Pages:
Author

Topic: A4 Dominator - Pre-Order Group Buy - 280mh, roughly 1000w, $1800 + shipping - page 31. (Read 122577 times)

newbie
Activity: 17
Merit: 0
My best guess is that its the top left chip in the pic posted earlier.
That would be the only chip that isnt easy to reprogram imo.

So if you want to replace a chip like that you'll need some soldering skills.
hero member
Activity: 833
Merit: 1001
based on inno's response it's a custom jig. probably some kind of tool to "trick" MCU into debug mode and then do firmware update? lightfoot you know this stuff better than most of us.. i think for you replacing that stm must be child's play. my suggestion to those who stay closer geographically to lightfoot its better if you engage his services or else once the board is messed up you'll have a hard time getting that warranty honoured by inno and end up paying more.

Ah, hm. I wonder if the "jig" is something like an AVR Dragon, like we used to reprogram the Atmel MCU's on the Jalapeno and Singles using a JTAG interface....

Is there a 10 pin diagnostic port on these things?
hero member
Activity: 952
Merit: 508

Hello Vinylwasp,

We have solved the restart and reboot problem in batch 3. No update in hardware but the firmware in STM32 MCU and SD card both have updates.
SD card firmware has been upgraded to v2.0.1. For ST MCU firmware, we will send you a testing jig (which is expensive) and user manual to update.
For smaller batch one and batch two buyers who bought less than 50 units, we will send them new STM32 MCU for local replacement, but will need them to find local expertise to execute that.
.


Anyone have any idea what is involved to swap out the STM32 MCU they are talking about?

x2 on this question -- I hope some information is put out there on what this thing even looks like.

In regards to allowing me to replace it myself vs sending it in for replacement; I'm glad they are going to give us a field replacement option, with my professional experience it shouldn't be an issue for me to do.

What I'm concerned about is if they are going to charge us for this ability.

I'd like to know where/how to get on the list for one of these.

hero member
Activity: 868
Merit: 517

Hello Vinylwasp,

We have solved the restart and reboot problem in batch 3. No update in hardware but the firmware in STM32 MCU and SD card both have updates.
SD card firmware has been upgraded to v2.0.1. For ST MCU firmware, we will send you a testing jig (which is expensive) and user manual to update.
For smaller batch one and batch two buyers who bought less than 50 units, we will send them new STM32 MCU for local replacement, but will need them to find local expertise to execute that.
.


Anyone have any idea what is involved to swap out the STM32 MCU they are talking about?
hero member
Activity: 868
Merit: 517
Well they said for us small batch purchasers that they would send a replacement STM and we would have to replace ourselves. 

It would be nice to know what is involved in either case, the replacement or the reflash.  I am all about the path of least resistance for the end user.  So if replacing the STM is easier then lets do that. 

Looking forward to what others think about this.  I would also like to see more data on the B3's to ensure this actually fixes the issue before we start jumping through hoops.  I would hate to see everyone rushing to do this and it doesn't correct the problems with dropped boards and the endless stream of restarts that these miners do on multipools with B1 and B2 miners.
copper member
Activity: 2898
Merit: 1465
Clueless!
Which is going to be easier for the end user?  Replacing the STM or reflashing it?  If it is easier to replace than to reflash then let them send the replacement STM to all of the B1 and B2 buyers.  That saves people shipping stuff all over to get flashed, and yet more risk in something going sideways in the process.  But I am certainly open to suggestions and willing to help anyway I can.

Not sure about electronics, but did they not say in above msg on such ( 5 msgs up) that this is your problem small order chumps?....or begone minions with your lame first adopter problems!
legendary
Activity: 3164
Merit: 2258
I fix broken miners. And make holes in teeth :-)
Ah, hm. I wonder if the "jig" is something like an AVR Dragon, like we used to reprogram the Atmel MCU's on the Jalapeno and Singles using a JTAG interface....

Is there a 10 pin diagnostic port on these things?
hero member
Activity: 868
Merit: 517
Which is going to be easier for the end user?  Replacing the STM or reflashing it?  If it is easier to replace than to reflash then let them send the replacement STM to all of the B1 and B2 buyers.  That saves people shipping stuff all over to get flashed, and yet more risk in something going sideways in the process.  But I am certainly open to suggestions and willing to help anyway I can.
copper member
Activity: 2898
Merit: 1465
Clueless!
They need to send longsnowsm a jig. Surely 50 plus on the group buy is enough reason on this
Lame imho support!

Have they gone back on open source as well on a side note as well?

Anyway they seem to not have run away from commitments on the
small orders ..... but seem to be backing away briskly for now

Edit:
So I gather they waited on this till your 45 day warranty expired?
member
Activity: 97
Merit: 10
Neat. Any chance someone can post pictures of the gen 1, 2 and 3 boards, along with the controllers. Curious if there are any board specific changes.
Lightfoot, I don't know what the root cause is, but I can tell everyone here that there's a fix coming from Inno.
Here's a slightly edited excerpt of an email I received from Chloe.

Hello Vinylwasp,

We have solved the restart and reboot problem in batch 3. No update in hardware but the firmware in STM32 MCU and SD card both have updates.
SD card firmware has been upgraded to v2.0.1. For ST MCU firmware, we will send you a testing jig (which is expensive) and user manual to update.
For smaller batch one and batch two buyers who bought less than 50 units, we will send them new STM32 MCU for local replacement, but will need them to find local expertise to execute that.
.

I replied suggesting that Chloe reach out to Longsnowsm, yourself and/or GenTarkin to discuss how to co-ordinate reflashing the group buy miners after sending you a jig.
I additionally suggested that yourself or GenTarkin could be utilized in your respective geographies to flash units belonging to small purchasers rather than sending them an updated STM32 MCU for local replacement, while offering my jig (once I receive it) and my services down under. I'm based in NZ.

I've been involved in the IT industry for the past 20 years and have seen numerous failures, so this is very familiar ground. Sales promise shipping date, orders are taken and paid for, development takes longer than anticipated (best case targets rarely come true), buyers start complaining about missed shipping dates, testing is shortened -> product has bugs.

So before everyone starts screaming about how terrible this situation is, remember the pressure applied to ship this product, the large investors who laid down $0.5-1M, and the fact that Inno have been working tirelessly to improve both the hardware and the software over the past 4-6 weeks with 3 versions of hardware and 6+ versions of software shipped in this period. Yes they can improve comms but they've got large enterprise customers who are screaming much closer to home in their native tongue and Chloe seems to be the only person in a comms role (btw: I checked with Chloe that it was ok to share the information above). This is the industry we've chosen to jump into for better or worse. I'm hurting too but I knew the risk before I laid down my money.

My personal experience is that the restarts are not the real problem, the problem is that some cards do not restart. To solve this, I've been monitoring each cards work and applying +2Mv to the miner and watching for 24 hours. It seems to be working. I now have miners which still restart all 4 cards most of the time. Not only that my B1's are hashing on Prohashing at around 250-260+ Mhs 2hr avg. and my B2's hash at 270+Mhs.

B1 = 822-824Mv at 1200Hz
B2 = 826-828Mv at 1224Hz

They're running between 45-52 C most of the time, but will drop into the 30's overnight. I do through a lot of cool air at them though so mileage may vary.

HTH
legendary
Activity: 3164
Merit: 2258
I fix broken miners. And make holes in teeth :-)
Neat. Any chance someone can post pictures of the gen 1, 2 and 3 boards, along with the controllers. Curious if there are any board specific changes.
legendary
Activity: 1834
Merit: 1080
---- winter*juvia -----
my A4s are online now at my ISP, Batch 3 with v2.01 firmware....

Tried Nicehash but the MHs are all over the place .... then I pointed them to Prohashing.com and seems stable... but let's see how it goes after 24 hours.

4 out 6 units are online, reading poolside, 350MH/s, 319MH/s, 250MH/s and 200 MH/s

Voltage: 815
Diff: 8192
Speed: 69.5 / 1200mhz across the board
Temps: 32-44c range (sampled from 4th rig - 200MH/s)

I will update again stay tuned.

Please keep us posted how your Batch 3's run.  I am surprised you took the risk given how badly the batch 1 and 2 miners are running.  Good to hear that they are working so far.

yes it was a big risk but calculated risk assuming that Innosilicon will eventually come up with a solution. My contacts in China shared with me that our rants in Bitcointalk is nothing compared to the "storm of fire and brimstone" from the A4 owners in China!

the rig setup are not ready or fully completed yet, because missing 1 more box from DHL which has not arrived. That box contained a controller, so to test the miners, we just use the available controllers we have and the MHs as spec should give this reading 400MHs, 400MHs, 280MHs and 400MHs for now. When the last controller arrives, the final setup with be 280MHs x 6. I will measure the actuals after we complete the setup.


hero member
Activity: 868
Merit: 517
my A4s are online now at my ISP, Batch 3 with v2.01 firmware....

Tried Nicehash but the MHs are all over the place .... then I pointed them to Prohashing.com and seems stable... but let's see how it goes after 24 hours.

4 out 6 units are online, reading poolside, 350MH/s, 319MH/s, 250MH/s and 200 MH/s

Voltage: 815
Diff: 8192
Speed: 69.5 / 1200mhz across the board
Temps: 32-44c range (sampled from 4th rig - 200MH/s)

I will update again stay tuned.

Please keep us posted how your Batch 3's run.  I am surprised you took the risk given how badly the batch 1 and 2 miners are running.  Good to hear that they are working so far.
legendary
Activity: 1834
Merit: 1080
---- winter*juvia -----
my A4s are online now at my ISP, Batch 3 with v2.01 firmware....

Tried Nicehash but the MHs are all over the place .... then I pointed them to Prohashing.com and seems stable... but let's see how it goes after 24 hours.

4 out 6 units are online, reading poolside, 350MH/s, 319MH/s, 250MH/s and 200 MH/s

Voltage: 815
Diff: 8192
Speed: 69.5 / 1200mhz across the board
Temps: 32-44c range (sampled from 4th rig - 200MH/s)

I will update again stay tuned.
legendary
Activity: 3164
Merit: 2258
I fix broken miners. And make holes in teeth :-)
This is actually an interesting line of analysis: Can someone stick one of these in a snowbank or a freezer? I'm curious if design issues in the power supplies or the board layout is resulting in heat plus resistance knocking down the SPI lines.

C
legendary
Activity: 1109
Merit: 1000
Hmmm ... Device temp is about 55 C. No rejected shares for 24 hours at litecoinpool. But noise from cooler fan is so strong, Does anybody have change cooler or have some ideas about it?
 And if i was change core speed down to 52Mhs which voltage i can set for stable work?

I had a bad fan on one of my cubes. Inno told me to replace with any fan CFM >= 200.
You could do a search on decibals per CFM and see if you can find a quieter fan.
Other options include duct-work to feed outside AC directly into cubes, similar to what they did for early AntMiners...
newbie
Activity: 3
Merit: 0
Hmmm ... Device temp is about 55 C. No rejected shares for 24 hours at litecoinpool. But noise from cooler fan is so strong, Does anybody have change cooler or have some ideas about it?
 And if i was change core speed down to 52Mhs which voltage i can set for stable work?
legendary
Activity: 3164
Merit: 2258
I fix broken miners. And make holes in teeth :-)
I dont have an IR gun, but the web page reports temps of 50-51 on all my cubes.
Fans are 200 CFM...
I was thinking more hotspots on the board itself. That's a long distance to freight x amps at low voltage, the issue can pop up where voltage instabilities between dies can lead to virtual ground loops on the SPI signal interface (to its' ground assuming it's not isolated).

The power supplies themselves look to be the same concept as the old jalapenos right down to the 1850 running two channels of FETs feeding the same supply line.

Interesting.
hero member
Activity: 952
Merit: 508
Had an electrician around to put a nice chunky cable in my mining room and 10 power ports.

Seems to be more stable however, much more stable... only one difference:

I'm now running each miner from a single power port...

where as before I had two miners into an extension lead, into the wall...

Miners are tipping just over 1000 watts usage - extension cables are usually rated at 2000 watts max...

Maybe this is a problem for other people?

Mine is running on a 208V - 2880W PSU and it has the restart issues. It's also in environment that the temperature reported by the miner is never higher than 50C

This is the longest its been running in days since my recent tests

legendary
Activity: 1109
Merit: 1000
Im running my A4's off a single 220V 50A circuit through a switched PDU which allows for remote power cycle on a per-port basis.
Each miner has it's own Seasonic PSU.

Still dropping about 1-2 boards per day.
Pages:
Jump to: