Post by Telemachus on Apr 26, 2020 13:28:05 GMT
Thought I should move the ramblings on an unrelated thread, to here:
oh so it’s moved forward then? Do keeps us informed how it’s going.Yes, I posted a lot of technical stuff on CWDF but you don’t go there! In summary I replaced the existing brush set / regulator module, with a cheapo one I got from ebay for a tractor. Snipped the wires to the regulator and modified so one brush was connected to the case, the other to a wire that went to my controller. My controller also connected to the W terminal, which provides an output from one phase (normally used for a tacho) and took the warning light wire off, connected to my board. And then obviously a battery and ground connection for my board.
In summary it did work, but with some issues:
The software is still in development but I checked the “alternator cooking” function by heating up the temperature probe with a heat gun. I’ve set it up so when it reaches 95C, the maximum field current is reduced to 2A (normally about 4A) which obviously reduces the alternator output. When it drops to 90C the higher field current is authorised. I have done it this way since for a given temperature, an alternator can produce more output at higher rpm because the fan is spinning faster, and the same is true of limiting the field current - reduces the output at lower rpms whilst allowing more output at higher rpms when the fan is going faster. So a tick for that one.
It can also receive battery SoC information over CANbus from my Mastervolt system. Tick.
Although this isn’t how I propose it will work (it will be automatic) I also included a means of typing in a modified target regulation voltage, in the range 10.5 to 16v via laptop/serial (RS232) connection. This also works, I type in the command, the regulated voltage changes accordingly. Tick.
I also noticed that the regulated voltage was held at fairly high outputs, it didn’t drop under load like it does with a conventional analogue regulator, in other words the voltage will be closer to the target voltage at high loads. Tick.
These regulator chips are designed to be in constant communication with a microprocessor (eg a car’s ECU) and if communication is lost, they revert to autonomous operation with “fallback” parameters. If I stop my microprocessor from repeatedly sending the target regulation voltage, after 3 seconds it goes to its default regulating voltage of 13.5v, which of course is “lithium safe”. Tick.
what was less good was that at higher outputs, there was some instability in the voltage regulations, enough fluctuation to have detectable if slight flicker in the incandescent lights. Looking at the diagnostic data showed big and rapid fluctuations in the field current, and the output voltage fluctuation by up to 0.5v. When I invoked the limited field current it all settled down, so there is no issue with controlling the field current, just the voltage regulation bit. Need to work on that. No tick!
These modern alternator regulator chips all have a thing call LRC - Load Response Control. The idea is that at low rpm (around idle), if a big load is switched on one doesn’t want the Alternator to instantly supply that load, because the sudden mechanical load can cause a modern light-flywheel car engine to stall. So at low rpm, when presented with a sudden big load, the alternator output ramps up over a few seconds, part of the load being supplied by the battery in the mean time. This gives the engine control/governor some time to increase fuel supply/raise rpm a bit.
The regulator chip I've chosen supports both specifying an rpm below which LRC is operating, and the number of seconds over which the alternator output will ramp up in response to a sudden heavy load. In fact I can have LRC permanently active. When LRC is in operation, my system totally settles down and is fine. So that is one solution, and I think for a boat it wouldn’t matter at all if the alternator was sluggish to respond to a big change in load (for the avoidance of doubt this only affects how quickly the alternator increases output, it will still respond rapidly to a sudden decrease in load). So a half-tick!
So I need to do some work on improving the stability in normal regulating mode.
I also made a cockup on the PCB design and had to “butcher” it, plus a couple of other minor changes needed including making it fit in an off-the-shelf box. So last night, a new PCB design sent off to Hong Kong. It is already in production, will be finished today but I used cheap shipping so probably 2 weeks to get it back. Total cost about £11 for 5 boards. You’ve gotta love Chinese slave labour!
Apr 26, 2020 9:17:05 GMT kris said:
And whilst I was there, test my prototype smart alternator regulator for LiFePO4 batteries,
In summary it did work, but with some issues:
The software is still in development but I checked the “alternator cooking” function by heating up the temperature probe with a heat gun. I’ve set it up so when it reaches 95C, the maximum field current is reduced to 2A (normally about 4A) which obviously reduces the alternator output. When it drops to 90C the higher field current is authorised. I have done it this way since for a given temperature, an alternator can produce more output at higher rpm because the fan is spinning faster, and the same is true of limiting the field current - reduces the output at lower rpms whilst allowing more output at higher rpms when the fan is going faster. So a tick for that one.
It can also receive battery SoC information over CANbus from my Mastervolt system. Tick.
Although this isn’t how I propose it will work (it will be automatic) I also included a means of typing in a modified target regulation voltage, in the range 10.5 to 16v via laptop/serial (RS232) connection. This also works, I type in the command, the regulated voltage changes accordingly. Tick.
I also noticed that the regulated voltage was held at fairly high outputs, it didn’t drop under load like it does with a conventional analogue regulator, in other words the voltage will be closer to the target voltage at high loads. Tick.
These regulator chips are designed to be in constant communication with a microprocessor (eg a car’s ECU) and if communication is lost, they revert to autonomous operation with “fallback” parameters. If I stop my microprocessor from repeatedly sending the target regulation voltage, after 3 seconds it goes to its default regulating voltage of 13.5v, which of course is “lithium safe”. Tick.
what was less good was that at higher outputs, there was some instability in the voltage regulations, enough fluctuation to have detectable if slight flicker in the incandescent lights. Looking at the diagnostic data showed big and rapid fluctuations in the field current, and the output voltage fluctuation by up to 0.5v. When I invoked the limited field current it all settled down, so there is no issue with controlling the field current, just the voltage regulation bit. Need to work on that. No tick!
These modern alternator regulator chips all have a thing call LRC - Load Response Control. The idea is that at low rpm (around idle), if a big load is switched on one doesn’t want the Alternator to instantly supply that load, because the sudden mechanical load can cause a modern light-flywheel car engine to stall. So at low rpm, when presented with a sudden big load, the alternator output ramps up over a few seconds, part of the load being supplied by the battery in the mean time. This gives the engine control/governor some time to increase fuel supply/raise rpm a bit.
The regulator chip I've chosen supports both specifying an rpm below which LRC is operating, and the number of seconds over which the alternator output will ramp up in response to a sudden heavy load. In fact I can have LRC permanently active. When LRC is in operation, my system totally settles down and is fine. So that is one solution, and I think for a boat it wouldn’t matter at all if the alternator was sluggish to respond to a big change in load (for the avoidance of doubt this only affects how quickly the alternator increases output, it will still respond rapidly to a sudden decrease in load). So a half-tick!
So I need to do some work on improving the stability in normal regulating mode.
I also made a cockup on the PCB design and had to “butcher” it, plus a couple of other minor changes needed including making it fit in an off-the-shelf box. So last night, a new PCB design sent off to Hong Kong. It is already in production, will be finished today but I used cheap shipping so probably 2 weeks to get it back. Total cost about £11 for 5 boards. You’ve gotta love Chinese slave labour!