(Possible) Effects of AC Load Power Ripple on Flavor?

[Raven-Knightly]

My mainstay mod is the “illustrious” Sigelei 30 Watt Mini (acquired dirt cheap; built like a tank). Have over time formulated a (for me preferred) Virginia/Perique/Cavendish NET flavoring recipe developed with Sigeleis (in Power control mode, which uses 30 Hz PWM). Have worked down to ~6 Watts (~3 Volts into ~1.5 Ohm single SS 316L coil in the case of the Magma RDA used in the tests described below; and ~2.5 Volts into a ~1.0 Ohm SS 316L coil in the case of Tobeco Kayfun 5 RTA that is my normal reference vaping rig). My e-juice used is ~1.5% (15 mg/ml) Nicotine base in ~60/40 PG/VG with ~20% NET extract flavoring.

Recently I acquired an eLeaf Pico 25 mod (mostly in order to have a Temperature Control - or, at least, a Temperature Limiting device to try out). In TC as well as in Power mode (as seen in PBusardo’s YouTube review of the similar Pico (75W) predecessor - albeit with an earlier chip-set inside - and including some nice oscilloscope views shown soon after the time where the video-link starts-out), there does not appear to be much of any observable AC ripple to speak of on the (nearly DC) drive-voltage (in Power or TC mode).

This (nearly zero AC ripple) drive-voltage all sounded well and good - until my tasting “palate” has firmly and repeatedly “told” me that - at least in the case of my NET flavorings developed using the Sigelei, and when using the Pico 25 in Power mode in order to keep things simple - the same NET-flavored juice in the same RTA tastes rather bland indeed when using the Pico 25. I do not think that the differences in heat sinking of the RTA base by the Sigelei and Pico mods are responsible for these repeatedly noted differences in taste.

Have measured the RMS AC voltage across the 1.5 Ohm coil with 3.0 Volts drive (6.0 Watts). The RMS ripple is a whopping ~1.9 VAC (with the fundamental frequency at the 30 Hz PWM repetition-rate). I have very nearly convinced myself that it is that AC Power Ripple that I like ! I do think that something “magical” happens (due to the AC ripple) that “clean” DC drive-levels simply do not seem to provide regarding taste.

Yes, the thermal time-constant of the coil-wick interface acts as a low-pass filter (thus, to some extent expected to attenuate the resultant “ripple” in the coil-wick interface temperature). However, it’s only 30 Hz, and ~1.9 VAC ripple on ~3.0 VDC is quite a lot of variation in the value of the (instantaneous) power that is dissipated in the electrical load of the coil assembly. Something is different - and I think that this may be it.

(Perhaps) temperature oscillations around the ~160 *C range at which several types of sugars decompose where the “interesting moments” occur on the temperature up-swings - but seem to fade once temperatures above that ~160 *C range exist. Such sugar(s) decomposition effects (may) not be unique to NET extracts.

Posting this in order to inquire as to whether others (may possibly) have formulated any similar impressions (regarding older, lower frequency Power PWM mode mods) ? Or perhaps, few mods in popular use these days use anything as “funky” as 30 Hz PWM ? Any observations by others that may be similar interest me.

[SuperFrog]

You are probably getting heat spikes that roast particulates in your NET or express different compounds. If you set the max power on the Pico a bit too high, it will overshoot if your coil is light enough

[SmilingOgre]

Any ripple that I’ve noted using an oscilloscope has been mitigated by the coil flux and the delta temp/time of the coil.

[Sprkslfly]

Definitely interested in this!
It’s something I haven’t thought about before. Thanks for an interesting thread!

[SuperFrog]

Even for lighter coils? For people using nails as heating elements, sure. But a .3mm Ss coil or worse a .1/.2 mm ribbon coil seems to be too fast for some of the resistance measurement control loops.

[SmilingOgre]

Just sharing my test results. I didn’t test every possible heating element on the planet.

[NatX]

First I have to point out that your repeated use of AC is incorrect, the output is in pulsed DC. I have noticed what you are referring to with the flavor not being the same. Temperature fluctuations at low frequency PWM are the most certainly cause. If you evaluate the temperature delta of the heating wire at 30cps at %50 duty cycle you will find it does swing enough to account for the flavor discrepancy. Even if the delta is only 40f at 420f. The on temperature is reaching ~440f. Flavor is highly temperature dependent and the temperature swing its much greater than you might think.

[NatX]

It is possible that the pulsing electromagnetic fields affect the flavor. Eliquid is chemically complex and a chemically delicate solution, aged over time, the flavor changes as the dissolved molecules reorganize (steep). With weak bonds, and the big swings in the fields around the heating wire that delicate origination may be changed. I have no idea if thats a reality but I have wondered/suspected it may be.

[SmilingOgre]

I’ve always thought it was the zubuzu fields caused by aliens as their space ships enter our ionosphere. As we all know zubuzo fields can cause reverse flux and hole movement in a coil. This makes quarks go crazy and gluons jealous. This is easily fixed however by wrapping your head in tin foil while you vape.

[SmilingOgre]

Actually, all kidding aside, I’m sure it’s cow farts causing global warming. Now your coil is no longer at the anticipated ambient temperature as assumed by the mod and obviously will have diminished flavor from overheating. I know every time I’m next to a cow that farts my vape tastes like crap.

[worm1]

Not an electrical engineer w/ ability to dissect op testing but flavors react different at different temps. If what u are describing it behaving more like a real pwm its averaging the watts and the fluctuation varies the temp tho slightly it may help you pick up diff tastes in the same mix. If picking up different tastes in same mix is more pleasant then a good tc mod then count yourself as a pwm mod fan and try one out.

[SmilingOgre]

I don’t think @Raven-Knightly knightly is talking about PWM. I could be wrong of course. I think he is referring to the unfiltered “ac ride” or noise produced from boosting circuitry. Every measurement of this ac ride I’ve made has been in millivolt ranges and it’s frequency is way beyond 30cps. I never bothered measuring the frequency but a 30cps signal would have a 38 foot wave length which would have been pretty obvious on the scope. It would appear as flat DC for the time interval of my test.

Here is an example where an Eleaf Pico X is boosting across a 1.03 Ω load. The ac ride appears as a wide a jagged line across the top of the screen shot. Although the time divisions do not show in the screen shot you can see the resolution by the 20ms rise time of the signal.

[SessionDrummer]

Good stuff @SmilingOgre

#OgreVision

[Raven-Knightly]

More research (and perhaps some detailed speculations) on my part to come (regarding low-pass filtering).

I apologize if my original post on this thread was unclear. It is the existing (and measured as 1.9 Vrms for a Sigelei-displayed coil-drive voltage of ~3.0 Volts DC into ~1.5 Ohms load) variations of the instantaneous coil-load power when using the Sigelei 30 Watt Mini - not the much more recently designed/manufactured eLeaf Pico 25 - that interests me (in regards to preferred flavor on the Sigelei when using the same RTA at the same time/conditions). There do seem to exist notable (and very much repeatable) differences (to me).

There is no doubt that the Sigelei 30 Watt Mini uses PWM modulation with an ~30 Hz (fundamental) repetition-rate. The mode itself is called “PWM” (on display and in the user-manual). Have a look at the oscilloscope display shown (in PWM mode) here in this PBusardo review (PWM Mode at ~5 Watts shown):

Since we may sometimes be measuring voltage as well as current, let’s (instead of “current”) call the “C” in “AC” and “DC” an alternate name of “component”. There is a “direct component” (DC) with energy at zero frequency (the “DC bias value”). There is (additional) energy that exists at the fundamental (lowest) frequency (in this case, ~30 Hz) as well as at odd-numbered (and as well potentially even-numbered) harmonic multiples of that fundamental-frequency. What an AC RMS meter measures is the effective heating value of the summation of all of the alternating components (at non-zero frequency). The “DC bias” component (if also present) is blocked by the AC RMS meter. Below are graphic diagrams of the frequency spectrum (including zero-frequency) of a 50% duty-cycle square-wave (with odd-numbered harmonic energy only) as well as a “pulsed” waveform (shown as having some energy at zero-frequency, which is “DC bias”):

The “DC bias” (zero-frequency, constant) component derives from (heavy) averaging of the (varying) data. This is what occurs when the DC Volts mode of a multi-meter is used. AC Volts mode (whether using average, or actual RMS method) measures only variations. RMS represents actual resistive heating value resulting from the variations of the time-domain coil-drive voltage/current (as opposed to a heavily averaged “DC bias” measurement, which reports an averaged approximation of any existing “DC polarity bias”).

The ~1.9 Volt alternating component that I measure which is present (across a ~1.5 Ohms resistive load) when the Sigelei displays ~3.0 Volt DC coil-drive is of a significant numerical value (relative to the ~3.0 Volts DC value displayed). ~2.407 Watts total of (alternating) components ((1.9)^2)/1.5) = 2.407. To be expected (using this 30 Hz repetition-rate PWM). The important thing is the coil-wick interface temperature variations.

I do not have a thermocouple with which to attempt to make temperature measurements on or around a wicked and operational (single) coil. Three issues would seem to exist as “confounders” with such pursuits:

Good thermal coupling between the thermocouple probe and coil under test is quite hard to achieve;

The mass of the thermocouple leads is significantly larger than that of the coil assembly measured;

The down-stream electronics in the test-gear is likely to (itself) include low-pass filtering (averaging).

.

(Of course), many readers may well “know all of the above facts” [about AC(rms) and DC-bias, etc].

Way over my knowledge-level where it comes to the “electromagnetic field” angle. From a tad of reading, Oxygen (when existing as O2, anyway) is “paramagnetic”. None of the other expected organic elements (Hydrogen, Carbon, Nitrogen) have near the magnetic susceptibility that Oxygen has. Can someone explain the proposed mechanism(s) involved ? If not, I’ll for now stay with coil-wick temperature over time analysis.

.

Any observations by others of noted differences between funky old (30 Hz, or some low frequency) PWM and snazzy new DC-DC conversion rigs (showing very low coil-voltage variations on oscilloscope scans) ? To keep things simple, I am referring to mod usage in Power Control mode only, utilizing same RTA/RDA.

[Raven-Knightly]

Went back to an earlier post of mine regarding 10-90% rise-times of temperature at the coil-wick interface, and to some data collected pertaining to some measured temperature increases under specified conditions:

Source: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5502764/

The -3 dB bandwidth (in Hz) for a first-order low-pass filter represented by a real pole in the S-plane, which is the case for all poles describing heat flow is approximately equal to [0.35 divided by (10-90%) rise-time].

Using a (roughly estimated) ~3.5 Second 10-90% temperature rise-time (in the case of my RTA, ~1.5 Ohm coil, ~6.0 Watts), one gets a (-3dB) low-pass filter bandwidth of ~ 0.1 Hz. This would produce an attenuation factor of ~300 (~50 dB) at the fundamental frequency of 30 Hz, with the attenuation-factor of the harmonic multiples of 30 Hz (also present) increasing further in proportion to the harmonic-number (ie, 2, 3, 4, … etc) - reducing estimated coil temperature variation magnitudes (at various individual frequency components involved, and/or when considered in summation in time-domain) to seemingly low numerical values, indeed.

Despite the previously mentioned difficulties and errors associated with thermocouple measurements, the above numbers would seem to strongly support SmilingOgre’s statement (appearing above in this thread):
Any ripple that I’ve noted using an oscilloscope has been mitigated by the coil flux and the delta temp/time of the coil.

.

That leaves me wondering. The truth must be out there. Meet you under Tinfoil Domes at dawn on Equinox. My internal nerd notes the only known remaining difference to be that of individual mod body thermal heat-sinking characteristics (much more heat-sinking mounted on the Sigelei than appears to exist when mounted on the Pico) - but that seems only distantly related to the energized coil-wick interface region temperatures. One could hypothesize that this heating-sinking keeps the e-juice in the RTA “cooler”. However, inter-mod differences noted seem immediate and constant in nature - and not a gradually evolving characteristic.

[SmilingOgre]

I admire your tenacity regarding research. I have also witnessed the differences in vape by simply changing mods… Can’t give you a cause. I’ve never pursued the issue. I can achieve a fairly decent connection between a thermal couple and the coil but the reaction time of the couple and meter negate any time measurement that would assist in working with this theory. IMHO your on the right track and considering the critical variables.

[anon84779643]

I want to say it i s a change between recipes and set ups… the way I mix, and you mix even tho they might be the same recipe… it is the how each of us do it… one drop could push a recipe one way the other… one coil vs the next… a simple loop or the squeeze of a screw…

[Raven-Knightly]

It’s simpler than that, Smoky. Same RTA at (what is essentially) the same time - mounted on two mods. The distinct differences in flavor between the (Sigelei and Pico) mods are immediately and repeatably evident.

[anon84779643]

@Raven-Knightly, of course… even like I said… a kink in a coil… manufactured juice is also just like diy ejuice… go over a bit… or under a bit on the recipe… nothing is exactly the same… wish it was…

plus you are talking 2 different mods… even at the same watts/tc what ever you are doing… you will still get different results…

[SmilingOgre]

I don’t think that is what @Raven-Knightly is experiencing but totally agree with you. Slightest change in protocol will effect the flavor per my experience.