| Electronics > Projects, Designs, and Technical Stuff |
| Low power, battery operated design with dual linear regulators |
| << < (4/4) |
| Srijal97:
--- Quote from: mariush on September 15, 2019, 09:06:03 pm ---Yes, you could use diodes. Once you get the high current regulator working and turn off the low current regulator, you could short out the diode on the high current regulator using a mosfet with low Rds(on) ... so you won't lose a lot of current on that diode. Or you could use a p channel mosfet to keep the output of that high current regulator disconnected until needed. Something else you could do would be to use specialized ICs for this task, but this would add more cost compared to a diode. For example here's an ideal diode : https://www.digikey.com/product-detail/en/texas-instruments/LM66100DCKR/296-53541-1-ND/10273271 Here's another: https://www.digikey.com/product-detail/en/maxim-integrated/MAX40200AUK-T/MAX40200AUK-TCT-ND/7599791 There's also controllers or multiplexers which allow you to switch between 2 or several inputs... but they're expensive. --- End quote --- Those ideal diodes are really nice! I see the "Dual Ideal Diode ORing of supplies" section in the LM66100 datasheet and it seems perfect for what I want to achieve. I didn't know that switching regulators are better even at such low currents, thanks for pointing it out. So I think the SC189(3.3V fixed, 1.5A) would be good for the high current part but I'm a bit doubtful of the low current section. The sleep currents would be about 5-10uA, would the switching regulator S-85S1P still be better than the linear TPS797 at that time? They mention efficiency for loads of 100uA, but I'm a 1/10th lower. Other than that, I think the TPS27081A PFET load switch would be good for turning the peripherals ON/OFF. I've attached the planned schematic for now, is it alright? |
| mariush:
for SC189 it would make sense to use it for higher currents (like 0.5-1A) as that's what it's optimized for. You can look in datasheet - https://www.semtech.com/uploads/documents/sc189.pdf - at page 5 and you see there two charts on the right side. 1. Efficiency vs Load Current (Vout = 3.3v) with a curve for 4.0v and a blue curve for 5v 2. Efficiency vs Load Current (Vin 5v , V out = 3.3v) with multiple curves depending on inductor chosen So you can see in first chart that with 4v (what would be typical for a lithium battery, let's ignore the 5v curve on that chart) , with the recommended inductor (or one with same properties) they think you're gonna get around 90% efficiency at 0.1A and peaks at round 0.3A ... 0.6A at around 93-94%and then above 0.6A the efficiency will slowly go down towards 90% Below around 0.1A, the efficiency goes down significantly, but it's still gonna be around 75% at let's say 25mA (the smallest tick on the x asis is 0.1A but let's estimate) So let's say your lithium battery's voltage right now is 4v and let's compare this regulator to linear regulator at these loads : The linear regulator's efficient is same across the whole range... voltage goes in, voltage goes out .. same amount of current + tiny current consumption of ldo. Vin ... 100% Vout = ? % => ? % = Vout x 100 / Vin = Vout x 100/4 = 25 x Vin = 82.5% The switching regulator will be 75% at 25mA, 90% at 0.1 , ~93% at 0.3..0.6A , ~90% at 1.5A So the linear regulator will only be more efficient at super low loads. If the battery voltage changes a bit, let's say it goes down to 3.6v .. the linear regulator becomes a bit more efficient efficiency is : 3.3v x 100 / 3.6 = ~91.6% The switching regulator may still be more efficient at 0.3..0.6A range, but be less efficient outside this range. However, keep in mind that the linear regulator will have a minimum voltage drop. For example, the linear regulator may have 0.1v drop at 0.1A, 0.25v at 0.3A and 0.4v at 1A or more. In this case, the regulator would output 3.3v at 0.1A, and will output 3.3v at 0.3A (because 3.6v - 0.25v = 3.35v) but at 1A, the linear regulator will only output 3.2v (because 3.6v - 0.4v = 3.2v) So at 3.6v, the linear regulator may become unusable. But let's look at that Torex regulator with a maximum 200mA output... See datasheet - https://www.torexsemi.com/file/xc9265/XC9265.pdf - Right on the first page, you will see a chart for the version of the chip with 1.8v output and you can see the curves for 2.7v, 3.6v and 4.2v - you an estimate similar curves for 3.3v output version But what's cool in that chart is that even at 0.1 mA the efficiency of the chip is already at over around 85-90% and stays relatively flat up to 100mA or so. So you could use this regulator for 3.3v < 100mA, and switch to the SC189 regulator from 100mA and higher. You'll have 85-90% efficiency below 100mA, and then you'd have 90% to 93-94% up to around 0.6A and then slowly go down towards 90% efficiency. A switching regulator will also be able to work with much lower input-output differential (linear regulators have to deal with that voltage drop on the internal transistor) Even at 1mA sleep power consumption, a linear regulator will be as low as ~78% efficient, while this Torex IC will be around 85-90% efficient, if they don't lie in datasheet. That extra 10% efficient is not something to sneeze at. |
| Srijal97:
--- Quote from: mariush on September 16, 2019, 05:50:02 pm ---for SC189 it would make sense to use it for higher currents (like 0.5-1A) as that's what it's optimized for. You can look in datasheet - https://www.semtech.com/uploads/documents/sc189.pdf - at page 5 and you see there two charts on the right side. 1. Efficiency vs Load Current (Vout = 3.3v) with a curve for 4.0v and a blue curve for 5v 2. Efficiency vs Load Current (Vin 5v , V out = 3.3v) with multiple curves depending on inductor chosen So you can see in first chart that with 4v (what would be typical for a lithium battery, let's ignore the 5v curve on that chart) , with the recommended inductor (or one with same properties) they think you're gonna get around 90% efficiency at 0.1A and peaks at round 0.3A ... 0.6A at around 93-94%and then above 0.6A the efficiency will slowly go down towards 90% Below around 0.1A, the efficiency goes down significantly, but it's still gonna be around 75% at let's say 25mA (the smallest tick on the x asis is 0.1A but let's estimate) So let's say your lithium battery's voltage right now is 4v and let's compare this regulator to linear regulator at these loads : The linear regulator's efficient is same across the whole range... voltage goes in, voltage goes out .. same amount of current + tiny current consumption of ldo. Vin ... 100% Vout = ? % => ? % = Vout x 100 / Vin = Vout x 100/4 = 25 x Vin = 82.5% The switching regulator will be 75% at 25mA, 90% at 0.1 , ~93% at 0.3..0.6A , ~90% at 1.5A So the linear regulator will only be more efficient at super low loads. If the battery voltage changes a bit, let's say it goes down to 3.6v .. the linear regulator becomes a bit more efficient efficiency is : 3.3v x 100 / 3.6 = ~91.6% The switching regulator may still be more efficient at 0.3..0.6A range, but be less efficient outside this range. However, keep in mind that the linear regulator will have a minimum voltage drop. For example, the linear regulator may have 0.1v drop at 0.1A, 0.25v at 0.3A and 0.4v at 1A or more. In this case, the regulator would output 3.3v at 0.1A, and will output 3.3v at 0.3A (because 3.6v - 0.25v = 3.35v) but at 1A, the linear regulator will only output 3.2v (because 3.6v - 0.4v = 3.2v) So at 3.6v, the linear regulator may become unusable. But let's look at that Torex regulator with a maximum 200mA output... See datasheet - https://www.torexsemi.com/file/xc9265/XC9265.pdf - Right on the first page, you will see a chart for the version of the chip with 1.8v output and you can see the curves for 2.7v, 3.6v and 4.2v - you an estimate similar curves for 3.3v output version But what's cool in that chart is that even at 0.1 mA the efficiency of the chip is already at over around 85-90% and stays relatively flat up to 100mA or so. So you could use this regulator for 3.3v < 100mA, and switch to the SC189 regulator from 100mA and higher. You'll have 85-90% efficiency below 100mA, and then you'd have 90% to 93-94% up to around 0.6A and then slowly go down towards 90% efficiency. A switching regulator will also be able to work with much lower input-output differential (linear regulators have to deal with that voltage drop on the internal transistor) Even at 1mA sleep power consumption, a linear regulator will be as low as ~78% efficient, while this Torex IC will be around 85-90% efficient, if they don't lie in datasheet. That extra 10% efficient is not something to sneeze at. --- End quote --- Yes, the SC189 is good for higher currents. For the lower currents, I just need to keep the ESP under sleep at 5-10uA. If you see the efficiency curves of the switching regulators (Page 30 of the S-85S1P datasheet), I'll be going from an efficiency of 80% to 60% from 10uA to 1uA. Meanwhile the TPS797 linear regulator is at an efficiency of 78% at full battery voltage of 4.2V, and this goes up to 90% as the battery discharges. So I think that the linear regulator is a better option here. However, the quiescent current for the switching regulator is significantly lower(300nA vs 1uA). Does this make it better? |
| mariush:
If you're sure your circuit won't use more than 50mA, then the linear regulator may make more sense. It's a less complex circuit and the output voltage will be smoother compared to the output of a switching regulator. Note though that there may be better options for linear regulators than that regulator you chose, which costs 1$ For example: MIC5317 (3v) or MIC5365 or MIC5366 (3.3v) - https://www.digikey.com/product-detail/en/microchip-technology/MIC5366-3.3YMT-TZ/1611-MIC5366-3.3YMTTZCT-ND/5701178 Less than 10 cents if you buy 25 or more. Has higher current at 30uA but has as little as 20mV dropout voltage at <1mA |
| Srijal97:
I just need to make 2 devices, so I'm ready to pay more for better performance here. Isn't that ground pin current just too high? |
| Navigation |
| Message Index |
| Previous page |