S7 controller has nine 18- pin jacks So what can we do with it?

[philipma1957]

I never had 3 machines on hand.  So I only know the batch 1 does 6 blades  .

when you said new firmware only did 6 blades I assumed the old did the full 9 blades.


No one has posted an s-7 with more then six blades.

As far as I know  six can be done.  I would love to see a photo of 7, 8 or 9 working blades  on 1 controller.

[1713556095]

1713556095

[1713556095]

1713556095

[1713556095]

1713556095

[Tupsu]

One useful solution for S7 controller.

It works with S5+ software to control S5+ single cube / does not have control of the fan

It works with S5 software to control S5+ single cube / with  fan control

60% = 3,840 rpm
50% = 3,360- 3,480 rpm  (depends on the fan )

Working with ports 2, 3, 6  (left to right, viewed  by the end of the network cable side )

[RichBC]

One useful solution for S7 controller.

It works with S5+ software to control S5+ single cube / does not have control of the fan

It works with S5 software to control S5+ single cube / with  fan control

60% = 3,840 rpm
50% = 3,360- 3,480 rpm  (depends on the fan )

Working with ports 2, 3, 6  (left to right, viewed  by the end of the network cable side )

Not sure how useful it would be but... has anyone tried an S5 controller with the now released S7 Firmware. I have loaded it and it runs, but do not have an S7 to test?


Rich

[yslyung]

just got my batch 3 s7, took me abt 20 mins or so to set everything up & to figure some other stuff out. running good for now.

fan set at 60%. fan 1 4080 rpm - fan 3 3840 rpm

chip freq set at 606

temps - 59-60-61

so far it has been running great.

compared to stock fans are around 4000-4200 rpm

it's a 600M sticker model v1.8

temps are 2 deg C cooler

with new 606 on the chips, 0 hw error hashrate is barely noticeable but a tad more. 600 on the chips (stock) 0.0001% hw error

1 thing i do notice that the hash rate does fluctuate quite abit regardless of chip freq settings. same on miner & poolside (antpool)

have 2 more batch 5 coming & hope that i can hook all 3 up but so far only 2 is reported successful. (darn should've ordered 1 more batch 5) so i can run 2 controllers - 4 miners - 6 blades each

gotta pm j4b to get the cables, psu and more pcie cables. i saw somewhere some guy power only 2 of 3 per board here in bct forum but i saw outside the forum , only powered 1 out of 3 pcie & it's working all week long. initially had some issues but when he moved the the pcie to the center instead of the sides & has been working flawlessly.

thx for the info filipma (name too long to type :p) ... let's keep each other updated.

[philipma1957]

just got my batch 3 s7, took me abt 20 mins or so to set everything up & to figure some other stuff out. running good for now.

fan set at 60%. fan 1 4080 rpm - fan 3 3840 rpm

chip freq set at 606

temps - 59-60-61

so far it has been running great.

compared to stock fans are around 4000-4200 rpm

it's a 600M sticker model v1.8

temps are 2 deg C cooler

with new 606 on the chips, 0 hw error hashrate is barely noticeable but a tad more. 600 on the chips (stock) 0.0001% hw error

1 thing i do notice that the hash rate does fluctuate quite abit regardless of chip freq settings. same on miner & poolside (antpool)

have 2 more batch 5 coming & hope that i can hook all 3 up but so far only 2 is reported successful. (darn should've ordered 1 more batch 5) so i can run 2 controllers - 4 miners - 6 blades each

gotta pm j4b to get the cables, psu and more pcie cables. i saw somewhere some guy power only 2 of 3 per board here in bct forum but i saw outside the forum , only powered 1 out of 3 pcie & it's working all week long. initially had some issues but when he moved the the pcie to the center instead of the sides & has been working flawlessly.

thx for the info filipma (name too long to type :p) ... let's keep each other updated.

My batch 2 is working like a true stud.
Freq 606
Errors 0.0001%

4910 gh

And boards are labeled 575.

My batch 2 is better then my batch 1.

I have found this gear wants hot psu's. By that 12.1 volts rather then 11.94
I have two evga 1600 psu's one is a plat and gives me 11.94 volts under a 1200 watt draw.

The other is a tit. Okay a titanium it is giving 12.1 volts  to the batch 2.

I think j4b has a way to do 2000 watt psu with over volts.

Maybe the s-7 thrives on 12.1 to 12.4 volts

[RichBC]

My batch 2 is working like a true stud.
Freq 606
Errors 0.0001%

4910 gh

And boards are labeled 575.

My batch 2 is better then my batch 1.

I have found this gear wants hot psu's. By that 12.1 volts rather then 11.94
I have two evga 1600 psu's one is a plat and gives me 11.94 volts under a 1200 watt draw.

The other is a tit. Okay a titanium it is giving 12.1 volts  to the batch 2.

I think j4b has a way to do 2000 watt psu with over volts.

Maybe the s-7 thrives on 12.1 to 12.4 volts

I think that is a correct assessment, everything we have seen Bitnain doing with PSU spec 12V-12.6V, board selection, Frequecy reduction and finally reduced chain length points at the S7 wanting more volts.

A bit of good news is that server PSU's can more easily have their voltage adjusted up than down, by using the sense feedback voltage, usually at least 13V can be achieved. A lot of info on the RC forums as they always want more Volts.

Rich

[mcbiv]

I just hooked two units together via the rails and then unmounted both control boards and place a single bb board and data port board in between the two miners held in place with  a little velcro.  The original cables the S7 ships with are plenty long enough when set up this way...
[img=http://s27.postimg.org/z2bmhxs3z/IMG_0128.jpg]
[img=http://s27.postimg.org/3u31km2dr/IMG_0120.jpg]

It would seem to me you could lash together two S7's with a single controller if you were to lay the miners on their sides, so that the "Top Side" were facing each other. Kind of awkward to wire up, but it shouldn't take any "long cables" to accomplish.

It would be interesting to know if the S7 cooling actually requires that the blades be vertical to work OK. While it's obvious what the "expected" orientation is, does it still behave well if it's laid on it's side?

point them upwards top to top that allows for  vertical and hot air raises.

stack them on a box fan  which allows for more air to help fans.  also if fans fail the box fan may save the day.

[yslyung]

I just hooked two units together via the rails and then unmounted both control boards and place a single bb board and data port board in between the two miners held in place with  a little velcro.  The original cables the S7 ships with are plenty long enough when set up this way...
[img=http://s27.postimg.org/z2bmhxs3z/IMG_0128.jpg]
[img=http://s27.postimg.org/3u31km2dr/IMG_0120.jpg]

It would seem to me you could lash together two S7's with a single controller if you were to lay the miners on their sides, so that the "Top Side" were facing each other. Kind of awkward to wire up, but it shouldn't take any "long cables" to accomplish.

It would be interesting to know if the S7 cooling actually requires that the blades be vertical to work OK. While it's obvious what the "expected" orientation is, does it still behave well if it's laid on it's side?

point them upwards top to top that allows for  vertical and hot air raises.

stack them on a box fan  which allows for more air to help fans.  also if fans fail the box fan may save the day.

thats what i did, works great.

thx for sharing the info guys but i noticed they do run slower than individual controller

[philipma1957]

I just hooked two units together via the rails and then unmounted both control boards and place a single bb board and data port board in between the two miners held in place with  a little velcro.  The original cables the S7 ships with are plenty long enough when set up this way...
[img=http://s27.postimg.org/z2bmhxs3z/IMG_0128.jpg]
[img=http://s27.postimg.org/3u31km2dr/IMG_0120.jpg]

It would seem to me you could lash together two S7's with a single controller if you were to lay the miners on their sides, so that the "Top Side" were facing each other. Kind of awkward to wire up, but it shouldn't take any "long cables" to accomplish.

It would be interesting to know if the S7 cooling actually requires that the blades be vertical to work OK. While it's obvious what the "expected" orientation is, does it still behave well if it's laid on it's side?

point them upwards top to top that allows for  vertical and hot air raises.

stack them on a box fan  which allows for more air to help fans.  also if fans fail the box fan may save the day.

thats what i did, works great.

thx for sharing the info guys but i noticed they do run slower than individual controller

how much slower?  4800 + 4800 = 9600  

are you dropping to 9500 or 9400 or worse?

I theorize the controller is running 4 fans vs 2  so that the pcie runs volts say 12.0 vs 12.1  and thus lower speeds.

I also think 2 s-7s on two controllers give two sets of psu readings  so if one psu is 12.1 and the other is 12.0  one s-7 will run better.

but 1 controller uses the lower reading so it reads the 12 and both s-7s run lower.

[Finksy]

Phillip, wouldn't you think that the voltage at the chips (powered separately than the connector powering the controller) would have more of an impact on performance?  Maybe it has more to do with the amount of connections between all the chips and controller or something? I thought I remembered sidehack touching on the idea of communications between chips and controller and some limitations.

[yslyung]

I just hooked two units together via the rails and then unmounted both control boards and place a single bb board and data port board in between the two miners held in place with  a little velcro.  The original cables the S7 ships with are plenty long enough when set up this way...
[img=http://s27.postimg.org/z2bmhxs3z/IMG_0128.jpg]
[img=http://s27.postimg.org/3u31km2dr/IMG_0120.jpg]

It would seem to me you could lash together two S7's with a single controller if you were to lay the miners on their sides, so that the "Top Side" were facing each other. Kind of awkward to wire up, but it shouldn't take any "long cables" to accomplish.

It would be interesting to know if the S7 cooling actually requires that the blades be vertical to work OK. While it's obvious what the "expected" orientation is, does it still behave well if it's laid on it's side?

point them upwards top to top that allows for  vertical and hot air raises.

stack them on a box fan  which allows for more air to help fans.  also if fans fail the box fan may save the day.

thats what i did, works great.

thx for sharing the info guys but i noticed they do run slower than individual controller

how much slower?  4800 + 4800 = 9600   

are you dropping to 9500 or 9400 or worse?

I theorize the controller is running 4 fans vs 2  so that the pcie runs volts say 12.0 vs 12.1  and thus lower speeds.

I also think 2 s-7s on two controllers give two sets of psu readings  so if one psu is 12.1 and the other is 12.0  one s-7 will run better.

but 1 controller uses the lower reading so it reads the 12 and both s-7s run lower.

voltage is minimum of 12.2V & i might up it soon (12.5V) after modding psu's.

i can run 3 s7's with 2 x 2000bb psu but i'm running 1 x 2000cc per s7 so power is not the issue.

i think finsky might be right about connection between chips & controller ?

after i've modded psu voltage to 12.5V or more i will try again & see how it goes.

tests will be at stock speeds. 12.5V & 12.8V will be tested & see if there's any difference.

lucky for us, no need the long cables like the s5's to merge 2 in 1.

more ideas & suggestions, results are welcomed.

[vg54dett]

sorry for getting this old post up,

but after an thunderstorm , I got 2 controllers out of service on batch 9 S7s..

I don't have old S7 with controllers able to drive 2x3 blades , so I 'll have to buy new 2 new controllers (1 by damaged S7), or maybe 1 old controller able to drive the 2x S7s.

Would you guys advice me ? the best place to get some ? (cheap and fast ?) maybe another idea ? thank you !

[killerjoegreece]

i would love to know what i can do with these pin connectors

[sidehack]

Can you test if it's the IO board or the BeagleBoner itself that is messed up? I believe the 'boner is the same on the S7 as is was for the S5 and S2 (but someone who can say for sure should verify that). If the IO board is okay, I've got a stack of the 'boners laying around.

[Tupsu]

Can you test if it's the IO board or the BeagleBoner itself that is messed up? I believe the 'boner is the same on the S7 as is was for the S5 and S2 (but someone who can say for sure should verify that). If the IO board is okay, I've got a stack of the 'boners laying around.

You can use B.B.B. from S5 together with S7 controller.

Even S5 whole controller works with S7 software. (look at the picture S7 Batch8 2 blade with S5 controller)
 But as it turned out, it shows the wrong temperature. always around 46 C

Click on the image

[NotFuzzyWarm]

Can you test if it's the IO board or the BeagleBoner itself that is messed up? I believe the 'boner is the same on the S7 as is was for the S5 and S2 (but someone who can say for sure should verify that). If the IO board is okay, I've got a stack of the 'boners laying around.

Here's an off the wall thought for using PI's and Boners...
I'm currently playing with CryptoGlance for monitor software and in a nutshell, it sets up a Command/Control server on whatever PCish thing you have around. It's open source and loves Debian Linux. They have pre-built RasPi images for it. So...

How about cobbling together some some RasPi or BB plug& play cryptoGlance servers with them?

[sidehack]

You want some to play with? I'm already booked solid on stuff I know what it is. And, to be fair, some of them came from your miners anyways. I figured on putting up a sale post sometime to get rid of 'em anyway, I got about a dozen.

[NotFuzzyWarm]

You want some to play with? I'm already booked solid on stuff I know what it is. And, to be fair, some of them came from your miners anyways. I figured on putting up a sale post sometime to get rid of 'em anyway, I got about a dozen.

Heh heh.. I already have 3 new-in-the-box Pi's from the A1/AMT days that I've yet to do more than open and look at.... Not to mention that I still have 6 more s5's and a couple s3's, s4's and even 2 s2's kickin' around. And the Demon of Time is at work here as well.

However, the cryptoGlance/Pi seems rather appealing so I WILL try it on at least one of mine. As for the pulls you got - sell'em off. You deserve the money.

[sidehack]

Completely off topic, but "deserve" is one of my least favorite words.

[NotFuzzyWarm]

Completely off topic, but "deserve" is one of my least favorite words.

Sorry. How about 'earned'?