Tool/software:
Hi,
My customer is evaluating TLV431AIDBVR devices from a post-process-changed lot (PCN#20241217004.1) and have identified notable degradation in performance compared to previous lots.
■ Observations
The 4M lot exhibits worse static load regulation (ΔVo):
From approximately +5 mV → +25 mV under identical conditions (VIN = 76 V, IOUT = 1.35 A).
This voltage deviation correlates with a change in the shunt reference voltage (ΔVref), where the REF pin shows a periodic voltage dip of ~200 mV, synchronized with the inverter’s switching cycle. The drop in Vref becomes more pronounced at higher load currents, indicating increased noise sensitivity.
Adding a 2200 pF capacitor between the K and A terminals successfully mitigates both Vref fluctuation and output voltage deviation—restoring performance close to the original lot.
○ ΔVo = Static Load Regulation [mV]
| Vin [V] | Before Change | 4M Lot | 4M + 1000pF | 4M + 2200pF | |
|---|---|---|---|---|---|
| 36 | 5.0 | 23.1 | 7.9 | 4.5 | |
| 48 | 4.3 | 23.4 | 7.7 | 4.5 | |
| 76 | 4.6 | 24.3 | 7.6 | 4.7 |
○ Shunt Reference Voltage ΔVref [mV]
| Vin [V] | Before Change | After Change | 4M + 1000pF | 4M + 2200pF |
|---|---|---|---|---|
| 36 | 1.6 | 4.9 | 1.7 | 1.3 |
| 48 | 1.5 | 4.9 | 1.8 | 1.3 |
| 76 | 1.7 | 5.5 | 2.0 | 1.4 |
■ Questions
1. Can you confirm whether any internal structural changes (e.g., ESD cell, internal compensation, output impedance) in the 4M lot could explain the increased noise susceptibility?
2. Given the above behavior, do you recommend using the K-A capacitor? If you have any countermeasures or comments, please let me know.
The capacitor addition is effective, but board space constraints make it impractical to apply this countermeasure in production.
Thanks,
Conor
