Topic: Official FutureBit Apollo BTC Software/Image and Support thread - page 40. (Read 49004 times)

Trying to cut power and already have the equipment and just trying to see if. having 88 chips hashing slow is better than 44 hashing at the same speed. i dont really know at this point im just trying things while i can.

The USB C probably has no way of getting the current where it needs to go without frying something. Or rather, It cannot power the hashboards at all.

But if you really have reduced the power per Apollo to 30W (how do you do that?), then maybe the daisy-chaining could work.

At 280W ( = 100% power) you would have about 280W / 12V / 2 connectors = 11.66A flowing through each connector on one Apollo.

At 2 x 30W = 60W  you would have 60W/12V = 5A going into the one connector on the first Apollo, and 30W/12V = 2.5A for the rest of the way to the second Apollo.

That seems feasible to me.

Why not run one Apollo at 120W instead?

Is it ok to daisy chain the apollos power connectors on the back of apollos? example 1 apollo gets 1 pci-e plugged into power and the other power port on that apollo feeds 1 power port on another apollo?

Or can i use usb C with like a 100w wall adapter to power both units? I have these really underclocked and only using 30w per board.

I believe batch 1-3 had orangepi-4 (not a version sold outside of futurebit to my knowledge) im not sure whats different on the newest mcu besides it only had 24 pins. The wires should be the same what makes the most difference is header pin count so long as it has 40 pins you should be good, the pi5 comes with 26 pins im not sure if that works or not. I have put armbian-bulleyes cli on mine an it has made the node a lot better in terms of resources used. I may try alpine next for fun after i built a full size node out of a normal x86.

make sure after flash and before you hook it to the internet you change root password and futurebit user password because ssh is wide open on the futurebit image.

Does anyone know if the SBC for this is the Orange Pi 4 and the newer one is the Orange Pi 4 LTS? I bought one used but there seems to have been a mishap before my buying it. The USB ports are dead. The SBC mine has the the Orange Pi 4, but I can't locate a replacement. But the Orange Pi 4 LTS can be bought. I also seen there was a picture to find out which SBC it has, and from the placement of that dastardly microphone it looks like it's the Pi 4 LTS variant, though the microphone for mine isn't in the same place as the LTS variant. In which case I should be able to just upgrade to that one, provided the wires and plug going from the GPIO to the hash board is the same and the wires connect to the same pins. Anyone willing to help di into this with/for me? Just need to compare note on the wire harness to make sure they are the same so I can just upgrade it.

This sounds great! Any idea on an ETA for this new software? Any plans to open it up to testing?

Alas you are still missing a few points about this

1) Block distribution is now much faster, size isn't exactly what is sent, so 'winning' the block size war has indeed been a VERY negative thing for bitcoin telling people to use other coins coz bitcoin is often high fees, and often too slow confirming transactions - all because blocks are too small - as you've said around 1.5Mbytes which I guess is somewhere near what they average.

2) That pool you like to mention is only a single node somewhere in the USA, with no worldwide block distribution at all.

3) A 3TH Apollo at the moment is 1/652058 per day chance, so hoping they wont hit the 1/600 (or 1/300 or 1/200 or worse) but that they will hit the 1 in 652058 seems rather a fun idea
Be aware that orphans still do happen on the big pools, just that marketing is what runs bitcoin mining, not facts.
And this is true even though none of the large pools confirm the transactions in the blocks they start building on until after they switch.

Yes this is one of the reasons we have not released it yet, we are building infrastructure on the backend that will allow each individual node to connect to the back bone of the main pools. Number of connections is not as important for new block propagation as how close your node is to the nodes of main pools.

Its in the interest of small miners to send the block to the main pools first, because as soon as the large pools node sees the Block they will stop mining on the previous block header and move work immediately to the next. If a small miner is on the fringes of the node network, by the time the block propagates to the pool nodes they could have very well found another block and will orphan yours.

This would be a horrible outcome, your device actually finds a 6BTC valid block but you lose out because of slow propagation. We are currently spinning up super nodes that are on fast connections and directly connected with major pools, in a future update all futurebit nodes will have our infrastructure nodes as a seed nodes, so once we enable solo mining direct to node if you do find a block the first "hop" will be to our backbone super nodes, so at worst your block will be two hops away from all the major pools.

This should significantly increase odds of small solo miners not orphaning blocks, but want to make it perfectly clear its still worse than solo mining directly to ckpool for example (and the degree of how much worse depends on your upload speed).

If your on dialup or DSL you should not be solo mining on your own node, and would not recommend anything less than a 30-40 mpbs upload speed and direct ethernet only (worst case 500ms delay with ~1.5MB blocks). Not bad but obviously not as good as a 30-50ms upload speed of a few kb pool share to ckpool. So for example to be on par with solo mining to ckpool you would need beat ~50ms share propagation to ckpool + ckpools 50ms block propagation (assuming his setup is near perfect which I would assume it to be), so 100ms for ~1.5MB you would need a 120-150 mbps upload speed.

I also dont want to discourage people and blow the issue out of proportion either, bitcoin has 10 minute blocks for a reason its extremely unlikely anyone will find two blocks within a few seconds of each other...for each second you delay uploading a block there is 1/600 chance you'll orphan the block.

This is also a good exercise in remembering why we won the block size war...if we did not no one on low end hardware or slow internet connections would ever have a good chance at solo mining...

18 connections? That's very low. I connect ONLY via Tor, no IPv4/v6 connections as all, and have 32 peers. I was reading a bit today about block propagation in Bitcoin network and apparently it's still an important factor, especially for big miners who hit blocks often. You must be extremely lucky to hit the block within your lifetime if you own one, or even 10, Apollos. Then, you must be extremely unlucky to have that block orphaned by not propagating fast enough, but it is possible.

Currently my node is connected with 18 other nodes and upload for a few kb nonces cmon u dont need 1Gbit. I just bought this full node package for mining the own node. Sure ck.pool is not bad and the only option right now not supporting the big boys like ant/via etc.

Solo mining to your own node may have the opposite effect. ckpool's VPS is in very well-connected paid datacentre, and most likely connected to maximum amount of 125 peers. ckpool can push the block to 125 connected nodes, and to the entire world, in a matter of few seconds. What is your home upload? Most likely in tens of Mbits or slower, no match for 1Gbit upload from datacentres. If your block is not propagated well, then 20ms delay you save by not connecting to ckpool means nothing if you propagate your block in 30000ms (30s) and by then someone faster pushes hit block and yours become an orphan.

I agree but it's the best we have right now

The point of mining solo directly on our node is it might be a little bit faster.
Any fraction of a seconds counts to get the edge.
ckpool is not the best with one block per 4 months and all the intermediate connections.
FutureBit told me solo mining was not ready for release yet.
Hopefully soon.

Your node wont show more than 10 connections until it is fully synced, so it's not a port forward issue.

As for wifi/ethernet, this is a know bug in the underlying system, if you enable wifi then try and connect ethernet it wont work. You need to disable the wifi interface in the system menu (not the dashboard, you need to connect to the system via HDMI), then ethernet will work again.

Good afternoon! I’ve had my Apollo running for just under three days now. I’m having two issues: the first, I’m having trouble setting up port forwarding, and I’m stuck on 10/32 connections. I’ve logged into my router admin page, and I updated the port forwarding section with my units IP, and TCP + UDP, with WAN Port 8333~8333, LAN host port 8333~8333.

Any ideas how I can remedy this?

My second issue is, this morning my internet connectivity dropped out (only on the Apollo, the home internet never stopped working), and on the dashboard, I was prompted with a “Disconnected” under eth0 network (direct Ethernet connection) - this happened without any apparent reason, and I was no longer able to access the admin page from the router. I’ve since did a hard reset on the router, but that isn’t solved the issue.

My only workaround for the second issue is to connect to wifi, which I’ve been able to do, but I would really like to sort out the hard wired connection issue. My Apollo won’t show an Ethernet connection at all even though the green light turns on when I plug the cable in to the port.

I’m also seeing a red  “internal error” message when I first launch the Apollo dashboard, right at the top of the page overtop of where I see the online status and TH/S.

I’ve emailed Futurebit, but thought I’d try here too.

Thanks!

Just use ckpool, that's what I'm doing

any updates on solo mining to our own nodes?  or any working workarounds ?  is it worth it or better to just use ckpool solo?

worked!  thank you

If i had any merit to give i would give it to you. the tips you provide help me progress in learning. i do agree i have learned a lot with mining and embedded systems not so much with the futurebit but with some of the older miners which i purchased prior to the futurebit. im just hitting a wall trying to figure out a definitive path to what i wanna do so many options. i lack hardware knowledge at a low level so this has been super challenging for me. and all of the scam talk and other things that go on are really discouraging.

Ooo the big farms are finally realizing plebs mining in masses are going to be a threat to their USD profits the future