I am not sure how much thermal grease should be applied. If anyone knows I'm all ears!
Soldered boards:
Inspected boards:
Heat sink:
Thermal paste:
Thermal paste:
Assembling hash unit:
Thermal paste:
Lots of screws...
Topic: Avalon Clone Assembly Service Europe - page 4. (Read 20425 times)
Starting to assemble more modules...
I am not sure how much thermal grease should be applied. If anyone knows I'm all ears!
Soldered boards:
Inspected boards:
Heat sink:
Thermal paste:
Thermal paste:
Assembling hash unit:
Thermal paste:
Lots of screws...
I am not sure how much thermal grease should be applied. If anyone knows I'm all ears!
Soldered boards:
Inspected boards:
Heat sink:
Thermal paste:
Thermal paste:
Assembling hash unit:
Thermal paste:
Lots of screws...
loshia: Absolutely 
Soldered chip:
Soldered chip:
Since the chips arrived I have been working on the assembly process, I have had to work out quite a few different problems. I have slowly been finding a process that gives consistent quality, at first a lot of chips were not soldered properly and I had to do rework. Still going a bit slow, I will have to ship units in the order the chips arrive, so far I have assembled around 500 chips.
hacko.bg: I received chips from zefir first batch.
hacko.bg: I received chips from zefir first batch.
Nice:)
When done please PM us tracking numbers
Since the chips arrived I have been working on the assembly process, I have had to work out quite a few different problems. I have slowly been finding a process that gives consistent quality, at first a lot of chips were not soldered properly and I had to do rework. Still going a bit slow, I will have to ship units in the order the chips arrive, so far I have assembled around 500 chips.
hacko.bg: I received chips from zefir first batch.
hacko.bg: I received chips from zefir first batch.
I have a tracking number for the first chips on the way to me, it seems they have arrived in Sweden so I expect them to arrive very soon!
You are talking about Avalon chips, right? When did you place the order with Avalon?
I have a tracking number for the first chips on the way to me, it seems they have arrived in Sweden so I expect them to arrive very soon!
All right! These are looking great. I wish my chips would hurry up and ship...
Wonderful, truly wonderful...
I really find it amazing that you managed to make an exact Avalon clone.
And within just a few weeks time.
Respect and congrats
I really find it amazing that you managed to make an exact Avalon clone.
And within just a few weeks time.
Respect and congrats
And so it begins... shipping the first (unassembled) units
magnets99: All components from my first batch is sold out. We have ordered a second batch which will arrive in a few weeks. I will reply to you in a PM.
Hi strombom,
I would like to purchase a controller board from you.
Is this possible? I'm not in a huge hurry, my chips don't arrive until September.
I am trying to make a water cooled Avalon.
Thanks
I would like to purchase a controller board from you.
Is this possible? I'm not in a huge hurry, my chips don't arrive until September.
I am trying to make a water cooled Avalon.
Thanks
Here are the timing settings I used in 159b:
40:300
39:305
38:310
38:315
37:320
36:325
36:330
35:335
35:340
34:345
34:350
33:355
33:360
32:365
32:370
32:375
31:380
31:385
30:390
30:395
40:300
39:305
38:310
38:315
37:320
36:325
36:330
35:335
35:340
34:345
34:350
33:355
33:360
32:365
32:370
32:375
31:380
31:385
30:390
30:395
I think that hw errors are counted as real work .
It was related to cgminer
The error count increased from 1.3% from your version519b to 1.7% with version703 for higher clock >360
for 445 from 0.7% to 1.2%
The shares submitted with version 703 have a lower pool speed variance .(I`m now constant at 57k 57.5k with version 519b variance was +-5k on btcguild)
It was related to cgminer
The error count increased from 1.3% from your version519b to 1.7% with version703 for higher clock >360
for 445 from 0.7% to 1.2%
The shares submitted with version 703 have a lower pool speed variance .(I`m now constant at 57k 57.5k with version 519b variance was +-5k on btcguild)
Update on HW errors...
When I was getting 25% HW errors I was using this setting: --avalon-options 115200:24:10:32:375
When I use my usual firmware on 245 MH/s I don't get many HW errors on the new hash units:
When I was getting 25% HW errors I was using this setting: --avalon-options 115200:24:10:32:375
When I use my usual firmware on 245 MH/s I don't get many HW errors on the new hash units:
2013-07-10: Update, I have started to do test assembly of the mechanics, there are some issues, for example the screw threads for the router are misaligned, I have to fix this somehow. Overall it looks good though
Regarding the HW error rate, I have not looked into it. A week ago I compared the voltage ripple at 1.2 V on the original hash unit and my hash unit, I could not see any noticable differences. I will make a more serious test with heat compound and the real heat sinks and fans later on to see if there are any issues.
What is the definition of a HW error? How does the system determine there was a HW error? What are the sources that can produce HW errors? Where to learn more?
Regarding the HW error rate, I have not looked into it. A week ago I compared the voltage ripple at 1.2 V on the original hash unit and my hash unit, I could not see any noticable differences. I will make a more serious test with heat compound and the real heat sinks and fans later on to see if there are any issues.
What is the definition of a HW error? How does the system determine there was a HW error? What are the sources that can produce HW errors? Where to learn more?
Oh shit I didnt realize it was you Johan
sorry
lol I thought i was flaming a n00b
carry on, great work
sorry
lol I thought i was flaming a n00b
carry on, great work
2013-06-29: I was supposed to have done this yesterday but now I have tested a full board with 10 chips. It seems to work well, running both modules at 390MHz for a few minutes is a good sign.
These are horrible numbers you are currently generating 48% hardware errors.
What made you think 390mhz was stable on your current design?
HW% = 100 * HW / (diff1shares + HW)
100 * 294 / (922 + 294) -> 24%
Still not good tho, should be less than 2%. Must be because of high freq of 390!
What HW% do you get at 300Mhz ?
I base my number on a direct ratio between accepted and HW errors.
This is beyond just not good. This is a travesty. Might as well attempt to mine with a GPU with those numbers
2013-06-29: I was supposed to have done this yesterday but now I have tested a full board with 10 chips. It seems to work well, running both modules at 390MHz for a few minutes is a good sign.
These are horrible numbers you are currently generating 48% hardware errors.
What made you think 390mhz was stable on your current design?
PCB not does not fit exactly to heatsink, no screws. To hot.
Heatsink is not cooled, the lack of air flow.
Efficiency of the inverter circuit can be reduced, coil used different from original Avalon.
2013-06-29: I was supposed to have done this yesterday but now I have tested a full board with 10 chips. It seems to work well, running both modules at 390MHz for a few minutes is a good sign.
These are horrible numbers you are currently generating 48% hardware errors.
What made you think 390mhz was stable on your current design?
HW% = 100 * HW / (diff1shares + HW)
100 * 294 / (922 + 294) -> 24%
Still not good tho, should be less than 2%. Must be because of high freq of 390!
What HW% do you get at 300Mhz ?
Jump to: