little. It runs at 70C in a warm room. Mudbankkeith's Chili is running cooler but i think the target temp is 70C-72C (I could be wrong) and I would rather it be a little lower profile than his but I'm comfortable with whatever you decide.
Steady at 57C and 32GH. Super quiet cooling (it's house friendly) after that it is down to price and personal preference.
Edit:- Coolermaster Hyper212 Evo CPU Cooler with backplate
2nd edit:- Hyper212 is £ 10 cheaper than Accelero on Amazon
November 03, 2013, 05:59:07 AM
Hello,
you got allready 16 chips from me. I've ordered with your form now 10 boards where 2 should be configured with all 2x8 chips, the other 8 pcbs I want to get without chips but with all parts.
Cheers...
you got allready 16 chips from me. I've ordered with your form now 10 boards where 2 should be configured with all 2x8 chips, the other 8 pcbs I want to get without chips but with all parts.
Cheers...
November 02, 2013, 01:08:56 PM
I have an Accelero Twin Turbo 2 cooler on mine sitting on four pegs or feet to raise it up a little. It runs at 70C in a warm room. Mudbankkeith's Chili is running cooler but i think the target temp is 70C-72C (I could be wrong) and I would rather it be a little lower profile than his but I'm comfortable with whatever you decide.
Steady at 57C and 32GH. Super quiet cooling (it's house friendly) after that it is down to price and personal preference.
Edit:- Coolermaster Hyper212 Evo CPU Cooler with backplate
2nd edit:- Hyper212 is £ 10 cheaper than Accelero on Amazon
November 02, 2013, 12:37:52 PM
I have an Accelero Twin Turbo 2 cooler on mine sitting on four pegs or feet to raise it up a little. It runs at 70C in a warm room. Mudbankkeith's Chili is running cooler but i think the target temp is 70C-72C (I could be wrong) and I would rather it be a little lower profile than his but I'm comfortable with whatever you decide.
November 02, 2013, 12:26:19 PM
I'd like to keep my orders with you for my 112 chips. PM me the details of the balance to be paid plz.
This is an unfortunate situation for both of us, and I place 100% of the blame on BFL. I mean, the purpose of sample chips is to have a solid reference to design against. The production chips are not supposed to be freakin' *DIFFERENT* in any way, but BFL went and changed them, and now it's costing both of us time and money.
Edit: So I guess there isn't any way to change the low-voltage-at-startup behavior of the existing boards, eh?
This is an unfortunate situation for both of us, and I place 100% of the blame on BFL. I mean, the purpose of sample chips is to have a solid reference to design against. The production chips are not supposed to be freakin' *DIFFERENT* in any way, but BFL went and changed them, and now it's costing both of us time and money.
Edit: So I guess there isn't any way to change the low-voltage-at-startup behavior of the existing boards, eh?
November 02, 2013, 08:06:42 AM
Well it will require new board for sure to fix this. Changing firmware didn't help... Chips are definitely different. They have engine 0 working... So that is a bonus but unfortunately this chip change by BFL is not compatible with this board(not engine 0 itself but the way they act at low voltage at startup(we think)).
I think the only way is Chili since redesign would take just too long... So I will go with backup plan... If anyone doesn't like that let me know and I will send him chips and refund(if he paid already).
So for thous interested in Chili board it will be done the same way as it was done by MrTeal
https://bitcointalksearch.org/topic/the-chili-30ghs-bfl-based-bitcoin-miner-assembly-304250
Batch 1 will have price of 320$ for any new order and 300$ for already received and paid orders, groupbuy buyers and chips donors. After that there will be small batches and price will got up to 350$.
Since I lost massive amount of money on this I need at lest 50 paid boards orders to start. I don't have enough to start without that. Need to pay licencing and new batch of boards.
If I can't get that many orders I will just refund anyone and send you chips. Sorry...
Timeline. After I get 50 boards paid I will paid for licence and get files. After that I need 5 to 10 days to get new boards and max 5 days to get them assembled... After that I will need some time to program and test then. I was told it is about 10 minutes per board. So I will need some time to do that as well. Don't know how many boards will be but will be processed in a order of arrived orders. I will do my best to do all ASAP but it will take till 11/15 do 11/18 in best case if I get enough orders over the weekend.
How to order:
You fill out this order form:
https://docs.google.com/a/prorang.si/forms/d/1SeHqK95mbnxvvGqWmIjLM2L1nTFd5DSQDrfFwhWt3Zc/viewform
Batch 1 orders will be open till boards will be available or send to assembly. After that you will need to wait for Batch 2 boards.
Possible additional parts:
Thermal pads for chips. Enough for 1 board (5$). http://www.frozencpu.com/products/16875/thr-161/Fujipoly_Extreme_System_Builder_Thermal_Pad_-_Full_Sheet_-_300_x_200_x_05_-_Thermal_Conductivity_110_WmK.html
PCB PC BRACKET STD PCI (3$). http://eu.mouser.com/ProductDetail/Keystone-Electronics/9203/?qs=sGAEpiMZZMtzcnMBgC2bs6gC270mlsjAcmixu2SULFc=
Will add some coolers when I get the idea what is needed. I have already bought 1 Chili from MrTeal so I will be able to test that soon...
I think the only way is Chili since redesign would take just too long... So I will go with backup plan... If anyone doesn't like that let me know and I will send him chips and refund(if he paid already).
So for thous interested in Chili board it will be done the same way as it was done by MrTeal
https://bitcointalksearch.org/topic/the-chili-30ghs-bfl-based-bitcoin-miner-assembly-304250
Batch 1 will have price of 320$ for any new order and 300$ for already received and paid orders, groupbuy buyers and chips donors. After that there will be small batches and price will got up to 350$.
Since I lost massive amount of money on this I need at lest 50 paid boards orders to start. I don't have enough to start without that. Need to pay licencing and new batch of boards.
If I can't get that many orders I will just refund anyone and send you chips. Sorry...
Timeline. After I get 50 boards paid I will paid for licence and get files. After that I need 5 to 10 days to get new boards and max 5 days to get them assembled... After that I will need some time to program and test then. I was told it is about 10 minutes per board. So I will need some time to do that as well. Don't know how many boards will be but will be processed in a order of arrived orders. I will do my best to do all ASAP but it will take till 11/15 do 11/18 in best case if I get enough orders over the weekend.
How to order:
You fill out this order form:
https://docs.google.com/a/prorang.si/forms/d/1SeHqK95mbnxvvGqWmIjLM2L1nTFd5DSQDrfFwhWt3Zc/viewform
Batch 1 orders will be open till boards will be available or send to assembly. After that you will need to wait for Batch 2 boards.
Possible additional parts:
Thermal pads for chips. Enough for 1 board (5$). http://www.frozencpu.com/products/16875/thr-161/Fujipoly_Extreme_System_Builder_Thermal_Pad_-_Full_Sheet_-_300_x_200_x_05_-_Thermal_Conductivity_110_WmK.html
PCB PC BRACKET STD PCI (3$). http://eu.mouser.com/ProductDetail/Keystone-Electronics/9203/?qs=sGAEpiMZZMtzcnMBgC2bs6gC270mlsjAcmixu2SULFc=
Will add some coolers when I get the idea what is needed. I have already bought 1 Chili from MrTeal so I will be able to test that soon...
October 31, 2013, 05:28:59 PM
He is making some progress but noting good at the moment...
October 31, 2013, 10:09:51 AM
Hello lucko
Any news from Form?
Regards
Any news from Form?
Regards
October 30, 2013, 10:19:39 AM
Tom (JJ) has mailed a check with tracking # to reimburse me so now I can pay Lucko directly for assembly.
Tom explained that he invested a lot of his own money in this project and had no intention of scamming anyone.
Tom explained that he invested a lot of his own money in this project and had no intention of scamming anyone.
I got my full reimbursement check from JJ and he sent my chips to Lucko who is trying to get them out of Slovenia customs. I was the first person to get bulk chips so I've lost significant hash time but JJ is out significant investment in time and money for parts and Lucko and his team are trying hard to fix their boards. So just to be clear, there is no evidence of a scam here.
October 30, 2013, 12:11:06 AM
Tom (JJ) has mailed a check with tracking # to reimburse me so now I can pay Lucko directly for assembly.
Tom explained that he invested a lot of his own money in this project and had no intention of scamming anyone.
Tom explained that he invested a lot of his own money in this project and had no intention of scamming anyone.
I got my full reimbursement check from JJ and he sent my chips to Lucko who is trying to get them out of Slovenia customs. I was the first person to get bulk chips so I've lost significant hash time but JJ is out significant investment in time and money for parts and Lucko and his team are trying hard to fix their boards. So just to be clear, there is no evidence of a scam here.
October 29, 2013, 11:16:02 PM
Batch 2 news
Quote
I will have it emended tomorrow. Please accept my apologies for the error.
Kind regards,
Jonathon
Kind regards,
Jonathon
October 29, 2013, 02:06:09 PM
Batch 1 and Batch 3 chips are now at Ljubljana.
Is there any news on the batch 2 chips?
Yes but lack of documentation is making it difficult to get them out of customs...Is there any news on the batch 2 chips?
I got no replay from BFL about batch 2....
Did they make out of customs? Any news on batch 2?
I also resend mail(John) couple of times and calling number in mail(John). No replay or answer... But my BFL account shows batch 2 as not paid... I did send link on Bitpay to John that show it was...
Jump to: