Fury/Blizzard tuning and mods

[SoreGums]

4-5%  is a normal error rate for a fury not 9%
9% you need to improve the cooling.
@381mhz i have 1.6Mh/s and less than 5% and very stable.

Fresh air is the key.

Yes, in a air-cooling room error rate drops

I'm sitting at <1% error rate, my room is cool as it is winter, no need to put the heater on, window left open. The Fury/Blizzards are not even warm, they are not as cold when turned off, still cool though. Temp outside is 11°c apparently, it is a bit cold in the room this morning :p

[1714025675]

1714025675

[Zeta0S]

4-5%  is a normal error rate for a fury not 9%
9% you need to improve the cooling.
@381mhz i have 1.6Mh/s and less than 5% and very stable.

Fresh air is the key.

Yes, in a air-cooling room error rate drops

I'm sitting at <1% error rate, my room is cool as it is winter, no need to put the heater on, window left open. The Fury/Blizzards are not even warm, they are not as cold when turned off, still cool though. Temp outside is 11°c apparently, it is a bit cold in the room this morning :p

My room temp is around 28°c at night and 35 day time

[bosondehiggs]

25°C in my room.

[Honeycutt22]

Hey guys - I have two blizzards (1.4mh each) running on fastcoin. I seem to be getting a pretty high rejection and stale rate with the following specs...

cgminer.exe --scrypt -o 127.**** -u u -p p --nocheck-golden --chips-count 6 --ltc-clk 328 -S //./COM3 -S //./COM6 --expiry 2 --scan-time 2 --queue 0

Fastcoin has a 12 s block time, so maybe it's normal because it goes so fast???

Does anyone know? Also - I am solo mining them, but had the same issues on the pools as well

Just bumpin this. Have not gotten any responses yet:(

[suchmoon]

Hey guys - I have two blizzards (1.4mh each) running on fastcoin. I seem to be getting a pretty high rejection and stale rate with the following specs...

cgminer.exe --scrypt -o 127.**** -u u -p p --nocheck-golden --chips-count 6 --ltc-clk 328 -S //./COM3 -S //./COM6 --expiry 2 --scan-time 2 --queue 0

Fastcoin has a 12 s block time, so maybe it's normal because it goes so fast???

Does anyone know? Also - I am solo mining them, but had the same issues on the pools as well

Just bumpin this. Have not gotten any responses yet:(

You didn't mention what you consider "high", but technically faster coins will have higher reject rates, so this is likely normal.

[jekv2]

I ordered

http://www.newegg.com/Product/Product.aspx?Item=N82E16803003418

&

http://www.newegg.com/Product/Product.aspx?Item=N82E16835100018

I will DIY hs for the voltage regulators.

I am going to take the HS's off the chips reapply with Céramique 2.

My pops has a drill press, we need to hook into the breaker box, thing is a old dark green bastard lol.

Then, ima drill press a 3" hole, unlike this but with 3" instead and use existing bolts and nuts, might have to take a few washers off to make it by a nich and place a 80mm fan. Gonna be some fun. Will use a dremel tool and tornado stone shaped bit to debur the crap off.

https://bitcointalk.org/index.php?topic=644921.msg7354916#msg7354916

[Honeycutt22]

Hey guys - I have two blizzards (1.4mh each) running on fastcoin. I seem to be getting a pretty high rejection and stale rate with the following specs...

cgminer.exe --scrypt -o 127.**** -u u -p p --nocheck-golden --chips-count 6 --ltc-clk 328 -S //./COM3 -S //./COM6 --expiry 2 --scan-time 2 --queue 0

Fastcoin has a 12 s block time, so maybe it's normal because it goes so fast???

Does anyone know? Also - I am solo mining them, but had the same issues on the pools as well

Just bumpin this. Have not gotten any responses yet:(

You didn't mention what you consider "high", but technically faster coins will have higher reject rates, so this is likely normal.

Sorry! By high I mean probably 40-50%.. I'm getting half the expected pay from the calculator.

[Ktulu1]

4-5%  is a normal error rate for a fury not 9%
9% you need to improve the cooling.
@381mhz i have 1.6Mh/s and less than 5% and very stable.

Fresh air is the key.

Exactly right. If you're running above about 350 and getting that many errors (9%), I would guess that you're running out of juice. Need more volts.

[Ktulu1]

4-5%  is a normal error rate for a fury not 9%
9% you need to improve the cooling.
@381mhz i have 1.6Mh/s and less than 5% and very stable.

Fresh air is the key.

Yes, in a air-cooling room error rate drops

I'm sitting at <1% error rate, my room is cool as it is winter, no need to put the heater on, window left open. The Fury/Blizzards are not even warm, they are not as cold when turned off, still cool though. Temp outside is 11°c apparently, it is a bit cold in the room this morning :p

https://i.imgur.com/QMddcgOs.png

I think you're calculating hardware error percent by dividing by diff1 instead of your accepted shares at diff. That version of cgminer shows accepted shares as diff1 instead of the difficulty your submitting them at.

Looking at your screen clips, the one on the left for example, you have a submitted diff1 of 1,595,368 and your worker diff is 512. 1,595,368/512=3,115.953125 (not a whole number because not all your accepted shares were at 512). Anyway... 218 (HW Errors)/3,116 = ~7% a much more reasonable number.

The 5% and 9% numbers quoted here were calculated using accepted shares at diff, not at diff1.

[Ruledo]

Btw, for voltmod...
Took both nst and j4b's advice.
Installed a 10K trimmer pot to replace the resistor and set at 7.5K.
Getting 1.68MH dev side and 1.5-2.0MH pool side with < 5-6% HW errors.
Too scared to measure watts though, because it may not be worth it and I had so much fun doing it.

[b3da]

Hey guys - I have two blizzards (1.4mh each) running on fastcoin. I seem to be getting a pretty high rejection and stale rate with the following specs...

cgminer.exe --scrypt -o 127.**** -u u -p p --nocheck-golden --chips-count 6 --ltc-clk 328 -S //./COM3 -S //./COM6 --expiry 2 --scan-time 2 --queue 0

Fastcoin has a 12 s block time, so maybe it's normal because it goes so fast???

Does anyone know? Also - I am solo mining them, but had the same issues on the pools as well

Just bumpin this. Have not gotten any responses yet:(

You didn't mention what you consider "high", but technically faster coins will have higher reject rates, so this is likely normal.

Sorry! By high I mean probably 40-50%.. I'm getting half the expected pay from the calculator.

from your config line i see that you are mining fst solo..
when you are mining solo, accepted are found blocks, rejected are orphans..
so, with 2 furys and diff 0.48, i think that 40-50% reject is good.. keep in mind that you have now about 2% of fst network hashrate.
maybe you will do better to try pool..
and try cgminer 4.3.5 - there are issues with 3.1.1 and low diff coins and some wall of rejects.

[jekv2]

4-5%  is a normal error rate for a fury not 9%
9% you need to improve the cooling.
@381mhz i have 1.6Mh/s and less than 5% and very stable.

Fresh air is the key.

Exactly right. If you're running above about 350 and getting that many errors (9%), I would guess that you're running out of juice. Need more volts.

I am running this off of a EVGA 750 G2. There is no way I am running out of power.

I was getting 9% at like I said above previous page, 939,950,954 i even tried 328 still got 9% HW error rate.

When ever my hole saw comes in I am tearing it apart cutting 3" hole slapping 80MM fan on, completely cleaning it, reapplying with better TIM etc, wait 30 hours on then power off. Do heat cycle thing. blah. we'll see I will report back.

[Borsht]

4-5%  is a normal error rate for a fury not 9%
9% you need to improve the cooling.
@381mhz i have 1.6Mh/s and less than 5% and very stable.

Fresh air is the key.

Exactly right. If you're running above about 350 and getting that many errors (9%), I would guess that you're running out of juice. Need more volts.

i added heat sinks to he ASICs and better thermal compound and found that the hash rate, errors, and rejects didn't show any improvement under 350. over 350 the ASICs are starving for power.

i just posted the wright up @ http://www.techunboxed.com/2014/07/gaw-miners-fury-power-supply-mod.html

[alienesb]

4-5%  is a normal error rate for a fury not 9%
9% you need to improve the cooling.
@381mhz i have 1.6Mh/s and less than 5% and very stable.

Fresh air is the key.

Exactly right. If you're running above about 350 and getting that many errors (9%), I would guess that you're running out of juice. Need more volts.

i added heat sinks to he ASICs and better thermal compound and found that the hash rate, errors, and rejects didn't show any improvement under 350. over 350 the ASICs are starving for power.

i just posted the wright up @ http://www.techunboxed.com/2014/07/gaw-miners-fury-power-supply-mod.html

Yeah I added heatsinks and better compound and found the same thing - no change at all.

[jekv2]

SO we have a few debates here?

One is 9% is normal.
Second is 5% is normal with cooling mod.

One. With cooling mod nothing changes.
Two, Without cooling mod nothing changes.

This is absurd.

I am sorry but who the fuck is getting 5% error rates without a volt mod?

[b3da]

with new tim, hs on chips and with another 80mm fan blowing from side, im getting 4-6% HW depending on changing room temp. (its better at night )
without voltmod, clk 350.. week 3 furies.

[Borsht]

stock under 350 - 5-7% errors is quite normal.

heat sinks and paste under 350 - 5-7% seems to be the norm as well.

obviously there are going to be outliers.

350 is an arbitrary number. i haven't tested 351, etc., but 354 shows more errors... around 9% with or without improved cooling.

[jekv2]

edit:
For the TIM I used was ceramique 2, 25 gram and used a spreader, spreaded nice and neatly on the back of the chip board.
edit:/

Whohoo.

You may add this to the cooling mod of the fury,

Same concept as happy daze.

I took a metal ruler aligned it crisscross of the 80MM fan holes diagonaly X marks the center with a scratchall"center punch will scratch as long as it has a point, its only aluminum"..  
I took and put the HS case in a vice with wood protecting the edges then I took 3" hole saw drilled with oil with a 19.2 cordless craftsman drill. I could not believe the arbor thing fit in the cordless but it did, thing is a monster "Milwaukee arbor with dewalt 3" hole saw"
I had to cut the washers off the existing fan bolts for the bolts to meet the nuts to hold the second 80mm fan on.

I am getting very good air flow now.

I could imagine you could if went to the hardware store get longer bolts and do a back to back inflow/outflow for two furys.

Well here are the results. Pretty.

https://imgur.com/a/N3oJl

[happydaze]

Nice job jekv2!  I know how hard it was to do mine.  You have good skills to keep it so professional looking.

[jekv2]

Nice job jekv2!  I know how hard it was to do mine.  You have good skills to keep it so professional looking.

Thanks, daze.

Man what I do love is how this C2 spreads. Like warm butter.