I was just looking at this page:
The Saber Touch
They say "Please Use 880mAh or Higher Rated Protected Batteries."
More than likely, the reason they think that this is the appropriate recommendation, is because they figure people will be using 3 Ohm atty's, which at 5V will result in ~1.67A to flow through the circuit (assuming a linear regulator). The assumption is that if a LiCo cell is rated at 880mAH then it should be able to safely deliver up to ~1.76A (2C rate for a true 880mAH cell if it holds this capacity at that load). Unfortunately, the recommendation is based on false pretenses.
False pretense #1: The cells are 880mAH
False pretense #2: The method of devising maximum discharge rate is to multiple capacity by "C" rate.
C rate is a function of time. A cell chemistry rated for a maximum discharge rate of 2C has a minimum complete discharge time of 30 minutes. Keep in mind that many cells are technically only supposed to be used up to ~1-1.5C rates. Any information given about an ultrafire product from the "manufacture" and passed down to the resellers should not be trusted (in case this wasn't already apparent).
Under a ~1.7A load, a typical RCR123 LiCo cell will have less than 500mAH of true capacity (probably ~400mAH). The UF880s have been independently tested to just barely eek out over 500mAH at a 1A discharge rate. The actual discharge rate on the saber touch would be around a 4C rate. Used in short bursts, (as ecigs are), the effective danger is theoretically lower than if the cell were exposed to these rates continuously since the cell never has a chance to warm up significantly. I still see it as less than ideal. LiCo chemistry cell data sheets I have read do not list safe burst rates. The assumption would be the chemistry should not be used in such a fashion regardless.
In conclusion, I would not personally use the saber touch with LiCo chemistry RCR123 size cells. Protected or not...
----------------------------------------------
IMR16340s can safely handle the load of the saber touch. They are high rate cells capable of safely delivering ~double the rate that the saber touch operates at. The label capacity of the IMR16340s is much closer to being accurate, and the lower cell resistance translates to less loss at higher loads. The result is that the actual available capacity (runtime) on IMR16340s is likely to be better.
All sounds too good right?
Well.... If an atty goes low resistance for some reason, does the saber touch have the ability to protect itself from over-current/over-heating? I would be concerned that the IMR16340s would have the ability to kill the regulation circuit if the atty fails low resistance. I'm not familiar with the frequency of atty's failing one way or another. I assume that most suffer from steadily rising resistance followed by a final open circuit much like a bulb burning out. Though I can envision circumstances where a wearing atty could collapse in on itself, fuse, and go low resistance.
Regulated devices make it difficult for the user to identify when cells are running low. There is no protection circuit on IMR16340s to prevent over-discharge or reverse charge. If the regulator is set to 5V, then it will operate steady all the way down to ~2.5V per cell. Since no 2 cells are perfectly matched, it's very possible to drag one well below 2V before noticing any change in vaping performance.
*[If the saber touch does not have a low voltage cut-off, then it will likely continue to allow vaping as the voltage of the cells drops below 5V. Some users may not notice a change until that last hit suddenly over-discharges the cell/s to well below 1V, or even potentially reverse charges a cell for a few moments. Over-discharge isn't a terrible problem if the cells are returned to the charger quickly. However, it's best to avoid it whenever possible.] The problem I see is that, unlike direct drive devices, the saber touch is apt to have you deep-discharging the cells on every cycle.
*I just read on the ST web page that the device does have over-discharge protection built in. I would want to confirm that it cuts in at an appropriate time for the cells.
With unprotected cells, it would be best to take some time to organize the cells into pairs of similar capacity. This would help reduce the chances of over-discharging any given cell repeatedly.
---------------------------------------------
Given those considerations, if I were using a Saber Touch, I would personally take my chances with the IMR16340s rather than the protected LiCo cells. I feel that they would be much safer and more reliable in this application.
A saber touch designed to operate on a pair of 17500s would have been much smarter. Could have used protected AW 17500s, which deliver a solid ~1.1AH performance. Perfectly safe to operate at ~1.7A. Protection to prevent over-discharge and reverse charge in the regulated application.
Eric