Power supply design doubts

[barbito]

Hello!
I am trying to design my first project and I am currently stuck on the power supply design phase.

My project is based on the Quectel MC60 that must be powered by a standard 12V automotive battery.
By looking at the hardware design datasheet (attached to this post) this chip requires 2 supply voltages:

Legend: (min - norm - max)

• VBAT (3.3V - 4.0V - 4.6V) - Max burst of 2A
• GNSS_VCC (2.8V - 3.3V - 4.3V) - At least 150mA

As specified by the datasheet, due to the large difference in voltages it's better to use a switching regulator, but unfortunately, all the designs that I found online or in the datasheet are based on LDO.

I've tried to find a suitable buck converter but I am not sure what are the parameters that I should look for. I guess that temperature range and low noise regulation are mandatory for automotive environments.
Moreover, I am not sure if using 2 buck-boost converters is a better choice than 1 buck converter (12V -> 4V) and 1 LDO (4V -> 3.3V).
The only requirement that I have is that the power consumption remains as low as possible since I don't want to make my battery car dead in few weeks.

Can someone help me?
Thank you

 

[KlausST]

Hi,

I guess you need to go deeper int automotive power requirements.
Usally there are several power states like
* active (When motor is running)
* low power (when motor is off)
* power down (to prevent from battery deep discharge)

My first idea is to use a step down to 4V followed by a LDO to 3V3.
Do some cacluations if this meets the expected power consumption.

I don´t see where a buck-boost makes sense.

Automotive:
* temperature range
* rugged in terms of overvoltage spikes and introduced noise
* I see no need for a dedicated low noise power supply design. (Like used for sensitive analog circuitry)
* and the power supply needs to be able to handle load current bursts ... while keeping output voltage within specifications.
* and it needs to handle input voltage variations .. while keeping output voltage within specifications.
(Typical voltage range may be 10.5V ... 14.4V ... plus overvoltage spikes and maybe less voltage when motor gets started.)

"As low as possible" is a term for infinite effort. Not useful. There always will be an engineer and a new technology to save a few nanoAmperes ;-)
It´s more useful to give absolute values, or "power supply efficiency expectations" (@ the various power states).

Klaus

 

[barbito]

Thank you for the reply!
Indeed I have no idea about the automotive power requirements, I will try to find more information online.
In your response you refer to a generic 12-4v step down converter. Are you referring to a linear converter?

 

[KlausST]

Hi,

I agree "step down" is not unambiguous. But very often used as switch mode buck converter.

Klaus

 

[BradtheRad]

VBAT (3.3V - 4.0V - 4.6V) - Max burst of 2A

• GNSS_VCC (2.8V - 3.3V - 4.3V) - At least 150mA

By making the 3.3V dependent on the 4.0v supply, it becomes possible for those 2 Amp bursts to disrupt the stability of both supplies.

Is the 3.3V circuitry sensitive to spikes on its supply wire? Does it reset when its power supply drops below a certain level? If so then you ought to install two independent buck converters.

 

[Easy peasy]

A buck converter is a good choice here, 12 to 4.4 and then a 2nd buck or a lin reg* to 3v3 as it's only 150mA*

 

[BradtheRad]

This capacitor-based voltage reduction might be suitable. It resembles a Nakagome topology. However by selecting an undersize Farad value, the duty cycle effectively becomes low. C1 charges and discharges, as current goes to the output stage during both halves of the cycle.
Supply 12V. The load gets 3.3V at 150mA. It's unregulated, although regulation can be done by manipulating the bias frequency and duty cycle.
A certain amount of resistive drop is involved.