8i20 DC power supply companion board.
Power components aren't my strong suit but assuming everything is specified for the max 2200w I'd like to get a couple suggestions on the following if anyone has any:
1) filter cap values and quantity
2) best way to handle initial inrush current of charging the caps on power on
3)on board brake resistors.
The brake resistors I'm not sure about. Trying to figure what a good minimum power resistor setup looks like and fit on a PCB. This would be heatsunk and have a thermostat or some other thermal device connected to an SSR to disconnect drive input power if the brake resistors were overheating.
There is one SSRs in the PSU box, one mechanical relay and two power resistors.
One resistor is on the incoming mains supply to limit inrush current. This has an SSR that bypasses it once the bus voltage reaches a threshold.
The mechanical relay has a 240V coil and is wired such that the NC contacts discharge the capacitor when the power goes off through the second power resistor. This relay is _not_ rated for the current / voltage.
As I recall, I use a single 10,000uF smoothing capacitor rated at 450V and with a low ESR. These are very expensive but I found a bunch on eBay and snapped them up.
A HAL component monitors the bus voltage. It turns on the surge resistor bypass relay when the bus voltage hits 90% max. It also will not allow the PSU to turn _on_ unless the bus voltage is < 10V. This is so that the discharge relay only ever has to "make" the circuit to discharge and is never asked to "break".
I got to playing with a schematic just so I can visualize the idea and came up with this. Missing some passives and probably a diode between the igbt and INA138 but I just wanted to get it on paper at the moment so I can get some advice. I'm not sure that I'll use any discharge circuit, waiting 5 minutes until the HV LED on the 8i20 goes out before playing with stuff hasn't really bothered me.
Thoughts.....or just plain wrong?
blazini36 wrote: I'm trying to avoid controlling the power supply with hal, other than an enable signal.
I built my first one self-contained, but then regretted it as I had no idea what was going on.
HAL is great for this sort of thing.
Any particular reason for the 10,000uf capacitor? From what I know about drives this size I would assume say 4 x 470uf caps is sufficient, which is the setup I was using previously.
I just checked, and the caps are 4700uF (RIFA PEH169UV447CQ ). I used them because I found a bunch of them cheap on eBay.
I do think that they might be over-sized. This web site can be useful: www.changpuak.ch/electronics/power_supply_design.php
as it graphs inrush and ripple.
> What value and power rating did you use for the inrush limiting resistor?
It's been a while, and I have forgotten. It was sized to take 10 seconds to reach full bus voltage (I have an RCD on the circuit and not tripping that was the main requirement). Then the power was calculated from that. And then reduced by a large factor to account for the veery low duty cycle.
waiting 5 minutes until the HV LED on the 8i20 goes out before playing with stuff hasn't really bothered me.
I have more capacitance, and why would the caps discharge at all with the drives disabled?
Important note: I am just sharing what I did. I have no qualifications to indicate that what I did was what should be done.
Just a philosophy thing. While there is nothing wrong with having hal control something like this, if hardware logic is not controlling it then hal must control it. Which means a configuration issue becomes a safety issue. Besides that it requires more wiring from the controller to the Power supply/drive area. Though I'm not exactly an electronics wizard myself I hope to come up with something small-ish and once the board is designed I slap it together and don't think about it much.
"I have more capacitance, and why would the caps discharge at all with the drives disabled?"
Caps aren't batteries, they cannot hold a charge indefinately. Just the passive circuitry on the 8i20 does in fact fully discharge the motor power caps...albeit slowly. The higher capacitance of your setup will take much longer to discharge but it will in fact discharge just sitting with the motor leads connected to the 8i20 (without the relay). There is a small benefit to the fact that if the caps have not discharged fully there is no inrush to deal with anyway.
Ignore that first drawing not sure what I was thinking putting the transistor before the rectifier.
"Important note: I am just sharing what I did. I have no qualifications to indicate that what I did was what should be done."
well me neither lol. I design boards as a hobby and have been relatively successful. I myself am not "qualified" to design a HV PCB but I will seek advice and play with it until I feel it is safe to test. I also seek advice for stuff like this on electronics forums and if we are lucky enough, maybe the grand wizard PCW himself will grace us with some advice.
Maybe slightly more proper schematic:
Both have the problem than if the load current gets very high the surge resistor comes back on. But then that might be a good thing, as it suggests a fault condition.
I'm looking at replacing the INA138 and shunt with a diode and IGBT gate driver which might take a bit of guesswork out of it. I use a Fet driver on another PCB I make but that's for switching speed wheres this is more for isolation.
blazini36 wrote: Still not quite sure how to derive the voltage reference to operate the gate of the IGBT but I'm working on that.
That might be why I decided to do it in HAL, as the 8i20 reports back bus voltage.
And again there's nothing wrong with that, just not what I have in mind for this particular thing. Main issue I have with that is if the hal config gets messed up and doesn't switch the bypass then the NTC or whatever just turns into a heater.
As I look into this I start to realize what the idea of the 8i20 probably was. If you have multiple motors/8i20's in a single system, you might throw together a large single DC bus supply. Control the input with a contactor and do exactly as you have done, using one of the 8i20's bus voltage pins to control the inrush bypass through hal. More or less how I was doing it previously but I never quite got to the inrush limiting device, just limited the bus caps so I could use a reasonable input fuse. I did come across PCB mount relays that can handle controlling an inrush current limiter on the AC side If I can't figure out the solid state device but I still need to derive the voltage reference to control the relay.