FEDEVEL
Platform forum

Eda003 , 09-16-2026, 01:55 PM

I’m debugging a boost converter section based on the LT8362 in a portable lamp project. I’ve attached the schematic.
The converter is supposed to boost from about 3.7 V (+SYSTEM) to roughly 35 V for the LED rail.
The problem is that when I enable the LT8362, the converter does not regulate. Instead:
- VLED oscillates roughly between 0.3 V and 3.7 V
- the lab supply current repeatedly ramps up to around 3 A
- it looks like the LT8362 is repeatedly hitting current limit / restarting
With the boost disabled:
- VIN = ~3.7 V
- EN/UVLO = 0 V
- INTVCC = 0 V
- +SYSTEM = ~3.7 V and stable
I have already checked the following:
- no short from VLED to GND
- no short from SW to GND
- Schottky diode DFLS260-7 is OK
- inductor is Vishay IHLP2525CZER6R8M01, 6.8 µH
- exposed pad / GND connection of LT8362 is ~0 Ω
- SW node continuity to the diode is good
- BIAS is tied to GND as allowed by the datasheet
- INTVCC is not shorted to GND
- FBX divider measures correctly:
- FBX → GND ≈ 48 kΩ
- VLED → FBX ≈ 1 MΩ
- RT → GND = 20 kΩ
- SS is not shorted
- SYNC/MODE is floating
The RT resistor is 20 kΩ, so the switching frequency should be around 2 MHz.
One thing I’m wondering about is whether 2 MHz is too aggressive for ~3.7 V → 35 V, causing the converter to hit current limit during startup. I’m considering temporarily changing RT to something like 45.3 kΩ (~1 MHz) to see whether behavior improves.

The PCB was fully assembled by JLCPCB, including the LT8362 and all surrounding passives, so I have fairly high confidence that the correct components were placed in the correct locations and that there are no obvious manual assembly mistakes.

, 09-16-2026, 02:03 PM

🤖 In the meantime, here’s a response from the [FEDEVEL AI HW Assistant]().
It helps to find relevant resources for your question, but it may make mistakes as it’s still in early development.

This is the response from the AI Assistant:
Based on the symptoms and your debug efforts, here are key considerations from similar scenarios in our hardware design videos:

1. **Current Limit Triggers During Startup**
Your observation of 3A current spikes suggests the inductor current exceeds the LT8362's 3.5A switch current limit (typical for this IC). At 2MHz switching frequency with a 6.8µH inductor, the **inductor's current ripple** might be too aggressive for this high step-up ratio (3.7V→35V).
- Inductor ripple current ΔI = (V_in * D) / (f_sw * L)
For D≈0.9 (extreme duty cycle), ΔI ≈ (3.7*0.9)/(2e6*6.8e-6) ≈ 2.5A
Peak current could reach I_load + ΔI/2, potentially exceeding 3A during startup ([Source](https://www.youtube.com/watch?v=zf-pvHysroM&t=5302))

2. **Bootstrap Circuit Verification**
Even though INTVCC isn't shorted, insufficient bootstrap capacitor charge can cause erratic switching:
- Check C_BST value (typ. 0.1-1µF) and its proximity to BST-SW pins
- Verify diode D1 in bootstrap path isn't reverse-leaking ([Example](https://www.youtube.com/watch?v=zf-pvHysroM&t=3812))

3. **Layout-Induced Issues**
JLCPCB assembly doesn't guarantee proper high-speed layout:
- SW node loop area must be minimized (<1cm²)
- Feedback path (FBX) should be routed away from SW/node
- Ground return paths for IC, input cap, and output cap must converge at single point ([Critical in EMC context](https://www.youtube.com/watch?v=fkNa-FejWsQ&t=2896))

4.

, 09-16-2026, 02:03 PM

**Soft-Start Configuration**
The SS pin's 10nF capacitor may be too small for this high ratio conversion:
- Try increasing to 100nF to slow startup current ramp
- Verify SS pin voltage with scope during enable ([Startup timing matters](https://www.youtube.com/watch?v=_DzpaKlDsWs&t=164))

5. **Frequency Adjustment**
Your plan to test 1MHz operation is sound:
- At 1MHz, inductor ripple current halves (ΔI≈1.25A)
- Reduces switching losses in the 35V output diode
- Confirms if frequency was causing premature current limiting ([Tradeoffs explained](https://www.youtube.com/watch?v=zf-pvHysroM&t=5302))

**Immediate Next Steps:**
1. Probe SW node with 10:1 probe (short ground lead)
- Look for clean switching waveform without excessive ringing
2. Check INTVCC voltage during operation
- Should be stable ~3.3V (not collapsing)
3. Temporarily add 10Ω resistor in series with RT pin
- Quick frequency test without desoldering

If these don't resolve, consider counterfeit IC possibility ([Example](https://www.youtube.com/watch?v=WGNlUFm0VV0&t=176)) - JLCPCB's parts sourcing can occasionally have issues.

You can continue chatting with the AI Assistant here:
https://fedevel.com/hw-assistant

Eda003 , 09-16-2026, 02:14 PM

The ripple-current calculation is off by about 10×. For 3.7 V, 6.8 µH, 2 MHz, and ~0.9 duty cycle, the ripple is roughly 0.24 A, not 2.5 A.
LT8362 does not use an external bootstrap capacitor/diode like a typical buck converter, so that part is not applicable :/

tomas.kaplan , 09-17-2026, 08:04 AM

Hi, have you tried to start the regulator unloaded or with just very small load of e.g. 50mAmps? I would also try the Burst mode with SYNC/MODE shorted to ground since majority of their application examples use this configuration.

How much current do you need out of the regulator?

QDrives , 09-22-2026, 07:56 PM

Vled should never be 0.7V when +System is 3.7V.
Have you checked SW with an oscilloscope?