|
This is the reason why im trying to put more value to mining than the dollar cost of what your mining vs electricity.
Mining was supposed to be a distribution method for crypto that anyone could partake in, not the centralized manipulated space it is now.
What needs to be done is actually counter intuitive (and what im trying to do with my miners). Get small ASICs in the hands of so many people that the hash power of that group of people > the hash power of the small centralized elite.
What happens then? People running 100 watt crypto platforms on there desk don't care about what it costs to run one lightbulb 24/7 because its negligible. The value of running a full hardware and software stack is much greater than the pure mining rewards for an individual that cares about crypto and its future. The large mining farms then start becoming unprofitable since all they see in crypto is pure profit and greed, and will start shutting down.
Thats the idealistic place I hope crypto gets to in the future.
|
|
|
|
With the case on and no heatsinks the Apollo stopped hashing at high temperatures (40 degrees C ambient).
A restart fixed the issue, but this happened to me a few times in ECO mode with the fan on auto.
I think the fan could not cool the Apollo enough at high temperatures as it is only rated to run up to 40 degrees C ambient.
Have you installed the latest firmware image? You were probably experiencing the MCU overheat issue, and the latest image fixes this issue. If you want to make sure your MCU does not overheat pull it out of its socket and put a heatsink on the ARM processor (the big black chip on the other side of the blue board). Thats probably the best use of all those heatsinks  The main heatsink/fan for the ASICs is way overbuilt. They can both dissipate over 300 watts of heat at max fan airflow. |
|
|
|
I have pimped up the Apollo with lots of heatsinks to reduce the noise further (fan runs below 1620rpm at 30 degrees C ambient) and I run it without the case. As a minimum I suggest adding a heatsink to the R44 power converter at it is quite hot (>60 degrees C). https://imgur.com/a/SB6A4puI use the following heatsinks: 20 off 20x20x10mm 11 off 8x30x8mm 2 off 6x4x4mm 1 off 14x14x5mm The 8x30x8mm heatsinks probably help cool the Apollo down the most as they are at the base of the main heatsink. Smaller 20x20x5mm heatsinks would probably work just as well and be cheaper at the top of the main heatsink by the fan. R44 is the inductor to the main power stage. This is designed to run a 90C+ ...60C for this part of the power stage is pretty cool. There is no need to put heatsinks on it, if anything the smaller heatsink right behind it probably helps the most, since thats the main MOSFET block and produces the most heat. These are rated for 125C but cooling them down helps with power efficiency and longevity of the mosfets..but again this is not needed. |
|
|
|
@jstefanop when can we expect the node and wallet function for the apollo ?
Hoping to have at least the full node feature released in the next month or two. Made a big mistake in hindsight and ordered the 256MB RAM version of the MCU instead of the 512, so trying to see if we can get the node to run stable with just 256MB of RAM, or if we need to swap the MCUs out. |
|
|
|
Guess ill have to point over some of my test Apollos here |
|
|
|
Got my Apollo today, however I keep getting connection issues. Not sure what to check. It was up and working for a minute or two in the beginning now I can't seem to get the connection to work. I can get to the IP address but it keeps saying that the miner isn't connecting.
Error: connect ECONNREFUSED 127.0.0.1:4028
This means you put in your pool credentials in wrong. |
|
|
|
I just got my Apollo with pre-burned image flash card installed. Plugged both PCIE power connectors in and LAN cable..
Just getting flashing yellow light on the front and don't see a new device on the LAN.
Any ideas for trouble shooting?
Thanks
Is the front yellow flashing slowly or fast? Do the leds around the ethernet connector show any signs? Is there a dhcp server in the LAN? How long have you waited for it to start? It's a fast flash. The LEDS come on initially then turn off. Yes there is a DHCP server on the network. Waited plenty Thats a network issue, either your local LAN is not providing an IP address or it cant connect to the pre-configured pool. Have you tried logging into your router and checking whether you see a new IP/Device show up there? |
|
|
|
This looks like a great form factor for home mining!
Do I understand correctly that the PCB houses 3x9 chips underneath that big heatsink?
I wonder how you lay out 27 chips in a square space...
Did you do the ASIC design in-house?
Any plans to branch out to other PoWs in future?
Several other memory oriented PoW have daily issuance comparable to scrypt,
like Equihash and Cuckoo Cycle.
Yes its 27 ASICs in a 3x9 Parallel block/serial configuration. It wasn't easy that's for sure. We are exploring whether we want to concentrate on scrypt or branch out to more algos so don't have anything to announce now. |
|
|
|
|
Check out my Apollo miner in my sig which was released a few months ago. It might not be the fastest, but its currently the most efficient scrypt miner out ~1 W/MH.
|
|
|
|
Hey everyone,
The fan on one of my moonlanders gave out a few days ago. I noticed that a few people have posted links to replacement fans earlier on this thread. I noticed however that these replacement fans do not come with a connector. That is, they come with bare wires on them. Would you happen to know where I can get the appropriate connector or where I can find a replacement fan with the correct connector?
Thanks!
You can message FutureBit on Facebook Messager and order fans that way. Thanks! I'll give that a try. I keep putting of listing replacement fans on shop.futurebit.io...ill try and put them up this week. |
|
|
|
Power question:
The original data posted on Futurebit says "Two Six Pin PCIE power connectors for wide-range of power draw (only one is needed in ECO mode)". I've been running in "Balanced" mode with just one PCIE connector used. I just felt the power wires at the Apollo back and they are a little warm, not hot at all. I'd like to leave this running in Balanced mode while I'm away a few days - should I worry? Do both power connectors need to be used for Balanced mode?
My PSU is very adequate (Raidmax Cobra RX-700AC-B 700W 80 PLUS) - runs quiet and cold. Anyone else running in Balanced mode with just 1 power connector?
I'd like to save the 2nd PCIE output for my 2nd Apollo.
As long as the wire is not hot it should be fine, but it comes down to the wire gauge used by that PSU, which I assume will probably be 18 gauge. Balance uses about ~150 watts, which is pushing it for a single 6pin PCIE. I personally would not run it on one cable unless its at least 16 gauge, and even then it comes down to safety and it will always be safer to run it on two. I obviously cant recommend anything other than having both plugged in but im sure others can chime in |
|
|
|
HELP! OK, with that out of the way. One of my Apollos seems to be having issues. bfg starts and then stops. I have replaced the SD card and removed all pools, twice. With just the donation pool looking at the sudo screen -dr miner I see bfgminer 5.5.0-41-gb0cd450 - Start: [2019-02-27 00:02:21] - [ 0 days 00:00:02]
[M]anage devices [P]ool management [S]ettings [D]isplay options [H]elp [Q]uit
Pool 0: us.litecoinpool.or Diff:1 +Strtm LU:[00:02:21] User:jstefanop.a1
Block: ...5f3733dbdb447146 Diff:8.25M (59.04T) Started: [00:02:21] I:?
ST:3 F:0 NB:1 AS:0 BW:[546/119 B/s] E:0.00 BS:0
--------------------------------------------------------------------------------
APL 0: | 0.0/ 0.0/ 0.0 h/s | A:0 R:0+0(none) HW:0/none
--------------------------------------------------------------------------------
[2019-02-27 00:02:14] Loaded configuration file /opt/bfgminer.conf
[2019-02-27 00:02:15] Probing for an alive pool
[2019-02-27 00:02:21] Pool 0 stratum+tcp://us.litecoinpool.org:3333 alive
[2019-02-27 00:02:21] Pool 0 is hiding block contents from us
[2019-02-27 00:02:21] Network difficulty changed to 8.25M (59.04T)
[2019-02-27 00:02:21] Stratum from pool 0 detected new block
[2019-02-27 00:02:21] API running in IP access mode on port 4028 Then something pops up quickly that I can't read and then. [screen is terminating] and I'm back at the prompt. It does the same if I put in my own pools. Anything I can look at? -Dave Is the dashboard loading fine? Whats the last state you had the miner in before this happened? |
|
|
|
I was playing around with making this thing even more quiet. My stock fan is a little noisy. Might have just gotten a noisier one out of the bunch. I strapped on a silent noctua fan onto a fan controller and threw it on top of the heat sink. Ran it on eco mode. Had it sitting at 60 Degrees pretty much completely silent. But If i connect the noctua fan connector onto the apollos board, the miner will shut off after 45ish seconds. Is that the apollo just telling me to stick with the stock fan? Lol. Just curious if i could make the switch.
Which Noctua fan model did you use and what was the fan speed setting in Apollo? The Apollo comes with an around 6k rpm fan producing a maximum airflow of about 93 m³/h (if I found the right specs). For comparison, the Noctua NF-A9 PWM is 2k rpm and about 78.9 m³/h. Most likely the shutdown you got is caused by a failsafe, possibly either the fan rpm not being connected, the fan rpm being too low to detect or the fan rpm being lower than some hardcoded minimum. Set the fan speed manually to 100% before swapping to the Noctua fan and see if that helps. If that keeps the miner running then you can start experimenting with lower fan speed settings up to the point that the failsafe kicks in again or the temperature gets too high. Yea the hardware controller is hardcoded with failsafes based on the stock fan. If you use anything below a 3-4k RPM fan the controller will shutdown the miner because it thinks something is wrong with the fan. I won't be supporting any third party fans, I could possibly but in a "disable fan monitoring" option, but I wouldn't include that for at least several more months. The stock fan is a 6k RPM 120 CFM fan. That noctua tops out at ~ 45 CFM which is probably enough for only ECO mode cooling. |
|
|
|
|
Yea sorry guys I know I need to update the binaries for the pi that fixes this issue...you can just pull the latest source off the repo and compile though.
|
|
|
|
Just got mine in today and it was a very easy setup. Thanks for bringing this tech home. Runs beautifully, as well as the 6 MoonLanders and 2 L3+ I've got in my mini setup. Thanks again, really really nice work.
Thanks for coming on here and posting! |
|
|
|
if it was at least a trade off between them, vega56 1800mhs, rvii 2800mhs on monero, rvii is 50% shorter on computing performance than it was supposed to be and yet it costs 50% more, not worth, brainwashed trolls will make it to be worthy, trust me.
Give it time, Mr Doom & Gloom. Unoptimised drivers, unoptimised miners - that's what we're mainly dealing with currently. Vegas were allegedly crap until somebody found out they weren't, remember? Well the Radeon VII isn't running on a new architecture. It's essentially a Vega 56 on a smaller node with higher bandwidth and memory bus. The smaller node allows it to clock at the 2-2.1Ghz range giving it that big boost in performance. I doubt drivers will do much more than we are seeing already. The performance numbers make sense based on what we know from Vega 10 (which is essentially Vega 20 at 14nm). This, nothing new about it, actually i think vega 14nm is way more powerful given is 14nm and i dont think drivers will help it either. If it was a pure die shrink then monero was supposed to have hashrates around 4000mhs which is not the case. I'm a lot more interested in navi. There is no reason why Radeon VII cant hit 4k monero. Its doing 90MH eth. Something in the cryptonight kernel is not properly optimized to take advantage of the full bandwidth. Only other thing I can think of is that the algorithm is way more compute bound, so reduced number of cores is having a significant bottleneck. |
|
|
|
|
Show Posts
