Vape Temp Profile Charts

[stickstones]

Ok, here is my first useful submission to this thread. I just charted a VapCap M being heated by @Pipes Jarhead induction heater.

I had to use a makeshift screen that I poked a hole into, but I got the sensor right about where I wanted it.

I then filled it with herb and ran the test. You'll notice that the first draw actually brings more heat into the bowl, while the second takes heat out of it.

I plan on running a similar test with the Ti Omni for comparison's sake.

Now we’re having fun! How do you easily get those labels on yours? You won’t believe the bullshit I went through last time.

 

[Stu]

How do you easily get those labels on yours? You won’t believe the bullshit I went through last time.

For simple stuff like that particular chart, I like to use FastStone Image Viewer as it has quick and dirty editing tools. It's free, too.

 

[rz]

Great thread!

I have some interesting charts to share too, but they are NOT of the actual temperature of anything. (/end disclaimer)

I'm using a melexis mlx90614 Infrared sensor to measure the temperature at the end of a VapCap tip.

I heat the tip with a PID controlled induction heater based on the common Royer oscillator.

The temperature measurement is heavily influenced by the VapCap tip, but is affected by a few things. Some points:

• The emissivity of the vapcap tip, which changes between caps largely depending on their prior exposure to any lighter. This would need to be calibrated for a specific cap for higher accuracy.
• The IR sensor I'm using now has a Field of View of 35deg so it doesn't get NEARLY as much stray IR as the previous 90deg ones I used, though I have a feeling the focal length is a bit further than the distance it's currently at from the cap, which may explain the little peaks seen as I insert/remove the cap and the distance increases. I need to experiment more with that. I have a 10deg FOV sensor I need to play with too.

So in the end, I get some uncalibrated number which doesn't REALLY bother me since it could even be on a scale from 0 to 0.100 , as long as it is precise enough and influenced most significantly by the actual object temperature that we're measureing to work well for this cause.. and it IS

Here are a bunch of measurements from a typical session. This one had 11 hits. I stretched the heating up steps out a bit, and I often start a bit higher. The experience is very 'spliff' like with the timing, I just place it back in between hits. I can finish a load in between about 3 and ~10 hits depending how hard I push. I typically do it in about 6~7.

There are a few other measurements thrown in which I've scaled up to fit in.

• Frequency of the Induction heater, about 24Khz. The circuit is self resonant, so the frequency changes when a capcap is inserted, this is used for insertion detection
• Object temp - ~ the VapCap tip. I zeroed the result when VC removed to reduce confusion. IR sensor continues to pick up IR from the coil and tube, like shown near the end where result wasn't zeroed when VC removed.
• Body temp - internal temperature of the IR sensor.
• Voltage*10 (ie, it maxed out at about 8.5Amps), and voltage*10, I started at about 7.5V.

Too many little things going on to get into it all here, but glad to discuss

 

[Stu]

Here's a chart of VapCap Omni in use:

Some observations contrasting this chart with the one for the M.

1. The Omni clicks a few seconds quicker than the M. Between 3-5 seconds (I've timed them separately several times to validate this). This makes sense as the M has more mass to heat up (5g version 3.5g for the Omni).
2. The M gets about 10°C hotter the Omni in similar usage tests. IMO this is due to the excess energy transferred to the M (stays in the IH heater longer).
3. The M stays hotter longer. I wasn't even able to do a second draw on the Omni before the cooldown click came. This I believe is a combination of the extra mass of the M holding the extra energy (point 2) and the fact that Ti conducts the thermal energy away to the surface allowing to radiate the heat away more quickly.

Both vapes work very well and I don't think one is "better" than the other, but I think it's helpful to understand what's going with them and why they perform slightly different.

Here's the M chart for comparison

 

[stickstones]

I did a time test on 5 different vapcaps a couple of weeks ago and they all came in within 8 to 10 seconds on the Portside, but they didn't all have the same 2-clicks. Some had a gap between both clicks and one was very close. I came away thinking each cap is a little bit different and I could probably find a favorite one based on how I like the click.

 

[rz]

I've found that when I heat up quickly(full throttle till click), the click is at a lower (measured by IR) temp. when heated slowly, measurement@click is quiet a few degrees higher. I usually start low and bump up in a few steps over hits, but same is true if I tweak PID to be slow or bump the setpoint while inserted, so I often get a pretty thorough extraction without ever clicking at all. I get better results on first cycle when heated slowly anyway, to let the whole tip get up to temp. It may be an IR thing too, so I wonder if you could reproduce with thermocouple (I'll have some in a few days) by say, a 6 second initial heat up, rest for 4 seconds, then till click and compare click temps. Or, alternatively, compare click temps within the same session to get a (from) cold click and (from) hot click. I wonder how pre-heating will change the post-click count strategy.

 

[Winegums]

Cool thread! This reminded me of the early post that @vtac made in The Herbalizer thread with some data logging graphs from an Herbalizer. I wish I could hack into my herbalizer and pull some data from it.

http://fuckcombustion.com/threads/the-herbalizer.9116/page-2#post-372825

 

[Flower of empathy]

I had to use a makeshift screen that I poked a hole into, but I got the sensor right about where I wanted it.

i would love to see a 2. measurement with the sensor touching the edge of the bowl
the middle should be the 'coolest' point of the vc tip...
Would be interresting in general to see the temperature differences/gradient inside the bowl of different vapes...

 

[jojo monkey]

Would be interesting in general to see the temperature differences/gradient inside the bowl of different vapes...

I am gonna try to make something that can do this. The rig will be using an arduino and a few probes.

The temp on the outside is also something I want to focus on. The MLX90621 has a 4X16 grid and costs under $50 usd. That's 64 measurements!

here is a guy messing with it on a car to give you an idea:

The MLX90640 will be 32x24. I can't wait for that to come out. Flir is $$$. $240-ish bucks gets me a flir that does 80X60 or 4800 measurements... crazy. The US gov caps them all at 9fps.

If I didn't care about recording the data, the seek thermal for android/iphone is tempting. https://www.amazon.com/dp/B00NYWAHHM/ref=asc_df_B00NYWAHHM5293151

Back to K probes, the response time of the probe is something to consider. I think we want very thin wire and no sheaths.

https://www.omega.com/techref/ThermocoupleResponseTime.html

My hope is that a series of fast probes and taking many measurements per second will paint a nice picture.

Just a side note... I did make a controller that measured near the heater and graphed in real-time. It was reading the temp 10 times a second.:

 

[rz]

gonna try.. MLX90621..

That looks like an awesome sensor!

I'm using an MLX90614 single point 35deg FOV to look at the tip of my VapCap(VC) as you see in the image a few posts above.

I'm finding it very difficult to get an accurate measurement of temperature, probably from a few different reasons. For instance, temperature gradient issues are mentioned a few times in the datasheets for these IR sensors, and my sensor is held by the glass thats touched both the IH coil and the VC, and the sensor rises to about 50~60(c) (~130F) by the end of a heavy session so there's obviously no equilibrium there... It used to hit 70~80 though I put a thin slice of cork cut into a ring to prevent direct VC cap contact with the sensor body. Another issue I have is that when I pull my VC up a about 1cm, I get a temperature reading that makes more sense - possibly due to focal length (which isn't mentioned in datasheets though - and there is no lens, so probably not focal length). When lifted by about 1cm, I'm getting reading about 20deg(c) higher (also visible in the graph I posted above).. which is odd. possible due to extra thermal energy bouncing around in there? I'm not sure. There's still loads of theory for me to learn, and I'm sure some math to redo, or get around to experimenting with the 10deg FOV version I have (and maybe thermally decouple the sensor by mounting it further back). Also from what I understand that these smaller FOV melexis sensors just have part of their view obstructed by shrouding, and are calibrated accordingly, and this may not work well when there is a temp gradient, which there usually is when in use.

I played with a thermocouple today, so now I have a few extra question marks in my head. I just used the tip+cap with no filter, and pushed the sensor down in there.

So I've come to the conclusion that the IR sensor in my current setting is NOT scientifically accurate, but that doesn't really matter either. The fact that there is some feedback element heavily dependent on the VC tip temp means I can set it to SOMETHING, and get pretty persistent results with minor quirks. My flow is to start at a lower temp to avoid surprises (the IR sensor has the largest temp gradient on it at this stage), and bump up a few degrees between hits, and I get the expected behaviour (albeit at an unexpected number). If I stop on a temp, It'll be persistent enough to stop producing vapor until I bump up, and I can taste when it's getting close to too much, which is pretty persistent reading between runs. I can really squeeze out a load without going overboard within 3~10 hits depending on my bump up pace, or blindly drive someone else's VC session. I haven't burnt a load unless my starting temp has been too high and it kind of overshoots, so I'm still REALLY happy that the IR sensor works as well as it does (I feared the IH would completely wonk it out)

Yesterday I played with a Piezo buzzer (inspired by this thread much?) - I used an elastic band to hold it agains the VC body, and by looking at my scope I could def. pick up a click. It was a pretty small signal which would need a bit of conditioning to avoid otherwise damaging my microcontroller from any little knock, but it could def. be built into a rig to graph out cap temps and clicks.

Meanwhile, I got loads of components delivered today (mouser and adafruit), and hope to have some PCBs (from two different projects I'm working on) in a few days time, so I'll probably be busy building for a while..

P.S I really find the temp gradianT along the VC tip very interesting. There are so many things going on in that seemingly simple tip . You could heat the cap to click quickly and close to the end, and get hardly a wisp of vapor by the time anything gets to your lips, or heat too slow from further back and completely char the load by the time the cap clicks. That tips thermal mass is very important to keep things hot so you actually get to use it, though the gradianTcan't be ignored. I'd love to get some IR array looking at that. Go @jojo monkey !

The gradient stands true for every vape I guess. @Winegums that Herbalizer post was interesting. Thanks.

 

[rz]

This evenings tests are juiced by a ~7.6V 2S pack with my all new set of VTC6s powering an IH.

I haven't depleted my first charge round with these VTC6 yet, but I can def. tell that my pack voltage depletion rate is much slower than my previous 25Rs. Anyway, this is not tonights topic.

I got around to comparing the M and and Ti Woody in my setup. A step response would have been better but this is not so far off and still has LOADS of interesting bits (depends who you ask I guess).

My setpoint was 95°c. I used the same cap in both cases to maintain the caps emissivity for the IR measurement.
Started both runs when all sensors settled at room temp, 26°c

28 seconds after insertion I turned the controller off and let things cool down for ~2.5 mins (it was actually 30 seconds after I turned things on, but the 2 seconds before insertion are removed).

Some of the observations that stood out to me:

The current is higher while using the M. I noticed this in use already.

Heat up time in my IH unit (and I would guess in other set ups too to some degree) does not differ drastically compared to Ti. The extra (thermal) mass that's heating up seems to weigh out from that extra current - this would not be the case with a lighter.

The SS is more prone to overshoot than the Ti for a given PID setting. This can be seen as controller in both cases performed almost identically, and the SS continues to heat up a little even after controller hit minimum.

Cooldown rates don't look significantly different in THESE measurement, they are slowed down by being in a warm glass tube, though the SS did remain hotter for longer, it does appear to have reached a slightly higher temp (due to that overshoot). Near the end of those first 28 seconds, the PID controller started to heat the Ti up again a little since it noticed it cooling, while the SS remained hot. since the SS was a few degrees hotter, you'd expect it to cool down faster due to larger delta T, though it's mass keeps the temp dropping rate similar to the cooler Ti

This really shows how the SS requires more energy to heat up, that the thermal energy takes time to spread through the tips (more so in the SS), and how the SS stays hotter for longer

Again, a step response up and down would have produced clearer results, but theres a lot of info buried in here Maybe next time.

Hm I'm finding it hard to compare how these worked with a lighter since I haven't used one for a few months (OK maybe once).

Crazy that It's been almost 10 years since my uni days, and about 10 years since I studied any thermodynamics, which I didn't specialize in anyway. I was much more interested in wave propagation stuff at the time. I somehow completed most of a physics degree, which took a turn towards electronic engineering and wireless/digital communications at some stage. I can hardly integrate for sheeet now, so I'm rusty to say the least and my analysis comes with a sack of salt. Glad to discuss though

P.S The current rising in the first ~8 seconds is NOT due to the PID controller - it is at 100% in both cases. It is probably mostly due to thermal characteristics of the MOSFETs and the chokes. The cause of differences in current consumed by IH between the SS and Ti is harder for me to figure out, I've ggot a bunch of ideas but confused since IH resonant frequency doesn't seem to change much between them. Still something to think about, but it's zzzzz time for now..

 

[jojo monkey]

I'd love to get some IR array looking at that. Go @jojo monkey !

I hope it works out. I'm excited to wire it up. It is very cool that there are a few code samples and libraries for it. The soldering should be the hardest part.

I'm getting reading about 20deg(c) higher (also visible in the graph I posted above).. which is odd. possible due to extra thermal energy bouncing around in there?

Do you do anything to compensate for the warming? Are you using the C version of the mlx90614?

from the doc:
"This effect is especially relevant for thermometers with a small FOV like the xxC and xxF as the energy
received by the sensor from the object is reduced. Therefore, Melexis has introduced the xCx version of the
MLX90614. In these MLX90614xCx, the thermal gradients are measured internally and the measured
temperature is compensated for them. In this way, the xCx version of the MLX90614 is much less sensitive to
thermal gradients, but the effect is not totally eliminated. It is therefore important to avoid the causes of thermal gradients as much as possible or to shield the sensor from them. "

I am also curious if you played with setting the emissivity?

 

[rz]

@jojo monkey yup, I'm currently using the BCC and have some BCD(10deg like the BCF but has a shorter shrouding) modules to test. I still have some tests to check effects of gradient (like heating VC in a different module) that I just haven't gotten around to. I think the solution may be to thermally decouple the sensor and any hot parts as much as I can. I need to create some (semi)transparent silicone sleeve (i have LEDs in a ring around the sensor on my next PCB that'll shine through the tube) that'll couple the sensor and tube on the other side since things need to be aligned pretty well. For now the thin slice of cork helps a bit. Fortunately this issue doesn't have much of a detrimental effect on the way I use the device, so it's not currently a priority just something I hope to address someday. I have played with emissivity but not extensively. I see the phenomena at various settings, but I haven't dived deep enough for finding best setting. Emissivity also varies a bit between caps, some of my caps already have a patina on the top. For accurate readings, I'd have to calibrate with a k-type thermocouple anyway. I could just map a typical target temp range to 0%100 and be done with all this but that's no fun

MLX90621 hookup looks identical to 90614 hookup. PM/mention me if you need any assistance with your proj.

 

[Stu]

Did some testing with the Flowerpot Showerhead last night. Please note that the probe was placed about 5mm below the screen, so these temps are not bowl temps, but just downstream of the bowl. Hence the lower temps.

First, here is a couple of hits at 680°F with no carb cap: Note: this test was performed with the probe inadvertently making contact with the inside of the bowl wall. The graph is interesting, but it must be noted that it is not tracking air temp, so take it for what it's worth.

Next, here's one hit with the carb cap put on mid hit, and the second hit with the carb cap on the whole time:

Notice how the temp increase slows down when the carb cap is applied. I speculate that the choked intake limiting draw velocity reduces the amount of heat energy going through the system. I think that the reduced temp is partially offset by the lowering of the pressure due to capping, so the vapor production doesn't necessarily suffer. That's my theory anyway...

Here is a chart of two hits both having the carb cap applied mid hit:

 

[jojo monkey]

MLX90621 hookup looks identical to 90614 hookup. PM/mention me if you need any assistance with your proj.

Thanks, rz. That is a great offer.

Your posts have definitely made me more conscious about how I set up the sensor. I appreciate the detail.

this test was performed with the probe inadvertently making contact with the inside of the bowl wall.

Still interesting stuff either way.

You have me thinking that measuring the screen itself is something to be done. They do get damn hot and are the only thing touching the herb.

 

[Stu]

I almost forgot that I temp charted the Solo on Saturday night. I think I consumed too many marijuanas that night.

Solo test, level 4:

Seems to match up pretty closely with @stickstones test of the Air 2 (surprise surprise).

 

[stickstones]

@Stu what was the set point on the solo? I’m wondering if you’re getting readings that are low similar to mine.

 

[Stu]

@Stu what was the set point on the solo? I’m wondering if you’re getting readings that are low similar to mine.

Solo test, level 4:

I don't know the set temp of level 4 off the top of my head.

I don't expect the herb chamber to necessarily match the device's temp readings since the probe they are using is outside the chamber. So they are either guestimating or maybe have an algorithm to calculate the temp?

 

[jojo monkey]

I am just getting prepped with the parts I have on hand.

The brains of the setup is called a Teensy. I am using a teensy 3.2(arduino family). The teensy is hooked to a usb port on my pc.

To read from the probes I am using 2 "Adafruit Universal Thermocouple Amplifier MAX31856 Breakout"'s. (XOXO Lady Ada.) This breakout board reads just about every kind of probe. I am getting readings out to 2 decimals with it. The breakout also has a thermometer in the chip so I get a reading at the port and the probe.

To read the air in a super fast way I went with the Omega CHAL-005-BW. The butt weld is the fastest probe they have. I have not received it yet. The specs say it should be able to go from 100F to 800F in still air in 1 second. The wire is extremely thin: 0.125mm (0.005inch). I am worried they will be easy to break. Fingers crossed. On the plus side you get 5 in a pack and they are made from the same lot so the results will be similar.

For now, I hooked a big-old metal K probe to a single MAX31856 breakout just to see it working. As a quick test I touched the end of the probe for a few seconds. The slow slug takes forever to go back to room temp. Well 10 seconds, but that would be no good for measuring what is going on in a bowl.

I was able to squeeze 10 measurements a second out of it. The units are in milliseconds. 5k=5seconds, 25k=25seconds,etc. This was just a test of me going A-Z.

This weekend I'll have most of the parts to play with.

 

[jojo monkey]

I got in the parts. The "Omega CHAL-005-BW" are crazy thin, but not as bad as I imagined.

I am still trying to figure out an easy way to rig a bowl.

For my first test: I positioned the probe in the middle of where the GS exhausts(red circle). The GS was placed on top of the adapter in the pic. The 3way is taped off where you see the yellow dot.

At the bottom of the pic is the teensy 3.2 and the breakout.

I put GS + probed 3way on top of a bowl and fc-ufo bubbler. I figured pulling with the fc-ufo is more real than anything. For the graph I take a 4 second hit and then a 5 second hit.

Want to zoom and see data? Public graph here: https://plot.ly/~symphony.vapor/4/

I am happy with the results. It will be neat to see what is happing in a bowl and to beat it up with many draws in a row.

 

[stickstones]

Damn! Those are beautiful graphs!

 

[Stu]

While I agree that it would be cool to see a bunch of vapes on one chart, I don't think it's practical, nor all that useful in the end.

It's not really practical because there is no way (that I could find) to export the data into a useful format so that the charts could be overlayed/combined into one chart. The charts that the software provide are great, but not very tweakable so customizing the temp range and or time elapsed isn't an option. @stickstones can correct me if I'm mistaken (it's been known to happen).

Embarrassing self-quote to admit that I am indeed a dumbass.

The software does actually allow exporting of data in csv format. You just have to save the file in the program - something I neglected to do. Armed with this knowledge I plan on doing some more testing and putting two different devices into the same chart for a nice visual comparison.

Sorry for the fake news earlier.

 

[OF]

@OF always on topic, but some new data is available
http://fuckcombustion.com/threads/vape-temp-profile-charts.25780/

Thank you @stickstones for the chart showing temperature of air2 while in use

Anyone can believe what anyone wants, facts are facts and words have a meaning, air2 is proved to be an hybrid (convection/conduction-radiation vape IMO)

Good on you for finding Stick's graph. Please note the discussions around it. Also note the temperatures. IMO it clearly shows conduction going on, not convection (where the core would heat from the airflow and heat the outside?).

I know guys really like the idea of convection, perhaps not really sure why except they've heard 'all over town' convection is better. Lots of sales types say so?

I agree, anyone can believe as they want. But if you're going to claim science on your side youi'd best be prepared to defend that claim in the face of 'peer reivew'? Thermodynamics (the FLOW of heat) tells us we need lots of energy (measured in either Joules or Calories, not degrees which is different). Thermodynamics studies that flow and has hard rules like 'heat flows from hot to cold' that must be obeyed.

Convection here, where the source of the heat (the cup) and the target temperature are basically the same (say 400F) conduction is the only mode that can possibly deliver enough heat energy to overcome the losses and make vapor. Like I said, if conduction was the rule you wouldn't need heat soaks? That is it would be like VG, Thermovape or other convection vapes such that you could put in a cold stem and draw hot air through it and make as much vapor as you want 'right away'. In convection superheated air would continue to flow in at the same rate and the hit would 'never collapse'. You wouldn't have to 'milk' slowly to 'chalk' your glass?

Convection is just not an effective way of adding energy to the load in olo

Perhaps the most telling is PIU agreeing, AFTER CAREFUL INSPECTION that there is no (useful) convection in play. He has every reason to keep up the ruse and sell more vapes, but chose honesty?

"The Arizer Air was widely believed to be a convection / conduction hybrid vape but after fully taking apart the upper chamber and heater we can safely conclude the Air and Air 2 and most likely the Solo and Solo 2 are purely conduction vapes without any convection heating."

OF

Edit: One other thought. Since 'heat flows from hot to cold' is a Law that means the cup (the hottest part) is the hottest part? That is it's at say 400F and as shown on Stick's graph 'everything else is cooler' than that?

Since energy is needed to make vapor, heat must flow into the load from somewhere hotter to support that. That too shows on the graph.

Air and vapor leave the load 'at vapor temperature', say 370 F? That's a couple dozen degrees loss at a maximum of all the air was 'fully tempered' (which of course it can never be since only a tiny fraction of the input air actually touches the edge of the cup on it's way in). So we have a pathetically small heat contribution available that way. The average air temperature of air entering through the vents is trivial for our needs.

To work you need LOTS of surface area to heat air (consider how much is in VG heat exchangers) and you need a much much hotter source to heat the air many degrees on average. Hundreds at least, not the very few possible here. VG uses a touch heating the exchanger during the hit, 1200C source, for 400F. The thermal chamber in TV ran around 1300F I was told (seriously glowing inside) and heated the air going through mostly by radiation, not conduction, in fact.

It's the way such things work, at least as I was taught........and in return taught in the days when I was paid to do so. Same as Randy was taught it seems.

No significant vapor made by convection (or radiation). For practical purposes all the heat energy input to the load is conduction through the hot metal cup. That's how the energy gets in, that's what makes it work so well for us.

OF

 

[Stu]

Here is a test I just did with the Haze V3. I took two draws.

Note that the temp drops during the hit as it does in Sticks' Flowermate chart. Contrast this with the temp increase during the draws with Arizer portables.

 

[KeroZen]

But why is it climbing higher after every hit? Didn't you wait for a stable temperature before the first hit and did you bump the temp between hits?

Looks like it's at 166°C before first hit and ends at 176°C at the end, that's a lot of change. Note that we also see the same trend in your other graphs.

PS: I thought we should summon @OF here but apparently his answer in the ArGo thread was moved here already...