133 - Building a 3kV DC power supply with a module from Start Spellman 

I actual did this (the plot, not a regression test) but did not show. It was pretty linear just from looking at the plot. At the end I discovered that the Vout can't handle much load and unless you use a >1MOhm meter, you get non-linearity just because of that. I really wanted to use a panel meter but they are all rubbish (and non-linear themselves) and terrible at showing low voltages. I don't want a meter showing zero if there is still >200V present. I felt those problems are more relevant in the video than linearity of Vmon to Vout. I settled for an analog meter which is even worse in linearity and accuracy than digital meters but at least shows if there is still voltage present. In hindsight, a panel meter with an additional "output not zero" LED powered from 10V but controlled by a transistor connected to Vmon would be a better solution.

@@TheHWcave Thanks for the update. Yes, the "invisible voltage" is really bad. What if you use a microcontroller either with its ADC or with a dedicated ADC for higher accuracy? Also, if that little current draw already changes the output, what is your application that this does not matter? Thank you for your good stuff. Currently using the TC66C next to me and logging data with your github code that we debugged in issue 4 to get it running on windows. Changed the code a bit since I dont want a command line tool. Got my hands on a 37V scooter battery that I want to use as a big power bank next month in the outback, hence testing the setup: Battery -> RD6006P -> XY-PSD100 (100 W USB-C PD) (-> TC66 for testing) -> device charging The other way, charging the battery, sadly very slow due to the needed boost converter being so low power: 12 V car -> XL6009E1 -> RD6006P -> battery

@@leocurious9919 Good stuff. On the boost converter, there isn't really much you can do. Have you looked if the scooter battery can be split (and isolated!) into banks that could be charged in parallel? On the HV power: Yes, a microcontroller with a display would solve that problem neatly... its tempting because that way I could add pre-selected test voltages, e.g. 100V, 750V, 1000V and even a lock "are you sure" button to stop me from going accidentally (much) beyond 1000V... Tempting as I said... but not short term due to lack of time.

@@TheHWcave The thing is, it is a neat battery. MOSFETs to turn the output on when a 3. battery contact is connected to ground (put an ON/OFF switch there). Nice balancing of the cells (all of the 10s4p voltages were withing ~1 mV when I got it). All water thigh in an extruded Aluminium profile. So I don't want to pull it all apart for this short term power bank. And I want 21...28 V for 20 V USB PD, so need the boost for charging regardless. Using it in a 3s configuration would also work, but that is too much work for now. Just have to figure out how to transport it in the airplane on the way back considering they have a hard limit on capacity per battery. So maybe I have to pull it apart just for that... making it far less safe in the process. I don't know what to do for now. About the µC: Use GPT 4.0 (costs a little) for the programming. It speeds things up a lot.