highperfocused/bitaxe
1
0
Fork 0
mirror of https://github.com/skot/bitaxe.git synced 2025-09-19 14:41:10 +02:00
Code Issues Packages Projects Releases Wiki Activity

updated readme, finished routing

Browse Source
This commit is contained in:
Skot Croshere
2022-06-27 16:04:36 -07:00
parent e6feaa443b
commit 349a0156a3
4 changed files with 4540 additions and 1632 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6129
bitaxePro.kicad_pcb
View File

File diff suppressed because it is too large Load Diff

BIN
doc/BM1397.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 1004 KiB

BIN
doc/render.png
View File

Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 552 KiB

After

Width:  |  Height:  |  Size: 324 KiB

43
readme.md
View File

@@ -1,54 +1,51 @@
# The bitaxe
![](doc/assembled.png)
an early-stage experimental bitcoin mining machine. Uses two Bitmain BM1387 ASICs.
an early-stage experimental bitcoin mining machine. Uses a single Bitmain BM1397 ASICs.
## BM1387
![bm1387 pic](doc/bm1387.png)
![BM1397 pic](doc/BM1397.png)
- The BM1387 is a undocumented SHA256 mining ASIC from Bitmain. It's mostly used in the Antminer S9
- Bitmain claims the BM1387 has 0.098j/Gh efficiency
- The BM1387 is available for around $1-2 each in low quantities from Aliexpress
- The BM1397 is a undocumented SHA256 mining ASIC from Bitmain. It's mostly used in the Antminer S17
- Bitmain claims the BM1397 has 0.045j/Gh efficiency
- The BM1397 is available for around $20 each in low quantities from Aliexpress
- The BM1397 has the same footprint as the BM1387, but a very different pinout.
## Current Status
- I have built several v1 PCBs
- v1 has a problem with the pin pitch on the BM1387 footprint.
- it's very unlikely *all* of the pins were making contact.
- this has been corrected on v2, but not verified yet. Still waiting for PCBs to arrive.
- The voltages all appear to be right
- I can't get any response over serial from the BM1387s yet.
- The current comsumption does change like I would expect when I send a serial command to change the frequency.
- I have not built up any or tested any of these PCBs yet. Hopefully soon!
## Hardware
- [BM1387B from random AliExpress seller](https://www.aliexpress.com/item/2251832867687077.html). Do these work?? Who knows!
- [BM1397 from random AliExpress seller](https://www.aliexpress.com/item/3256802274958527.html). I got the "AG" variant. Not really sure what the difference is.
- [40x40mm heatsink and 5V fan](https://www.aliexpress.com/item/2251832861666365.html) from a random AliExpress seller. At least half of these arrived broken in some way. But they are cheap and the working ones do keep the BM1387's nice and cool when used with some thermal compound.
- The serial port is 1.8V. The [FTDI TTL-232RG-VREG1V8-WE](https://www.digikey.com/en/products/detail/ftdi,-future-technology-devices-international-ltd/TTL-232RG-VREG1V8-WE/2441359) USB adapter does the trick, although it's pretty expensive.
- [jim.sh Micro1v8 FT230X](https://www.amazon.com/dp/B076B9YRMP) adapters also work.
- Neither the FTDI cable or the jim.sh adapter breakout any of the CBUS pins that could be used for the reset line.
- for that I'm going to try the [UMFT231XE-01](https://www.digikey.com/en/products/detail/ftdi-future-technology-devices-international-ltd/UMFT231XE-01/4487117) breakout board.
- I added level shifters for the interface with the [ESP32-C3](https://docs.espressif.com/projects/esp-idf/en/latest/esp32c3/hw-reference/esp32c3/user-guide-devkitc-02.html#user-guide-c3-devkitc-02-v1-header-blocks) (Because it has 3.3V GPIOs)
- I added a NCT218 so that the BM1397 core temperature can be read out over I2C
- [KiCad 6](https://www.kicad.org) design files
- All of the parts on the board are listed in the KiCad BOM
## Software
Check my [BM1387 Scripts](https://github.com/skot/bm1387_scripts)!
- still TBD
## Connections
![](doc/render.png)
| Pin Name | Description |
| ----------- | ----------- |
| VIN | ASIC VIN - powers both ASICs in series. Should be around 0.9V for 0.450V at each ASIC. Will need at least 3A, maybe more! This needs more experimentation. |
| +5V | 5V input gets regulated to 1.8V and 0.8V for the BM1387. Also 5V for Fan output. Not much current required. |
| +5V (Fan) | 5V output for fan |
| SPD (Fan) | Speed input from fan. Doesn't currently go anywhere |
| 1V8 | 1.8V output. regulated from 5V. Use as a IO reference, or leave unconnected. |
| VIN | ASIC VIN - ASIC core voltage. Rumor has it this should be around 1.5V for the the BM1397. This needs more experimentation. |
| +5V | 5V input gets regulated to 1.8V and 0.8V for the BM1397. Also 5V for Fan output. Not much current required. If the ESP32 is powered by USB, you don't need to power this. |
| +5V (Fan) | 5V output for fan . |
| SPD (Fan) | Speed input from fan. |
| VIO | ESP32-C3 GPIO voltage (3.3V) |
| RXI | 1.8V Serial input |
| TXO | 1.8V Serial output |
| RST | Inverted reset pin to the first BM1387. Unsure how necessary this is. |
| RST | Inverted reset pin to the first BM1397. Unsure how necessary this is. |
| GND | Common ground |
## Goals
- Figure out the BM1387 communication and put all of that in an onboard microcontroller like the ESP32-C3 (RISC-V, yum).
- Connect up the BM1387 internal temperature to the onboard microcontroller.
- Connect up the fan speed sensor to the onboard microcontroller.
- Develop some firmware for the ESP32 to connect to a stratum server/pool over WiFi and get work for the mining ASIC
- Firmware should monitor fan speed
- Firmware should monitor BM1397 core temperature (via NCT218)
- Develop some software in Rust to getblocktemplate from a bitcoin node and communicate work to a fleet of bitaxes.
- add an onboard power sensor so we can set the hashrate (ASIC requency) and therefore current to extract maximum power from a solar panel.