As for noise, it's pretty quiet. 80mm 160mA 12V case fan. I don't have a decibel reading but it's quite nonintrusive from more than a couple feet away.
Sound is negligible, quiet as hell and I even mounted a second fan on the back.
|
|
|
|
|
I built it with the potential for thermostatic fan control but didn't put it in the firmware because the fan's pretty quiet already so throttling it back wasn't necessary. Also the BM1387 run beautifully when cold and have a pretty high temperature coefficient so even at the bottom end the extra airflow is a benefit.
If someone ever wants to put in a jacked-up fan and have it throttle at low temps it would't take long to change the firmware. Most of the mechanism is already in there. You'd have to re-flash it yourself though, which will require some PIC ISP device.
|
|
|
|
|
Mostly for convenience. Right now we're all 10 on one tester for 7.5TH
Don't forget, these can be ordered direct for a slight discount. See the first post for contact info.
|
|
|
|
  Finishing and testing is going well. We'll start shipping on Friday. |
|
|
|
|
Cool. There's one running 544 instead of 550, and one that's down closer to 500 which I'm really disappointed with (could be a junk chip) but the rest are running well. My end says 10.4GH, but this isn't VH's bleeding-edge build. It's around a week old.
|
|
|
|
|
There are currently five R606 prototypes in the wild, so any of those guys could do a review. Most of them are already talking in this thread. When I ship gear to people it's because I want their honest opinions, not their noses in my butt. If it's got problems I want to know early so I can fix 'em.
I am looking forward to what Phil has to say. He's pretty thorough. If I had more prototypes to send out, or more time to do it, I probably would have given him one already.
I don't know yet when we'll open international sales, but apparently there's quite a waiting list. If we get all the first "batch" out and don't see a lot of trouble off the bat, we can open everything up. I just want to make sure there aren't issues with shipping, setup and overall reliability first, and by limiting sales to physically and politically contiguous locations with predominantly English speakers, it's much easier to handle exceptions.
Probably sometime in May.
|
|
|
|
|
Second fan? Why does everyone think these things need a second fan? Just because the friggin' S9 has two fans to clear 1.5KW from a case the size of an adult shoe doesn't mean everything needs two fans.
Seriously. The highest heatsink temperature I've yet marked was about 60C on full blast with the provided fan. Two fans is only really necessary when you need high pressure and this guy's heatsinks are nowhere near as constrictive as the ones used in the horrendous cooling designs we've been convinced to call "normal" now.
Looks like the lag right now is power bricks, so I may go ahead and ship pods so non-kit orders can be filled and then send bricks behind them as soon as they're arrived. Tired of waiting.
|
|
|
|
|
Best thing about this pre-order is we're not waiting on the miners, just the accessories. The PCBs are already built, the heatsinks are drilled out, the fans are here. Right now the only thing keeping me from shipping out miners on Monday is waiting for USB cables (broker sent the wrong ones last week), power bricks and the sweet logo decals (which are coming from a print shop across the street).
As for noise, it's pretty quiet. 80mm 160mA 12V case fan. I don't have a decibel reading but it's quite nonintrusive from more than a couple feet away.
Real-world power draw is going to depend on your power supply efficiency. I believe the bricks we have coming in are rated for 87% or so; 419mining is the go-between there so he can tell you the specs better. I spent about 10 hours the other day getting efficiency data from several units so once I finish charting the last ones I can put up some good info for different frequencies at different voltage settings, with margins for deviation and the like. But in general you're looking at a machine-level range from about 75 to 100W/TH depending on volt setting (and other factors like board temperature).
|
|
|
|
|
If the rest of the year's plans pan out, which mostly hinges on the success of the R606, I might finally get that S1/S3/S5 refit kit built. The S3 was pretty much the last full-scale miner designed to be quiet so hopefully a lot of them still exist to be upgraded.
I think a lot of gear is now coming with built-in PSUs because power supplies in the required power range aren't readily available at quantity, and if the PSU is integrated with the controller, board-level string voltages can be adjusted without an inline regulator which gives a 5-10% efficiency bump in the overall machine.
|
|
|
|
This would be nice, I never had one but heard when they worked they were great. Unfortunately it doesn't make sense for manufacturers to cater to the home miner anymore; if they did a run of something like this I imagine the price would be ridiculously expensive in Cost/TH.
At least one manufacturer caters exclusively to the home miner, but you're not wrong about the cost/TH being too high. Work in progress. |
|
|
|
Also, looks like I was mistaken about how many - or more specifically, the total expected hashrate - which can be run off one hub. Looks like frames to the hub can combine multiple device frames, and one frame can hold six work packets, so the peak theoretical hashrate for hub'd miners would be six times the peak theoretical for a single connection, or roughly 15TH. Of course I'll be testing this as soon as I have enough devices. And then I'm gonna giggle like a little girl and post awesome screenshots. Until then, y'all will have to be satisfied with this magnificent six-hour runtime of six pods on one GekkoScience Base Hub and powered by a GekkoScience DPS1200 PSU kit.  |
|
|
|
Tester 3, all plugged into PC via single Plugable 7-port hub, volt setting 7 [2019-04-18 07:31:24.324]
Summary of runtime statistics:
[2019-04-18 07:31:24.324] Started at [2019-04-17 10:22:34.010]
[2019-04-18 07:31:24.324] Pool: stratum+tcp://pool.ckpool.org:3333
[2019-04-18 07:31:24.324] Runtime: 21 hrs : 8 mins : 50 secs
[2019-04-18 07:31:24.324] Average hashrate: 3346651.5 Mhash/s
[2019-04-18 07:31:24.324] Solved blocks: 0
[2019-04-18 07:31:24.324] Best share difficulty: 74.5M
[2019-04-18 07:31:24.324] Share submissions: 30384
[2019-04-18 07:31:24.324] Accepted shares: 30323
[2019-04-18 07:31:24.324] Rejected shares: 61
[2019-04-18 07:31:24.324] Accepted difficulty shares: 57910187
[2019-04-18 07:31:24.324] Rejected difficulty shares: 45241
[2019-04-18 07:31:24.324] Reject ratio: 0.2%
[2019-04-18 07:31:24.324] Hardware errors: 12345
[2019-04-18 07:31:24.324] Utility (accepted shares / min): 23.90/min
[2019-04-18 07:31:24.324] Work Utility (diff1 shares solved / min): 45287.42/min
[2019-04-18 07:31:24.324] Stale submissions discarded due to new blocks: 0
[2019-04-18 07:31:24.324] Unable to get work from server occasions: 0
[2019-04-18 07:31:24.324] Work items generated locally: 53185659
[2019-04-18 07:31:24.324] Submitting work remotely delay occasions: 0
[2019-04-18 07:31:24.324] New blocks detected on network: 123
[2019-04-18 07:31:24.324] Summary of per device statistics:
[2019-04-18 07:31:24.324] GSH 0 (5s):1.023T (avg):1.076Th/s | A:18679068 R:16773 HW:3506 WU:145
[2019-04-18 07:31:24.324] GSH 1 (5s):1.151T (avg):1.177Th/s | A:20294999 R:13634 HW:5142 WU:159
[2019-04-18 07:31:24.324] GSH 2 (5s):1.095T (avg):1.092Th/s | A:18936120 R:14834 HW:3697 WU:147 Tester 4, all plugged directly into PC USB ports, volt setting 7 Summary of runtime statistics:
[2019-04-18 11:26:52.780] Started at [2019-04-17 16:46:25.805]
[2019-04-18 11:26:52.780] Pool: stratum+tcp://pool.ckpool.org:3333
[2019-04-18 11:26:52.780] Runtime: 18 hrs : 40 mins : 26 secs
[2019-04-18 11:26:52.780] Average hashrate: 3348909.2 Mhash/s
[2019-04-18 11:26:52.780] Solved blocks: 0
[2019-04-18 11:26:52.780] Best share difficulty: 37.7M
[2019-04-18 11:26:52.780] Share submissions: 24117
[2019-04-18 11:26:52.780] Accepted shares: 24073
[2019-04-18 11:26:52.780] Rejected shares: 44
[2019-04-18 11:26:52.780] Accepted difficulty shares: 51235035
[2019-04-18 11:26:52.780] Rejected difficulty shares: 26090
[2019-04-18 11:26:52.780] Reject ratio: 0.2%
[2019-04-18 11:26:52.780] Hardware errors: 10830
[2019-04-18 11:26:52.780] Utility (accepted shares / min): 21.49/min
[2019-04-18 11:26:52.780] Work Utility (diff1 shares solved / min): 45465.61/min
[2019-04-18 11:26:52.780] Stale submissions discarded due to new blocks: 0
[2019-04-18 11:26:52.780] Unable to get work from server occasions: 0
[2019-04-18 11:26:52.780] Work items generated locally: 47074080
[2019-04-18 11:26:52.780] Submitting work remotely delay occasions: 0
[2019-04-18 11:26:52.780] New blocks detected on network: 111
[2019-04-18 11:26:52.780] Summary of per device statistics:
[2019-04-18 11:26:52.780] GSH 0 (5s):1.158T (avg):1.105Th/s | A:16758662 R:7586 HW:3354 WU:1500
[2019-04-18 11:26:52.780] GSH 1 (5s):1.097T (avg):1.146Th/s | A:17689480 R:10536 HW:4074 WU:155
[2019-04-18 11:26:52.780] GSH 2 (5s):1.074T (avg):1.097Th/s | A:16786893 R:7968 HW:3402 WU:1490 I'm re-running them now, all still at V7 and plugged into Tester 3 via a single GS 7-port hub. Should let us know what's the practical max. |
|
|
|
|
6 NewPacs is 12 ASICs, like the R606. However the pod doesn't require a high-powered USB hub, or an external fan, or a screwdriver and meter for voltage adjustment, and the main regulator is rated for higher current which means the potential for more speed.
|
|
|
|
|
Could be worthwhile to get some outside verifying. My bench test station already points to kano (you can see ongoing testing under the 1BURGER address, worker _r606) but the main tester scripts got switched to ckpool sometime back when we were having connection issues to kano servers (my ISP's upstream provider is CenturyLink and they suck balls forever).
|
|
|
|
Just grabbed a screenshot from one of the testers. I pulled the first six production pods to run them through the wringer and get statistics on max hashrate per voltage setting so we can get a better idea of performance and deviations. This is the ones on Tester 3 (connected through a Plugable 7-port hub), voltage setting 7 (max), on VH's latest r606 branch build, powered via 6-pin from a DPS1200 PSU and one of our breakout boards (the PSU is comfortably powering 7 pods off 120V right now) Note the frequencies and hashrates listed are the absolute max for these things.  I'll have the final numbers on Tester 4 in a couple hours, and then we'll re-run the tests with a new batch of 6 for comparison. |
|
|
|
you have now got a buyer of your products for life! keep it up!
Funny you should mention that. *ahem*
GekkoScience Terminus R606 up to 1THAnyways hey, glad stuff is working for you. The --gekko-serial flag VH put in there is really handy for isolating sticks. As kano mentioned, cgminer already had stuff for separating out items on the USB bus but the serial flag simplifies things, though it only works for GekkoScience hardware. Course anymore, that's about all there is. Other than jstefanop's Litecoin gear and some Chinese knockoffs of my designs (some of which are legit fire hazards), there haven't been many non-Gekko USB miners made since about 2014. |
|
|
|
|
Went straight from the ground in front of the icy staircase to the hospital. Fortunately I hadn't eaten anything in about 20 hours anyway so they operated that evening to install a metal rod in what became of the tibia (the fibula chunks, since it doesn't carrying much of a load, were left to fend for themselves). It was technically load-bearing from the get-go but asking one screw driven through a bone to carry my not-unsubstantial mass without extreme pain, and asking the same of the knee they opened up to do the install, was a little much. I didn't drive for a full month, and didn't go from crutches to a cane until a few weeks ago. Now I can walk around relatively well unassisted, at least for a while. Did not have a cast at all, interestingly enough.
It's a good thing my TV was already tied to a PC because I did almost all of the PCB design for this pod from the sofa during the first few weeks. All in all, this project has been a lot of work, but a lot of fun work, and I'm really excited about the finished product. It's easily the best thing I've ever built and VH's code work has been top-notch as well.
|
|
|
|
New build (7a48c8f)
...
Add new testing parameter --gekko-serial
Accepts double quoted, comma seperated list of Serial Numbers to target.
Available for all gekko miner types. Especially the 2pac type.
--gekko-serial "GS-10001010,GS-10001011"
|
|
|
|
|
Show Posts
