73833 / 259 charger behavior

Breakout boards, sensors, other Adafruit kits, etc.

Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.
User avatar
mangocats
 
Posts: 38
Joined: Thu Aug 18, 2022 7:23 am

73833 / 259 charger behavior

Post by mangocats »

Hi, I'm back viewtopic.php?t=194692 with a new from package product id 259 charger and a pid 2011 2000 mAh battery. For relative simplicity, the load is just a Raspberry Pi Pico W running a uasyncio web server and a pid 987 amplifier in standby mode. 259 charger is powered from a 2A USB charging plug.

Observation: on initial power up, 259 shows red and amber lights. Battery voltage reads 3.95v on the BATT pads of the 259. Pico website is responding.

8 hours later: all the same, Pico website is responding, battery voltage reads 3.96v, red and amber lights on the 259. Remove power to the Pico and lights don't change, BATT voltage pops up a bit to 4.030v, plug the Pico back in and voltage drops a tiny bit to 4.026v.

Question: how long should it take a 500mA charger to charge up a 2000mAh battery (in real life)?

I don't have a good way to measure load current, but guesstimate the Pico draw around 50mA in standby webserving like it is now. Wouldn't I expect the 2000mAh battery to run in "green" mode on the 259 for at least 8-10 hours after a full charge?

User avatar
adafruit_support_bill
 
Posts: 88099
Joined: Sat Feb 07, 2009 10:11 am

Re: 73833 / 259 charger behavior

Post by adafruit_support_bill »

Hypothetically, if the entire charge cycle could run at 500mA it would charge completely in 4 hours. In practice however, the constant current 'fast charge' mode is only the middle phase of the charge cycle. There is a constant voltage pre-conditioning phase as well as a top-off phase at the end. The timings of these are dependent on the battery charge characteristics.

The other significant factor here is that you are charging with the load attached. The #259 charger does not have a load-sharing circuit. The load is just connected in parallel to the battery. This means that a portion of the charge current is actually going to the load, so the charge will take longer.

User avatar
mangocats
 
Posts: 38
Joined: Thu Aug 18, 2022 7:23 am

Re: 73833 / 259 charger behavior

Post by mangocats »

So, at what point would no green light be a concern? Still amber and 4.07v now several hours later.

User avatar
adafruit_support_bill
 
Posts: 88099
Joined: Sat Feb 07, 2009 10:11 am

Re: 73833 / 259 charger behavior

Post by adafruit_support_bill »

Hard to say without knowing the current requirements of your load, since any current consumed by the load is current that is not charging the battery. In fact, if the load were to pull more than 500mA, you would actually end up with a net discharge.

With no load and a 500mA charge rate, I'd expect a full charge withing 5 or 6 hours.

User avatar
mangocats
 
Posts: 38
Joined: Thu Aug 18, 2022 7:23 am

Re: 73833 / 259 charger behavior

Post by mangocats »

O.K. so I finally got the ammeter in-line with the total load, and my guess was pretty good, it runs about 42mA with pretty frequent spikes to 49mA and I'm guessing the manual garbage collection cycle is what drives it up around 60mA once in a great while. If I had to guess an average from what I'm seeing on the DMM, it would be about 44 or 45mA.

Light still amber only, reading BATT with Load connected, no other power in to the 259 (all lights out), I get 4.04v. Add my USB charger, Red and Amber lights come on, but BATT still reads 4.04V?

User avatar
adafruit_support_bill
 
Posts: 88099
Joined: Sat Feb 07, 2009 10:11 am

Re: 73833 / 259 charger behavior

Post by adafruit_support_bill »

The charger will remain in constant-current mode until the battery voltage reaches 4.2v. At that point the cell is about 80% charged. It will be 4.2v at the battery terminals - but the charge will not have leveled out across the whole cell.

To top-off the cell, the charge controller will transition to constant-voltage mode and remain in that mode until the current goes to zero indicating that the cell is fully charged. However, with a load attached, the current will never go to zero, so you will probably never see the full charge indicator.

User avatar
mangocats
 
Posts: 38
Joined: Thu Aug 18, 2022 7:23 am

Re: 73833 / 259 charger behavior

Post by mangocats »

So, at 450mA net into a 2000mAh cell, simple theory says: less than 4.5 hours to fully charge.

At 16 hours, I'm still hovering at 4.030v on the BATT/Load terminals, DCIN reads 5.067v, amber and red lights - no hint of green.

I have a similar unit deployed in the yard with a solar power setup, it has been running for months with similar Pico + Amp hardware and software - the current version of software was installed in the solar unit a week ago and has been running 24-7 since. The solar charger is set at default 500mA limit and there have been plenty of cloudy days. So, I don't think the Pico + Amp is draining anything unusually more than what I observed during the current measurement.

If 42mA is enough to prevent the amber light from going out, I guess I can still work with that using the 4712 fuel gauge... it's just, unlike the solar setup, I have responsibility for deciding when to give power to the system for charging the battery. The Pico can signal IFTTT to switch on the landscape light power, so I want to figure out when it needs to do that, and when it can / should shut off.

User avatar
adafruit_support_bill
 
Posts: 88099
Joined: Sat Feb 07, 2009 10:11 am

Re: 73833 / 259 charger behavior

Post by adafruit_support_bill »

So, at 450mA net into a 2000mAh cell, simple theory says: less than 4.5 hours to fully charge.
But simple theory only holds for the constant-current part of the cycle. 4.0v is likely about 90% capacity. You will never get all the way to 100% with this setup because the load is continually consuming your charge current.
The solar charger is set at default 500mA limit
Which solar charger are you using? Both of our solar chargers have load-sharing circuitry. Assuming adequate input power, they will supply full charge current to the cell while simultaneously powering the load.

User avatar
mangocats
 
Posts: 38
Joined: Thu Aug 18, 2022 7:23 am

Re: 73833 / 259 charger behavior

Post by mangocats »

adafruit_support_bill wrote: Tue Sep 27, 2022 1:12 pm Which solar charger are you using?
The 4755. I thought about using it in this scenario and just feeding it 5V instead of the solar panel power, but the: CHRG DONE indicators seemed like they would be easier to work with for deciding when to charge and when I can shut off the input power. Edit to say: oops, I see the 4755 is defaulted to 1000mA... sometimes I'm easily confused.

Being outdoors, the variations in temperature make reading LiIon voltages pretty challenging for deciding when the battery is really low or full on charge.

User avatar
mangocats
 
Posts: 38
Joined: Thu Aug 18, 2022 7:23 am

Re: 73833 / 259 charger behavior

Post by mangocats »

adafruit_support_bill wrote: Tue Sep 27, 2022 1:12 pm You will never get all the way to 100% with this setup because the load is continually consuming your charge current.
I don't really agree that that should be the case: the load is only consuming a bit less than 10% of the available charge current, if the 73833 were really entering "Constant Voltage" state it should be more than capable of maintaining 4.2V while feeding 42 to 60mA to the load, and the battery takes whatever it needs at 4.2V. Section 4.5 of the datasheet also says that "Constant Voltage" state will terminate on a timer timeout, which seems to be 8 hours or less... depending.

So far, the 73833 datasheet is still ambiguous to me about "Precondition" state, whether that is 10, 20, 40 or 100% of the PROG current (500mA in this case) it's probably specified somewhere in there, but the fonts of their .pdf aren't available in Ubuntu, so I'm stuck looking at the datasheet on my phone screen.... The flowchart indicates that "Fast Charge" can return to "Precondition" mode, but the text of section 4.4 doesn't indicate how this might happen. Nonetheless, it's the only thing that makes sense with how the lights are behaving, somehow it seems to be cycling back to, or stuck in, "Precondition" because the battery voltage is not up near 4.2v (like it would be in "Constant Voltage" state) and it's not climbing like it should be if it were receiving 450+mA of charge current.

I do agree that I'm not getting to 100% charge, but I'm not sure that the 73833 is supposed to be getting stuck in "Precondition" mode either.

User avatar
adafruit_support_bill
 
Posts: 88099
Joined: Sat Feb 07, 2009 10:11 am

Re: 73833 / 259 charger behavior

Post by adafruit_support_bill »

Microchip makes 2 variants of the chip. The MCP73833 is the one used in the #259 charger. The MCP73834 replaces the Power Good pin with a Timer Enable pin and is probably better suited for your configuration.

From the datasheet:
3.6 Timer Enable Input (TE)
MCP73834 Only
The timer enable (TE) input option is used to enable or
disable the internal timer. A low signal on this pin
enables the internal timer and a high signal disables
the internal timer. The TE input can be used to disable
the timer when the charger is supplying current to
charge the battery and power the system load.
The TE
input is compatible with 1.8V logic

User avatar
mangocats
 
Posts: 38
Joined: Thu Aug 18, 2022 7:23 am

Re: 73833 / 259 charger behavior

Post by mangocats »

Well, the goals for my application here are: 1) nothing catches fire, 2) battery isn't too abused with over-discharge or over-charge activity, 3) Landscape lighting isn't powered up more than necessary to keep the battery happy, but some flexibility is certainly available within reason.

The main challenge is: outdoor installation with temperatures (in the box) ranging from 0C up to 50C depending on time of day / season.

So, most of those goals can be met by guesstimating that the landscape lighting needs to run 68+ minutes a day (charger capable of 1000mA) to feed what the Pico consumes on a regular basis. The thing that's bugging me a bit is not knowing how much more than 68 minutes I'm going to need while the charger diddles around in other modes that still aren't making a whole lot of sense to me. Since the device has the capability to call out for power when it needs it, I thought that would be extra cool for it to make those decisions, extending charging time as needed based on observed battery state, automatically compensating for weather and aging of the cell... But, basic functionality, yeah, it can be set to charge for 75 minutes a day and probably be just fine for a couple of years.

User avatar
mangocats
 
Posts: 38
Joined: Thu Aug 18, 2022 7:23 am

Re: 73833 / 259 charger behavior

Post by mangocats »

adafruit_support_bill wrote: Tue Sep 27, 2022 11:51 am To top-off the cell, the charge controller will transition to constant-voltage mode and remain in that mode until the current goes to zero indicating that the cell is fully charged. However, with a load attached, the current will never go to zero, so you will probably never see the full charge indicator.
Morning update: around 8pm last night I removed the load, only the pid2011 2000mAh LiIon battery is connected to the 259 along with a USB charger. Charger was removed for a minute then reconnected. BATT voltage read
4.07 at the time. This morning, 12 hours later it is still in some kind of slow charge mode and voltage has made it up to 4.17v. Lights are same as ever, red and amber.

User avatar
mangocats
 
Posts: 38
Joined: Thu Aug 18, 2022 7:23 am

Re: 73833 / 259 charger behavior

Post by mangocats »

2pm update: lights remain amber and red. Voltage at 4.221.

User avatar
adafruit_support_bill
 
Posts: 88099
Joined: Sat Feb 07, 2009 10:11 am

Re: 73833 / 259 charger behavior

Post by adafruit_support_bill »

12 hours does seem like more than enough time. The chart in fig. 12-17 shows the top-off phase taking almost twice as long as the fast-charge phase. This will vary somewhat from cell to cell, but I don't see a reason why a 2000mAh cell would take that long.

If the charger is not going to be suitable for your needs, we can issue a refund. You can contact [email protected] with a link to this thread.

Locked
Please be positive and constructive with your questions and comments.

Return to “Other Products from Adafruit”