GekkoScience Terminus R808 Miner Official Support Thread

[sidehack]

By the way - an update regarding changes to the Terminus sales process.

[mwin58747]

Somebody had asked how to set up more than one Terminus on a single power supply. Here is a pic of mine running on a 200 watt adjustable supply using the PCIe connectors (very solid connection). Testing with a small fan mounted over the Regulator on the right one to keep it cool.
237.50Mhz / 98-103GH/s each.
https://imgur.com/a/FPfMe

hello,

how are you getting more gh/s than me when i clock mine higher than yours? https://i.imgur.com/4NoB85b.jpg

is it just my luck on the with the chip draw or is it something of my doing?? thanks in advance

[sidehack]

They're getting choked out somewhere. That's about the speed one would expect from 200MHz. Try setting them at 200MHz and see if it stays about the same or goes down.

My guess is, you'd have to divide them amongst two controllers or something in order to get full throughput (and about 107GH) to all devices.

[QuarterMaster]

Somebody had asked how to set up more than one Terminus on a single power supply. Here is a pic of mine running on a 200 watt adjustable supply using the PCIe connectors (very solid connection). Testing with a small fan mounted over the Regulator on the right one to keep it cool.
237.50Mhz / 98-103GH/s each.
https://imgur.com/a/FPfMe

hello,

how are you getting more gh/s than me when i clock mine higher than yours? https://i.imgur.com/4NoB85b.jpg

is it just my luck on the with the chip draw or is it something of my doing?? thanks in advance

I did notice a slight increase when I switched to the PCIe connector instead of the barrel connector. More stable current to the chips maybe?
It also does fluctuate for me but not that low and the fact that all of  yours are around the the high 80GH/s to low 90GH/s would not mean "luck of the draw". I would try to separate just one miner on it's own system/controller to see what it ramps up to. That way you can rule out that it's the POD but maybe the controller.

[mwin58747]

They're getting choked out somewhere. That's about the speed one would expect from 200MHz. Try setting them at 200MHz and see if it stays about the same or goes down.

My guess is, you'd have to divide them amongst two controllers or something in order to get full throughput (and about 107GH) to all devices.

I have the six connected to a powered usb hub. i removed 4 now there are 2 connected to that hub and they seem to perform better. im guessing the hub's throughput is the bottleneck.

when i get home i will plug another 2 directly into the old laptop that i use as a controller and see whats going to happen. ill also try using 2 cheap unpowered usb hubs and see whats going to happen. (like the ones at '$2 shop' or something')

I did notice a slight increase when I switched to the PCIe connector instead of the barrel connector. More stable current to the chips maybe?
It also does fluctuate for me but not that low and the fact that all of  yours are around the the high 80GH/s to low 90GH/s would not mean "luck of the draw". I would try to separate just one miner on it's own system/controller to see what it ramps up to. That way you can rule out that it's the POD but maybe the controller.

i do run them via a PCIe connector. Corsair GS800 to be more specific, although i still dont know if i should buy of those kits from parallel miners that comes complete with the server psu, breakout board and cables. i mean whats the difference right? even if on paper it can deliver more watts to the pods, the pods themselves cant handle more maxing out at >250mhz (in a stable manner) :thinking:  aside from other things like longevity of the power supply and all that other stuff

[QuarterMaster]

Sounds like the right way to go and that power set up should be more than sufficient. I clocked at just under 60 watts each around that frequency. Let us know what you find and I have 2 plugged into a USB controller then the Windows server.

[QuarterMaster]

Question for Sidehack and I think I know the answer (regulator limit).

So the S5 with the BM1384 chips has a setting for the frequency and it goes all the way up to 500Mhz! I run one but very underclocked.  That must mean the 1384 chip can handle that with proper cooling and enough stable voltage/current. The terminus for me starts to become unstable at 250Mhz and I am guessing that is because of the voltage regulator. I mean that sucker gets hot at 225Mhz but if we had a way to cool it could we go higher or is there a specific limit to the current it will allow which is why the posted power table stops at 275Mhz.

Just curious is all.
Thank you

[Ecnad]

Just ordered a 5-pack of these for some mini-mining fun. Don't want to spend the electricity money or deal with the heat of a big miner, so these are perfect to just has away for a few cents a day.

[lok8nusa]

Stress test, 4 hours.  EVGA 600B psu. These things are awesome!
https://i.imgur.com/A4Pwbdd.png

[minerbobbert]

...
Failed to create restart thread errno=12 in cgminer.c restart_threads():5063
... It was up to being the 229th device in cgminer.  ...

It's unhealthy for that number to continually climb.   
The system will eventually prevent new resources from being allocated..

Are you referring here to the new GSE device number the miner is assigned? Of my three miners, two are tanking along just great and the third is steadily going up. I get about 30 in a day.

[Ecnad]

Just ordered a 5-pack of these for some mini-mining fun. Don't want to spend the electricity money or deal with the heat of a big miner, so these are perfect to just has away for a few cents a day.

And they arrived already!

3 of them hashing away at 200mhz - truly was plug and play with the recommended build of cgminer. Last 2 are waiting on more usb cables.

Using a 550w PSU I bought for a PC build that didn't happen; and a powered USB hub just to avoid any bottlenecks (Probably not a thing, but whatevs)

The pods are smaller than I expected (about a 4 inch by 4 inch cube) which is nice. You could fit quite a few of these on a shelf/rack. Mine are on a bakers rack so there is airflow below the boards as well.

[avulsionist]

Just got mine in the mail and playing with them at work before I go home and hook them into my setup.

2 Questions:
1. The Core Voltage Adjuster is full clockwise turn. Is that Minimum or Maximum?
2. Looking through the comments I don't see anything about where to measure the voltage. I found it for the 2pac's but can't figure it out for the R808


So far running off low end hub and a macbook at 200MHz with the supplied barrel power bricks I am getting a lot of resets back to 100MHz and recovery from Zombies but have to say, so nice to see the recovery!

Have to follow all the debugging in the beginning of the thread for one at a time, different hubs, direct connection but this is very exciting.

[sidehack]

It could be that the brick is garbage. If you can power it off a known reliable 12V source and the resets go away, let me know and I'll send you a replacement brick.

Let it be known, I'm pretty disappointed with the bricks I bought. 200MHz draws about 3.4 amps, and these are 6A bricks but still crap out to often. I'll be getting better ones in the future.

The voltage knob works just like a volume knob, or the knob on the 2Pac. Clockwise means higher.

[avulsionist]

It could be that the brick is garbage. If you can power it off a known reliable 12V source and the resets go away, let me know and I'll send you a replacement brick.

Will do.

The voltage knob works just like a volume knob, or the knob on the 2Pac. Clockwise means higher.

So full clockwise is how is arrived. Maybe i am starving them with default 200 MHz?

What is the spot to measure the voltage?

[sidehack]

"starving" what now? Whatever it is you're thinking is likely not the case. They're all tested at 200MHz stable for at least 1 hour before I ship them.

Each node has a small testpad for measuring voltage. The voltage between ground and Vc3 will give you the full string voltage, or you can measure across Vc pads to get node voltages.

[avulsionist]

Great thanks. Ill find the right pads.

Yeah I don't know much if anything about electoronics.
All conjecture here. Thinking that if I need to increase the voltage for higher MHz like on a 2Pac, then too much voltage with not enough MHz would also be bad.

[avulsionist]

Did some isolation testing with my two. Ran into issues with one.

Unit One just runs no problem or errors.
Unit Two has some issues. I used the same power and USB connection from unit one on unit two to make sure that was not it.

Steps:
Run unit One with no error for 4 hours.
shut down CGMiner and swap usb and power to Unit Two
Run CGMiner with no extras
cgminer -o stratum+tcp://us-east.stratum.slushpool.com:3333 -u avulsionist.terminus -p x

Results:

Code:

 [2018-03-30 11:47:54.833] Started cgminer 4.10.0
 [2018-03-30 11:47:54.836] Loaded configuration file /Users/bryankiechle/.cgminer/cgminer.conf
 [2018-03-30 11:47:55.737] GSE 0: Terminus BM1384 Bitcoin Miner (10030157)
 [2018-03-30 11:47:57.057] Probing for an alive pool
 [2018-03-30 11:47:57.167] Pool 2 difficulty changed to 4000
 [2018-03-30 11:47:57.222] Pool 2 message: Authorised, welcome to solo.ckpool.org 3B6LTrJmkEVhiY5p
i9TzmKf4Ah9dDoYcaH!
 [2018-03-30 11:47:57.223] Switching to pool 2 stratum+tcp://solo.ckpool.org:3333 - first alive po
ol
 [2018-03-30 11:47:57.342] GSE 0: setting frequency to 100.00MHz
 [2018-03-30 11:47:58.034] GSE 0: open cores @ 100.00MHz
 [2018-03-30 11:47:58.060] Network diff set to 3.46T
 [2018-03-30 11:48:06.309] Pool 0 difficulty changed to 8192
 [2018-03-30 11:48:09.357] Pool 0 stratum+tcp://us-east.stratum.slushpool.com:3333 alive, testing
stability
 [2018-03-30 11:48:09.357] Switching to pool 0 stratum+tcp://us-east.stratum.slushpool.com:3333
 [2018-03-30 11:48:10.083] Stratum from pool 0 requested work restart
 [2018-03-30 11:48:10.499] Pool 1 difficulty changed to 8192
 [2018-03-30 11:48:22.146] GSE 0: setting frequency to 200.00MHz
 [2018-03-30 11:48:22.845] GSE 0: open cores @ 200.00MHz
 [2018-03-30 11:48:33.266] GSE 0 failure, disabling!
 [2018-03-30 11:48:33.266] GSE 0: setting frequency to 100.00MHz
 [2018-03-30 11:48:38.261] GSE 1: Terminus BM1384 Bitcoin Miner (10030157)
 [2018-03-30 11:48:39.588] Hotplug: GekkoScience added GSE 1
 [2018-03-30 11:48:39.874] GSE 1: setting frequency to 100.00MHz
 [2018-03-30 11:48:40.548] GSE 1: open cores @ 100.00MHz
 [2018-03-30 11:48:57.709] Pool 0 difficulty changed to 1638
 [2018-03-30 11:49:02.701] Stratum from pool 0 requested work restart
 [2018-03-30 11:49:04.660] GSE 1: setting frequency to 200.00MHz
 [2018-03-30 11:49:05.362] GSE 1: open cores @ 200.00MHz
 [2018-03-30 11:49:15.698] GSE 1: start work @ 200.00MHz
 [2018-03-30 11:49:23.538] Accepted 261d0b1f Diff 1.72K/1638 GSE 1 pool 0
 [2018-03-30 11:49:27.583] Pool 0 difficulty changed to 512
 [2018-03-30 11:49:45.893] Stratum from pool 0 requested work restart
 [2018-03-30 11:49:46.119] Accepted 16f47483 Diff 2.85K/1638 GSE 1 pool 0
 [2018-03-30 11:50:00.709] GSE 1: asic(s) went offline
 [2018-03-30 11:50:00.734] GSE 1 failure, disabling!
 [2018-03-30 11:50:00.734] GSE 1: setting frequency to 100.00MHz
 [2018-03-30 11:50:00.908] GSE 2: Terminus BM1384 Bitcoin Miner (10030157)
 [2018-03-30 11:50:02.222] Hotplug: GekkoScience added GSE 2

Let me know if I can provide any other info.

[goose1072]

I having the same issue from one of the ones I just got back, It starts up at Zombie and stays there.

I have tried multiple computers / psu /usb cables to run it on and still no luck.

I ran the ones I got back on there own on different computer with there own psu, one is working beautifully the other not so much.

I'm going to have to sent this one back again.

[sidehack]

Check the note on it that tells you exactly how long I ran it during testing, after fixing the problem, with zero further issues.

I'm about ready to dynamite the whole Terminus project. I just spent an entire day fixing problems with boards that passed my initial bench test but then failed a burn-in test. I isolated the flaws, repaired them, then they passed another bench test. And freaking half of them failed the burn-in again. Literally, half. Well, 47% but only because I tested an odd number.

I don't know if the chips are garbage, or the assembly process is flawed, or what. I don't want to doubt the assembly process because it's the same process I've used to make 17,000 2Pacs with pretty fair reliability.

[goose1072]

Check the note on it that tells you exactly how long I ran it during testing, after fixing the problem, with zero further issues.

I'm about ready to dynamite the whole Terminus project. I just spent an entire day fixing problems with boards that passed my initial bench test but then failed a burn-in test. I isolated the flaws, repaired them, then they passed another bench test. And freaking half of them failed the burn-in again. Literally, half. Well, 47% but only because I tested an odd number.

I don't know if the chips are garbage, or the assembly process is flawed, or what. I don't want to doubt the assembly process because it's the same process I've used to make 17,000 2Pacs with pretty fair reliability.

I saw the note on it and thought everything would be resolved, the second one has been running since I got it back with no issues but you found something wrong with that one.
This is the one you couldn't find a problem with, for some reason I can't get it to work for me.