Hi Zim,
You could save yourself a further 5uA and possibly some battery cells by eliminating the regulator.
Not sure about your best battery option for such cold temperatures but there are a few options that would not need a regulator.
3.7v LiPo (eg 18650) via a 1N4001 diode. (4.2v fully charged)
3.2v LiPO4 Direct connection. (3.6v fully charged, also available in 18650 size)
3 * 1.2v NiMh or NiCd via 1N4001 (4.3v fully charged) [Maybe even only 2 cells direct connection]
2 * Regular 1.5v disposable cell, direct connection.
The ESP will be OK down to as low as 1.8v. Don't know about the compass though.
I'm using 3 * NiMh which are currently at -3 'C and working fine.
Wireless Remote Propane Tank Level Gauge
-
BeanieBots
- Posts: 344
- Joined: Tue Jun 21, 2022 2:17 pm
- Location: South coast UK
- Has thanked: 182 times
- Been thanked: 112 times
Re: Wireless Remote Propane Tank Level Gauge
Hi BeanieBots
It seems to me the only affordable batteries that withstand -45F to some degree, are alkaline. My theory is having a 6v battery with a regulator gives me a 2 volt buffer zone to lose before brownout. Lantern batteries are cheap like me! I'll try that first. Then I'll try 3 "D" batteries with a diode in series (no regulator) and see how that performs. Its all fun!
Thanks
Zim
It seems to me the only affordable batteries that withstand -45F to some degree, are alkaline. My theory is having a 6v battery with a regulator gives me a 2 volt buffer zone to lose before brownout. Lantern batteries are cheap like me! I'll try that first. Then I'll try 3 "D" batteries with a diode in series (no regulator) and see how that performs. Its all fun!
Thanks
Zim
-
BeanieBots
- Posts: 344
- Joined: Tue Jun 21, 2022 2:17 pm
- Location: South coast UK
- Has thanked: 182 times
- Been thanked: 112 times
Re: Wireless Remote Propane Tank Level Gauge
If your supply can provide enough current, you should not get any brownout until less than 2v.
It's probably the regulator collapsing due to insufficient head room.
Also, make sure you have adequate decoupling. It is one of the most neglected subjects by hobbiests
It's not simply about having large caps, it's about having the right type and several decade values in parallel and at the right location on the power rail.
eg 100uF + 10uF + 100nF of low ESR type. Larger caps eg 1000uF have higher ESR and will not catch very high short current spikes.
Please be careful if you are thinking of using 1.5v Alkaline batteries. (3*1.5 -0.6) = 3.9v That is enough to kill the ESP.
You could use just two cells and no diode. That will give you a good solid 3v. ESPs really do work well even down at 2v. >3.6v kills them!
They do not HAVE to be at 3v3 to work well.
Just a thought. How about rechargeable topped up by an old garden solar lamp?
It's probably the regulator collapsing due to insufficient head room.
Also, make sure you have adequate decoupling. It is one of the most neglected subjects by hobbiests
It's not simply about having large caps, it's about having the right type and several decade values in parallel and at the right location on the power rail.
eg 100uF + 10uF + 100nF of low ESR type. Larger caps eg 1000uF have higher ESR and will not catch very high short current spikes.
Please be careful if you are thinking of using 1.5v Alkaline batteries. (3*1.5 -0.6) = 3.9v That is enough to kill the ESP.
You could use just two cells and no diode. That will give you a good solid 3v. ESPs really do work well even down at 2v. >3.6v kills them!
They do not HAVE to be at 3v3 to work well.
Just a thought. How about rechargeable topped up by an old garden solar lamp?
Re: Wireless Remote Propane Tank Level Gauge
Hi BeanieBots[Local Link Removed for Guests] wrote: [Local Link Removed for Guests]Tue Jan 16, 2024 11:30 am Also, make sure you have adequate decoupling. It is one of the most neglected subjects by hobbiests
Thanks for the reply. Please explain decoupling.
That would have made my project so much easier, but the area is totally shaded.Just a thought. How about rechargeable topped up by an old garden solar lamp?
Thanks again!
Zim
-
BeanieBots
- Posts: 344
- Joined: Tue Jun 21, 2022 2:17 pm
- Location: South coast UK
- Has thanked: 182 times
- Been thanked: 112 times
Re: Wireless Remote Propane Tank Level Gauge
Decoupling:
Whenever a transistor switches state, it has to move a bunch of electrons from one location to another. Very similar to discharging a capacitor.
That results in a very brief yet high current. What is not appreciated is that the time can be sub nano second or even pico second and the current can be as much as 10's of amps. It is not possible to see this even with the best o'scope but it IS there.
These 'spikes' can cause all manner of issues, not only to the device which is switching but also to other nearby devices. Especially when analogue and digital are mixed on the same board.
Memory devices have very fast spikes, an IO changing state is much slower but the current is on for much longer.
A simple analogy would be like comparing hitting a nail with a hammer compared to nudging the garage door with your car!
So, what to do about it.
The answer is to have something that can supply that current for the amount of time required.
If capacitors and the PCB were perfect, it could simply be solved by fitting a capacitor.
Unfortunately, all tracks, connections, component legs and even devices have both resitance and inductance. (as well as capacitance)
Although these values are very small, when we are talking about amps and sub nano second time scales, they become significant.
All capacitors have a characteristic ESR (effective series resistance). It is a compound value made up of resistance and inductance.
As a general rule, the smaller the capacitor, the lower the ESR. Or more to the point, large value capacitors tend to have a large ESR.
Therefore, fast switching needs low value capacitance and slow switching requires larger capacitance to be effective.
Unfortunately, it doesn't end there.
As leads etc. also have inductance and resistance, where the capacitor is physically placed and how far away it is from the device it is trying to decouple can have a large effect. It needs to be fitted as close as possible to the device. If it is at the end of some long tracks or even leads, fitting a decoupling capacitor can actuall make things worse. That is because you now have inductance + capacitance + current spike = tuned circuit. You've now made a radio transmitter!
Summary:
Use several decade values of capacitor in parallel as close as possible to each switching device. 10uF - 100uF + 10nF - 100nF will fix most issues.
Regulators are another subject. Most REQUIRE capacitance on input and output to actually work. The datasheet will advise.
Sorry about the long ramble but it really is a missunderstood subject. More than happy to give any further explanation if required.
Whenever a transistor switches state, it has to move a bunch of electrons from one location to another. Very similar to discharging a capacitor.
That results in a very brief yet high current. What is not appreciated is that the time can be sub nano second or even pico second and the current can be as much as 10's of amps. It is not possible to see this even with the best o'scope but it IS there.
These 'spikes' can cause all manner of issues, not only to the device which is switching but also to other nearby devices. Especially when analogue and digital are mixed on the same board.
Memory devices have very fast spikes, an IO changing state is much slower but the current is on for much longer.
A simple analogy would be like comparing hitting a nail with a hammer compared to nudging the garage door with your car!
So, what to do about it.
The answer is to have something that can supply that current for the amount of time required.
If capacitors and the PCB were perfect, it could simply be solved by fitting a capacitor.
Unfortunately, all tracks, connections, component legs and even devices have both resitance and inductance. (as well as capacitance)
Although these values are very small, when we are talking about amps and sub nano second time scales, they become significant.
All capacitors have a characteristic ESR (effective series resistance). It is a compound value made up of resistance and inductance.
As a general rule, the smaller the capacitor, the lower the ESR. Or more to the point, large value capacitors tend to have a large ESR.
Therefore, fast switching needs low value capacitance and slow switching requires larger capacitance to be effective.
Unfortunately, it doesn't end there.
As leads etc. also have inductance and resistance, where the capacitor is physically placed and how far away it is from the device it is trying to decouple can have a large effect. It needs to be fitted as close as possible to the device. If it is at the end of some long tracks or even leads, fitting a decoupling capacitor can actuall make things worse. That is because you now have inductance + capacitance + current spike = tuned circuit. You've now made a radio transmitter!
Summary:
Use several decade values of capacitor in parallel as close as possible to each switching device. 10uF - 100uF + 10nF - 100nF will fix most issues.
Regulators are another subject. Most REQUIRE capacitance on input and output to actually work. The datasheet will advise.
Sorry about the long ramble but it really is a missunderstood subject. More than happy to give any further explanation if required.