forked from sinara-hw/assembly
Compare commits
5 Commits
f4fd9ac0a1
...
9929476733
Author | SHA1 | Date |
---|---|---|
mwojcik | 9929476733 | |
Egor Savkin | 401c65d4f1 | |
Egor Savkin | 65d73c6cff | |
morgan | b174819fbb | |
morgan | 4b03c538a1 |
|
@ -2,6 +2,7 @@
|
|||
|
||||
- [Build and test firmware](./build_test_firmware.md)
|
||||
- [Hardware](./hw/hardware.md)
|
||||
- [Sinara Kasli](./hw/kasli.md)
|
||||
- [Sinara Kasli-SOC](./hw/kasli_soc.md)
|
||||
- [Sinara 4624 AWG Phaser (Upconverter/Baseband)](./hw/phaser.md)
|
||||
- [Sinara 4456 synthesizer Mirny / Sinara 4457 Almazny Mezzanine card](./hw/mirny_almazny.md)
|
||||
|
|
|
@ -0,0 +1,9 @@
|
|||
# Kasli
|
||||
|
||||
## Mounting fan onto heatsink
|
||||
![](../img/kasli_fan.jpg)
|
||||
1. ⚠️ Verify the fan has the **correct polarity (powering on with wrong polarity will burn the MOSFET in series💥)**
|
||||
2. Place the fan on a heatsink
|
||||
3. Tap 3 threads on the heatsink using M2.5 pointy tapping screws (e.g. front panel screws)
|
||||
4. Replace the tapping screws with M2.5x14mm screws
|
||||
5. Verify the fan is secure
|
|
@ -5,4 +5,12 @@
|
|||
Check the BOOT mode switches - they both should be at SD if the Kasli-SoC going to be shipped to customer.
|
||||
POR jumper needs only for JTAG mode.
|
||||
|
||||
![](../img/kasli_soc.jpg)
|
||||
![](../img/kasli_soc.jpg)
|
||||
|
||||
## Mounting fan onto heatsink
|
||||
![](../img/kasli_soc_fan.jpg)
|
||||
1. ⚠️ Verify the fan has the **correct polarity (powering on with wrong polarity will burn the MOSFET in series💥)**
|
||||
2. Place the fan on a heatsink
|
||||
3. Tap 3 threads on the heatsink using M2.5 pointy tapping screws (e.g. front panel screws)
|
||||
4. Replace the tapping screws with M2.5x14mm screws
|
||||
5. Verify the fan is secure
|
||||
|
|
|
@ -30,6 +30,22 @@ Synchronization requires Kasli and Urukul to be clocked from the same oscillator
|
|||
why this feature is disabled by default.
|
||||
There is no intrinsic impact on Urukul output phase noise and the synchronization process is quick and reliable when done correctly.
|
||||
|
||||
### One-EEM mode
|
||||
|
||||
Users may choose to use only one EEM port, if they want more cards to be in their crate. However following features
|
||||
will become unavailable:
|
||||
* SU-Servo
|
||||
* Low-latency RF switch control
|
||||
* Synchronization
|
||||
|
||||
RF switches are still available but the commands need to go over the SPI bus so it's higher-latency and lower-resolution.
|
||||
|
||||
### Urukul 4412
|
||||
|
||||
Urukul 4412 has higher frequency resolution (47 bit against 32 at Urukul 4410), however lacks such features:
|
||||
* SU-Servo
|
||||
* Synchronization
|
||||
|
||||
## Testing
|
||||
|
||||
After running `artiq_sinara_test`:
|
||||
|
@ -132,4 +148,25 @@ matches real clocker source.
|
|||
ValueError: Urukul AD9910 AUX_DAC mismatch
|
||||
```
|
||||
|
||||
Ensure it is the AD9910 and not the AD9912. Also check SUServo pins are set up respective to the JSON description.
|
||||
Ensure it is the AD9910 and not the AD9912. Also check SUServo pins are set up respective to the JSON description.
|
||||
|
||||
### Jagged signal with 1GHz external clock on AD9910
|
||||
|
||||
By default, on AD9910 external clock signal is divided by 4, while it should be not divided at all with PLL disabled.
|
||||
Change the ``clk_div`` parameter to the CPLD in the device_db file:
|
||||
|
||||
```python
|
||||
device_db["urukulX_cpld"] = {
|
||||
"type": "local",
|
||||
"module": "artiq.coredevice.urukul",
|
||||
"class": "CPLD",
|
||||
"arguments": {
|
||||
"spi_device": "spi_urukul0",
|
||||
"sync_device": None,
|
||||
"io_update_device": "ttl_urukul0_io_update",
|
||||
"refclk": 1000000000.0,
|
||||
"clk_sel": 1,
|
||||
"clk_div" : 1 # <--- add this line
|
||||
}
|
||||
}
|
||||
```
|
|
@ -55,4 +55,20 @@ Press ENTER when done.
|
|||
### High-freq audible noise and output values all near -0.1 on Zotino v1.4.2
|
||||
|
||||
This may happen when power-cycle is too short. Power down the crate, wait at least 30 seconds, and power up again.
|
||||
[Issue](https://github.com/sinara-hw/Zotino/issues/37).
|
||||
[Issue](https://github.com/sinara-hw/Zotino/issues/37).
|
||||
|
||||
### Zero voltage output on Fastino
|
||||
|
||||
Some Fastino may not output any voltage during testing, usually that means it has no gateware.
|
||||
|
||||
Another common symptom of no gateware is that no LEDs are lit up. Whereas if the gateware has been flashed, the PG and FD LEDs will be lit green.
|
||||
|
||||
You can flash the gateware with a standalone Kasli/Kasli-SoC:
|
||||
|
||||
1. Download the latest `fastino.bin` release from [quartiq/fastino](https://github.com/quartiq/fastino/releases)
|
||||
2. Run `git clone https://github.com/quartiq/kasli-i2c.git` and place `fastino.bin` in the kasli-i2c directory
|
||||
2. Connect the Fastino's EEM0 to any available Kasli/Kasli-SoC EEM port (**do not hot-plug**)
|
||||
3. Power on the standalone Kasli/Kasli-SoC
|
||||
4. Run `nix-shell -p python311Packages.pyftdi`
|
||||
5. Run `cd kasli-i2c; python flash_fastino.py 0 EEM<number> write fastino.bin` where `<number>` is the EEM port number on the Kasli/Kasli-SoC side
|
||||
6. If PG and FD LEDs are lit green, the Fastino is ready.
|
||||
|
|
Binary file not shown.
After Width: | Height: | Size: 134 KiB |
Binary file not shown.
After Width: | Height: | Size: 216 KiB |
Loading…
Reference in New Issue