I have recently gotten a Vapcell S4+ to charge and test my flashlight batteries. I’ve been using it to test a few cells, but it looks like it consistently underestimate my cells by about ~10%. I have some lightly used 18650s rated to 3450 mAh, but it only measured around 3000-3100, and some 21700s rated to 4000 and 5000, but the charger only measured 3500 and 4700 (at 0.5A, instead of 1A with other cells) respectively. I’m not completely sure of the 21700s, but the 18650s should be from reliable sellers. Anyone else used the cap test feature on this charger? Does it overestimate or underestimate? Have you cross referenced it with another instrument? The seller said this charger doesn’t discharge the batteries fully, which seems plausible, but I do see it discharging them to 2.5V which seems to be in accordance with the battery datasheets.
The S4+ was in my short-list a few months ago (ended up getting a SkyRC MC3000 instead, should reach me in the next few days, yay!)
I remember reading it had issues with measuring internal resistence (which is one of the reasons I didn’t buy it as IR is rather important to me).
I would not be surprised if his other measurements are not very precise either.
PS: you know there are 3 versions of the S4+ already, right? Which one is yours?
It is very common for cells to have slightly lower capacity than advertised, often even large cell manufacturers use “typical capacity” and not minimal capacity as the number they advertise, so some cells end up having lower capacity than rated.
When not new, even “lightly used” capacity will be a bit lower too.
Discharge current is important too, ratings are usually for 0.2C (700mA for 3.5Ah cell), if you discharge with higher current you get slightly lower capacity.
Then… did you verify voltage with multimeter? Not sure about specific version you have, but typically chargers do not discharge to 2.5V exactly, but to 2.7V or even 3V, “reducing” capacity even more.
All this can add up and result in lower capacity than expected. Some measurement error is bound to exist too.
Considering all this 10% lower is not that bad…
Well, this is a fair bit below even the claimed minimum capacities (3900 and 4900), and noticeably less than the other reviews I see on the web. Most cells are tested at 1A, except the INR21700-50E cells at 0.5A so this can account for some loss, though I wouldn’t expect it to me be much based on reviews. The capacity only starts to drop noticeably with much larger currents. The batteries are already getting worrying hot without a fan to cool it though. The temperature climbed over 60 C after a while just at 1 A. I used a fan to blow at it to stop it from overheating. The charger did say it discharged to 2.52V in one instance, but I didn’t measure it with a multimeter that time, and the rest I also didn’t measure as it took many hours, so I did it overnight and it starts to recharge soon after. I guess that can account for a noticeable difference here. In any case the batteries I’m getting are likely to be legit.
It says it is the 2.0 version.
Chargers in generally can’t measure IR very well. You should use a dedicated device if you care about it much, I think. You need a device that supports measuring with four wires. While inaccurate it seems somewhat informative, though. Larger batteries have much lower IR, and high drain batteries less than the high capacity ones.
Now that’s interesting! Where would the two other wires go, since every cylindrical cell I ever saw only has two poles?
Thanks for the answer, I’ve heard about it before but that’s not the case here, right? The battery still has only two poles. Unless we’re talking about the voltage probes touching a different part of the pole surface? But then (as the charger terminal that makes contact with each pole is also quite large) each voltage/current pair could also do the same at the charger terminal (inside the charger), no?
Correct, you want two points of contact on each pole to correct for the contact resistance.
No, because voltage and current would be measured after the extra resistance of the contact.
So what would be the solution when measuring a battery? drill holes in it to insert the voltage probes in a different path than the current probes? 
Seriously, In the wikipedia page you linked, please see the diagram “Layout example of a Kelvin connection” with the voltage probes being attached to the “inside” of the same connectors used to measure current. I think this could be pretty much approximated just by having two different wires connected internally to each of the contacts on the charger, a thicker one for current and a thinner for voltage, instead of sharing the same cable. That would be as good as it gets with a battery (unless one’s willing to drill holes in it ) or is there a better way?
The vapcell discharges batteries further than most chargers do, 2.5v like you said. Usually it’s like 2.8v. So idk why someone told you that.
As for accuracy. It doesn’t rest cells like the manufacturer tells you to, so there’s that. And it’s not a precision instrument. They’re all just estimates. But it’s as good an estimate as any of these thing.
The numbers you got are what I would expect. The manufacturer rating is for a cell that’s been cycled less than 3 times and under really specific conditions, and much lower than 1A. Ideal temperature, ideal humidity, they even have a heat sink on the whole cell for complete uniform temperatures
Probes with two contacts as shown in the other post. Or two probes at each side. It’s just important that voltage and current aren’t measured with the same probe.
Isn’t it going to be quite hard to make contact with the positive pole of a button-top that way? Specially to get a good enough contact with the current probe to be able to draw 10 amps or more as the more capable batteries allow?
Isn’t it better (or at least more doable) to have a pair of single, large, as-low-resistance-as-possible contacts touching the battery poles, and then two different sets of wires soldered at different points on these contacts – thick wires for the current and non-necessarily thick ones for voltage? I would bet that’s what good chargers do to get reliable IR measurements.
Thanks for the answer and sorry if I still see problems with that – see the response I just posted to Jeff.
Oh, ya something like that would be good too. I think what you would want, and I’m not an expert I’m just speculating, would be the voltage wires attached just before when the probes contact the battery at each pole, for the probes to just be for current, then have the two voltage wires attached to the base of each probe like a half inch before where it contacts the battery. Doesn’t have to be soldered, could just be bolted down or something. I think the main thing is that they just don’t carry any current.
Btw the probes in that pic are spring loaded
No, as @Parametrek has shown in ##flashlight.
No. This doesn’t account for the contact resistance that is always present and cannot be predicted (except when soldering, probably). The current wire doesn’t have to be thick, it’s resistance doesn’t matter much anymore (as long as it can has low enough resistance that it reaches the target current).
What? 
On his own site (http://lygte-info.dk/) member @HKJ has a description of a charger setup. What I would like to point out, the picture shows a battery holder that has four wires coming from it. This holder could be bought from Fasttech, but they stopped their activities. Maybe somebody else can offer us a clearer picture of the battery holder.
He hasn’t actually show that in fact But he and I did agree that we were talking different stuff, with he meaning “pro-level equipment” doing things like a donut adjusted to the button for the current probe and a pogo-pin going through the aperture for the voltage probe, and I meaning “consumer-level equipment” (even very good ones like the SkyRC MC3000 charger) doing things the way I described (with a single large good contact for each battery pole showing externally, and separate wires internally for the current/voltage probes).
We agree on that. My point is that (at least for “consumer-level equipment”) the same applies for the separate-probes-all-the-way-to-the-battery vs the single-external-good-large-contact-with-separate-probes-internally approach as described.
Just so we can see that in practice, my SkyRC MC3000 has just arrived in Chile and should be in my eager hands in a couple of days. I will then do some testing and see how reliable/repeatable its IR measurements are; from all I’ve read, I predict I will see less than 10% variation on these measures, which is good enough for my use case (@parametrek on our ##flashlight discussion thought I was talking about 0.1%, which for me belongs to the realm of big corps and mad scientist labs )
I’ve been reading @HKJ reviews and using them extensively for my purchase decisions since basically forever, and have never bothered to check his exact charger testing setup, thanks for the link to that pic.
But please see my last resposnse to @SammyHP – I think this kind of setup is not practical for a consumer-level charger… and actually not required unless one needs sub-1% precision, which is not my case (10% precision is good enough for my case)
I’ve seen results from the MC3000 as high as 50-100 mΩ off. Wiggle the battery a little bit and the result will change. If it is enough for you to know whether the battery has 30 mΩ, 200 mΩ or more it should work for you. For more precise results you need better measurement devices.
Even in professional applications you use specialized devices for different tasks if you need precision, even when other high quality devices are available. They might be good, but not perfect. It all depends on the level of precision you need.