LNK3294G 5V 80mA Buck-boost topology support
In new design I'm using LNK3294G IC (165V to 400V AC application). Design will be running on either 240VAC or 400VAC continously, required output voltage is 5V, current draw will be switching between 20mA and 80mA every ten or so minutes.
Can such low power output be achieved with LNK3294G? While design is based on PIXI, design guide (used 725V, could not find 900V version, assume most of guidelines apply) and reference designs, resulting design is working sufficiently well but there are some points I would like assistance with.
1. Inductor selection for such low output is a bit problematic. PIXI suggested value is 570uH (very low), at the moment two boards with 820uH and 1000uH are running, PIXI suggests to "Verify SOA limit is not violated or increase L". Can you explain a bit on what parameters to check?
Also are 820uH and 1000uH inductance values suitable? Going higher than 1000uH results in clearly audible acoustic noise (magnetic resin SMT inductors are used).
2. Is reduced current limit or standard current limit better (provided this application is feasible)? And can R2 be disconnected since voltage from output voltage after diode is going to be lower than bypass pin voltage?
评论
Changed in board C4 to 1uF (RED current limit)
Attaching PIXI sheets. Two variants - one suggested by default and one updated to our design.
Attaching voltage waveforms across switch (LNK3294G) and across inductor pins (files are named) with differential probe.
Also added some 1000uH and 1500uH waveforms (switch).
Waiting for your suggestions.
From our perspective - it looks like 5V 20mA output is somewhat troubling to reach as device operates in low frequency, mosfet on-time is near datasheet minimum.
12V with same output current would probably be better option and easier to achieve
| Attachment | 大小 |
|---|---|
| Default solution suggestions based on parameters (468.83 KB) | 468.83 KB |
| Edited some parameters in solution (468.99 KB) | 468.99 KB |
| 820uH_20mA_inductor_420VAC_1wfs.jpg (84.62 KB) | 84.62 KB |
| 820uH_20mA_inductor_420VAC_5wfs.jpg (84.52 KB) | 84.52 KB |
| 820uH_20mA_inductor_420VAC_100wfs.jpg (102.55 KB) | 102.55 KB |
| 820uH_20mA_switch_420VAC_1wfs.jpg (83.2 KB) | 83.2 KB |
| 820uH_20mA_switch_420VAC_5wfs.jpg (83.64 KB) | 83.64 KB |
| 820uH_20mA_switch_420VAC_100wfs.jpg (103.09 KB) | 103.09 KB |
| 820uH_80mA_switch_420VAC_1wfs.jpg (83.3 KB) | 83.3 KB |
| 820uH_80mA_switch_420VAC_5wfs.jpg (83.12 KB) | 83.12 KB |
| 820uH_80mA_switch_420VAC_100wfs.jpg (100.14 KB) | 100.14 KB |
| 1000uH_80mA_switch_420VAC_1wfs.jpg (83.52 KB) | 83.52 KB |
| 1000uH_80mA_switch_420VAC_5wfs.jpg (84.17 KB) | 84.17 KB |
| 1500uH_80mA_switch_420VAC_1wfs.jpg (82.9 KB) | 82.9 KB |
| 1500uH_80mA_switch_420VAC_5wfs.jpg (85.27 KB) | 85.27 KB |
Hi Darius,
Yes you are correct, the frequency is very low so the regulation is probably not very good. Transients may be a problem as well. The controller will not behave brilliant in close to no-load power condition, with close to the minimum on-time.
The SOA is internal protection checking if the current reaches the limit too quick after initial blanking time. If this happen for some number of pulses the controller will auto restart assuming short. However, during the blanking time where the parasitic capacitors are charging/discharging there is no limit. I cannot see auto-restarting in the pulses. The noise probably comes from the low frequency and jittering.
I am not familiar with the spec, but will you load work with 12V? You can just increase the artificial load with R5.
Best Regards,
Tested R5 replacement with electronic load across full current range but performance we are looking for is near current limit, heat dissipation could become a problem.
Since there are reference designs available for 12V and 16V, we'll update our design to be able to operate with 12V outputs.
Is something like XT2 900V would be better (flyback) for very low power non-isolated output? Non-isolated because we need positive output referenced to mains to operate triac in quadrants 2 and 3.
Hi Darius,
If 12V is possible as output for your application it will be easier to achieve better operation point. You are correct, flyback is even better because you can vary the turns ratio.
Because of the non-isolated output the transformer can use the same wire as the primary and the coupling will be better than the TIW.
Best Regards,
Hi Darius,
Thank you for the question.
I need more information in order to get myself familiar with your design. Could you please post here
- the design PIXLS spreadsheet.
- some inductor waveforms (2-5 switching cycle interval and 100+ switching cycle interval)
- Switch voltage waveforms as above
for the 820uH inductor case.
Best Regards,