Sabvoton 7280 regen brake throttle question

[RIGeek]

I recently got a Sabvoton 7280 controller and noticed that there is a few missing connections and wondering if I can just solder in the missing wires. One of those options would be the 'brake throttle' option. I do plan to visit some mountains with this bike, have a LiFePO4 pack that can accept up to 50A charge current so I figure, make use of it.

The other option that I do not see is the 3-speed/boot button. I am very new to the e-bike world so I'm unsure if that was just installed because I have an LCD display with that option or not.

Thanks all.

 

[amberwolf]

Depends on whether the model you have normally supports those functions, and then if the firmware and hardware installed inside the version you have also supports them.

Most likely you'll have to simply try it and see if it works, unless you can get Siaecosys (or whoever made your version of Sabvoton) to look up your specific model and version and firmware version and tell you if it works as desired. If you're not sure which pads to solder to, that is also a guessing game unless you run across pictures of internals of that model / version that show what you need (I didn't see one in a quick search of this forum, but a detailed search might turn one up).




Regarding the charge current...does the battery's BMS have a common charge/discharge port? Or separate ports? If they're separate, then you'd be using the discharge port for regen braking charging, and it would be uncontrolled (meaning if something went wrong inside the pack the BMS could not disconnect it from the controller and prevent problems).

However...if you're generating that much regen current and it did disconnect the battery, the votlage would almsot certainly spike so high at the controller before it could do anything about it that it would damage the controller and anything else connected to the main power bus (DC-DC, etc). So...rather than have the BMS cut power for that event, you might want to wire it up so that the gate signal for the BMS FETs or contactor control instead of disconnecting things just disengages the controller's braking and throttle inputs (or grounds them if the brake and throttle outputs can handle that; a small resistor between brake / throttle and the inputs would help with that).

 

[RIGeek]

Depends on whether the model you have normally supports those functions, and then if the firmware and hardware installed inside the version you have also supports them.

Thank you!

Most likely you'll have to simply try it and see if it works, unless you can get Siaecosys (or whoever made your version of Sabvoton) to look up your specific model and version and firmware version and tell you if it works as desired. If you're not sure which pads to solder to, that is also a guessing game unless you run across pictures of internals of that model / version that show what you need (I didn't see one in a quick search of this forum, but a detailed search might turn one up).




Regarding the charge current...does the battery's BMS have a common charge/discharge port? Or separate ports? If they're separate, then you'd be using the discharge port for regen braking charging, and it would be uncontrolled (meaning if something went wrong inside the pack the BMS could not disconnect it from the controller and prevent problems).

A bit of both. My pack has a dedicated 10A charge port for balance charging but can accept up to 50A charge through the common port.

However...if you're generating that much regen current and it did disconnect the battery, the votlage would almsot certainly spike so high at the controller before it could do anything about it that it would damage the controller and anything else connected to the main power bus (DC-DC, etc). So...rather than have the BMS cut power for that event, you might want to wire it up so that the gate signal for the BMS FETs or contactor control instead of disconnecting things just disengages the controller's braking and throttle inputs (or grounds them if the brake and throttle outputs can handle that; a small resistor between brake / throttle and the inputs would help with that).

Yes, I would set up the controller to not allow regen over a specified voltage to protect for that.

 

[amberwolf]

A bit of both. My pack has a dedicated 10A charge port for balance charging but can accept up to 50A charge through the common port.

Just to be clear--the common port vs separate ports I'm referring to are at the actual BMS board; it's the way the FETs are wired up to the P- and C- ports. If P- and C- are actually physically the same solder or bolt connections on the PCB (not just connected together by wires or solder, which are not functionally the same), then that's a common port BMS. If they are electrically separate connection points, it's a separate-port BMS.

Connectors on the outside of the pack may or may not be connected separately or in common at the BMS board inside, and may be rated very differently from the actual BMS itself.

For instance, you could have a charge connector like an XLR that's rated at say, 8A, but the actual BMS port inside might have FETs only able to handle 5A. You can also have a discharge connector like an Anderson SB-50 that could handle 50-70A, but the BMS discharge FETs only able to handle 20A.

Or the other way around (FETs able to take more than the associated wiring and connectors).

Yes, I would set up the controller to not allow regen over a specified voltage to protect for that.

The regen overvoltage setting is to protect against the controller sending so high a voltage to the pack that it would cause the pack to shutdown due to HVC...but if the pack shutsdown *first* for some other reason, it is unlikely to be able to react fast enough to prevent a "high voltage excursion" that can easily be above what the components (FETs, LVPS, etc) can handle and damage things.

This kind of failure has happened to people, though often they dont' understand the situation that caused it--just riding around and suddenly the controller blows up. Analyzing their actions and the situaiton leading up to it then may find the pack was in a state that could shut off (cut out) under the conditions about to happen, and then they braked hard with regen either as or after the pack did cut out, and POOF. Sometimes the situation is pretty clear cut, sometimes it's a best guess for the cause, and sometimes we just don't know...but there are some that we're sure that is what happened.

 

[RIGeek]

Just to be clear--the common port vs separate ports I'm referring to are at the actual BMS board; it's the way the FETs are wired up to the P- and C- ports. If P- and C- are actually physically the same solder or bolt connections on the PCB (not just connected together by wires or solder, which are not functionally the same), then that's a common port BMS. If they are electrically separate connection points, it's a separate-port BMS.

Connectors on the outside of the pack may or may not be connected separately or in common at the BMS board inside, and may be rated very differently from the actual BMS itself.

For instance, you could have a charge connector like an XLR that's rated at say, 8A, but the actual BMS port inside might have FETs only able to handle 5A. You can also have a discharge connector like an Anderson SB-50 that could handle 50-70A, but the BMS discharge FETs only able to handle 20A.

Or the other way around (FETs able to take more than the associated wiring and connectors).

Great info. Thank you. I will have to dig into my controller. I did have this pack custom built and one of my requirements to the builder was the ability to charge with up to 50A through the common connector. I do not technically have the battery yet, it's on a boat somewhere on the Pacific Ocean. I am in the waiting and planning phase still. I was told that the BMS and cells can support 80A constant, 160A 10 sec burst and 50A charge from the output connector. There is a C10 (I think) connector for up to 10A dedicated charge.

Once I get it, I will pull it apart to find out what exactly is inside of it.

The regen overvoltage setting is to protect against the controller sending so high a voltage to the pack that it would cause the pack to shutdown due to HVC...but if the pack shutsdown *first* for some other reason, it is unlikely to be able to react fast enough to prevent a "high voltage excursion" that can easily be above what the components (FETs, LVPS, etc) can handle and damage things.

This kind of failure has happened to people, though often they dont' understand the situation that caused it--just riding around and suddenly the controller blows up. Analyzing their actions and the situaiton leading up to it then may find the pack was in a state that could shut off (cut out) under the conditions about to happen, and then they braked hard with regen either as or after the pack did cut out, and POOF. Sometimes the situation is pretty clear cut, sometimes it's a best guess for the cause, and sometimes we just don't know...but there are some that we're sure that is what happened.

I will have to ponder this for a bit. I see the issue you're outlining here but now I have to think of a failsafe to protect from this. Typically not a hard thing using an ESP or Arduino controller but having a load bank to dump 1kw into will not be ideal on a bicycle... I may just have to extensively test the BMS for it's exact shutdown and drop the over voltage protection in the Sabvoton a volt or more less. If the BMS will kick out at say 80%, set the Sabvoton to disable e-brake over 70-75%.

Thank you again. The info you supplied has likely helped me smoke a controller as it all does make loads of sense. I do have a great deal of electronics experience but this is my first dabble into an e-bike.

 

[amberwolf]

Great info. Thank you. I will have to dig into my controller.

Just to be sure we're talking about the same things: The BMS is normally inside the battery, not the controller. The controller is normally oustide thebattery, and connects between it and the motor.

I will have to ponder this for a bit. I see the issue you're outlining here but now I have to think of a failsafe to protect from this. Typically not a hard thing using an ESP or Arduino controller but having a load bank to dump 1kw into will not be ideal on a bicycle... I may just have to extensively test the BMS for it's exact shutdown and drop the over voltage protection in the Sabvoton a volt or more less. If the BMS will kick out at say 80%, set the Sabvoton to disable e-brake over 70-75%.

The controller safety will help for normal situations, but the specific situation in which the damage can occur is literally anytime the BMS limits are exceeded and it shuts the output off. If it does that, for any reason, while regen is occuring, stuff happens fast enough that an MCU (the controller's, or something external like the Arduino/etc) simply doesn't have time to react.

You might be able to design a purely hardware solution for it; there are surge protection devices that can do it for really high voltage, but I don't know what would work for something below the FET operational voltage range; you don't want it to trigger on intermittent tiny spikes that happen just from flyback currents in the motor during switching, but you do want it to trigger on battery disconnect flyback spikes.

The only *certain* way to prevent this is to prevent the BMS from disconnecting the output under *any* conditions, and instead of letting it do that, you use the control signal the BMS would have used to turn it's output on and offf ot instead enable or dsable the controller. In this design, you could have a switch or battery disconnect or bypass cable that simply shorts around the FETs while riding, but is disconnected / turned off when not riding so the BMS can truly disconnect itself from the output for safety when it needs to then.

Thank you again. The info you supplied has likely helped me smoke a controller as it all does make loads of sense. I do have a great deal of electronics experience but this is my first dabble into an e-bike.

Well, we're really here to help keep people from smoking parts (unless that's their goal :lol: ), but....

Most of the electronics on ebikes work just like anything else, except that almost all of them are designed with zero margin and are advertised with specs that they can barely handle (and almost anything unusual in operation can cause problems or failures with some of them). And documentation is either nonexistent or very confusing / poorly translated, etc. Cheap stuff may be made with counterfeit or salvaged components, and should be expected to not work correclty anywhere near the advertised ratings. If a certain name on something seems to sell more or for more money, a bajillion copycats slap that name on things that aren't that original thing, making it difficult to tell if you're getting what you think you are until you have it and can test it, unless the original maker / seller can verify it's real (iff you can even tell who that is).

(many of those things are true of many other kinds of devices as well...but it's nearly impossible to avoid the problem with typical ebike stuff).

 

[RIGeek]

Well, we're really here to help keep people from smoking parts (unless that's their goal :lol: ), but....

Most of the electronics on ebikes work just like anything else, except that almost all of them are designed with zero margin and are advertised with specs that they can barely handle (and almost anything unusual in operation can cause problems or failures with some of them). And documentation is either nonexistent or very confusing / poorly translated, etc. Cheap stuff may be made with counterfeit or salvaged components, and should be expected to not work correclty anywhere near the advertised ratings. If a certain name on something seems to sell more or for more money, a bajillion copycats slap that name on things that aren't that original thing, making it difficult to tell if you're getting what you think you are until you have it and can test it, unless the original maker / seller can verify it's real (iff you can even tell who that is).

(many of those things are true of many other kinds of devices as well...but it's nearly impossible to avoid the problem with typical ebike stuff).

Yeah, lots of experimenting to do now. BMSs are not rocket science too, I might be able to reverse engineer one that could activate a buzzer 30 sec before a disconnect. It's a lot of effort for the cause but I will test the regen on a lightly charged battery to see if it's worth it first.

I did open my Sabvoton (likely knock-off) controller and found some of the cheapest capacitors I've seen in a while. I'm pondering the investment of $35 to rip out the trashy 470uf 100v ones in there for some quality 1000uf 120v rated ones. I'm not looking to shunt mod or go past the rated current limit but I think replacing those caps might make life easier on the FETs, especially when one of the caps suddenly fails under load. The battery I have on order says just about 87v max charge so those 100v caps I worry about. The next thing was the raw aluminum plate that is used as a heatsink inside. It's not even machined flat, it's raw stock. I'm itching to remove that and put in a piece of machined copper or at least just machining that one flat.

I've not looked at the FETs. Almost scared to now. LOL

 

[amberwolf]

Yeah, lots of experimenting to do now. BMSs are not rocket science too, I might be able to reverse engineer one that could activate a buzzer 30 sec before a disconnect. It's a lot of effort for the cause but I will test the regen on a lightly charged battery to see if it's worth it first.

sure...but you won't have 30 seconds if something goes wrong in ther and the bms reaches a limit. itt will disconnect immediately unless you are hacking it so it can't control it's FETs, or wiring around it.

also, as noted, it isn't just hte regen that could trigger a bms shutodwn. it is any limit in the battery that can cause it to disconnect the output. that could be temperature, voltage of any cell / group not just low or high, but also the difference between any two cells / groups, current, etc. could even be a software crash in the bms mcu that ccould cause a disconnect, etc.

testing the regen with a lightly charged battery is only going to test the regen itself, it will not test any of the failure modes nor will it test in the situations most likely to cause batteyr shutoff.

and if a bms shutodwn doesn't cause a voltage spike that nukes parts once / under one set of conditions, it doen'st mean it won'thappen at some other time with slightly different conditions.

there's no guarantee that any thing would ever fail or blow up...but it has happened often enough to know that it is a problem....

I did open my Sabvoton (likely knock-off) controller and found some of the cheapest capacitors I've seen in a while. I'm pondering the investment of $35 to rip out the trashy 470uf 100v ones in there for some quality 1000uf 120v rated ones.

i'd use rubycon or panasonic, etc. ther'es some capacitor discussions around that discuss various brands and properties.

for main power bus caps you usually want the lowest esr you can get, and the highest temperautre rating because thigns get hot in there and the higher the rating the more hours theya re guaranteed for near/at that temperature.

more capacitance isn't always better, but lower esr will help them do their job with less internal heating.

higher capacitance makes biger caps and so does higher voltage, so both together make even biggger caps. probably big enough they wont' fit vertically on the board anymore, and then you end up with long leads from end of horizontal cap down into the board adn those leads can then vibrate and break, or they can heat up with the current flow and fail, or at least impede the filtering effect they're there for, etc.

The next thing was the raw aluminum plate that is used as a heatsink inside. It's not even machined flat, it's raw stock. I'm itching to remove that and put in a piece of machined copper or at least just machining that one flat.

lapping it to the case surface it mounts to will help with heat transfer-by how much depends on the amount of actual metal contact they make now, and how much air is between them instead, vs how much metal contact they could have if they were properly lapped.

iff you go with copper y ou might as well go all the way and extend the bar out as a full heatsink, slotting the case to feedthe full size of the copper out inot the airflwo and fin it appropriately all as one piece. this will mean one less thermal gap to cross.