diy solar

diy solar

Class T vs NH00 Fuse

In principle yes, but depends on the fault impedance. You hope to get a hard short in order to blow the fuse. And hope your battery soc is high enough that the resistance of your battery pack is low enough. In most situations it will probably be just fine. But in some it might not

A typical overload situation that is not catastrophic like a short is also handled by the BMS: your BMS should cut off before the fuse in that situation.
 
A typical overload situation that is not catastrophic like a short is also handled by the BMS: your BMS should cut off before the fuse in that situation.
Did you ever check your BMS to verify the trip time for a dead short?
 
Did you ever check your BMS to verify the trip time for a dead short?

On the todo list. One of the reasons I didn't yet is because I want to make sure the other issue with the BMS is solved first (so I have enough BMSes in stock if things go wrong with testing that are not having other potential issues interfering).

All I know right now is that at some point I did the short circuit test, the BMS shut down and survived. Once that other issue is dealt with, I'll get back to this.
 
Did you ever check your BMS to verify the trip time for a dead short?
Well mosfet don't really have this 10us time within which you can turn them off, like igbt do (current limiting to 8xIn). Maybe you habe enough cable impedance or inductance in your circuit so that the bms can react, but for a dead short I also have my doubts. Especially if the bms was preheated to max temperature at nominal current.

I'm not trying to scare anybody. Maybe you will be fine in most cases. It's for those that you are not fine that you must planN. Luck cannot replace good design and engineering. No solution is perfect, that's why I would suggest a backup overload protection. MCB, contactor or gS or similar fuse.
 
On the todo list. One of the reasons I didn't yet is because I want to make sure the other issue with the BMS is solved first (so I have enough BMSes in stock if things go wrong with testing that are not having other potential issues interfering).

All I know right now is that at some point I did the short circuit test, the BMS shut down and survived. Once that other issue is dealt with, I'll get back to this.
To be clear I didn't mean a functional test although that is certainly required.
I just meant to check the configuration to see what the default trigger timer is set to.
 
Well mosfet don't really have this 10us time within which you can turn them off, like igbt do (current limiting to 8xIn).
I don't think I'm getting your full meaning here.
I may understand better after I read this.

Maybe you habe enough cable impedance or inductance in your circuit so that the bms can react, but for a dead short I also have my doubts. Especially if the bms was preheated to max temperature at nominal current.

I'm not trying to scare anybody. Maybe you will be fine in most cases. It's for those that you are not fine that you must planN. Luck cannot replace good design and engineering. No solution is perfect, that's why I would suggest a backup overload protection. MCB, contactor or gS or similar fuse.
My overkill(jbd) bms is configured to trip in 135 microseconds, if it can actually do that I don't think any fuse including the the BS88 can beat it in a game of bloody knuckles.
Its my understanding that the fastest BS88 will trip in ~10-20 milliseconds.

Because of this I see the BMS as sacrificial.
If it fails to do its job its already compromised and the very fast fuse should break the circuit before the wires melt or the cells vent.
The problem is determining if the BMS needs to be replaced.
Since the FETS are in parallel 1 or more of them can fail open or closed and it could be difficult to know that any damage occurred.
 
The FETS have a pretty low on resistance, and, IIUC, should be turned off if the BMS is disconnected. Any reason a FET fail on state couldn't be tested by putting a ohmmeter across the BMS?

I just tested a brand new JK B2A24S20PC I have near me. One direction I get >50MΩ, the other I get 1.197MΩ, according to my Fluke 87V. Seems like a FET fail on should look very different.
 
I have replaced hundreds of bipolar power transistors, diodes, and some power fets over the years. Dead Short is the typical failure mode. Sometimes the Diodes body turns to charcoal and falls apart, then it's an open. Inside the transistor the (gold) bonding wire could melt or detach from the pads.

My BMS (overkill solar) has a Off Setting for Charge and Discharge.

BTW: If you have an Amplifier or Power Supply with parallel output transistors: Replace One, Replace them All.
 
Last edited:
Ah, ok - I'll check next time if I remember this.
Andy's latest video shows the relevant Heltec parameters.
The options for "SCD delay" are enumerated to 100, 200 and 400 microseconds.

This link show the JK SCP delay set to 1500 micro seconds.
Its presented as a range.
I wonder what the lower bound of the range is?
I would guess that the quicker the BMS can disconnect the short, the more likely it is to survive.

So even the slow JK is still an order of magnitude faster than the fastest fuse.
Do you agree?
 
Last edited:
Back
Top