Theoretically, charger could start at low current, see how fast cell voltage is rising, estimate capacity from there, then raise current if necessary.
Olight UC (Universal Charger) Magnetic & USB input
- Thread starter keithy
- Start date
Help Support Candle Power Flashlight Forum
Theoretically, charger could start at low current, see how fast cell voltage is rising, estimate capacity from there, then raise current if necessary.
Such marketing hype has already been deconstructed in the thread on Nitecore's "Superb Charger SC2" with claims of "Active Charging with Infinite Intelligence ... that automatically detects battery capacities".
It is sad that the culture of grossly exaggerated (capacity) claims is spreading to chargers too. It appears that Nitecore has recently scaled back their exaggerated claims a bit in their latest brochures. Maybe the critiques here (and elsewhere) are having some impact. If only the same were true for 10000mAh Ultraxxx cells.
Last edited:
I agree about marketing hype, I am just saying it's not impossible to make such a charger.
To successfully develop automatic optimum-charging current heuristics would require significant R&D - something very unlikely to come from a small consumer-level charger company. It would require access to extensive databases of chemistry parameters etc in order to be sure to achieve an effective balance of safety and performance across the entire spectrum. Such databases do exist (e.g. TI has large databases for configuring fuel gauges). Mining this data might yield some some decent heuristics. But I am not aware of any work along those lines that has percolated down to consumer-level devices.
Last edited:
You are over complicating it. Any charger that shows charged percentage is already estimating cell capacity.
Not true. It is innately complex due to the very wide spectrum of chemistry, capacity, health, etc that a general charger must handle. Those "guesstimates" of capaciy that you refer to are generally so far off the mark that they are mostly useless (check the many complaints on chargers that do such guessing).
Such guesses are nowhere near accurate enough to universally automagically infer a safe and efficient charging current. If the guess is too conservative then it will yield painfully slow charges on some cells. If it's too agressive then it will damage some cells (and possibly be unsafe). To implement an appropriate balance that is reasonable for most all scenarios is much more tricky than it may seem at first glance.
Last edited:
Perhaps a little history will prove illuminating. There once was an RC hobby charger by a widely respected (American) manufacturer (Astroflight 109) that claimed to be able to automagically guess the number of cells in LiPo packs. This did not always work correctly, and led to many fires, e.g. see the RCGroups thread AstroFlight 109 Charger Fire, AGAIN! The designer initially refused to admit this was a design error, but eventually (after many requests, liability issues, etc) they issued a revised firmware chip (user swappable) that forced the user to confirm that it had guessed the correct number of cells before proceeding with the charge.
This specific problem only affects pack chargers (cells in series), but there are related safety problems for non-pack chargers. Namely, if the charger incorrectly guesses cell capacity then it may end up charging a very low capacity cell at very high charge rate, surely damaging the cell, and possibly leading to venting (with flames).
It is a nontrivial task to devise heuristics that are both safe and efficient. This is why such heuristics have yet to appear in such chargers (despite marketing snake-oil claims to the contrary). Advanced charging algorithms will likely percolate down to the consumer level in the not-too-distant future, but we are not quite there yet.
This specific problem only affects pack chargers (cells in series), but there are related safety problems for non-pack chargers. Namely, if the charger incorrectly guesses cell capacity then it may end up charging a very low capacity cell at very high charge rate, surely damaging the cell, and possibly leading to venting (with flames).
It is a nontrivial task to devise heuristics that are both safe and efficient. This is why such heuristics have yet to appear in such chargers (despite marketing snake-oil claims to the contrary). Advanced charging algorithms will likely percolate down to the consumer level in the not-too-distant future, but we are not quite there yet.
Last edited:
You keep talking about some mythical "safe and efficient charging voltage", while most Li-Ion chargers use CC/CV algorithm with predetermined voltage levels. Are you sure you sufficiently understand how a charger works to discuss this subject?
Many current chargers show charging percentage pretty accurately. There is nothing extraordinary about it - they just measure the speed of voltage rise in the cell.
The above discussion concerns automatic selection of charging current. That goes beyond standard CC/CV algorithms. And, yes, I have expertise in such matters, having consulted on the design of some charging and fuel gauge algorithms used in professional chargers.
This is false. There are no consumer-level chargers that can quickly and accurately measure SOC in general, i.e. for arbitrary cells in arbitrary states. To do so requires requires access to information that is not accessible to the charger (precise chemistry, IR profiles, etc). Read the literature on fuel gauge algorithms if you desire to learn more about such matters, e.g. look up the TI impedance tracking algorithm for a start.
Last edited:
No one is talking about measuring state of charge. That is complete unnecessary for the purpose of automatic charging current selection. Please re-read what I said from the very beginning. All that is required is to measure speed of cell voltage rise, once the charging starts. This allows to relatively accurately estimate remaining capacity until cell will be full and select appropriate current. It does not allow to accurately estimate total cell capacity, but that is completely unnecessary for the purpose.
Examples:
500mAh cell, with 50mAh remaining charge:
charger starts at 0.25A, by the speed of voltage rise in the cell estimates charging at 0.25A will take about two hours, stays at 0.25A, finishes in about two hours.
3500mAh cell, with 3000mAh remaining charge:
charger starts at 0.25A, by the speed of voltage rise in the cell estimates charging at 0.25A will take about two hours, stays at 0.25A, finishes in about two hours.
3500mAh cell, with 50mAh remaining charge:
charger starts at 0.25A, by the speed of voltage rise in the cell estimates charging at 0.25A will take about 14 hours, raises current to 2A, finishes in about two hours.
There are many ideas on how to implement automatic charge current selection - some simple, some very sophisticated. None of those ideas have yet to yield a safe and efficient algorithm that works under general conditions at the consumer level. If the problem had a simple solution then that feature would already be implemented in consumer level chargers.
Last edited:
brighterthanthesun
Enlightened
For what it is worth, Aggressor's statement is pretty much what I was told when I contacted Olight. I have ordered one on the hopes that it works the way they told me it would. I will wait patiently for a review by HJK for final approval on this charger. But for $6, this would be a great portable and flexible charger.
It's snake oil marketing hype - just like Nitcore's claims to do the same.
Beautiful argument! The statement is wrong, because if it were right, someone would have made it earlier
Consumer level chargers for standalone Li-Ion cells are a tiny part of overall Li-Ion cells and chargers market. Most Li-Ion cells are used in integrated packs so chargers already know the optimum charging parameters.
There may simply never been a strong enough incentive to develop automatic current selection. Let's hope this will change soon. Because there is nothing complex about it. Chargers already estimate remaining charging time accurately enough, and that's all the info they need to implement this feature.
StandardBattery
Flashlight Enthusiast
Got one of these in the mail a couple days ago. A little bigger than I thought, especially the magnetic ends. It does not come with a portable carrying case, but I found a near perfect little long rectangular tin to put it into for storage or if I want to put it in a to-go bag of some sort or pocket.
Did a little test.
Based on this short test. It would work OK for 16340 & 18350 cells, and would be not too bad for 18650 when you are not in a hurry. Of course this is just a very simple preliminary test so I look forward to hearing from others on complete charge cycles with this charger. I've got it completing the AA charge now to see if it might pickup steam at some point.
Update 1: after an additional 3.5hrs of charging the AA NiMH battery the input current was unchanged and input energy had reached 3.764Wh so I stopped the test. Battery voltage was 1.43V (essentially Full), the battery was not warm at all. The charger was slightly warm, I didn't measure temp, but it was mild. So no expected completed charge signal, but I'm curious on why the input power was still so high.
Update 2: after an additional 78min the 18350 Terminated correctly with green light (4.205V). USB monitor reported 3.95Wh of energy had been consumed. Input current decreased as the charge reached the end of the cycle as expected for a LiIon charge algorithm.
Update 3: It was able to complete the charge on the 18650 an properly terminated the charge 4.2V I didn't monitor the time. Input power was 6.25 Wh. So I'm happy with LiIon charging, but AA and AAA charging needs further investigation.
Update 4: I had it complete the charge on the AAA cell in 71min reported input power was 893mWh. The cell was a bit warm.
Update 5: I had it charge another AA cell and it did terminate correctly, but reported input power was 5000mWh which seem to indicate very poor efficiency, but I was not able to record the time it took to charge. Still not liking this charger for AA cells. I may try a more detailed test, but I'm not likely to use it for AA or AAA anyway.
Did a little test.
| Battery | Initial Voltage | USB Input I @ 5.1V (Initial) | USB Input I @5.1V (after 1min) |
| AAA NiMH | 1.245V | 122mA | 240mA*** |
| AA NiMH #1 (eneloop) | 1.178V | 115mA | 240mA*** |
| 18350 LiIon (XTAR 850mAh) | 2.958V | 650mAh* | 660mAh (690mA** @ 5min) |
| 18650 LiIon (TF 2500mAh) | 3.786V | 739mAh* | 739mAh |
- Quiesent USB input current (i.e. no connection) was 0.0055mA.
- Note measured current in Input current from the USB power source, not the battery charge current.
- Only did test on each cell for 5min to get summary initial charge stats.
- ***Current after 1min was fluctuating between 200mA and 240mA, I didn't scope it but there may be some pulsing. Unless indicated current was the same @ the 5min mark.
- * reading occurred very quickly ~10-12sec
- ** current was still increasing
Based on this short test. It would work OK for 16340 & 18350 cells, and would be not too bad for 18650 when you are not in a hurry. Of course this is just a very simple preliminary test so I look forward to hearing from others on complete charge cycles with this charger. I've got it completing the AA charge now to see if it might pickup steam at some point.
Update 1: after an additional 3.5hrs of charging the AA NiMH battery the input current was unchanged and input energy had reached 3.764Wh so I stopped the test. Battery voltage was 1.43V (essentially Full), the battery was not warm at all. The charger was slightly warm, I didn't measure temp, but it was mild. So no expected completed charge signal, but I'm curious on why the input power was still so high.
Update 2: after an additional 78min the 18350 Terminated correctly with green light (4.205V). USB monitor reported 3.95Wh of energy had been consumed. Input current decreased as the charge reached the end of the cycle as expected for a LiIon charge algorithm.
Update 3: It was able to complete the charge on the 18650 an properly terminated the charge 4.2V I didn't monitor the time. Input power was 6.25 Wh. So I'm happy with LiIon charging, but AA and AAA charging needs further investigation.
Update 4: I had it complete the charge on the AAA cell in 71min reported input power was 893mWh. The cell was a bit warm.
Update 5: I had it charge another AA cell and it did terminate correctly, but reported input power was 5000mWh which seem to indicate very poor efficiency, but I was not able to record the time it took to charge. Still not liking this charger for AA cells. I may try a more detailed test, but I'm not likely to use it for AA or AAA anyway.
Last edited:
In order to refute this highly dubious claim we went straight to the source, Olight. On October 2, we sent queries to all of the support email addresses for Olight (US, Germany, and China), viz. [email protected], [email protected], [email protected]. We quoted the above claim, and asked for the specific current(s) used to charge Li-ion cells. We received no reply from the US contact. The German CSR forwarded our email to China, and we received replies only from someone named Shirley in China. The first few replies from Shirley simply quoted vague information already available in their brochures. Finally, after many requests for specifics, she supplied some test results (see below). The data seems to imply that the charger does not adjust the output using some capacity heuristics (as claimed above). Rather, it appears to use a fixed current of about 750mA for all Li-ion cells (except it precharges at about 250mA for cells below 3.0V). Of course such tests results are by no means definitive, but this is the best info we could obtain from Olight after many rounds of emails.
It is bizarre that no one from Olight could supply a definitive answer. They had to run tests to attempt to infer what charge currents are used. I don't expect a company like Olight to have in-house expertise on Li-ion charging, but they should be able to obtain that info by contacting the OEM who designed the charger. Answers to questions like those we posed are essential knowledge required in order to know if the charger can be used safely with smaller cells.
Below is the test data supplied by Olight.
Last edited:
markr6
Flashaholic
- Joined
- Jul 16, 2012
- Messages
- 9,258
I've been using this for a few days now just to see if I like it. I love it! Li-Ion and NiMH, tiny, simple. I sure wish there was a voltage readout, but this will do. Too cheap and small NOT to have one somewhere for emergencies or just to top off that EDC or single cell light.
StandardBattery
Flashlight Enthusiast
Seems to align with my test for Lithium. AA charge time seems crazy though in how long it takes to charge. And my test showed it used basically the same rate for AAA and AA. I think it's just easier to carry a single cell charger from XTAR and forget NiMH or get a Lii 100 and get more flexibility with an increase in size, but not that much and you get 2 cell capability which is good for NiMH and no messing with magnets.
I need to play with it more, but I think it's worth the $1 on sale although I would not pay $10 for it.
I need to play with it more, but I think it's worth the $1 on sale although I would not pay $10 for it.
Olightworld
Enlightened
- Joined
- Dec 11, 2014
- Messages
- 579
The charger can identify the polarity of the battery so it does not matter what end you use to plug it to. It can detect polarity and battery type so that the battery is not overcharged and damaged but the speed and output will not adjust.
rookiedaddy
Enlightened
Bought a unit, I really like its ultra portable size...
only thing is I wish they up the charging current to 1A instead of 750mA... well, one can always hope...
only thing is I wish they up the charging current to 1A instead of 750mA... well, one can always hope...
