BlueSeas
100 W
I have a hypothesis on possibly why there have been so many, what seem like premature failures of Sabvoton controllers. The hypothesis may be true or false, but the problem I'm going to point out is real. Makes reviewing your setup worth the time if using regen of any fashion (variable, glide or push button) with any controller, but especially Sabvoton.
One of the "known" issues with regen on Sabvoton is the "over volt" parameter on the "basic" tab of the programming software. The general advice I've seen is this should be 10V over the fully charged battery pack voltage. This because it's known that regen on Sabvoton will not work if the pack voltage is somewhere near the "over volt" setting. Does anyone know at exactly what internal algorithm the Sabvoton uses to decide when to cut off regen when approaching the "over volt" setting?
10V over pack voltage can lead to potential self induced suicide of the controller. Why? If you use a BMS, and most do, it's protection circuit is based on the high cell voltage, not pack voltage. If regen charging causes the BMS to disconnect the battery pack due to a cell over voltage setting, the Sabvoton suddenly has nowhere to dump the charge current. Without the battery load, the voltage spikes, potentially frying and letting the smoke out of both the controller and BMS FET's.
On a 72V 20S Polymer pack. 10V is .5V per cell. I'm new to this chemistry, but my first eBike using 20S fresh off the charger actually operates at about 83V of the 84V max charge voltage. There is nowhere near 10V of headroom before the BMS will disconnect the pack. LFP packs get to divide by 24S and the actual operating voltage per cell drops very quickly, so there more headroom. But at 10V, 3.3V per cell still only has to get to 3.7V and boom. Use a more common 48-60V pack and this math gets much worse.
On a brand new, perfectly balanced pack, regen won't normally actually push a cell in the pack high enough to trigger a disconnect, but the problem will worsen over time as the cells internal resistance increases. It is a potential ticking time bomb. Especially using "over volt" = 10V over max pack voltage.
This problem probably applies to other controllers as well. My only other controller was the Phaserunner. But it ramped down regen, didn't actually shut it all the way off. Probably would have at 84V, but there wasn't any reason to artificially inflate that controller setting.
I will try and narrow down a more exact behavior of the Sabvoton and regen as it relates to this setting. But it must be set to inhibit regen before the BMS disconnects with some safety margin.
One of the "known" issues with regen on Sabvoton is the "over volt" parameter on the "basic" tab of the programming software. The general advice I've seen is this should be 10V over the fully charged battery pack voltage. This because it's known that regen on Sabvoton will not work if the pack voltage is somewhere near the "over volt" setting. Does anyone know at exactly what internal algorithm the Sabvoton uses to decide when to cut off regen when approaching the "over volt" setting?
10V over pack voltage can lead to potential self induced suicide of the controller. Why? If you use a BMS, and most do, it's protection circuit is based on the high cell voltage, not pack voltage. If regen charging causes the BMS to disconnect the battery pack due to a cell over voltage setting, the Sabvoton suddenly has nowhere to dump the charge current. Without the battery load, the voltage spikes, potentially frying and letting the smoke out of both the controller and BMS FET's.
On a 72V 20S Polymer pack. 10V is .5V per cell. I'm new to this chemistry, but my first eBike using 20S fresh off the charger actually operates at about 83V of the 84V max charge voltage. There is nowhere near 10V of headroom before the BMS will disconnect the pack. LFP packs get to divide by 24S and the actual operating voltage per cell drops very quickly, so there more headroom. But at 10V, 3.3V per cell still only has to get to 3.7V and boom. Use a more common 48-60V pack and this math gets much worse.
On a brand new, perfectly balanced pack, regen won't normally actually push a cell in the pack high enough to trigger a disconnect, but the problem will worsen over time as the cells internal resistance increases. It is a potential ticking time bomb. Especially using "over volt" = 10V over max pack voltage.
This problem probably applies to other controllers as well. My only other controller was the Phaserunner. But it ramped down regen, didn't actually shut it all the way off. Probably would have at 84V, but there wasn't any reason to artificially inflate that controller setting.
I will try and narrow down a more exact behavior of the Sabvoton and regen as it relates to this setting. But it must be set to inhibit regen before the BMS disconnects with some safety margin.
