• Resolved

TPS61165: URGENT issue: entering Easyscale instead of PWM mode

Expert 5145 points

Part Number: TPS61165

Dear team,

On of my customer is observing production faillure with TPS61165 which is going into easyscale mode, despite the fact their 0.2% @ 5kHz PWM should keep it in PWM mode.

The generated pulse is 400ns, which is way longer than the minimum documented in the application note (50ns). http://www.ti.com/lit/an/snva769/snva769.pdf


This was not solve using 1% pulse (2us, 5kHz), it started showing positive results with 2% pulses (4us)… which is almost 100 times above the recommendation.


Can someone help here ?

Thanks a lot,

Best regards,


  • Hi Pial,
    In which situation, your customer has entered into ES mode? During the start up or dimming period? Have you prepared the pwm dimming waveform with Figure 9 in datasheet? Since the ES detect delay and ES detect time both longer than the 400ns.
  • Hi PA,
    The minimum “Easy Scale detection time” is about tes_det_min=213us*(1-25%)=160us. You need to make sure the first off pulse’s width is less than 160us. For your 5kHz with 0.2%, the off width is about 1/5k*99.8%=199.6us, which is larger than 160us.
    So there is a simple method to solve this problem: increase the switching frequency, the minimum switching frequency with 0.2% should abouve: 1/(0.16/99.8%)=6.24kHz, so you can change the dimming frequency to 6.5kHz or higher.

    The first detection delay time 100us is not important, because the internal bandgap will enable when it receive a rising edge on CTRL pin, 100us was the safet time for internal bandgap start ready. The 160us calculation results is based on the maximum varivation, may be the corner data for your customer is about 196us, I’m not sure, but please considering the worst case as I mentioned to choose the safety dimming frequency.
    For ES mode detection, it will need a rising edge to turn on the internal bandgap, and it will start counting 256 times when it detects falling edge on CTRL pin for ES mode detection. That make sense for you?
    For Actia application, the first rising edge may has triggered the internal bandgap, when it detect the falling edge on CTRL pin, then after 199.us off time, it entered into ES mode, maybe the second triggered the internal count due to internal bandgap need some time to start up ready.