FutureBit Apollo III - Next Gen Home Mining

[heslo]

Flashed the new update aaaaannnnddd.......

NOTHING

Same issues before Definitely seems like a hardware failure and I'm pretty bummed about it but it is what it is

[philipma1957]

I just did check yes (just woke up) and I'm giving that a shot. Hopefully that's the issue and will save me from sending back the unit all the way from Australia. Will report back soon!

hey mine went from Howell , New Jersey to Brooklyn, New York under 60 miles and instead of a day will take at lease 4 to get there.

[jstefanop]

Broken pull tabs seems to be almost on every shipment at this point. We made this easy to tear off, but apparently they are way too easy so even small ship drop will cause them to tear.

Will have better ones coming but not much we can do, these are mainly designed for bulk shipments to our distributors/retailers so not a big deal.

Looking into the cause of the dead boards...seems to be early manufacturing defects we are not catching at the factory. We increased our cycle/test time, but so far we have only had 3 units with this issue (2 here and 1 over normal support channels) and so wont know more till we get them back.

[jstefanop]

I just did check yes (just woke up) and I'm giving that a shot. Hopefully that's the issue and will save me from sending back the unit all the way from Australia. Will report back soon!

hey mine went from Howell , New Jersey to Brooklyn, New York under 60 miles and instead of a day will take at lease 4 to get there.

Support/order team has instructions to send out new units, we are not waiting to get the old one back. Had to pause on shipments this week to look further into testing and firmware update we need to reflash on all units at the factory...replacements and more shipments going out early next week.

[heslo]

Wasn't fussed about the tabs, just something I pointed out as if you put them there for a reason then at least you know they're not working.

The dead board isn't fun, but it is what it is; no point me complaining about it. Just the pain of dealing with an RMA is never fun, especially since you guys can't issue a label for me being I'm not inside the US. I hope when I sort out the shipping back I don't need to deal with any of the tariff dramas but lets see how it goes. Currently waiting on a reply from support with all the RMA deets I need and then I'll start to get the ball rolling on sending this back. Hopefully they can pickup from me, the nearest FedEx drop point is near the airport which is 45 minutes away lol

Appreciate you guys sending units back straight away without waiting for returns; great customer service!

[philipma1957]

I just did check yes (just woke up) and I'm giving that a shot. Hopefully that's the issue and will save me from sending back the unit all the way from Australia. Will report back soon!

hey mine went from Howell , New Jersey to Brooklyn, New York under 60 miles and instead of a day will take at lease 4 to get there.

Support/order team has instructions to send out new units, we are not waiting to get the old one back. Had to pause on shipments this week to look further into testing and firmware update we need to reflash on all units at the factory...replacements and more shipments going out early next week.

first off thanks for being professional.

Shit happens.

I remember when I was selling upgraded 2005 Mac minis with upgrades to a 500gb hdd 10 of 11 seagate drives failed .  It sucked bigly. I can imagine it is somewhat like that for you.

[heslo]

Yeah their support has been great!

While it sucks to have a broken unit I really can't ask for more with how they've responded with support; top stuff guys!

[jstefanop]

I just did check yes (just woke up) and I'm giving that a shot. Hopefully that's the issue and will save me from sending back the unit all the way from Australia. Will report back soon!

hey mine went from Howell , New Jersey to Brooklyn, New York under 60 miles and instead of a day will take at lease 4 to get there.

Support/order team has instructions to send out new units, we are not waiting to get the old one back. Had to pause on shipments this week to look further into testing and firmware update we need to reflash on all units at the factory...replacements and more shipments going out early next week.

first off thanks for being professional.

Shit happens.

I remember when I was selling upgraded 2005 Mac minis with upgrades to a 500gb hdd 10 of 11 seagate drives failed .  It sucked bigly. I can imagine it is somewhat like that for you.

Well if 90% of the units were failing that would be a shitshow.

Looks like we are under 1%, but still early I was expecting around a 2% field failure rate with our new design since our old design was a parallel coms and each bank of chips had its own IO chip which was a cool design...dont think we had more than a dozen boards fail over thousands shipped.

This new design is a serial chain design like almost every other bitcoin ASIC, which means if one chip blows up the whole chain is dead.

[philipma1957]

I just did check yes (just woke up) and I'm giving that a shot. Hopefully that's the issue and will save me from sending back the unit all the way from Australia. Will report back soon!

hey mine went from Howell , New Jersey to Brooklyn, New York under 60 miles and instead of a day will take at lease 4 to get there.

Support/order team has instructions to send out new units, we are not waiting to get the old one back. Had to pause on shipments this week to look further into testing and firmware update we need to reflash on all units at the factory...replacements and more shipments going out early next week.

first off thanks for being professional.

Shit happens.

I remember when I was selling upgraded 2005 Mac minis with upgrades to a 500gb hdd 10 of 11 seagate drives failed .  It sucked bigly. I can imagine it is somewhat like that for you.

Well if 90% of the units were failing that would be a shitshow.

Looks like we are under 1%, but still early I was expecting around a 2% field failure rate with our new design since our old design was a parallel coms and each bank of chips had its own IO chip which was a cool design...dont think we had more than a dozen boards fail over thousands shipped.

This new design is a serial chain design like almost every other bitcoin ASIC, which means if one chip blows up the whole chain is dead.

yeah my old Mac min upgrade business was 3 or 4 pieces a week.  I got a case of seagate 500gb hdd's no one had them but me so I sold 11 units  of Mac minis in a week and did 10 rmas the next week

[crypto_curious]

our old design was a parallel coms and each bank of chips had its own IO chip which was a cool design...dont think we had more than a dozen boards fail over thousands shipped.

This new design is a serial chain design like almost every other bitcoin ASIC, which means if one chip blows up the whole chain is dead.

And now you’re sharing this? If I had known this before buying, I would seriously have reconsidered purchasing the device.

My old Apollo II has been showing a 97% error rate for longer than I can remember, probably over a year. I assume that means one of the 44 chips has failed. Yet the remaining 43 chips continue to work perfectly well, and the unit is still usable.

With the new design, one chip fails and the entire device becomes useless.

What a waste.

[philipma1957]

our old design was a parallel coms and each bank of chips had its own IO chip which was a cool design...dont think we had more than a dozen boards fail over thousands shipped.

This new design is a serial chain design like almost every other bitcoin ASIC, which means if one chip blows up the whole chain is dead.

And now you’re sharing this? If I had known this before buying, I would seriously have reconsidered purchasing the device.

My old Apollo II has been showing a 97% error rate for longer than I can remember, probably over a year. I assume that means one of the 44 chips has failed. Yet the remaining 43 chips continue to work perfectly well, and the unit is still usable.

With the new design, one chip fails and the entire device becomes useless.

What a waste.

It is not so bad as you think if the board and chips are properly fixed installed soldered.

But a 2% immediate fail rate is frightening.

Let go  off topic and talk about the bitmain s21xp i have 9 of them 20 months old flawless for this entire time.

And let's talk about the t21 in three years 4 units 12 boards and  7 failed.


So if this Apollo runs like the s21xp we will all like it.

But if it runs like the t21we will hate it as 7/12 fails in 3 years is terrible.

I hope to get tracking for my replacement unit. So far no tracking for the replacement

[heslo]

I don't think we will get shipping numbers until Monday US time

[crypto_curious]

I really hope Futurebit hasn’t sacrificed the robustness of the old design just to save a few dollars per unit. If my Apollo III dies out of warranty because of this, it will probably be my last Futurebit.

[heslo]

They've always stood behind their hardware before; I don't think that will change now (at least I hope not)

For full transparency, this RMA process has been absolutely stress free and they're shipping a replacement out to me as well before they even receive the dead unit. I'll keep everyone updated here with how it goes

[jstefanop]

I really hope Futurebit hasn’t sacrificed the robustness of the old design just to save a few dollars per unit. If my Apollo III dies out of warranty because of this, it will probably be my last Futurebit.

I have never skimped out on components/design for cost, and never will as long as im making the engineering decisions.

This is a design change around the ASIC itself, nothing I could do about it...every 3nm ASIC now handles IO in serial, because no one wants the added complexity around parallel IO anymore.

Failure rates will be higher with this design, no way around it when you have 21 chips in series, and one chip or IO voltage rail/ LDO fails the whole chain dies.

Our board designs at this point are better than most so we mitigate most of those issues.

Regarding the few failures we have had so far they actually seem to be correlated with a firmware bug that we fixed, but some units that have shipped out in the last few weeks had the old firmware on it. So thats a plus.

[philipma1957]

STILL waiting for someone to receive my rma

[huntingthesnark]

Just walking through setup, have updated to the latest firmware. Weirdly, the node just refuses to start. Good wifi and ethernet connection. <edit: A restart has cured that, it seems!>