### Lab Update *2025-01-13 -> 2025-02-10 | [[#Pics|Skip to Pics]]* [[Lab Update - 2024-11-03|Previous Lab Update]] | [[Lab Update - 2025-03-04|Next Lab Update]] Busy couple weeks. I threw the rat rig into an 8'x4' grow tent, then added a space dehumidifier, new carbon filter, 5kg spool dryer, overhead light, etc. Now the whole thing is climate controlled and in the garage where I don't have to listen to it or worry about tripping a breaker. I've got most of the SBR badges out, and am moving into producing the maglevs now. All the instruments arrived and I dove right in with Joule Thief. Been doing a lot of back and forth learning how battery management systems work, how alkaline cells work, LTSpice, and actually laying stuff out: - [[!Joule Thief]] 1. **COURSE:** [[Battery Boot Camp Course]] 2. [[Misc. Joule Thief References]] 3. **SIM:** [[LTSpice - Rechargeable Battery Simulation]]]] 4. **BUILD LOG:** [[Joule Thief Build Day - Pi Mount and Schmitt Trigger Design]] 5. [[Power Constraints - How to Kill a Battery]] I've been reviewing and second-guessing a lot of aspects about the design, but that's fine. The previous draft was just me throwing part numbers together more or less at random with minimal consult from chatgpt/deepseek. I've been working the chicken-egg problem from both ends and am pretty close to meeting in the middle. Currently my main obstacles are 1. drain a AA Alkaline cell as fast as possible, or rather, get the highest stable average power delivery out of it 2. efficiently re-conditioning that energy for efficient storage into a LiPo's BMS I talked myself out of finishing the BMS course I was watching because I thought I might be able to just throw together a functional MVP with off-the-shelf parts, but I found that both extraction and storage are too inefficient for it to be considered functional. With the inductor-based Joule Thief, I'm pulling 100mA and delivering 20mA. It's not even enough for the stock BMS to enable charging mode. I don't think it's even enough to power the LEDs. After fiddling with the AA cells under various load conditions and seeing what's possible, I'm becoming increasingly convinced it actually is worthwhile for me to roll my own discrete component BMS so I can ensure the charging circuit is appropriate for the application & not wasting a bunch of energy on blinking LEDs/always-on microcontrollers. The MVP I'm chasing currently is: Use a AA Alkaline cell to charge a single 18650 through a single-cell BMS. - My bench tests show I can get ~400mW from a AA. - My LTSpice sims show that a 400mW input to the LTC3105 can yield ~160mW out. - My bench tests with the single-cell BMS show that an 18650 will charge with 160mW in. (also no idea why the efficiency is so bad in sim when the LTC3105 datasheet brags 70-90%) So next steps are to redo the breadboard layout with the LTC3105 when it arrives, see what numbers come out with that, and then continue the BMS course. Along the way, I'm going to continue looking into novel AA torture test methods (negative voltage, hf pulsing, etc.) to increase power delivery. ### Pics ![[IMG_3475.webp]] ![[IMG_3499 1.webp]] ![[image0 2.webp]] ![[IMG_3013 1.webp]] ![[IMG_2715.webp]] ![[Pasted image 20250210062710.png]] ![[image 6.webp]] ![[Joule_Thief_Sketch_4.webp]]![[image 7.webp]] ![[IMG_2709.webp]] ![[Pasted image 20250210063015.png]]