ZTEX USB-FPGA Modules 1.15x and 1.15y: 215 and 860 MH/s FPGA Boards

[ztex]

I also experience these "hashrate drop errors", but only with BTCMiner. With cgminer 2.4.1 I've never got such an error. The hashrate drop code is present in cgminer. So what does it mean? The hashrate drop errors are present, but cgminer fails to detect them. Or, something is wrong within BTCMiner. One observation I've made... I get the hashrate drop errors in BTCMiner always around "new block" announcments.

These "hashrate drop errors" occur if the frequency is reduced too much because the error rate increases. If cgminer uses an other error rate detecting algorithm (e.g. slower sampling rate) it may oversee this errors.

I some cases (e.g. if ambient temperature increases by more then 10°C) the default setting may be a little bit to sensitive, try out "-oh 0.06" or "-oh 0.08", but not higher because that would break the overheat detection mechanism. If the frequency drops by more than 8% there are problems on the hardware side (cooling, power supply stability)

[1714328558]

1714328558

[1714328558]

1714328558

[1714328558]

1714328558

[1714328558]

1714328558

[ztex]

I assume you're only seeing the errors on the quads (I had only seen them on the quads). Are you driving them (1 or 2) on 1 psu? How are you wiring your +12V and GND wires? I had these errors initially where over half of my quads eventually were disabled over the course of a week. I rewired the power cables and haven't seen them for 2 weeks now.

One quad with its own PSU. Five singles on the splitter thingie and one Single on its separate wall plug.

No its not only the quad. Today for example i came home, three of the five on the splitter and the one on the separate wall plug singles had an LED on (indication that they weren't programmed or stopped working). This kinda happened when emc pool did go offline today (i assume!)

These PSU's become unstable (interruptions) if they get to hot. Try to connect four singles to one 60W PSU,

[ztex]

ztex can we please, with sugar on top, get a -f XXX commandline flag that hardcodes the frequency so that we can underclock the chips this summer?

Wee already discussed this: https://bitcointalk.org/index.php?topic=40047.msg802603#msg802603

[mrb]

ztex, can you please upgrade the AOZ1025 buck regulator on your designs to another able to output more than 8A to each FPGA, and more efficient? Multiple groups (eldentyrell, bitfury, etc) are writing bitstreams pushing the LX150 beyond 8A... And the AOZ1025 is not super efficient.

Perhaps the IR3840MPbF (86% efficient at 12A output at 1.2V) ?

[rupy]

ztex can we please, with sugar on top, get a -f XXX commandline flag that hardcodes the frequency so that we can underclock the chips this summer?

Wee already discussed this: https://bitcointalk.org/index.php?topic=40047.msg802603#msg802603

Ok, could you add a zero error rate command... that lowers the frequency until error rate is = 0.00% forever?

I understand it's bad for business if the cards never break, because then we wont buy new ones, but we would like to be ABLE to control the heat. It should be easy no?

Basically can you try to add anything for the passive community to control longevity.

[BR0KK]

No it would be Bad(er) for business if the boards tend to break ....

I only invest into things that are durable and covered by warranty.

[rupy]

I think you need to wakeup, every product in your home has been designed to break after a certain time after your warranty expires.

[BR0KK]

Do u honestly belive in "build in time to fail counters"?

:d

[rupy]

It's a fact, lightbulbs used to last 2500 hours, then the manufacturers agreed they would guarantee 1000 hours... Ink jet printers for example have electronic chooking once the pad soaking up the ink gets wet, you can install software that ignores it. I have stopped buying "consumer" products. Professional gear doesn't have fail counters and are built to last.

[BR0KK]

Were not talking about printers or lightbulps here; these ztex boards are custom designed and not cheap...
I wouldnt count them as a consumer product, more likely a enterprise solution.

Hardware degeneration is normal and to be expected. Like some Parts of your car need to be changed from time to time.

[rupy]

I know, but I would like to be able to control this custom designed enterprise solution to suit my setup.

[vv01f]

I know, but I would like to be able to control this custom designed enterprise solution to suit my setup.

You might ask to license the layout and built it on your own or with partners.

I personally think ztex is quite qualified to judge his work and how he does it. Or did you hear/read of one badly built boards by him?
I didnt. But I read about things normally boards doant outlive that well.

But if youre right in your position, I doubt ztex wont listen . Just be patient.

[sadpandatech]

ztex, can you please upgrade the AOZ1025 buck regulator on your designs to another able to output more than 8A to each FPGA, and more efficient? Multiple groups (eldentyrell, bitfury, etc) are writing bitstreams pushing the LX150 beyond 8A... And the AOZ1025 is not super efficient.

Perhaps the IR3840MPbF (86% efficient at 12A output at 1.2V) ?

I can't disagree on wanting available a higher possible amp and not to sound like an Aosmd fan boy, but it's a shame they don't appear offer any regs at higher amperage. :/

the current one is rated 95%
http://www.aosmd.com/res/data_sheets/AOZ1025DI.pdf



On the one you linked, is the output voltage adjustable? And does thaty company by chance make a higher eff rated reg?

[mrb]

I can't disagree on wanting available a higher possible amp and not to sound like an Aosmd fan boy, but it's a shame they don't appear offer any regs at higher amperage. :/

the current one is rated 95%
http://www.aosmd.com/res/data_sheets/AOZ1025DI.pdf

No. The current one is 75% efficient for a 1.2V@8A output, see page 6 of http://www.aosmd.com/res/data_sheets/AOZ1025DI.pdf

Mine is 88% efficient for the same output, see page 9: http://www.irf.com/product-info/datasheets/data/ir3840m.pdf (output voltage is adjustable just like the AOZ1025, by changing resistors).

[sadpandatech]

I can't disagree on wanting available a higher possible amp and not to sound like an Aosmd fan boy, but it's a shame they don't appear offer any regs at higher amperage. :/

the current one is rated 95%
http://www.aosmd.com/res/data_sheets/AOZ1025DI.pdf

No. The current one is 75% efficient for a 1.2V@8A output, see page 6 of http://www.aosmd.com/res/data_sheets/AOZ1025DI.pdf

Mine is 88% efficient for the same output, see page 9: http://www.irf.com/product-info/datasheets/data/ir3840m.pdf (output voltage is adjustable just like the AOZ1025, by changing resistors).

Thanks for taking te time to flip through those specs to point that out. Probably a good call to migrate off Aosmd anyhows then. ;p  They rating 95% with reg at the lowest input then? jerks.

[ztex]

ztex, can you please upgrade the AOZ1025 buck regulator on your designs to another able to output more than 8A to each FPGA, and more efficient? Multiple groups (eldentyrell, bitfury, etc) are writing bitstreams pushing the LX150 beyond 8A...

The current design can deliver up to 240 MH/s (tested).

New designs will not appear within the next three months

And the AOZ1025 is not super efficient.

How do you determine that? The efficiency mainly depends from the external components, namely the lower mosfet and the inductor.

Perhaps the IR3840MPbF (86% efficient at 12A output at 1.2V) ?

I made some bad experiences with that series.

[ztex]

Ok, could you add a zero error rate command... that lowers the frequency until error rate is = 0.00% forever?

Yes, a error rate limit this can be added easily. I put it on the todo list for the next release (in June).

[ztex]

But if youre right in your position, I doubt ztex wont listen . Just be patient.

I listen, but I do not reply to such kind of statements

[yohan]

Manufacturers of regulator chips usually headline the highest efficiency achieveable which is usually at the smallest input/output difference. It's not unusual for a 12V to 1.2V regulator circuit to be around 75-80% even if the headline says 90%+. Best efficiency is rarely at maximum current either.

Yohan

[luffy]

BTC payment is not accepted?