Swedish ASIC miner company kncminer.com

[soy]

And here is a tip for everyone.  You know those two pins you jump to make the PSU work?  Do you know what those are used for?  

When you press the OFF switch your motherboard opens those two pins, telling the PSU to power everything down safely.  That is the only correct way to shut a system down - pull the jumper, then power off the PSU.  

If you just flip the PSU switch off, you leave a "whole lot of amerage" suddenly looking for the quickest way to get home.  And electrons aren't fussy, if they can find a shorter path across a component instead of thru it, they'll use that.  If there are unseen micro droplets of solder on the solder mask that present a shorter path, they'll use that path, carbonizing the solder mask and making it an even BETTER shortcut - resulting in nasty smelling smoke and burned solder mask...

Just flipping the PSU switch to off might not break anything the first time or even 100th time you do it, but eventually you'll pay the price, and the magic smoke gets out.  There's a very valid reason why you don't just turn off your computer by yanking the plug out of the wall socket - and flipping the PSU switch is the equivalent of doing that.  Do that to your desktop computer a few times and see how well it runs as a result...

+1

I, for one, would like to see that shown experimentally.  I personally believe there is no fault in pulling the plug.

More accurately, I believe there is no fault in pulling the plug with the miner attached to the power supply.

Still, that doesn't have a bearing on potential buildup on the power supply lines after disconnecting from the miner, plugged in with switch off or unplugged from the wall with power supply switch on.  The possible problem would be the power supply outputs looking at a open given that the leads would be unplugged from the miner.  The longer unplugged the higher the potential potential buildup. KnC does not know the design of every model of every power supply manufactured.  If a power supply had been powered on for hours or days the potential for voltage potential buildup on the filter capacitors while leads are disconnected may very well exist in some miners.

 In any case, I have two 12 volt fans here I bought to mount on a Blade but difficulty made me decide that it can't pay itself off in a reasonable period.  So, those two fans are going on to exhaust the Mercury, which should have no problem with heat unless I buy modules to upgrade to Saturn then Jupiter, if possible and cost effective.  These fans will always be plugged in first and then the miner, draining off any potential before firing up the supply - and the supply will have been shut down via an external power strip switch and back on the same way.

soy

[DeathAndTaxes]

Well, if the exploding caps are caused by voltage buildup on disconnected leads after the miner is disconnected, even with the power supply off, the voltage buildup due to voltage memory of the capacitors, voltage buildup that the power supply designers decided would be drained by the motherboard, then the jumper is signaling a condition, motherboard connected, that does not exist.

The jumper isn't signalling a motherboard is connected it is signalling the power supply should supply power to the rails.

Overvoltage into a motherboard is just as bad as overvoltage into any other device.  You are replacing one device which will be destroyed by overvoltage with another device which will be destroyed by overvoltage.  Motherboards contain filtering capacitors on their input voltage leads, just like mining boards do, just like GPU do, just like just about every electronic device ever made does.

Any regulating powersupply is designed to do exactly that.  The sole purpose of the power supply is to ensure that the output remains within a range (usually <3%) around 12V nominal.  Please show me in the ATX power spec where a power supply is suppose to shunt overvoltage into the ATX connector.  Please show me in the ATX spec where a motherboard (a device vulernable to overvoltage) is designed to be the dump for excessively high voltage.

It is complete and utter nonsense.  No power supply is designed to dump dangerous overvoltage into the ATX connector to let the motherboard "handle it".  What would be the point?  The POWER SUPPLY is the place to regulate voltage.  If there is excessive voltage it is the responsibility of the powersupply to either regulate it or disconnect the load (connected devices).   Yes power supplies contain overvoltage, overcurrent, and overheat protection circuits.   The powersupply (not the motherboard) monitors the voltage, current, and its internal temp and the powersupply not the motherboard kill the current if it is out of spec.     If the motherboard was going to regulate the voltage then there would be no need for a power supply you could just connect your power cable directly into the motherboard.

[Cablez]

I bought this (sry for bad pic)
I didnt take much notice at the time, but the jumper is set to pin 4 and pin 6 unlike the setup suggested in the manual of pin 4 and 5
I also have the 8vrm boards and the 6 connection controller board

http://thumbs4.ebaystatic.com/d/l225/m/mgjVeAWKwqzEQE7SmKpr4Jg.jpg[/img]

I bought that too. Is there something wrong with the configuration? Should i change the pins? Do you have problems with yours?

There is nothing wrong with that jumper, the pinout looks to be correct.  Its not as pretty or as functional as mine but hey.  

[soy]

Well, if the exploding caps are caused by voltage buildup on disconnected leads after the miner is disconnected, even with the power supply off, the voltage buildup due to voltage memory of the capacitors, voltage buildup that the power supply designers decided would be drained by the motherboard, then the jumper is signaling a condition, motherboard connected, that does not exist.

Where would the overvoltage go in the motherboard.  You are replacing one device destroyed by overvoltage with another device destoyed by overvoltage.  Motherboards contain filtering capacitors on their input voltage leads, just like mining boards do, just like GPU do, just like just about every electronic device ever made does.

Any regulating powersupply does exactly that REGULATE the voltage.  The sole purpose of the power supply is to ensure that the output remains within a range (usually <3%) around 12V nominal.  

No power supply is designed to dump dangerous overvoltage into the ATX connector to let the motherboard "handle it".  What would be the point.  The POWER SUPPLY is the place to regulate voltage.  If the motherboard was going to regulate the voltage then there would be no need for a power supply you could just connect your power cable directly into the motherboard.

The motherboard isn't expected to be normally unplugged.  The voltage wouldn't build up with the motherboard connected, instead the capacitors, filter capacitors, are discharged thru the motherboard.  No buildup.

Looking at the switching supplies recently that are used in USB hubs, switching supplies' voltage is regulated while the circuit is a "current mode" device".  The flow of current allows the voltage to be accurately assessed and regulated.  When the switching supply is looking at an open it may not actually measure or control the voltage.

soy

[operador]

FYI, my Jupiter day 1 customer 1-500 is running hosted since 2 days now at an underwhelming 100 GH/s average and my order still says PAID, never went to "processing" or anything else.

[DeathAndTaxes]

The motherboard isn't expected to be normally unplugged.  The voltage wouldn't build up with the motherboard connected, instead the capacitors, filter capacitors, are discharged thru the motherboard.  No buildup.

Now you are just going round and round in circles.  A miner isn't normally unplugged either.  However both a routinely are connected and disconnected over the lifecycle of the device and a powersupply needs to handle that.  It is the sole purpose of a regulated power supply.  It is what they get paid to do.

If the scenario that disconnecting and reconnecting a mining board would allow overvoltage to occur the same would happen when disconnecting and connecting any powered device.  People do upgrade computers you know.   If the bogus scenario you believe was true then a high portion of motherboard upgrades would result in destroyed motherboards.  Of course that doesn't happen because power supplies are designed to regulate the voltage.  The power supply has both a power-on and power-good signal.  I doesn't energize the output rails until it verifies the voltage is regulated.   There never is a scenario where a power supply lacks an active circuit.  It has control of the input power and the ability to power internal circuits for the sole purpose of continually monitoring and regulating the power output.

[xstr8guy]

Order 29XX, Jupiter
Paid June 28
Changed status from "Paid" to "In progress"

35xx in progress too.  Seems they're moving at a good pace now.

Are either of you hosted or are you both shipping?

Shipping.

[soy]

The motherboard isn't expected to be normally unplugged.  The voltage wouldn't build up with the motherboard connected, instead the capacitors, filter capacitors, are discharged thru the motherboard.  No buildup.

Now you are just going round and round in circles.  A miner isn't normally unplugged either.  However both a routinely are connected and disconnected over the lifecycle of the device and a powersupply needs to handle that.  It is the sole purpose of a regulated power supply.  It is what they get paid to do.

If the scenario that disconnecting and reconnecting a mining board would allow overvoltage to occur the same would happen when disconnecting and connecting any powered device.  People do upgrade computers you know.   If the bogus scenario you believe was true then a high portion of motherboard upgrades would result in destroyed motherboards.  Of course that doesn't happen because power supplies are designed to regulate the voltage.  The power supply has both a power-on and power-good signal.  I doesn't energize the output rails until it verifies the voltage is regulated.   There never is a scenario where a power supply lacks an active circuit.  It has control of the input power and the ability to power internal circuits for the sole purpose of continually monitoring and regulating the power output.

Not all computer switching supplies are without capacitor bleed down resistors.  The power supplies in question here are quite powerful - not your everyday supply.

When you disconnect a motherboard, do you also disconnect all drives, fans, every power supply lead out of the power supply?  Sometimes, maybe.  Usually no.

Mine is a supposition.  This is a somewhat unique application.


(current mode, current mode, current mode, current mode)

[xstr8guy]

Maybe this whole "PSU zapping the machine" thing can also explain why the SD card on my Bitfury rig scrambles every time I do a power cycle.

[amencon]

Probably a stupid question:
Is anyone's Jupiter actually working reliably at 450+GH/s?

After some more fiddling and letting it run I seem to be averaging close to 500 steady for now.  Thinking ill do like other and buy some cheapo fans for the top of the open case.

[amencon]

I bought this (sry for bad pic)
I didnt take much notice at the time, but the jumper is set to pin 4 and pin 6 unlike the setup suggested in the manual of pin 4 and 5
I also have the 8vrm boards and the 6 connection controller board

I bought that too. Is there something wrong with the configuration? Should i change the pins? Do you have problems with yours?

I bought the same and it's worked fine for me so far.

[trepex]

FYI, my Jupiter day 1 customer 1-500 is running hosted since 2 days now at an underwhelming 100 GH/s average and ...

Same problem here with a Saturn:



I created a topic at the KnC forum.

[Bitcoinorama]

Been away from the screen, as I was on this forum so late last night. Those who were around at that point know this. I've just been getting to grips and learning more about what components do what and when upon the board.

Been trying to catch up here, but am still confused as to the purpose of the opening of the casing and top loading of fans. It makes zero difference, it just appears to be a bizarre form of planking.

http://en.wikipedia.org/wiki/Planking_(fad)

The temps ascertained by the BertMod (great work that man), reveal the voltage regulator temps, and not the ASIC. The voltage regulators can handle in excess of 100 degrees Celsius - ambient (surrounding environment) - so blowing a fan on them is pointless.

Likewise the ASICs are barely above 50 degrees when i was playing with a laser temp gauge. You can actually touch the heatsink bracket above the ASICs and hold your finger there. The heatsink can dissipate above 320 Watts of heat, and is a totally OTT margin. So there are no heat issues whatsoever. The unit hashing at the Atlanta conference had everyone stating the device barely generates any heat. Those present are welcome to testify to this fact. It was running for several hours as verifiable via Eligius.

Some of the armchair engineering is amusing, but I'm the only one that's skimmed it, as previously stated this is all firmware issues. For those in despair that the whole world is falling apart and that they are entirely unprofessional; grow up, you've knowingly bought an unrefined product that is undergoing refinement. In effect a prototype revision and this was always the fact. You didn't want to wait, but at the same you can't expect blind perfection. If the margins did not exist, you wouldn't be seeing promised specs met. I'm not sugar coating, this is an indisputable truth, but that said you have nothing to fear, just be aware no one knew what these chips could actually do 10 days ago. The fact we've seen multiple instances of significantly +400 Gh/s demonstrates what you should be reaching, and shall be. It's only ever been firmware.

In standard product design the refinement between receiving chips in hand and consumer product in around 12-18 months, and not 24 hours. However, due to the nature of the reward being greater in Bitcoin - the earlier you can play with the kit, the greater the reward.

You all chose to take the route of a product that is shipped proven to achieve a declared spec, and have to accept that it will improve overtime.

As such one can only assume you'll be relieved to upgrade to 0.95, this includes the hosting facility...

I've seen what the next update is capable of. You'll have that to play with shortly. From what i've seen CGMiner may be asked to restart every now and then if performance begins to degrade over time. Then you'll have a further revision the otherside of the weekend. The freak 4 module boards sapping unnecessary wattage will also be rectified as well, but everyone should be happy hashing. Any inconsistencies, keep them posted. Feedback is useful, whinging is not.

The blown caps appear to be linked to a particular type of PSU, safety in the HX850 is triggered, the reset involves removing ATX cabling and re-inserting it. When this happens the current from the power supply applied is too great, it does not gradually load as when usually switched on. The cap blows as a result. There has been a specialist from General Electric's (GE) Critical Power Division that has flown in from the States especially here the last couple of days working on determining what the fault is and how to best proceed, hence the silence until there is a definitive conclusion. He couldn't make it earlier due to the fact he was at a conference. I am not an expert from General Electric's Critical Power Division and as such can only convey what I've understood from a brief conversation, as it is more important these guys work together on this. There is nothing wrong with the board design from what he can see and it adheres to all the standard recommendations GE outlay. That part I did understand. The fact he is British helps too, no loss of translation by me there...

EDIT: I don't know if this applies soley to this particular Corsair PSU, but for sure as it is a popular one that has been causing it and the GE guy confirmed the reasoning behind it, it's best avoided. Also if you are super worried I think he said it takes like 30 seconds for the PSU to get this out of it's system before you can turn it on and off again without it delivering a high load. So maybe if your PSU does trip, just disconnect the PSU, reset the ATX cable without the miner attached, and run it without the miner attached for 30 seconds, before turning the PSU back off, attaching the miner and without the power supply switched on and then turn on again, just to get this freak load out of it's system. I will check again with GE's guy tomo when he has a moment to confirm this edit here makes sense. Although I won;t go out of my way to disturb him while he's busy testing kit.

[Phoenix1969]

FYI, my Jupiter day 1 customer 1-500 is running hosted since 2 days now at an underwhelming 100 GH/s average and my order still says PAID, never went to "processing" or anything else.

FYI, my Jupiter day 1 customer 1-500 is running hosted since 2 days now at an underwhelming 100 GH/s average and ...

Same problem here with a Saturn:



I created a topic at the KnC forum.

Id say that saturn needs to be cooled better... are you in their pool too?  difficulty level on it?

[FractionalReserve]

The new asic bords with only 4 VRMs [which were designed to use 8] are using to much power. KNC will release a bug fix for this in next (0.95?) Firmware hopefully arriving in short (Today?)  A Jupiter in this Config will use app. 890W. So it can be difficult to start with a 850W ATX. (You can still run them with only 3 boards conected).

What happen to margins on margins.  The boards were designed to have 320A of DC regulators (which drop the 12V supply down to the 0.75V used by the chip).  Now if BFL cheaped out I could see that but KNC has talked for months and months and month about margins on margins so when they actually get to the finish line and start shipping products at the very last minute they decide to yank out all the margins and run the hardware at close to the theoretical limit and hope everything works.  

So the solution to an overheating and overloaded power system is to push the power system even harder.  To take 160A (which was designed to be 320A) of capacity and run it at 25% over the redline.  

So what is is this negative margins on negative margins now?

The VRMs are also running over spec with app. 50+ A current each! They probably won't burn, but if you can cool them extra until the new firmware is realeased, do it!!
 The Asics are tunning to hot (+70 deg C) as they get app, 0.9V instead of 0.7V as they are supposed to. Cool them as much as you can!!!  The Problem is due to the VRMs not working according to spec but KNC will be able to fix it with new firmware (we did our own patch for our miners yesterday).

How about just start shipping the proper 8 VRM design and offer replacements to affected customers (who you shipped underperforming units).

 Easiest fix to cool properly is to provide cool air, belo 20 deg C is a good idea. Open case + big fan will also work.

Yeah that seems like the solution.  Open case, giant fan, hardware running overspec, and power consumption through the roof.  Throw in the cost of high AC load to maintain 20 deg C cooling.  Too bad there is no simpler solution like I don't know ...shipping the product as designed, promised, and sold.

PS, this does not apply to 8 VRM asic boards. No worries here.

So wouldn't the best solution to ship the boards with 8 VRMS as you initially promised they would?

Hi, I see you quoted me from the KNC forum. Just to get thing straight, I am not from KNC!
We just run a huge bunch of Jupiters in our private mining op. We discovered the problems and let KNC now of them asap.

In the meantime we made our own mod where we adjusted the VRMs voltage offset (by software), the 4 VRM boards are running just as good as the 8 VRM boards and poerconsumtion at wall is below 550W.
All that is needed is the next FW uppdate that will fix this. I talked to KNC development (not support!!!) and they are really rushing this!

No need to ship with 8 vrms as there are no need for them, only need for correct FW! Don´t see any point for them to go back to 4 VRMs when it can be easily adjusted in software.

0.94 has a bug that shuts down parts of the ASICS but never restarts them, thats wy 0.94 performes badly, that will also be in the next FW update.

insuuficent cooling will cause HW errors and shut down cores => bad perfomance

The low performance on some ASICs even if cooling is sufficient comes from another problem they have regarding the internal ASIC communication. As I understand they are on a goood way to fix this. next FW will be better but there will still be more perfomance gains in future FW versions. now they are trying to rush out a 0.95 that will fix the above issues and hopefully give some performance gains and not only solve the above issues! We will se and we all hope for the better.

[naRky]

@Bitcoinorama


some info about slow shipping? why its so slow? and why is KnC silent about everything, i mean official status on website etc.

[Phoenix1969]

Hi, I see you quoted me from the KNC forum. Just to get thing straight, I am not from KNC!
We just run a huge bunch of Jupiters in our private mining op. We discovered the problems and let KNC now of them asap.

In the meantime we made our own mod where we adjusted the VRMs voltage offset (by software), the 4 VRM boards are running just as good as the 8 VRM boards and poerconsumtion at wall is below 550W.
All that is needed is the next FW uppdate that will fix this. I talked to KNC development (not support!!!) and they are really rushing this!

No need to ship with 8 vrms as there are no need for them, only need for correct FW! Don´t see any point for them to go back to 4 VRMs when it can be easily adjusted in software.

0.94 has a bug that shuts down parts of the ASICS but never restarts them, thats wy 0.94 performes badly, that will also be in the next FW update.

insuuficent cooling will cause HW errors and shut down cores => bad perfomance

The low performance on some ASICs even if cooling is sufficient comes from another problem they have regarding the internal ASIC communication. As I understand they are on a goood way to fix this. next FW will be better but there will still be more perfomance gains in future FW versions. now they are trying to rush out a 0.95 that will fix the above issues and hopefully give some performance gains and not only solve the above issues! We will se and we all hope for the better.

I'm running 0.94, and it enables/disables cores just fine... depending on the temps... and you say it's going to get better?  YEAH BABY!

[Bitcoinorama]

@Bitcoinorama


some info about slow shipping? why its so slow? and why is KnC silent about everything, i mean official status on website etc.

As I said above you need feedback from multiple parties, once there is mutual agreed feedback from situations then something official can be said.

Common sense.

These are multiple issues, as to be expected one has to confirm, and then react, as and when they appear.

You'll be getting just that with the update momentarily.

The guy from GE flew in especially. The factory is no longer staffed by KnC. The manufacturing facility has taken over their role. KnC/ORSoC have concentrated on testing issues, refining and releasing patches, and hosting. So Sam and Andreas have given people a chance to ascertain complications and propose resolutions.

Now that is in hand you can be briefed on how this is moving forward for the following week ahead.

[BenTuras]

...

Thanks for the update. Please keep doing this daily!

Now about the sequence of orders, it seems that hosting orders are on halt, for example I have 280x and 281x and both are still on paid while higher numbers are already being processed. Both were paid within minutes after ordering. Hosting was ordered and paid later, but that shouldn't matter IMHO.

[nightengale]

So is there a confirmed "safe" brand of PSU to go with, or just anything but Corsair?