timzilla
New Member
To make a long story short, I took possession of a Goal Zero Yeti Pro 4000 that had some "issues". Due to the nature of how I got it, Goal Zero has declined to offer me any options to get it repaired or replaced, so I threw it on my workbench and got to pulling out the battery pack from the bottom of the unit chassis.
The issue the unit was experiencing was that even when sitting on wall power for days, and reporting a 100% SoC, it would start to decline precipitously under any sort of load reaching a zero state of charge after without even drawing anywhere close to its official capacity. Fully cycling the battery to low then high would temporarily put it in a fully functional state but any sort of extended idle, unused state would throw the unit in a bad state. Also the unit would report a low voltage cell alert if left on idle for days. Based on that I suspected that one of the cells probably was defective and self-discharging at a very high rate, but I had no idea what the cells were, how it was constructed and all that so lets fuck around and find out.
The BMS bears an Iconergy brand. This seems to be a outfit that GoalZero works with, I've seen that same brand on my six year old Yeti 1000 that I took apart at a different time. But lets take the BMS off and look at the battery pack itself.
GZ's product specs mention it is Lifepo4 - and to be honest, I had been expecting to find prismatic cells but found what I think are 40135 C40 cells with 20Ah each. They have no markings that are visible to me. They are organized in a series of 16, with 4 parallel cells each. 64 cells total.
I took initial readings of each parallel cell grouping which you can see I have noted on the blue masking tape. Most of the cell groups were between 1.7 to 2.1 V but one particular cell group was at a shockingly low 0.5 V. I believe that the defective cell is among those four.
Using a bench power supply, I have been trickle charging each cell group back up to around 3.3V. My goal is to get the entire battery pack back in a state where the BMS can be re-attached and take over charging. But I know I cannot get this back in a reliable and useable state as long as that defective cell isn't replaced.
NOW THE PROBLEM ...
I cannot split the battery case. There are four M6 screws holding the two halves of the battery case together but even after removing them, i cannot pull the case apart. I suspect that the metal tabs on both sides have been soldered to each end of each of the cylinders. So the case has essentially been fused together permanently. I could try desoldering each tab from each cylinder but I really have no desire to try to desolder 128 tabs just to try to pull apart the case.
However...... The group of cells that is self-discharging is at one of the external corners. I have been looking at it and thinking I could, with careful application of an oscillating saw or a hot soldering iron, I could cut away enough of the case at that corner to expose that group of four battery cells and remove them from the pack and replace them with one (or four) new cells. That would mean I'd only have to desolder 8 tabs. I'm also pretty well equipped with 3D modeling and a printer, I could create an stand-in to replace the part of the case I removed.
But before I do all that - am I missing anything? Is there anything ridiculous about what I'm doing (other than trying to hack around with a battery pack that clearly isn't meant to be user serviceable).
The issue the unit was experiencing was that even when sitting on wall power for days, and reporting a 100% SoC, it would start to decline precipitously under any sort of load reaching a zero state of charge after without even drawing anywhere close to its official capacity. Fully cycling the battery to low then high would temporarily put it in a fully functional state but any sort of extended idle, unused state would throw the unit in a bad state. Also the unit would report a low voltage cell alert if left on idle for days. Based on that I suspected that one of the cells probably was defective and self-discharging at a very high rate, but I had no idea what the cells were, how it was constructed and all that so lets fuck around and find out.
The BMS bears an Iconergy brand. This seems to be a outfit that GoalZero works with, I've seen that same brand on my six year old Yeti 1000 that I took apart at a different time. But lets take the BMS off and look at the battery pack itself.
GZ's product specs mention it is Lifepo4 - and to be honest, I had been expecting to find prismatic cells but found what I think are 40135 C40 cells with 20Ah each. They have no markings that are visible to me. They are organized in a series of 16, with 4 parallel cells each. 64 cells total.
I took initial readings of each parallel cell grouping which you can see I have noted on the blue masking tape. Most of the cell groups were between 1.7 to 2.1 V but one particular cell group was at a shockingly low 0.5 V. I believe that the defective cell is among those four.
Using a bench power supply, I have been trickle charging each cell group back up to around 3.3V. My goal is to get the entire battery pack back in a state where the BMS can be re-attached and take over charging. But I know I cannot get this back in a reliable and useable state as long as that defective cell isn't replaced.
NOW THE PROBLEM ...
I cannot split the battery case. There are four M6 screws holding the two halves of the battery case together but even after removing them, i cannot pull the case apart. I suspect that the metal tabs on both sides have been soldered to each end of each of the cylinders. So the case has essentially been fused together permanently. I could try desoldering each tab from each cylinder but I really have no desire to try to desolder 128 tabs just to try to pull apart the case.
However...... The group of cells that is self-discharging is at one of the external corners. I have been looking at it and thinking I could, with careful application of an oscillating saw or a hot soldering iron, I could cut away enough of the case at that corner to expose that group of four battery cells and remove them from the pack and replace them with one (or four) new cells. That would mean I'd only have to desolder 8 tabs. I'm also pretty well equipped with 3D modeling and a printer, I could create an stand-in to replace the part of the case I removed.
But before I do all that - am I missing anything? Is there anything ridiculous about what I'm doing (other than trying to hack around with a battery pack that clearly isn't meant to be user serviceable).
