Design for power supply from source >12V low quiescent current

Moderators: grovkillen, Stuntteam, TD-er

Post Reply
Message
Author
Wiki
Normal user
Posts: 413
Joined: 23 Apr 2018, 17:55
Location: Germany

Design for power supply from source >12V low quiescent current

#1 Post by Wiki » 15 May 2020, 06:54

I am using my car only now and then. Sometimes when I want to use the car I am faced with the fact, that the battery is on low voltage. So I want to build a battery surveillance with ESPEasy in deep sleep mode using a low quiescent current LDO voltage regulator to minimize the power consumption of the circuit to get a warning from Domoticz on low battery voltage.

Currently I am trying to use a LT 1763-3.3 voltage regulator. The datasheet talks about an input voltage up to 20V. Works fine with input voltage below 12.5V. But with a battery voltage >12.5V the voltage regulator gets thermal problems, is not able to feed the Wemos D1 without overheating and cuts off the power supply after some seconds.

I've tried everything, including cooling cap.

Anyone any idea, which voltage regulator is able to work with an input voltage between 10.6V and roughly 14.5V (peak)) and feed my Wemos with 3.3V?
Last edited by Wiki on 15 May 2020, 18:26, edited 1 time in total.

Code: Select all

pi@raspberrypi:~ $ man woman
No manual entry for woman
pi@raspberrypi:~ $

Tecumseh
Normal user
Posts: 37
Joined: 30 Sep 2016, 09:42

Re: Design for power supply from source >12V

#2 Post by Tecumseh » 15 May 2020, 08:42

I am using a LM2940 for a LEDstrip module. It should be capable of input voltage until 26V.

http://www.ti.com/lit/ds/symlink/lm2940 ... 9524769520

Do make sure to add the recommended capacitors though otherwise it won't give you a stable 5V source.

Wiki
Normal user
Posts: 413
Joined: 23 Apr 2018, 17:55
Location: Germany

Re: Design for power supply from source >12V

#3 Post by Wiki » 15 May 2020, 13:19

Thank you for the tipp. But unfortunately these kinds of regulators are not useable for me because of their high quiescent current.

I should have pointed out, that a quiescent current below 1mA is a prerequisite for me, sorry.

Code: Select all

pi@raspberrypi:~ $ man woman
No manual entry for woman
pi@raspberrypi:~ $

User avatar
ThomasB
Normal user
Posts: 1064
Joined: 17 Jun 2018, 20:41
Location: USA

Re: Design for power supply from source >12V low quiescent current

#4 Post by ThomasB » 15 May 2020, 20:25

With a Linear LDO you're looking at 2-3 watts dissipation when the ESP is running, so it's going to get hot. Instead, consider a Buck DC-DC switching regulator.

For example, I have some Mini MP1584EN Adjustable DC-DC Buck Converter boards that are 250uA quiescent. About $1 on Aliexpress, but even Amazon sells them:
https://www.amazon.com/MP1584EN-DC-DC-C ... B01MQGMOKI

- Thomas

Wiki
Normal user
Posts: 413
Joined: 23 Apr 2018, 17:55
Location: Germany

Re: Design for power supply from source >12V low quiescent current

#5 Post by Wiki » 15 May 2020, 23:43

Tha k you, Thomas. Somewhat like them is exactly what I was looking for.

They tell "don't use without or with load less than 10% of output rating". Any experience what happens, if the Wemos goes to sleep?

Code: Select all

pi@raspberrypi:~ $ man woman
No manual entry for woman
pi@raspberrypi:~ $

TD-er
Core team member
Posts: 8643
Joined: 01 Sep 2017, 22:13
Location: the Netherlands
Contact:

Re: Design for power supply from source >12V low quiescent current

#6 Post by TD-er » 16 May 2020, 00:12

I use LM2596 or equivalent, whatever JLCPCB has in stock when I order my PCBs with them including them soldering the tiny parts for me :)
Ali Express also has lots of boards using those.

I don't know what the quiescent current is of those, but I think on a car battery you don't have to be that picky on that.
Even if it takes 10 mA as quiescent current (which is a lot), then a typical car battery will last months before it may have difficulty starting.
My unit mounted in my car, which is always on + has several sensors and GPS takes roughly 1 Watt (which is quite a lot)
That depletes the car battery in roughly 10 days to 11.6V which I think is still able to start my (diesel) car, but I let it warn me at that point just to be sure.

10 mA is roughly 8 times less, so that should last over 2 months, if no other items in the car draw current (e.g. stand-by of the car stereo)

What I usually do is DC/DC to 5V (because of some sensors) and then use a AMS1117 to get 3.3V
That last step is a linear regulator, which gives the cleanest output, but isn't that efficient.
On the other hand, if you only have it draw significant power for a short period, then the linear efficiency isn't that much of a problem. (1.5 times almost nothing is still almost nothing)

User avatar
ThomasB
Normal user
Posts: 1064
Joined: 17 Jun 2018, 20:41
Location: USA

Re: Design for power supply from source >12V low quiescent current

#7 Post by ThomasB » 16 May 2020, 00:29

They tell "don't use without or with load less than 10% of output rating". Any experience what happens, if the Wemos goes to sleep?
That shouldn't be a problem for you. Here is what the MP1584EN data sheet says about no-load conditions:

Code: Select all

At no load or light load, the converter may
operate in pulse skipping mode in order to
maintain the output voltage in regulation. Thus
there is less time to refresh the BS voltage. In
order to have enough gate voltage under such
operating conditions, the difference of VIN –VOUT
should be greater than 3V. For example, if the
VOUT is set to 3.3V, the VIN needs to be higher
than 3.3V+3V=6.3V to maintain enough BS
voltage at no load or light load.
- Thomas

TD-er
Core team member
Posts: 8643
Joined: 01 Sep 2017, 22:13
Location: the Netherlands
Contact:

Re: Design for power supply from source >12V low quiescent current

#8 Post by TD-er » 16 May 2020, 00:34

Such pulsating operation is also a good reason to use a linear regulator as the last step to 3.3V

User avatar
ThomasB
Normal user
Posts: 1064
Joined: 17 Jun 2018, 20:41
Location: USA

Re: Design for power supply from source >12V low quiescent current

#9 Post by ThomasB » 16 May 2020, 00:46

Such pulsating operation is also a good reason to use a linear regulator as the last step to 3.3V
Using a linear regulator on the output of the switching regulator is a good idea on traditional switchers. But the MP1584EN runs at about 1MHz. So I find that the ripple (typically <30mV) can be removed with a typical LC filter (small inductor and cap). Painless and cheap.

- Thomas

happytm
Normal user
Posts: 107
Joined: 15 Aug 2016, 17:53

Re: Design for power supply from source >12V low quiescent current

#10 Post by happytm » 16 May 2020, 01:43

This could be solution to your problem. It uses 380 nA current. Threshold voltage of 8.5 V or 10.8 V. Reset output below threshold is maximum 5.5V. Cost less than $1.

http://www.ti.com/lit/ds/symlink/tps384 ... 9585425146

Thanks.

Wiki
Normal user
Posts: 413
Joined: 23 Apr 2018, 17:55
Location: Germany

Re: Design for power supply from source >12V low quiescent current

#11 Post by Wiki » 16 May 2020, 02:29

@ThomasB: I notice, it would be preferrable to read before posting - I will do so in the future. You're right, the named behaviour won't hurt me in my application in any way. Thank you for clarifying. You're my best and made my day with your suggestions - I think.

@TD-er: I am used to unsolder the original voltage regulator of the Wemos D1 and replace them with something less power consuming. As we all know, the original 4A20 voltage regulators used typically on the clone Wemos D1 boards are not the very best ones. So I trust in the power management of the board itself and don't care about ripple voltage in an area of 30mV. Maybe I am wrong, but up to now every voltage regulator I used didn't do any harm to my applications at all. But anyway, Thanks for your thoughts.

@happytm: Interesting link. But I don't want to monitor but to power my application with the lowest energy consumption as possible. Monitoring of the battery is done by using an INA219, feeded by an GPIO pin of the Wemos, so it doesn't count during slep time. My problem is, that the smallest current ever is able to drain a battery of what capacity ever - only a question of time. And I don't want to waste my worthy battery energy by using a power supply which drains >1mA quiescent current - and this is not impossible.

Code: Select all

pi@raspberrypi:~ $ man woman
No manual entry for woman
pi@raspberrypi:~ $

Wiki
Normal user
Posts: 413
Joined: 23 Apr 2018, 17:55
Location: Germany

Re: Design for power supply from source >12V low quiescent current

#12 Post by Wiki » 16 May 2020, 02:45

TD-er wrote: 16 May 2020, 00:12 .
.
.
10 mA is roughly 8 times less, so that should last over 2 months, if no other items in the car draw current (e.g. stand-by of the car stereo)
.
.
.
Sorry, I've seen your post too late.
My problem is, that the car sometimes isn't used for two month or so. A quiescent current of 10mA would discharge the battery by roughly estimated 15AH during this time. Taking this in mind, reaching the goal of a quiescent current below 1mA makes sence. The car itself is 18 years old, so the reduction of power consumption implemented in modern cars isn't present in any way. Additionally its an american model and these ones normally don't care about power consumption. In Amiland a person walking by foot is a person on its way from or to the car......

So any mAh counts for me.

Code: Select all

pi@raspberrypi:~ $ man woman
No manual entry for woman
pi@raspberrypi:~ $

Wiki
Normal user
Posts: 413
Joined: 23 Apr 2018, 17:55
Location: Germany

Re: Design for power supply from source >12V low quiescent current

#13 Post by Wiki » 16 May 2020, 03:20

Additional information about my application to whom is interested:

I made the same modifications on the Wemos board as already described here. The used INA219 for monitoring battery voltage is powered directly by a GPIO, high on awakening, low on sleep. Works so far.

I robbed a small alloy heatsink from an unused Raspberry and put it on the LT1763. Using deep sleep for 600 secs, awake time of 5 secs works so far now up to a voltage of 14.2V - normal using cycle of a lead acid 12V battery. 5 secs awake seem to be not sufficient to heat up the LT1763 equipped with a heatsink up to thermal shutoff. But at the moment I have a battery charger running, which raises the charging voltage >15.6V if a sulfation of the battery is detected (and it did so at the moment) after a check routine - and now the LT1763 shuts off after a while.

[edit]
...and btw the whole circuit draws a current of 280µA in the current setup during deep sleep
[/edit]

Code: Select all

pi@raspberrypi:~ $ man woman
No manual entry for woman
pi@raspberrypi:~ $

happytm
Normal user
Posts: 107
Joined: 15 Aug 2016, 17:53

Re: Design for power supply from source >12V low quiescent current

#14 Post by happytm » 16 May 2020, 04:44

The solution I suggested does exactly what you want to achieve by using only few nanoamps in power consumption.

Let me be little more clear. If I understood right you want to monitor low voltage on your car battery and send you warning on your smart phone when battery goes below certain threshold. If that is your goal then this IC is perfect for it. There is no deep sleep required. When car battery is more than 10.8 volts the whole circuit including ESP8266 is powered off and uses only few nanoamps (very negligible) power. As soon as car battery capacity goes below 10.8 volts it send current to reset pin which can be connected as a trigger for a regulator which will turn power on for your microcontroller. Once ESP8266 is turned on it can immediately send MQTT message to you warning your car battery is low.

The concept is described very well here https://github.com/gitpeut/IRF7317-selfextinction . Also watch video of Ralph Bacon linked on that repo.

Thanks.

Wiki
Normal user
Posts: 413
Joined: 23 Apr 2018, 17:55
Location: Germany

Re: Design for power supply from source >12V low quiescent current

#15 Post by Wiki » 16 May 2020, 06:11

OK, now I understood what you were talking about. And in theory, you have posted a perfect solution for getting an alarm on bad battery conditions - useful for a stationary installation in a lot up in the mountains with solar panels and attached battery.

But: no, I don't want to get only a warning if battery voltage is <10.8V (or whatever). If I get a warning in this case, it may be too late. Depending on the current climate conditions I want to get an idea, when it would make sense to attach my battery charger. The goal is, to measure ambient and engine temperature (which could differ in some circumstances) and battery voltage to have an idea if the car would start or not depending on additional parameters like i.e. the health state of the battery. Maybe in summertime, when temperatures are around 25°C, the battery will be able to start the engine at a voltage of 11.5V. At wintertime with temperatures around zero or below I will need at least somewhat around 12.5V to start the engine. And an idea about the health state of the battery you get only by permanently monitoring its discharge behaviour.

For this project I would like to have a deep sleep configuration for awakening and sending the data only once a day to my Domoticz server, but thats not possible due to the limitations of the ESP8266 - except TD-er now tells something different, which he wouldn't do.

Code: Select all

pi@raspberrypi:~ $ man woman
No manual entry for woman
pi@raspberrypi:~ $

happytm
Normal user
Posts: 107
Joined: 15 Aug 2016, 17:53

Re: Design for power supply from source >12V low quiescent current

#16 Post by happytm » 16 May 2020, 22:23

OK I understand your goal now. This is not solution for your goal but I found an interesting video of LT431 (adjustable zener) here https://www.youtube.com/watch?v=7zE9SjdAAdk. LT. This chip is one of the highest selling chip in the world.Unfortunately it can sink maximum of 100 ma of current. Datasheet here : https://www.onsemi.com/pub/Collateral/TL431-D.PDF. Page 7 to 10 of datasheet list wide range of application examples.

Just put it here for reference so it can be used for some other use cases.

Thanks.

Wiki
Normal user
Posts: 413
Joined: 23 Apr 2018, 17:55
Location: Germany

Re: Design for power supply from source >12V low quiescent current

#17 Post by Wiki » 17 May 2020, 01:36

Now I've had a deep look into the datasheet of the MP1584EN linked by ThomasB which names a quiescent current of 100µA. I stopped already the tests with the LT1763-3.3 and ordered some of the MP1584EN, seems as if it will be my favourite DC-DC power supply in 12V applications/projects.

Just for comparism: In my 12V-projects I do use such small adjustable mini DC/DC bucks with MP2307DN. They draw a quiescent current of around 5.5mA which is in no way acceptable in a battery driven project. Even if the circuit with the MP1584EN won't reach the ultra small quiescent current of a HT7333 or takes three to five times more than the LT1763x it will be the best choice for me if I can confirm the ~250µA quiescent current mentioned by ThomasB.

So once again @ThomasB: Thank you

Code: Select all

pi@raspberrypi:~ $ man woman
No manual entry for woman
pi@raspberrypi:~ $

TD-er
Core team member
Posts: 8643
Joined: 01 Sep 2017, 22:13
Location: the Netherlands
Contact:

Re: Design for power supply from source >12V low quiescent current

#18 Post by TD-er » 18 May 2020, 00:06

Just one more thought to keep in mind.
When monitoring the battery voltage, one of the most common setups is to use a few resistors as voltage divider.
But those also consume some steady current.
So maybe you also want to put those in series with some FET so you can disable the current through those also when not needed.
Switching this FET may need an extra GPIO pin, or you must be able to make some fancy setup where you switch it on when the ESP is also powered.

Wiki
Normal user
Posts: 413
Joined: 23 Apr 2018, 17:55
Location: Germany

Re: Design for power supply from source >12V low quiescent current

#19 Post by Wiki » 18 May 2020, 01:08

Thank you for your thoughts. I am measuring the voltag using an INA219 which is powered by a GPIO directly. Works fine and shuts off the power supply of the INA during deep sleep. Easy going, 100% efficient way. In another project I am using a pretty energy consuming ultrasonic device. This one is being switched on/off by using a transistor - works as well.

In theory, you are right. But following Ohm's law and having a look at the onboard voltage divider *1 we are talking about a resistor to ground of 220kOhm, which results in a current drain of about 55µA using a 12V circuit. My experience: compared with cutting off the CH240 off the board (I personally am just lifting the ground pin to be able to reanimate it if its ever needed) has an effect of a factor >120 concerning the unneeded power consumption of a voltage divider, unsoldering the onboard voltage regulator and replacing it by a low quiescent current one doubles almost energy saving. So in my opinion the story about energy drain by measuring the voltage using a voltage divider being responsible for short battery life belongs to the overall present packages of fairy tales, pls correct me if I'm wrong.

[edit]
*1 of a Wemos D1
[/edit]

Code: Select all

pi@raspberrypi:~ $ man woman
No manual entry for woman
pi@raspberrypi:~ $

TD-er
Core team member
Posts: 8643
Joined: 01 Sep 2017, 22:13
Location: the Netherlands
Contact:

Re: Design for power supply from source >12V low quiescent current

#20 Post by TD-er » 18 May 2020, 11:21

Well it was more like 50-ish uA purely for the voltage divider can be significant if you're looking to get the quiescent current down to the order of uA's.
This voltage divider is connected all the time, so it can become a huge part of the total quiescent current if you're looking at voltage regulators needing 1 uA or less during sleep.
Whether that's an issue or not depends on the type of power source used.
If it is a coin cell, then 50 uA is a lot (will be less of course due to the lower voltage, but you get the idea), but looking at a car battery then it is probably less than the self-discharge of the battery.

That's all I tried to mention.

Wiki
Normal user
Posts: 413
Joined: 23 Apr 2018, 17:55
Location: Germany

Re: Design for power supply from source >12V low quiescent current

#21 Post by Wiki » 18 May 2020, 12:29

OK, yes, your'e right, I am wrong. I should have had in mind that depending on the project the energy source may be much smaller than the ones I use.

Code: Select all

pi@raspberrypi:~ $ man woman
No manual entry for woman
pi@raspberrypi:~ $

Post Reply

Who is online

Users browsing this forum: No registered users and 30 guests