Try lowering frequency to 150mhz to see if that brings the chips back online. Then work your way back up to 200MHz.
Otherwise, ask for compensation from seller.
I'll do that. Also if I move around the board connectors to see if the issue follows the board or the connector is that ok to do? Yep, I was going to suggest that if playing with the clock speed didn't help. Ok so at 150 it's stable with no X's but still missing 3 Asic chips on that board. I'll keep adding speed till it fails then back off one. After that I'll swap board connectors. Thanks for the help I'd open her up and verify all data and power connectors to make sure nothing got loosened during transport. |
|
|
|
The problem is, they put the software that knows how to communicate with the cubes, on an encrypted chip sitting on the controller board. Most ASICs (not just miners) have controllers that communicate with the hardware much faster than a computer's available bus will allow.
Having the controller not be part of the host allows the titans to be independent and their efficiently won't be reliant on the host computer (raspberry pi in this case). The controller is definitely needed when dealing with these kinds of things.
If you compare and contrast with the BFL monarchs, they directly connect to your host PC, but rely on modified BFGminer binaries.
Each cube needs more bandwidth to the controller than USB 2.0 could provide to a PC? I'm surprised. So, I'm ignorant about this but... Is there no way to attack the FPGA to figure out how to communicate with the cubes? I had a EE friend of mine explain to me how hard it would be to sniff the signals off the ribbon cable and reverse engineer it. He said it would be much better to try and reverse engineer the software. I really, really, seriously, doubt that a cube would be too much even for USB 1.0. |
|
|
|
Let me know if you have any to sell. Looking for a few of them. Please identify the revision and date code when contacting me. For example:  0610 represents assembly during the 6th week of 2010. I'm looking for as late a build date / revision as possible. |
|
|
|
Try lowering frequency to 150mhz to see if that brings the chips back online. Then work your way back up to 200MHz.
Otherwise, ask for compensation from seller.
I'll do that. Also if I move around the board connectors to see if the issue follows the board or the connector is that ok to do? Yep, I was going to suggest that if playing with the clock speed didn't help. |
|
|
|
|
Try lowering frequency to 150mhz to see if that brings the chips back online. Then work your way back up to 200MHz.
Otherwise, ask for compensation from seller.
|
|
|
|
I've been experimenting too.  I didn't measure at the wall yet, but I did measure DC power coming from the PSU. Each blade is burning 500w, so 1500w / 5770gh = 0.259w/gh at the PSU. What's being pulled at the wall will depend on the quality of your PSU, obviously. I don't see why you're saying efficiency is bad, it's incredibly good considering it's running almost 1TH/s faster than stock. To compare, stock 600MHz each blade pulls 412w. 1236w / 4860gh = 0.254w/gh at the PSU. I'm overvolting the S7. PSU is outputting 12.58v unloaded, 12.48v loaded. running 600MHz = 12.4v at the blade, 737 = 12.33v at the blade. I don't suggest anyone attempt this. It's definitely not good for the S7, and you risk killing it. |
|
|
|
Thanks for the reply. Firmware is updated to the latest version and using different cgminer revisions yielded the same results. It will run all day without issue and the fans are working perfectly. I guess this S3 is just a little special! The 0 Best Share is a long standing bug with bitmain hardware. Best share wont show up if your share difficulty is a power of 2. Change your share difficulty and it'll start working. Try 500 instead of 512. |
|
|
|
hi guys ... anyone get batch 1 refund information ?  when oct still get a email reply will get soon... but now is dec still no any respond from bitmain  Don't hold your breath, you won't get it. Look into the S4 psu "compensation" promises from Bitmain. Yes it sucks....nothing but a broken promise Par for the course, unfortunately. I wish they'd just outright say "Too bad, it's late, you're SOL. Thank you for your business." instead of making promises they know they won't keep. |
|
|
|
S7 B7  Could you post a screen shot of a 24+ hour run time please? |
|
|
|
Those of you who need a bridge connector for your titan controller, i'll be having them manufactured and sent to me to test. If you need one it'll be 0.5 BTC.
I've increased the copper thickness on the bridge, so if you dont need a bridge but want your controller to last longer, than you'd still want one of these. PM me for more details.
A bridge connector? What is that? the only real piece someone can make between the raspberry pi and the knc ported board....supposedly it is a common port of failure of controllers I will likely get one ..just to have stuff around in case something goes boom on 1 of my 2 titan controllers....got a raspberry pi B+ I could use in a pinch ...getting stuff together for the re paste of a cube or two...with fans and heatsinks (Swedish guy mod on youtube do a search of knc titan he pops right up) glen trakin's firmware for the knc titan 3rd party stuff know how to make a replacement ribbon or what to do if a voltage plug goes 'pooey' on the cube proper to replace so with the bridge (comes out to about 80 bucks piece of mind per titan) wtf.....the only point of collapse after that point in time is the titan data port board itself..that folk have said would have to be reverse engineered anyway.... (i've spent 0.5 btc on 'dumber stuff" anyway my logic in for all this anyway (damn i'm an 'asic prepper' need to get a camo sticker to stick on my titan cubes...sorry they are 'evil' they will brick themselves will i be prepared  Yeah, Searing is pretty much correct. The bridge connector is the PCB connecting the RPI to the bottom Titan board. If you're using any other PSU than the one they recommend, that bridge will for sure burn out (6 burned out on me). I've designed some better ones with thicker copper junctions to stop the PCB from burning out when the PSU gives it a jolt. I think you've got another issue if you've got PSUs giving jolts that fry hardware, friend. I've measured power consumption of the entire Titan controller with 5 cubes and the LCD... 2.5w. Nowhere near enough to burn anything out. |
|
|
|
hi guys ... anyone get batch 1 refund information ? when oct still get a email reply will get soon... but now is dec still no any respond from bitmain Don't hold your breath, you won't get it. Look into the S4 psu "compensation" promises from Bitmain. |
|
|
|
wow, thats a sweet rig. sweeeeeeeeeeeeeeeeeeeeeeeeeet rig. Well done. Do you manage to pull out more than it costs in electricity ? I'm guessing yes. is it X11 or x13 that uses like 40 watts a card ? Thanks. I used to have about 60 7970s back in the scrypt glory days (late 2012 to mid 2013ish), needless to say that they are all paid off many times over. I've sold off about half of them, and now use the rest (29 gpus) to solo mine ethereum. It's worth it for me since I have cheap (~6c usd) power and already had all the required hardware. They were all sitting in a drawer waiting to be sold until ethereum came around. I wish they were consuming only 40 watts! Each card is burning 200 watts. Gotta say I had fun setting up my gpus again. ASIC are so boring! No skills required. |
|
|
|
Here's a better quality pic of my eth rig I posted last time:  |
|
|
|
|
I don't understand anyone who supports Chinese pools. If you want to mine on a big pool, why not move your power to Slush? It's 42 Ph/s and growing. I've always been happy there, and I'm mining there currently. Kano's pool is probably the best choice if you don't mind the variance that comes with a 2.5 Ph/s pool.
|
|
|
|
Anyone OCed his Batch 7 to 650 Mhz ? is it performing well ??
Playing around with different clock freqs. Here's a snapshot from my current test.  Looking at the RPM's on those fans, you must have them set around 30% manual. It looks like you need to bring up the fans to at least 80% to get those HW errors and temps down. I am in Switz. There's a communal law on the amount of noise one can generate esp. at night. Also the miner pulls cold air from the outside. Have to live with these temps. Here's the perf with 687MHz clock speed.  Very nice overclock! What power supply are you using, and what voltage are you measuring at the hash boards? I can't seem to get my units happy over 662. |
|
|
|
|
Closing in on 40Ph/s now. Thread's title should be updated to reflect this.
|
|
|
|
|
Another perfect S7 (B7) delivered to me today. Thanks, bitmain!
|
|
|
|
|
I have a few sets I don't need.
|
|
|
|
Colour change is pointless IMO, readability is what matters most and stock is perfect. No need to rice it out.
Updating bfgminer also seems like a waste of resources. Should only be updated if there's a specific fix for Titans in a release, such as decreasing slush time or whatever. If it ain't broken, don't fix it.
KNCMinion I haven't really used in the past, don't think I'll use it much in the future either. I vote remove if it makes your lives easier with releases.
My viewpoint exactly on bfgminer updates but ... for the sake of "keeping it updated so people think they are running the 'latest and greatest'" ... Ive always at least coded in the stock 5.1.0 (works best w/ titans) and whatever latest build there is for bfgminer. =) ... so far, Ima leaning towards keeping it like that. Juggling more versions then that around ... is chaotic and nothing I intend to do. KNCminion is a free thing(I dont have to change stuff w/ it when I make other webgui changes) , doesnt require any additional work for me to keep it in the webgui. So, Im probably not gonna worry about taking it out. Cool, I agree that the last official release from KNC (5.1.0) should always be a part of your firmware. Really no need for anything else. Having the option to use a newer version is just icing on the cake. Keep up the good work! |
|
|
|
|
Show Posts
