Statistical Analysis of Bloog Cartomizer Ohm Ratings

[Valentine Michael Smith]

Hey all you bloogers,

In one of the more popular threads on this forum a gentleman by the name of ManicMaurice posted some data about the ohm ratings of Bloog Cartomizers he had recieved. Since I am a statistics tutor I thought I would crunch the numbers and take a look at the results. I will do my best to break this down into usable information and leave out all the stuff that us math geeks like to look at when data sets are involved.

Disclaimer: With 55 data points that Manic Maurice posted I can be accurate around 87% of the time. Which means that if someone repeatedly ordered 55 or more carto's from Bloog my numbers would hold up around 9 out of 10 times (87 out of 100 to be more precise).

The analysis is as follow:

Mean (average) ohm rating: 2.68 ohms

Median (middle data point) and Mode (Most commonly seen data point): 2.6 ohms

The mode (2.6 ohms) occurred in 7 out of 55 Cartomizers. Or 12.7% of those tested. This is actually not bad statistically speaking and I was somewhat surprised that many data points were exactly on the Mode. And that the median and mode were the same value.

A note about median and mode: Median and Mode as measurements of central tendency (what one would expect to see) are less sensitive to extreme values such as the one cartomizer that measured at 4.3 ohms. I would need to crunch more numbers to be sure that the 4.3 ohm carto is indeed an outlier (something WAY WAY beyond what one might expect to see).

Standard Deviation of the Data set: .49855 ohms

What all this mean:
68% of Cartomizers should fall between 2.18 and 3.18 ohms
95% of Cartomizers should fall between 1.68 and 3.68 ohms
99.974 % of Cartomizers should fall between 1.18 and 4.18 ohm

One thing probably worth mentioning is that the lowest data value present was 2.0 ohms and as such the lower end of the spectrum seems to be more consistent than the upper end of the spectrum. A theoretical lowest value we would ever expect to see (given the limited data) of 1.18 ohms is quite naturally unrealistic as I would think the heating element would just burn out right away(?). We did see a value of 4.3 ohms which we would expect to see less than 0.013% of them time according to this very small data set. That particular cartomizer was quite literally a statistical anomaly and I was very surprised to see that data value.

Conclusion: If I were you I would advise quality control to focus more on the upper end of the ohm scale. The Lower end seems to be much more resilient to wild fluctuations.

If I had more data, I could run the numbers again and be more accurate. If by some miracle of nature we collect 1,000 data points I could get the margin of error down to 3%. The more numbers I can get the better.

Thanks again ManicMaurice for providing this small data set.

And they all said I would never use this stuff. HA!

Math is fun.

I hope this was helpful.

Cheers,
VMS

 

[ManicMaurice]

Thanks again ManicMaurice for providing this small data set.

No problem, glad to help.

Just a little more info, I believe that Leaford has stated that the new cartomizers are within the range of 2.6 to 2.8 ohms. I'm still working thorough my last order, so I don't really have a need to place another order, but when I do I you can bet that I'll be posting more #s.

 

[dragginfly]

You may already know this, Manic. But to be sure... calibrate or "zero out" your ohm meter.

All 5 of my blanks from the recent batch measured right at 2.5 ohms. (The meter shows
2.6 ~ 2.7 due to resistance in the meter leads.)

Valentine, I agree with your interpretation of the mean and deviation from the sample (assuming
the measurements are also accurate). But don't forget the measurements themselves can add
in their own fluctuations. It also seems like a very small sample to extrapolate out to the ends
of the distribution curve (ie. 95% of Cartomizers should fall between 1.68 and 3.68 ohms).

If quality control snips the ends of the curve, the distribution should look more like a square
than a standard bell.

 

[redempti0N]

NERD! but seriously thank you for taking your time to run these tests

 

[jayvap33]

You may already know this, Manic. But to be sure... calibrate or "zero out" your ohm meter.

All 5 of my blanks from the recent batch measured right at 2.5 ohms. (The meter shows
2.6 ~ 2.7 due to resistance in the meter leads.)

Valentine, I agree with your interpretation of the mean and deviation from the sample (assuming
the measurements are also accurate). But don't forget the measurements themselves can add
in their own fluctuations. It also seems like a very small sample to extrapolate out to the ends
of the distribution curve (ie. 95% of Cartomizers should fall between 1.68 and 3.68 ohms).

If quality control snips the ends of the curve, the distribution should look more like a square
than a standard bell.

• Most multimeters don't zero, so you can just short the leads and subtract that resistance.
• But, I work in Test Equipment and unless you have a calibrated Micro-ohmmeter or Milli-ohmmeter or LCR Bridge you probably won't get accurate resistance readings. Not under say 10 Ohms or so anyway. You can, however, get relative consistency measurements if you have good quality test leads.

 

[Valentine Michael Smith]

dragginfly,

Your interpretation of what the curve would like with tighter quality control is correct. As Leaford mentioned in an earlier thread that can also be accomplished on the production end of things (i.e. use better materials, get less fluctuation).

As far the 95% confidence interval is concerned, I don't think most people realize how difficult it would be to get 19 out of 20 (95%) of ANYTHING to fall within a strict standard.

While the sample size is small, it does exceed the 30 samples that statisticians require to be able to even confidently run interval testing. Having a margin of error of 13% like I said is not really acceptable in terms of "risk" and I would love to have about 1,000 more data values and give that nice little 3% margin of error.

One of the points I should have perhaps emphasized more clearly is that even with such a small data set, the numbers are still lookin' pretty good. There was actually less fluctuation and more stability than I thought would be present in the data set.

I don't claim to have much knowledge of electricity, so as far the numbers fluctuating because of the accuracy of measurement equipment, I can't say. I totally believe you that only so much accuracy can be expected though.

 

[leaford]

I luv techies!!

Great analysis VMS!

It fits my observations, although I'm too lazy to run the math, that the majority (rough estimate 85% or so) are between 2.4-2.6 with about twice as many below 2.4 as above 2.6.

And accuracy is precisely why my manufacturer prefers to measure the aamperage under load; that equipment gives a nice steady readout, whereas the multimeter readings always fluctuates, so where to call it comes down to judgement and experience.

 

[dragginfly]

And accuracy is precisely why my manufacturer prefers to measure the aamperage under load

Resistance also varies under load (as the coil heats up).

On another note, design and configuration of the coil probably plays a bigger
role than resistance. I have observed MORE vapor from a Boge 3.0 ohm carto
than from a Kanger 2.8 ohm carto on the same battery.

The new Bloog carto blows them both away, but it's not just the lower
ohms. With the Bloog, far more of the heat is transferred to the juice. The
carto stay cooler.

so where to call it comes down to judgement and experience.

Exactly the reason I hang pictures on the wall by sight. I don't necessarily
want them level... I want them to look right.

My new MF Carto are JUST RIGHT!

 

[SteelJan]

Another way of finding the carto/atty resistance would be a microcontroller ciruit inside the pv. Feedback the amps and volts, let the ucontroller display the ohms.

 

[leaford]

I actually looked into step-down chips when trying to design a cig-sized 5v baty. The engineers I was working with at the time couldn't find one small enough to fit. Plus there were heat issues, apparently you need some sort of heat sink to dump the waste heat generated by the added resistance. But that was a different factory and different engineers, maybe it's ti,e to revisit that idea.

 

[dragginfly]

I actually looked into step-down chips when trying to design a cig-sized 5v baty. The engineers I was working with at the time couldn't find one small enough to fit.

What did not fit? The chip or the battery?

It was my understanding (from reading in the ECF) that batteries
with higher voltage are all too big. If you have one to fit in a 9mm
casing, don't worry about the voltage. Let it run at 6 or 7.4 volts,
then design a workhorse HR carto... say around 5 ohms.

Imagine... it would be twice the "size" in terms of heat transfer.

Think of it like 2 batts and 2 coils in series rather than parallel.

EDIT: I'm thinking in term of voltage drop across each half of
the coil as being 3.7 volts. Total drop 7.4 volts. If each half
has 2.5 ohms resistance, total resistance is 5 ohms. Instead of
about 5.5 watts, it would generate about 11 watts of power.

The idea is you have no need to drop the voltage when you
can design an appropriately sized coil.

 

[leaford]

It was the voltage regulator that wouldn't fit. The battery cell would have been two 3.7v in series, like mods. Basically it would have been an XL size, with a shortie's runtime, but at 5v.

But no, I don't forsee making a 6 or 7v batt. My vision is to have a variety of battery options, but all based around the same cartomizers. I want things to be modular and interchangeable, not have a dozen different systems and need different cartos, etc for each.

 

[dragginfly]

Understood.

However, at 5 volts you still probably need another line of cartos at a higher ohm rating.

Imagine the vapor production if a carto had two coils in it with your new coil design.

All you need is twice the voltage... and a way to stack the coils into the carto.