EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: Psi on August 18, 2021, 06:02:57 am
-
I'm using a AP62301 switchmode regulator to convert 6V down to 3.2V at 5-30mA but am not having much lucky with any efficiency.
2.2uH inductor CIGT252008LM2R2MNE
low current test - input 10mA @ 6V = output load 0.5mA @ 3V
High current test - input 30mA at 6V = output load 22mA @ 3V
6.6uH inductor (three 2.2uH in series)
low current test - input 5mA @ 6V = output load 0.5mA @ 3V
High current test - input 19mA at 6V = output load 22mA @ 3V
My current suspicion is either
- The inductor is not suitable for some reason.
The reg runs at 750 kHz and the inductor is rated for 2.2uH at 1mhz with enough current headroom so i'm not sure why it wouldnt be suitable. I did try a 1.8uH that I also had but it was even worse.
The inductor is a little physically smaller than the datasheet shows on the PCB layout, but i only need 50mA max, not 3A so i picked a smaller one.
- Maybe by bottom layer track for the reg input is forming a turn too close to the inductor and causing some sort of feedback issue.
Anyone have other theories? Any help appreciated
The PCB layout and SCH is pretty much just copied from the datasheet.
Regulator datasheet
https://www.diodes.com/assets/Datasheets/AP62300_AP62301_AP62300T.pdf (https://www.diodes.com/assets/Datasheets/AP62300_AP62301_AP62300T.pdf)
Inductor datasheet
https://www.digikey.co.nz/product-detail/en/samsung-electro-mechanics/CIGT252008LM2R2MNE/1276-6925-1-ND/7041325 (https://www.digikey.co.nz/product-detail/en/samsung-electro-mechanics/CIGT252008LM2R2MNE/1276-6925-1-ND/7041325)
-
Yea that look's about right with that regulator. Have a look at Fig 3 in datasheet, you don't see 70% till your load current is >100mA
-
Yea that look's about right with that regulator. Have a look at Fig 3 in datasheet, you don't see 70% till your load current is >100mA
It says "Up to 83% Efficiency at 5mA Light Load" on page 1
So I was expecting something above 80% at my high current test of 22mA. but it took 19mA @ 6V to make 22mA at 3.2V which is 62%
But if anyone knows of a better chip for efficiency at this sort of current please let me know. (needs to be under $0.80 in 100 qty)
-
hm.. i see the problem
I have the AP62301 which is the shit version. I need to get AP62300.
The datasheet doesn't make that clear on page 1 statement about 83% at 5mA.
At least it's an easy fix and cost looks the same.
-
hm.. i see the problem
I have the AP62301 which is the shit version. I need to get AP62300.
The datasheet doesn't make that as clear on page 1 statement about 83% at 5mA.
At least it's an easy fix and cost looks the same.
That's the one ;^)
Gotta read those finer details
-
hm.. i see the problem
I have the AP62301 which is the shit version. I need to get AP62300.
The datasheet doesn't make that clear on page 1 statement about 83% at 5mA.
At least it's an easy fix and cost looks the same.
It's not about better or worse, they are different versions of the same chip with different operating modes, with different tradeoffs. The 301 version operates in forced CCM, which generates much less output ripple below Fsw, at the cost of low-load efficiency. The 300 version goes to pulse-skipping mode at light loads, greatly improving efficiency but at the cost of subharmonic ripple on the output, which can be difficult to filter out due to its low frequency content and very wide spectrum. This can especially be problematic in audio applications, because the subharmonics change with loading and often end up in the audible frequency range.
-
Why are you not using a linear regulator at such low currents? They could be MORE efficient and have no "burst mode" operation at the operating points you want.
-
Agree with Phoenix
And shouldn't have inductor some specifications about losses? Maybe I just don't see it, but there is nothing about it. At 750 kHz it can be huge compared to your load.
-
I'm trying to get maximum battery life out of two CR2032's running a 3.3V device
Also need to prevent 3.3V reg drop-out as the battery ESR starts to go up towards end of life.
I can either put them in parallel and boost to 3.3V or put in series and buck down to 3.3V.
My PCB does have a secondary footprint for a linear reg option, just in case switchmode doesn't work out.
That was the main reason I choose to put the cells in series and do boost. To keep a fallback linear reg option
-
I'm trying to get maximum battery life out of two CR2032's running a 3.3V device
Also need to prevent 3.3V reg drop-out as the battery ESR starts to go up towards end of life.
There are parts specifically designed for that:
https://www.ti.com/product/TPS63900 (https://www.ti.com/product/TPS63900)
or even simpler:
https://www.n-redc.co.jp/en/products/dc-dc-switching-regulator/spec/?product=rp604 (https://www.n-redc.co.jp/en/products/dc-dc-switching-regulator/spec/?product=rp604)
-
yeah, but I don't really need low current standby.
The circuit has a pfet which totally disconnects the battery until you hit a button to wake up the MCU which holds the pfet on until its done.
-
Low current standby often goes hand in hand with high efficiency at low power (Iq operating etc), look at the efficiency curves on those right down to the left. Buck-boost can be efficient.
-
With 6V in and 3.2V out, the best efficiency I would expect from a linear regulator would be 53%. And that does not include any quiescent current in the device itself which would only make it worse. He is already at 62% ("...but it took 19mA @ 6V to make 22mA at 3.2V which is 62%") so unless I am missing something, he can not gain anything by going to a linear.
Care to elaborate on that?
Why are you not using a linear regulator at such low currents? They could be MORE efficient and have no "burst mode" operation at the operating points you want.