GekkoScience Compac BM1384 Stickminer Official Support Thread - page 78.

notlist3d

legendary

Activity: 1456

Merit: 1000

Quote from: sidehack on December 17, 2015, 05:59:40 PM

The specific information I posted is that cgminer-gekko does not currently recognize the EU sticks. Stock cgminer also doesn't recognize the EU sticks as Compacs but enumerates as U3, which has limited frequency options and defaults to a higher clock speed than the factory stock setting. BFG works better, but stock cgminer should still work.

Stock CGMiner works fine as icarus. I have set it up on my RPI to have a original and a compac-cgminer. So I run 2 instances which it is very capable of doing and you can select miners.

That is workaround I have used so far. Seems to do fine.

sidehack

legendary

Activity: 3374

Merit: 1859

Curmudgeonly hardware guy

The specific information I posted is that cgminer-gekko does not currently recognize the EU sticks. Stock cgminer also doesn't recognize the EU sticks as Compacs but enumerates as U3, which has limited frequency options and defaults to a higher clock speed than the factory stock setting. BFG works better, but stock cgminer should still work.

bix0

newbie

Activity: 15

Merit: 0

Quote from: Biodom on December 17, 2015, 03:53:32 PM

Quote from: sidehack on December 17, 2015, 09:17:18 AM

Anyone trying to use cgminer-gekko with bitshopper.de sticks, it's not going to work until we put out an update. So stop fretting and just download BFG 5.4

Anyone looking at that one guy's spreadsheet for power estimates, stop. For the seventh time, that's current into the ASIC, not into the USB. For the seventh time, take the current and voltage into the ASIC, multiply them to get the power into the ASIC, divide by 90% to get buck losses, then divide by your USB voltage to get USB current. P=IV, which means 5Vx1.4A gets you 7W. So the chip is pulling about 6.3W (90%), which for 240MHz probably means about 700mV about 9A. Which is way too high and you need to turn your voltage down, because these will run 300MHz off about 6.7W USB power.

as ^^^, bfgminer 5.4.1 works fine (and the only one that works on EU sticks), and it DOES work on sticks connected to usb 3.0 hub, which is then connected to unpowered usb 2.0 hub, which is then connected to rasp pi.

I wouldn't vouch for direct connection of ALL usb 3.0 powered hubs and pi, but SOME do work (one example that was confirmed multiple times is 7 slot 70W usb 3.0 Superbpag).
here are super detailed instructions (you can just read comments below) re bfgminer and pi:
https://www.youtube.com/watch?v=hOLDrSiI_ic

Quote from: Kevinthijs on December 17, 2015, 05:57:52 AM

Having the same problem as most users
Stick is connect to power USB 3.0 hub - connect to 2.0 (unpowered) hub - connect to rasberry (latest raspbian)

is dectected with driver but when starting CGminer, receive error: no usb devices detected.
With the option U & L, I can see the usb device as "known"
But cannot activate it.

Any suggestions?

Cgminer works fine too on euro sticks. I have my euro one running now correctly. Wich doesn't work fine is the modified version of gekko.

Biodom

legendary

Activity: 3892

Merit: 4331

Quote from: sidehack on December 17, 2015, 09:17:18 AM

Anyone trying to use cgminer-gekko with bitshopper.de sticks, it's not going to work until we put out an update. So stop fretting and just download BFG 5.4

Anyone looking at that one guy's spreadsheet for power estimates, stop. For the seventh time, that's current into the ASIC, not into the USB. For the seventh time, take the current and voltage into the ASIC, multiply them to get the power into the ASIC, divide by 90% to get buck losses, then divide by your USB voltage to get USB current. P=IV, which means 5Vx1.4A gets you 7W. So the chip is pulling about 6.3W (90%), which for 240MHz probably means about 700mV about 9A. Which is way too high and you need to turn your voltage down, because these will run 300MHz off about 6.7W USB power.

as ^^^, bfgminer 5.4.1 works fine (apparently, cgminer works on EU sticks as well listing them as U3 ~~and the only one that works on EU sticks~~), and it DOES work on sticks connected to usb 3.0 hub, which is then connected to unpowered usb 2.0 hub, which is then connected to rasp pi.

I wouldn't vouch for direct connection of ALL usb 3.0 powered hubs and pi, but SOME do work (one example that was confirmed multiple times is 7 slot 70W usb 3.0 Superbpag).
here are super detailed instructions (you can just read comments below) re bfgminer and pi:
https://www.youtube.com/watch?v=hOLDrSiI_ic

Quote from: Kevinthijs on December 17, 2015, 05:57:52 AM

Having the same problem as most users
Stick is connect to power USB 3.0 hub - connect to 2.0 (unpowered) hub - connect to rasberry (latest raspbian)

is dectected with driver but when starting CGminer, receive error: no usb devices detected.
With the option U & L, I can see the usb device as "known"
But cannot activate it.

Any suggestions?

zOU

hero member

Activity: 728

Merit: 500

★ these are stars ★

Quote from: RichBC on December 14, 2015, 04:31:14 PM

The spreadsheet is expecting to see the ASIC current which there is no easy way to measure. You have entered the USB current and in that case the only sensible thing to multiply it by is the USB voltage which will give a measurement of the wattage of the complete stick.

Rich

I created the spreadsheet but honestly I had no idea what was supposed to be entered in the "Current (A)" column.

I just re-used the datasheet table.

But yes, what you say makes sense.

I'll update the spreadsheet which is now unshared.
I'll open it again when I get something usable/reliable.

zOU

hero member

Activity: 728

Merit: 500

★ these are stars ★

Quote from: sidehack on December 17, 2015, 11:36:42 AM

If you're seeing the flat at about 9 o'clock, you're probably running more like 750mV. That's more than you need for 240MHz, so the 7W reading is starting to make sense. For 240MHz you probably want the flat at more like 6 o'clock. 200MHz I see about 4W draw, and 300MHz is about 6.5, so 240MHz you'll probably see around 5W when it's set for stable and efficient running.

you were right and my memory was flawed.

I just measured Vcore at 750mA.

Anyway.

I'll start again on monday as I now have the USB doctors.
Hopefully I'll have some quiet time so I can figure the relation between the figures in the BM1384 datasheet (I understand they're for the chip only, so I'll have to add 10% for the rest of the stick components)

0.75VCore = 1.4Amp on the display => 5Vx1.4A=7W

sidehack

legendary

Activity: 3374

Merit: 1859

Curmudgeonly hardware guy

If you're seeing the flat at about 9 o'clock, you're probably running more like 750mV. That's more than you need for 240MHz, so the 7W reading is starting to make sense. For 240MHz you probably want the flat at more like 6 o'clock. 200MHz I see about 4W draw, and 300MHz is about 6.5, so 240MHz you'll probably see around 5W when it's set for stable and efficient running.

zOU

hero member

Activity: 728

Merit: 500

★ these are stars ★

It's been a while since I set them up, so i'll check VCORE tomorrow.

Flat bit of the pod is at 90degres from the top

sidehack

legendary

Activity: 3374

Merit: 1859

Curmudgeonly hardware guy

If you're running 240MHz, you're either not running 600mV or getting errors up the butt. They take 640mV to run stable at 200MHz. And if you're running at 600mV and drawing 1.4A you're not running at 240MHz; it might be 1A at that speed. I've worked with the ASICs since about April and I've had literally every GS Compac made so far on a meter in front of me at one time or another, all of which compounds to tell me either you're reporting information wrong or you're doing something very very weird.

zOU

hero member

Activity: 728

Merit: 500

★ these are stars ★

Quote from: sidehack on December 17, 2015, 09:17:18 AM

Anyone trying to use cgminer-gekko with bitshopper.de sticks, it's not going to work until we put out an update. So stop fretting and just download BFG 5.4

Anyone looking at that one guy's spreadsheet for power estimates, stop. For the seventh time, that's current into the ASIC, not into the USB. For the seventh time, take the current and voltage into the ASIC, multiply them to get the power into the ASIC, divide by 90% to get buck losses, then divide by your USB voltage to get USB current. P=IV, which means 5Vx1.4A gets you 7W. So the chip is pulling about 6.3W (90%), which for 240MHz probably means about 700mV about 9A. Which is way too high and you need to turn your voltage down, because these will run 300MHz off about 6.7W USB power.

My sincere apologies sidehack.

I did not mean any harm I was just trying to be helpful despite my obvious lack of knowledge.

The spreadsheet is not viewable anymore.

P=IV, which means 5Vx1.4A gets you 7W ==> that I get.
which for 240MHz probably means about 700mV about 9A. => that I don't get how you came to these numbers

Which is way too high and you need to turn your voltage down, because these will run 300MHz off about 6.7W USB power. => I do not understand why 0.6Vcore would be too high for 240Mhs ?

My apologies.

sidehack

legendary

Activity: 3374

Merit: 1859

Curmudgeonly hardware guy

Anyone trying to use cgminer-gekko with bitshopper.de sticks, it's not going to work until we put out an update. So stop fretting and just download BFG 5.4

Anyone looking at that one guy's spreadsheet for power estimates, stop. For the seventh time, that's current into the ASIC, not into the USB. For the seventh time, take the current and voltage into the ASIC, multiply them to get the power into the ASIC, divide by 90% to get buck losses, then divide by your USB voltage to get USB current. P=IV, which means 5Vx1.4A gets you 7W. So the chip is pulling about 6.3W (90%), which for 240MHz probably means about 700mV about 9A. Which is way too high and you need to turn your voltage down, because these will run 300MHz off about 6.7W USB power.

zOU

hero member

Activity: 728

Merit: 500

★ these are stars ★

Quote from: hedgy73 on December 14, 2015, 04:23:12 PM

Ok great, multimeter shows 0.69v between gnd and vcore pads.

USB meter shows 4.89v and 1.4a to stick.

Frequency is 300MHz.

If I enter these figures it doesn't look right.

It looks like the 1.4a the USB meter shows is incorrect compared to the other readings in your table.

I agree !!!

I received my USB doctor and I plugged them in today.

My sticks are running at 240Mhs/0.6Vcore and readings are between 1.36A and 1.55A on my 6 sticks, where they should be closer to 6A !!

I'm confused...

Unless W = 5V x Vcore x Amp reading ?

so 5x0.6x1.4=4.2W, which seems better.

but then I do not know where the Amp values from the datasheet come from...
Current(A)
12.31
9.2
7.44
6.6
4.46
3.43

AJRGale

hero member

Activity: 767

Merit: 500

Quote from: Kevinthijs on December 17, 2015, 05:57:52 AM

Having the same problem as most users
Stick is connect to power USB 3.0 hub - connect to 2.0 (unpowered) hub - connect to rasberry (latest raspbian)

is dectected with driver but when starting CGminer, receive error: no usb devices detected.
With the option U & L, I can see the usb device as "known"
But cannot activate it.

Any suggestions?

same as everyone else, the rPI doesn't do usb3 hubs well (if at all), plug the stick into the usb2 hub.

you can pull up a terminal and test to see if the system sees it by typing:

Code:

lsusb

you should see a list of USB items and hubs plugged in, including internal, what you are looking for is:

Code:

Bus ### Device ###: ID 10c4:ea60 Cygnal Integrated Products, Inc. CP210x UART Bridge / myAVR mySmartUSB light

Kevinthijs

newbie

Activity: 1

Merit: 0

Having the same problem as most users
Stick is connect to power USB 3.0 hub - connect to 2.0 (unpowered) hub - connect to rasberry (latest raspbian)

is dectected with driver but when starting CGminer, receive error: no usb devices detected.
With the option U & L, I can see the usb device as "known"
But cannot activate it.

Any suggestions?

hedgy73

legendary

Activity: 1414

Merit: 1077

Quote from: AJRGale on December 17, 2015, 02:57:15 AM

Quote from: hedgy73 on December 17, 2015, 02:43:16 AM

Quote from: AJRGale on December 16, 2015, 09:22:45 PM

~~Well, I'll attempt to recreate the "issue" and I'll get another screen-cap.~~ I thought it was funny watching the little thing pulling 150GHs, I was waiting for smoke to bellow out!

Since the thing is sitting on a EEEPC701, to much voltage pulled would shut the system down, not make it pop a fuse.. I can see the green light dropping intensity when voltage starts getting pulled..

I haven't seen that either but then I'm not overclocking that much or using hubs as powerful as yours.

I cant believe you wanted to re-create it, perhaps you really are as crazy as the guy in your avatar Cheesy

.

Well, that was my face when it first happened, that surprised disbelieved look

Ha ha I can believe that Grin

.

AJRGale

hero member

Activity: 767

Merit: 500

Quote from: hedgy73 on December 17, 2015, 02:43:16 AM

Quote from: AJRGale on December 16, 2015, 09:22:45 PM

~~Well, I'll attempt to recreate the "issue" and I'll get another screen-cap.~~ I thought it was funny watching the little thing pulling 150GHs, I was waiting for smoke to bellow out!

Since the thing is sitting on a EEEPC701, to much voltage pulled would shut the system down, not make it pop a fuse.. I can see the green light dropping intensity when voltage starts getting pulled..

I haven't seen that either but then I'm not overclocking that much or using hubs as powerful as yours.

I cant believe you wanted to re-create it, perhaps you really are as crazy as the guy in your avatar Cheesy

.

Well, that was my face when it first happened, that surprised disbelieved look

hedgy73

legendary

Activity: 1414

Merit: 1077

Quote from: AJRGale on December 16, 2015, 09:22:45 PM

~~Well, I'll attempt to recreate the "issue" and I'll get another screen-cap.~~ I thought it was funny watching the little thing pulling 150GHs, I was waiting for smoke to bellow out!

Since the thing is sitting on a EEEPC701, to much voltage pulled would shut the system down, not make it pop a fuse.. I can see the green light dropping intensity when voltage starts getting pulled..

I haven't seen that either but then I'm not overclocking that much or using hubs as powerful as yours.

I cant believe you wanted to re-create it, perhaps you really are as crazy as the guy in your avatar Cheesy

.

AJRGale

hero member

Activity: 767

Merit: 500

~~Well, I'll attempt to recreate the "issue" and I'll get another screen-cap.~~ I thought it was funny watching the little thing pulling 150GHs, I was waiting for smoke to bellow out!

Since the thing is sitting on a EEEPC701, to much voltage pulled would shut the system down, not make it pop a fuse.. I can see the green light dropping intensity when voltage starts getting pulled..

RichBC

hero member

Activity: 588

Merit: 500

I have just run mine at 400MHz overnight without problem, Vcore at 0.75V. I have a big fan with a temperature controller with a sensor on the heatsink and keep the temp at 45 Deg C.

Tried 425Mhz but too many HW/ error and am having trouble getting the core voltage up much more. Pot does not seem to like being too near the end of it's travel and runs for a while but then drops to 0.65V and everything stops. Need to sort out the 5V supply as I am adjusting on load for 5V out of the Hub, but off load it rises to 5.5V.

However not seen the runaway effect described.

Rich

garpsky

newbie

Activity: 10

Merit: 0

Hi
I just oc one of mine to 400mhz on bfgminer running on an old whs v1 HP Datavault x510 its hovering around 24gh/s with a temp of 55 degree's C ( not sure what temps are safe?)
I have took a couple of photos but is there an easy way to post them using the forum as I had a mare the last time with bucket!

getting back to yours I hope shes still working after what happened?
this is the first time I have had a miner and these sticks are great to learn with.

I would love to use CGminer on the server but I cant get it to work ( not very good with software ) but at least I got it to work with BFGminer

Cheer's

Topic: GekkoScience Compac BM1384 Stickminer Official Support Thread - page 78. (Read 268027 times)