Unfortunately I have no experience with your generator, though I may be able to help with some of the other questions. On the topic of temperature, you're correct on the Celsius scale. The temp limit is likely a control measure for evaporation and the protection of instrumentation etc. If you make CS without a current limiter, you will have no way to calculate your final ppm and/or the overall size of the particles or ion clusters in the solution. Granted, there's always the possibility of stopping the process before it becomes excessive, though there isn't much one can do to control particle size beyond that. - this will however, produce your colloidal silver solution as fast as theoretically possible. And so there's that to consider. I suppose. Interestingly enough, I've found that the smallest particle size didn't make the most effective solutions against infection. Which I found quite interesting. IOW. it appears as though the slightly larger ones are more effective than the smallest. This was observed using a very large 99.9999 anode at optimum load at .95mAh @ room temp. Which took nearly 36 hours to process. - But didn't prove to be very effective against the heat processed 1 hour batch produced at 8.4mAh (which is light yellow in color)- Both of which measured very similar in terms of ppm. On the topic of ions, I'd add that while they do repel each other, that the field which makes this possible can collapse allowing them to combine. Which in turn, creates ion clusters. As for the color of your solution(in this case) is almost always as a result of light being blocked(or reflected) through your solution. Which almost always happens as a result of high particle content. Though the type of particles are not always the same. ie, oxides will give turn a solution yellow or gold long before ions will. Likewise, the addition of any product such as a reducing agent that will allow the ions to bind will change the color of a colloidal silver solution. That being said, if you're using a generator which uses alternative polarity(polarity reversal) so as to extend the need for electrode cleaning in the process. Then you can be sure that the oxides are left in your solution. And I don't mean this to attack any specific product as I know there are many out there who use this approach. Though this has been my observation on matters.

If you heat your water during processing, this will affect the conductivity of your solution which will likely throw off your readings.

Though TDS/PWT meters certainly have there uses, I'd say the most accurate way of measuring ppm would that of combining a weight differentiation with theoretical potential. ie. collecting a min. sample of 3 batches weighing your anode before and after and verifying against the common faraday calculation found here.

In the end, it won't be easy(perhaps not even possible) to make a 20ppm colloidal silver solution using conventional methods due to the nature and limitations of the solution. Though you should be able to reach between 16 and 18ppm on average without much trouble. Anything more, and you won't likely have enough stability for the solution to remain colloidal. ie, when you heat your solution, you raise the saturation point(due to the absence of oxygen). Which will effectively allow you to reach 20ppm. However, once the solution cools, the properties of the solution return to its initial state and you'll find the silver plating and/or pooling(silver puddles) as a result. Now if you're studious, you might even employ rapid cooling which can seemingly bend the laws of physics(brownian motion), but only for a slightly longer time. :) Though I'd add that achieving 18ppm is very very good in my view. And has proven to be very effective. Hope this helps