I really did not mean to shoot your idea down at all! I think it would be great to be able to use GC-FID for all of this, since it is faster, easier, and cheaper than HPLC, but I have quite a bit of experience with analytical chemistry, GC, and even SRI units in particular. One of the biggest problems with GC, though, is the high rate of varying amounts of thermal decomposition… whether that is decarboxylation, reduction by the presence of H2, or some other mode of reactive chemical change. That is why HPLC-UV (single wavelength or most often DAD) is the standard analytical instrument for cannabis resin. However, even HPLC can be fooled, especially in the hands of operators without sufficient education or the will and workplace ethics and tools that allow them to approach perfection of analysis through careful examination of all available data (e.g. cross analyses via DAD-UV, FT-IR and nIR spectral detections to weed out spurious identification and/or convolution of peaks).
In my opinion, if you want to do in-house analytical, you should use a well-validated and majority operator-error-reducing method of detection that is entirely different than that available from 3rd-party analytical labs. For example, using the Sage Profiler II I just got, cross analyzing a purchased batch of CBD with it (and a conversation with Alex, the PhD engineer who designed and built the device), I was able to suss out an answer to a long-standing curiosity in my 3rd-party HPLC-UV-DAD analysis of the material. The seller of the CBD told us it was extracted from hemp and worked up to 99+% pure (of course), but 3rd party analysis came back with a very weird result of 96% CBD and 1% THCa! As you probably know, it is not possible with our current methods of extraction and purification to end up with THCa in our extracted CBD… especially with no detection at all of THC! I went to the lab and sat down at the computer for the HPLC-DAD that was used to analyze my sample and had the lab director bring up my sample. The chromatogram definitely had a small peak in the ROI (region of interest) for THCa, but they only use 1 wavelength in the UV range to display the relative absorption of all the standardized cannabinoids… so I had him show me the full DAD spectrum (the full range of simultaneously collected UV wavelengths) of this “THCa” peak, and although it absorbed strongly in the single wavelength used for the chromatogram, the spectrum was shaped nothing like THCa or any other cannabinoid in their database! Basically, that just told me that the 1% THCa was a false positive.
Well, because of the way the Sage Profiler works on algorithmic modeling of thousands of actual pure and impure (unintentionally and intentionally oxidized, reduced, light & heat decomposed, isomerized, and otherwise degraded) samples of plant-extracted CBD-containing materials, or iow CBD with all kinds of cannabinoid, terpene, quinoid, plant matter, lipids, lecithin, sugars, etc. in the background, this CBD could not even be measured on it! Instead, the analyzer gave me a 60 to 70% “fit error”, meaning that although the sample fit the algorithm based on all these genuine sample models by 60-70%, there was something highly absorbing of nIR light that made it 30-40% unlike any sample model ever used to create the algorithm. Combined with the information from the DAD spectrum, I now knew that this CBD was either intentionally cut with something, or it was synthetic CBD that still contained at least 4% reagents and possibly some catalyst. Since the “cut” was only 4% and the CBD completely dissolves in any solvent I used to dissolve it with, among other evidence I collected, my guess is the latter; it is simply synthetic CBD that had not been completely purified.
I am not trying to sell the Sage Profiler units with this story, although they did decide to make me a distributor when I bought mine. Since the Profiler only analyzes THC, CBD, and their carboxylic acid counterparts, it would not completely fulfill your needs, anyway. Near infrared (nIR) like the Profiler uses is great for accurate quantitation against complex backgrounds, but it is not best for identification because the peaks are more like broad hills and valleys. I am just explaining how important cross analysis with various detection methods can be; and since every 3rd party lab uses HPLC with UV and/or MS detection, you should consider some other form of detection… like FTIR (Fourier-transform infrared), for example. Granted, although IR makes numerous sharp peaks based on each individual bond in the molecule, giving its EM range the nickname the “fingerprint region”, it best for pure or nearly pure materials rather than mixtures. I am hoping that some day soon there will be a company creating cannabis resin mixture fingerprinting algorithms for infrared detection technology.
For now, if you are willing to deal with cost of consumables and a bit of complexity in sample preparation, and you need something affordable but effective for various common cannabinoids, you should consider the Orange Photonics Light Lab. Unfortunately, I know of no “affordable” terpene analytical systems. You just have to get a HPLC system and a boat-load of standards.