 Show Posts
|
|
Pages: [1] 2 3 4 5 »
|
So, the updated firmware is at https://github.com/zipiju/klondike/tree/master/firmware/K16/v1.0. You have to change two lines in CGM driver - driver-klondike.c lines: 419: (unsigned int)K_NONCE(kline->wr.nonce) - 0x180); 861: uint32_t nonce = K_NONCE(kline->wr.nonce) - 0x180; I was talking to Kano the other day, and there are plans to include those changes to the upstream CGM. But I need a little more time to do that. And also, the firmware will have to be slightly changed again, because now the new K16 is recognized as the old one. And one more thing, in driver-klondike.c change line 86 to #define LATE_UPDATE_MS ((int)(1 * 1000)) And do not use klondike-options in cgminer.conf for now. Also, you can change the freq. in klondike.h firmware file. There is the DEFAULT_HASHCLOCK value, you can change that one down to 600. It is possible, that core regs will be shutting down when freq. will be set to 900 because of the power draw. I'll be back in about 11 hours. Bye. |
|
|
|
So, the updated firmware is at https://github.com/zipiju/klondike/tree/master/firmware/K16/v1.0. You have to change two lines in CGM driver - driver-klondike.c lines: 419: (unsigned int)K_NONCE(kline->wr.nonce) - 0x180); 861: uint32_t nonce = K_NONCE(kline->wr.nonce) - 0x180; I was talking to Kano the other day, and there are plans to include those changes to the upstream CGM. But I need a little more time to do that. And also, the firmware will have to be slightly changed again, because now the new K16 is recognized as the old one. |
|
|
|
Just an update, still working on it. I've removed all the ASICs. The short is no longer present, so now I have to solder the spares back on.
Hopefully, it will be completed tomorrow (I'm leaving to work in about 45 minutes, so can't work on it now).
I also have modified firmware in hand, so I'll be able to test it right away
@zipiju I have made two PCB(ver. K16s-1.0.1)  I found mismatch Schematic and Partlist. There are some errors in the Partlist. Part Schematic PartList ------------------------------ R21 10K 390R R45,R46 390R 10K R22 100R 49K9 R24,R25 1K 2K2 ------------------------------ R45,46 should be 390R. R24,25 should be 1K. R22 should be 100R and the R21 should be 10K. Basically, always look on the schematics, because BOM can be wrong. And is it working? I mean no magic smoke escaping? I'll upload the modified firmware to the Git in a while tomorrow morning (in about 8 hours). BTW, I wanted to finish mine today, but 2 hours that I had free today wasn't enough to do so. It wil l be the same tomorrow and friday too. I'm really sorry, but nothing I can do about this. I will have to wait till the weekend to finally complete it. |
|
|
|
|
Just an update, still working on it. I've removed all the ASICs. The short is no longer present, so now I have to solder the spares back on.
Hopefully, it will be completed tomorrow (I'm leaving to work in about 45 minutes, so can't work on it now).
I also have modified firmware in hand, so I'll be able to test it right away
|
|
|
|
@zipiju : still alive ? Hope you didn't put you on vfib on 3.3V rail  I'm still here, but lacking any free time. So please just wait till the weekend. |
|
|
|
14 ohm across the 3.3 volt supply seems kind of normal.
That is only 235ma of current draw. Seem like that shouldn't burn up the regulators.
You're right. The continuity test on MM here doesn't display resistivity. I've measured it again in R mode, and it shows 0 Ohms (when I substract resistivity of the MM probes). |
|
|
|
Zipiju,
That thar is some nice looking PCB work. Just to be clear, those are Gen2, 55nm Avalon chips AND you're chasing down a short somewhere .... right??? Correct. Short on a 3.3V rail to the GND. Solder bridge shorts are a true PITA to work out. On replacing the Avalon chips, I've done tight package replacements using just a hardware store heat gun (on high). I made an aluminum foil shield to cover the components around the BGP package device I was replacing (so as not to melt their solder) and then ever so gingerly heated the part I needed to replace to a point where it would release from the board using my chip suction device. It's not quite a hot air re-flow station BUT it does the trick. To replace the part I tinned the pads on the PCB, put the part on the board, held it in place with a long rigid wire (because I didn't want to toast my fingers in the process) and again used the heat gun to re-flow the tinned pads. If don't want to pre-tin the pads and you happen to have solder paste laying around you can dab a small amount (and I mean a pin head amount) onto the device pads and then, using your chip suction device, line the device up with the PCB pads and drop it onto the board. The heat gun process applies once the part is stuck to the board. On a final note, I always part with a bit of advice. Do NOT drop that hot heat gun on your junk  . In fact, try to NEVER do electronics work wearing only a Speedo. Safety first  On the topic of solder bridges that can happen with REALLY tight pitched electronics used on the K16S-55nm board have you considered building yourself one of these -> http://www.instructables.com/id/DIY-Soldering-Reflow-Oven/ . I did and it makes doing tight pitch packages as ell as BGM type packages SUPER easy (using a stencil and solder paste). H@shKraker I've been reworking Avalons on the first few K16, so I have a practice  . The issue here is, that it can really be any component, from a capacitor to Avalons itself. Another thing is, that I was in a work from 1PM till now (11PM), so I'm not in a state to start working on that now. I'm sorry, but I'll try tomorrow morning. But as it looks, it won't be probably fixed before weekend  . |
|
|
|
It seems to me that the first generation chip(avalon).
Is right?
Those are Gen2, but Avalon made a mistake with that batch, so they are marked as A256. I was told that by the guys from the Avalon when they were sending them to me. |
|
|
|
zipiju I sure hope you can find that short with a visual inspection.
Otherwise it's going to suck.
How low of resistance is the short?
It's really low, like 14 Ohms. I'll be working on that later today. But it seems, that I will have to remove all components that are using 3.3V rail, one by one, because no visible bridge is there. EDIT: Maybe that burned 3.3V reg just punched through one of the ASICs, or NOR or something. |
|
|
|
Guys, just an update. I've assembled K16s v1.0. As a precaution, I've measured conductivity between powerrails and grounds as usuall, but unfortunatelly, 3.3V reg burned out anyways . It was working for a while, so solder bridge probably caused that (it happened to me before). I've replaced it, but now the 3.3V rail is shorted to the ground somewhere, so I'm debugging that now. I've tripple checked the schematics and the datasheets today, and it's 100% correct, so this is really just an issue with solder bridges. I just hope that the short is not beacuse of burned out ASIC, beacuse replacing those it's pain in the a**. As soon as I'll have something new, I'll post it ASAP. can we has a pic?  |
|
|
|
Guys, just an update. I've assembled K16s v1.0. As a precaution, I've measured conductivity between powerrails and grounds as usuall, but unfortunatelly, 3.3V reg burned out anyways . It was working for a while, so solder bridge probably caused that (it happened to me before). I've replaced it, but now the 3.3V rail is shorted to the ground somewhere, so I'm debugging that now. I've tripple checked the schematics and the datasheets today, and it's 100% correct, so this is really just an issue with solder bridges. I just hope that the short is not beacuse of burned out ASIC, beacuse replacing those it's pain in the a**. As soon as I'll have something new, I'll post it ASAP. |
|
|
|
Yes, Zipiju is waiting on PCBs to do initial Avalon Gen2 assembly and bring up testing.
H@shKraker
Guys, sorry for the delays, and my absence here. Things are happening here, so I'm short on the free time. I'm going to pick the PCBs up tomorrow (today) morning. They are already cleared by the customs. Unfortunatelly, I'm not at home, so can't do the assembly tomorrow. I'm returning home friday night, and will be assembling the PCBs at saturday. So expect some (hopefully good) news this weekend. So until then, hold tight . |
|
|
|
Thank you for imidiate answer. One more thing. Is it any KIT available with all parts and PCB (except avalon chips) paid by us $ or euro's. Thank you again I don't know about any. |
|
|
|
Hello
in K1 schematic there are 2 NOR PORT (U6) and 2 NOT PORT (U7).
Why not use the other 2 NOR in U6 transforming it in NOT PORT?
Thanks
The two NORs are used to extract the clock signal from data signal and delay it behind the data one. The two NOTs are used just to sharpen them up. It's basically the same design as in the K16. I'm not sure you can do the same with two NOTs instead of two NORs. Never thought about it, since it's working OK. |
|
|
|
Is it possible to inform me where i can find the latest bugless version of this K16 project? It's diifficult to search so many threads. Also i want to know if someone has it as a KIT. I have 18 avalon chips to test TIA The up to date versions are at https://github.com/zipiju/klondike |
|
|
|
How is it going with testing of last PCB version for 55nm Avalon?
Is anybody have assembled K16 Gen.2 version?
DHL called me like 2 hours ago, that the PCBs are in the country. Now the package has to go through the customs, so I'm expecting them delivered at the beginning of the next week. That's great. BTY, Can I ask something about klondike firmware? In klondike.h, at define section. there is an constant "21.33uS" in the comment. It is 0.00002133 sec. what means that? Where 0.00002133 is come from? I would say it's based on a PIC freq. (48MHz) and some prescaler value, so basically something like 1/(48x10 6/some value). And it would be the time it takes the Timer0 in the PIC to do one click. |
|
|
|
How is it going with testing of last PCB version for 55nm Avalon?
Is anybody have assembled K16 Gen.2 version?
DHL called me like 2 hours ago, that the PCBs are in the country. Now the package has to go through the customs, so I'm expecting them delivered at the beginning of the next week. Sounds like music in my ears :-) Is the BOM fully updated sow i can order the Mouser parts ?.. And if anyone would like to buy Gen2 Chips please sent me a PM ( can buy them for around 6$ ex shipping and paypal fee) and if u guys know a cheaper way please do pm as well Zipiju Thanks again bro. Well, I'll be probably changing the values of two resistors, so hold on for one more day. Meanwhile, I'll update the schematics and BOM. |
|
|
|
How is it going with testing of last PCB version for 55nm Avalon?
Is anybody have assembled K16 Gen.2 version?
DHL called me like 2 hours ago, that the PCBs are in the country. Now the package has to go through the customs, so I'm expecting them delivered at the beginning of the next week. |
|
|
|
Hello
on 3256 datasheet in recommended operating condition section we have:
a) Supply Core voltage DVDD 1.08 1.2 1.32 V
b) Supply 1.2V analog voltage VDDA_PLL 1.08 1.2 1.32 V
c) Supply I/O voltage OVDD 3.0 3.3 3.6 V
My question is:
How many current need any voltage?
Thanks
Nobody really knows, but: If you want only a slight overclock, or no overclock at all, Core current should be about 2A per chip. If you want to overclock them hard, make that 3 or more. In the K16, the VDDA supply was about 75mA per chip (600mA LDO for 8 chips), and it was working OK. For the I/O, count the same (75mA/chip). |
|
|
|
Hello
I had received 10 3256 chip from Avalon.
I must realize 10 K1 usb stick or I can also build k2 with only 10 chip?
Thanks
Hi, No need to build 10 K1's. You can assemble 10 chips on a K16 (5 to each side), and then just change the Chipcount value in firmware to 10 and you will be good. |
|
|
|
|
