L
Admin
Diode-ring (and many passive) mixer datasheets quote conversion loss and IP3 at a single LO drive. On the bench, LO level is a knob that trades **conversion loss, linearity, port isolation, and spurs** — and the "+7 to +10 dBm" habit is only a starting point.
**What actually moves**
1. **Conversion loss vs LO**
Too little LO → diodes not fully switching → higher conversion loss, worse isolation, soft compression.
Enough LO → loss flattens near the datasheet number.
Excess LO → diminishing returns on loss, more LO leakage / heating, and sometimes worse IF noise floor from LO noise sidebands.
2. **IP3 / spur behaviour**
Linearity usually improves as LO rises through the recommended window, then flattens or even degrades if the LO chain itself is dirty or the mixer is driven into a region the datasheet never showed.
Habit: sweep LO in 1 dB steps across ±3 dB of the recommended drive and plot **conversion loss, IF output IP3 (or two-tone IMD), and LO–RF leakage** on the same axes.
3. **LO purity matters as much as LO power**
A "+10 dBm" LO with poor harmonic content or high broadband noise can look fine on a power meter and still trash close-in spurs or noise figure. Pad then amplify (or filter then amplify) so the mixer sees a clean drive at the target level.
**Minimum bring-up notes to file**
- Recommended LO from datasheet **and** the LO level you actually used (at the mixer LO pin, not at the synth output).
- Conversion loss at that LO (CW RF in → IF out).
- Two-tone IMD or output IP3 at the same LO / RF power.
- LO leakage at RF and IF ports.
- Whether an LO filter / buffer / isolator sits between synth and mixer.
**Common traps**
- Measuring LO at the generator and ignoring cable + pad + balun loss into the mixer.
- Comparing IP3 across boards at different LO levels and calling one "better".
- Raising LO to "fix" a spur that is actually an external pickup or a bias issue.
**Discussion**
On your last diode-ring or passive mixer bring-up, did conversion loss or IP3 settle first as you swept LO — and at what drive did LO leakage become the binding constraint?
— rfforum 编辑部 / Editorial Team
**What actually moves**
1. **Conversion loss vs LO**
Too little LO → diodes not fully switching → higher conversion loss, worse isolation, soft compression.
Enough LO → loss flattens near the datasheet number.
Excess LO → diminishing returns on loss, more LO leakage / heating, and sometimes worse IF noise floor from LO noise sidebands.
2. **IP3 / spur behaviour**
Linearity usually improves as LO rises through the recommended window, then flattens or even degrades if the LO chain itself is dirty or the mixer is driven into a region the datasheet never showed.
Habit: sweep LO in 1 dB steps across ±3 dB of the recommended drive and plot **conversion loss, IF output IP3 (or two-tone IMD), and LO–RF leakage** on the same axes.
3. **LO purity matters as much as LO power**
A "+10 dBm" LO with poor harmonic content or high broadband noise can look fine on a power meter and still trash close-in spurs or noise figure. Pad then amplify (or filter then amplify) so the mixer sees a clean drive at the target level.
**Minimum bring-up notes to file**
- Recommended LO from datasheet **and** the LO level you actually used (at the mixer LO pin, not at the synth output).
- Conversion loss at that LO (CW RF in → IF out).
- Two-tone IMD or output IP3 at the same LO / RF power.
- LO leakage at RF and IF ports.
- Whether an LO filter / buffer / isolator sits between synth and mixer.
**Common traps**
- Measuring LO at the generator and ignoring cable + pad + balun loss into the mixer.
- Comparing IP3 across boards at different LO levels and calling one "better".
- Raising LO to "fix" a spur that is actually an external pickup or a bias issue.
**Discussion**
On your last diode-ring or passive mixer bring-up, did conversion loss or IP3 settle first as you swept LO — and at what drive did LO leakage become the binding constraint?
— rfforum 编辑部 / Editorial Team
