Hi all,
My Logic MSO works fine in Logic 2, but every capture through the saleae-mso-api Python package fails during device initialization on macOS. Posting the details in case someone has seen this or Saleae can point me at what’s different between Logic 2 and the SDK’s recorder.
Summary
mso_api.MSO()connects andmso.part_numberreturns correctly.- Any
mso.capture(...)then fails after about 60 s withMsoCommandError. - In the recorder log, initialization reaches 100% (firmware, FPGA, DDR, ADC and JESD all succeed). The failure is the last step:
MsoDeviceInterface::SetEnabledSmartCablePorts, which sendsScPdProcessingEnable, enabled=0x1f, disabled=0xffe0. That USB control transfer fails with:[0xe00002eb] (iokit/common) operation aborted; DeviceRequestTO failed(first cable), or[0xe00002ed] (iokit/common) device not responding; DeviceRequestTO failed(second cable).
- The bulk read stream then aborts, the device re-enumerates (
Device connected while awake), and the recorder reportsFailed: 'MSO not ready'.
Environment
| Item | Value |
|---|---|
| Device | Logic MSO, MsoPod_2ch_100Mhz_9b_100Ms |
| Boards | MsoMain_1_5_GB Rev0B; AFE MsoAfe_2ch_9b Rev0E |
| Firmware | 1.1.0, build date Nov 12 2026, info 76f1ac3 (the SDK logs “Firmware is up-to-date”) |
| Host | macOS 27.2.1, Apple Silicon (arm64) |
| Python | 3.14.7 (also tried 3.12 with 0.5.6) |
| SDK | saleae-mso-api 0.5.8 from downloads.saleae.com/pypi. 0.5.6 fails identically. 0.5.0 and 0.4.0 have no macOS build. |
| Logic 2 | 2.4.50, quit completely before running the SDK. The MSO captures normally in it on the same machine, port and cable. |
| USB | Enumerates at 5 Gb/s. The recorder reports a 20 V / 1.5 A supply and “PD contract established”. |
What I’ve ruled out
- Logic 2 holding the device (fully quit; with it open I get a different “exclusive access” error)
- Reboot
- Plugging the scope directly into the Mac, no hub or dock
- Two different USB cables
- With and without a probe on channel 1
- Capture settings: defaults, settings matched to Logic 2 (AC, 1x, 2 V range, 12.5 MS/s, 100 mV rising trigger, 10 ms window), and a 1 s
TimedCapture - The unmodified example from the Getting Started guide fails the same way
- Firmware: reported up-to-date by the SDK itself
Reproduction
from pathlib import Path
from saleae import mso_api
mso = mso_api.MSO()
config = mso_api.CaptureConfig(
enabled_channels=[mso_api.AnalogChannel(channel=0, name="clock")],
analog_settings=mso_api.AnalogSettings(sample_rate=100e6),
capture_settings=mso_api.TimedCapture(capture_length_seconds=0.1),
)
mso.capture(config, save_dir=Path("my-captures") / "00") # fails after ~60 s
Questions
- Is
saleae-mso-api0.5.8 supported on macOS 27 / Apple Silicon with the 2-channel 100 MHz MSO? Which OS and SDK versions have you validated? - Is firmware 1.1.0 (build Nov 12 2026) the version 0.5.8 is meant to work with?
- Why would the
ScPdProcessingEnablecontrol transfer succeed in Logic 2 but time out in the SDK’smso_data_recorder? - Is there a debug option or environment variable to get more detail from the recorder?
- Is the
logic2-automation1.2.0 headless preview with MSO support the recommended path on macOS whilemso_apiis being replaced?
Full recorder logs are attached (device serial, board IDs and my username redacted; the repeated DDR polling lines are collapsed and marked):
log1_raw_sdk_capture.txt: raw SDK capture (1 MS/s, 4 V range, 0 V rising trigger, 1 ms window)log2_second_usb_cable.txt: settings matched to Logic 2, second USB cablelog3_getting_started_example.txt: the Getting Started example, unmodified
Thanks!
log1_raw_sdk_capture.txt (21.0 KB)
log2_second_usb_cable.txt (20.8 KB)
log3_getting_started_example.txt (20.6 KB)