I just received my unit and the results have been outstanding so far! I am very impressed with how this technology is moving along and look forward to seeing the analyzer get closer and closer to GC accuracy.
I came across this on their website as well:
https://www.cdxlife.com/accuracy-specs/
Well, I like those accuracy numbers. I guess I'll give the new one a try out. It has just been sitting on my desk for a few days. I was wary when it showed up, spent a lot of time with the first one with very disappointing results.
Their "Accuracy Specification" is quite entertaining...
How can they specify a percentage of any of the components this devices supposedly analyzes without knowing how much mass has been placed in the sample insert?
"Approximately 40-50 mg of flower sample was ground against the rim of a MyDx Sample Insert and filled to the Brim" SERIOUSLY? that qualifies as analytical sample preparation?
"approximately 40-50mg" gives a 25% error window to start with, and to have any hope of actually reading a percentage the device should have an entry field for weight of sample.
...but they claim 9% variance for THC. When there is little THC present in the flowers, its primarily THCA which gets converted to THC when heated.
At the very best the "sensor" might detect the presence of terpene vapor. That might get compared against a database and a randomizer spits out approximate guess of THC numbers.
I'd like to see it tested with leaf or maybe crap quality weed to see what numbers it generates.
Or put a chunk of shatter in the sample tray and see what comes out... Or a sample of THC distillate (Clear) which has very little terpene profile...
THC or THCA has very little vapor pressure so I'm very skeptical their sensor can detect it at all without high heat.
They also claim "our data will only be as accurate as the reference database" which suggests all they are doing is manipulating known testing data and presenting it as actual analysis...
If there was a decent technical description of how it worked it might be worth taking seriously. But there is absolutely nothing other than essentially: "the sensor detects stuff, our algorithm does stuff and the result are accurate"
Obviously if there was proprietary tech involves they would want to protect that but there isn't even a vague description of how this thing can quantify components of a sample without any entry of the sample mass, any description of how the vapor flow is analyzed, any description of sensor response or linearity or even the most basic description of how the sensor is calibrated or responds to known standards.
Not to mention Zero mention of what terpenes it is actually detecting. Lots of semiconductor sensors have been developed, but they are usually highly specific to detecting one chemical.
Flashy website and press releases though...
...
From my experience the device is just guessing what you insert in your test tray. I was trying it with different samples of THC-low industrial hemp and also different THC-rich strains and extracts.
The device always showed me a THC content of between 15 and 20%. ...
Yup. highly glorified random number generator. I'm sure if you put lawn clippings in this thing and entered Blue Dream into the app it would still come back with a number between 15 and 20%.
They raised a bunch of money recently so a long con seems to be the business model they are pursuing. Tons of press releases on the web but zero product reviews other than here.
For the money get a thin layer chromatography kit.
RB
I'm guessing the sensor works on NIR (near infrared) wavelengths. NIR analyzers don't need much sample prep (no extraction necessary, no weighing), they are fast, and given a good enough database, are good enough to give the tolerances the manufacturer is trying to attain.Their "Accuracy Specification" is quite entertaining...
How can they specify a percentage of any of the components this devices supposedly analyzes without knowing how much mass has been placed in the sample insert?
"Approximately 40-50 mg of flower sample was ground against the rim of a MyDx Sample Insert and filled to the Brim" SERIOUSLY? that qualifies as analytical sample preparation?
"approximately 40-50mg" gives a 25% error window to start with, and to have any hope of actually reading a percentage the device should have an entry field for weight of sample.
...but they claim 9% variance for THC. When there is little THC present in the flowers, its primarily THCA which gets converted to THC when heated.
At the very best the "sensor" might detect the presence of terpene vapor. That might get compared against a database and a randomizer spits out approximate guess of THC numbers.
I'd like to see it tested with leaf or maybe crap quality weed to see what numbers it generates.
Or put a chunk of shatter in the sample tray and see what comes out... Or a sample of THC distillate (Clear) which has very little terpene profile...
THC or THCA has very little vapor pressure so I'm very skeptical their sensor can detect it at all without high heat.
They also claim "our data will only be as accurate as the reference database" which suggests all they are doing is manipulating known testing data and presenting it as actual analysis...
If there was a decent technical description of how it worked it might be worth taking seriously. But there is absolutely nothing other than essentially: "the sensor detects stuff, our algorithm does stuff and the result are accurate"
Obviously if there was proprietary tech involves they would want to protect that but there isn't even a vague description of how this thing can quantify components of a sample without any entry of the sample mass, any description of how the vapor flow is analyzed, any description of sensor response or linearity or even the most basic description of how the sensor is calibrated or responds to known standards.
Not to mention Zero mention of what terpenes it is actually detecting. Lots of semiconductor sensors have been developed, but they are usually highly specific to detecting one chemical.
Flashy website and press releases though...
This is reeeeediculous on so many levels. don't you think if they could easily test things like, ohhh, food, water, and air in this revolutionary new way it would be bigger news? that kind of device might appeal to roughly 6 billion people for survival, not just trying to identify the terps in their recent dime bag pickup.
If it worked on any level to provide anything of use, why wouldn't every lab in the country replace $100k machines with this thing?
I give this company 6 months before they are forgotten. Even if they could measure something and it correlated across strains (doubtful) they will never ever be able to sell enough devices to hit some sort of critical mass for it to be of any use.