FL5160 Dimmer Issues

[machinehum]

I'm building a device for a 120V, 60Hz dimmer application. I'm having some severe issues with Onsemi's FL5160. I'm making this thread ideally for help but also to let others know, I believe there's might be something wrong with this chip.

Some notes on the operation.

• The DIM_Control pin takes an analog voltage and dims the output as a function of that voltage. It also sources 10uA.
• I'm injecting a DC voltage directly into TP1, so the driving circuitry/LPF on the left can be ignored. JP3 is closed and C3, R10 and R4 are populated.
• JP4 is jumped shut, so the low power mode of the chip can be ignored. Q4, R6 are populated

The failure mode is as follows. When the DC voltage (into TP1) is turned on and then the AC power comes up, the device works properly. However, if the AC comes up first, then the DC control voltage, the device will not enable the output, regardless of the DC voltage.

Datasheet: https://www.mouser.com/datasheet/2/149/FL5150-964157.pdf



Any ideas or suggestions would be highly appriciated   

[moffy]

"I'm injecting a DC voltage directly into TP1, so the driving circuitry/LPF on the left can be ignored. JP3 is closed and C3, R10 and R4 are populated."

I am not sure how that injected voltage is grounded, but there are all sorts of possible problems in using an external supply because of possible switch on surges. If it has to be external use a battery that will have no coupling to the mains, but preferably use the VDD supply to control the DIM_Control, and if you use a potentiometer treat it as live, use a plastic pot preferably, but again treat it as live, also treat the external battery as live if you use one, safety first.

[machinehum]

"I'm injecting a DC voltage directly into TP1, so the driving circuitry/LPF on the left can be ignored. JP3 is closed and C3, R10 and R4 are populated."

I am not sure how that injected voltage is grounded, but there are all sorts of possible problems in using an external supply because of possible switch on surges. If it has to be external use a battery that will have no coupling to the mains, but preferably use the VDD supply to control the DIM_Control, and if you use a potentiometer treat it as live, use a plastic pot preferably, but again treat it as live, also treat the external battery as live if you use one, safety first.

Thanks for the reply. The injected voltage is grounded via the pin5 or GND of the FL5160. If you're talking about earth ground, the supply isn't tied to earth ground. I will try to remove the DC supply from the equation and use the VDD of the FL5160, thanks for the suggestion.

[machinehum]

I've made a striking new development. Things start working when connecting the bench power supply common (not earth ground) to the GND of my circuit and power on. All that is required is this common connection during power up.

[moffy]

That is excellent news, congrats.

[machinehum]

That is excellent news, congrats.

Well, not really, I don't have a bench supply in the final design. And after some inspection, it's not repeatable. I'm going to have to go with the two-wire mode, this is because if the voltage at the DIM_Control pin is over 4V is will force a POR. Forcing this POR is the only way I've been able to reliably get the device to work...

[moffy]

Well, not really, I don't have a bench supply in the final design. And after some inspection, it's not repeatable. I'm going to have to go with the two-wire mode, this is because if the voltage at the DIM_Control pin is over 4V is will force a POR. Forcing this POR is the only way I've been able to reliably get the device to work...

Forcing a reset after powering on is good practice. It means you start from a known state, not something indeterminate.

[machinehum]

Well, not really, I don't have a bench supply in the final design. And after some inspection, it's not repeatable. I'm going to have to go with the two-wire mode, this is because if the voltage at the DIM_Control pin is over 4V is will force a POR. Forcing this POR is the only way I've been able to reliably get the device to work...

Forcing a reset after powering on is good practice. It means you start from a known state, not something indeterminate.

Makes sense, but as it is, there is no way to reset the chip in 3wire mode. I'm switching to 2wire mode because, as per the datasheet "If the DIM Control voltage is >4 V a 100% duty cycle is selected and the DRV Gate will be on 100%. However, a 100% duty cycle can only occur for a 3-wire application. If a 2-wire application is used and the DIM Control pin voltage is >4 V a POR will occur"

This POR is required to fix the improper operation I am seeing above.