SMPS: What topology for 3kW 36 V to 360 V converter?
Storken, Thu Dec 29 2016, 12:44AMHello,
Being somewhat experienced with buck and boost converters I have now encountered a challenge where these topologies are far from ideal.
Basically I'm charging a big li-ion battery (20 kWh 360 Vnom) from a 36 V bus (Solar/generator/4kWh li-ion stabilizing battery).
The specifications:
Input voltage: 30 to 42 V
Output voltage: 280 to 430 V
Power: 3 kW
The easy:
Both the load and source are low impedance batteries. This makes control loops less critical.
The hard:
Input side will see quite high currents. Parallelled modules might be a solution.
Safety. Lethal voltage. When not connected, converter should not be hot.
If transformer topology, making a efficient transformer.
I'm thinking I have to choose between flyback, push-pull or half bridge. Flyback seems easy, but lacks the power unless very big magnetics is used. Push-pull needs more attention regarding flux imbalance, but my understanding is that with current control - this will be fine. I know half bridge from SR-Buck converters and one 3PH ebike inverter I built some time ago, but connected to a transformer - I can't readily grasp the challenges.
Relevant literature: Isolated Supply Overview and Design Trade-Offs
Relevant topic: here on 4hv
What would you do? Any help or pointers to similar work is greatly appreciated.
Re: SMPS: What topology for 3kW 36 V to 360 V converter?
johnf, Thu Dec 29 2016, 06:06AM
Current fed current mode full bridge would be the topology of choice.
bullet proof in this sort of situation.
johnf, Thu Dec 29 2016, 06:06AM
Current fed current mode full bridge would be the topology of choice.
bullet proof in this sort of situation.
Re: SMPS: What topology for 3kW 36 V to 360 V converter?
Steve Ward, Thu Dec 29 2016, 06:51AM
Given that your input voltage is so low, a full-bridge would be my choice.
Since the output voltage is rather high, a passive set of rectifiers + filter inductance (as seen in most "push-pull" converters) would be pretty straight forward. SiC schottky diodes might be of some benefit.
Push-pull gets harder at higher power levels as leakage inductance naturally grows with size and consequently voltage spikes across the switches become annoying to deal with. Full-bridge is pretty much the "go-to" for high power conversions.
Steve Ward, Thu Dec 29 2016, 06:51AM
Given that your input voltage is so low, a full-bridge would be my choice.
Since the output voltage is rather high, a passive set of rectifiers + filter inductance (as seen in most "push-pull" converters) would be pretty straight forward. SiC schottky diodes might be of some benefit.
Push-pull gets harder at higher power levels as leakage inductance naturally grows with size and consequently voltage spikes across the switches become annoying to deal with. Full-bridge is pretty much the "go-to" for high power conversions.
Re: SMPS: What topology for 3kW 36 V to 360 V converter?
johnf, Fri Dec 30 2016, 06:52PM
Steve
that is why I went for the current fed current mode topology as there is overlap in the full bridge to control the leakage inductance spikes
johnf, Fri Dec 30 2016, 06:52PM
Steve
that is why I went for the current fed current mode topology as there is overlap in the full bridge to control the leakage inductance spikes
Re: SMPS: What topology for 3kW 36 V to 360 V converter?
Hazmatt_(The Underdog), Sat Dec 31 2016, 06:56AM
I would try to split up that task to 2 U-core transformers, and not try to get one transformer to handle the full load.
The U-cores are fairly inexpensive on ebay, and with 2 transformers you should be able to get just enough of a winding window to handle 4x 90V windings.
I think you will need 16 AWG teflon wire to handle 10A though. Don't bother with copper magnet wire, it will have more losses.
On my small converter (18W) I needed maximum efficiency possible because its portable, and I ran into a problem where I could not source enough current, so I have to use 2 identical windings and rectify them center-tap grounded full wave.
If you do a small-scale experiment and find you cannot source the current you need reliably, you might also need to double-up each secondary for more current.
Hazmatt_(The Underdog), Sat Dec 31 2016, 06:56AM
I would try to split up that task to 2 U-core transformers, and not try to get one transformer to handle the full load.
The U-cores are fairly inexpensive on ebay, and with 2 transformers you should be able to get just enough of a winding window to handle 4x 90V windings.
I think you will need 16 AWG teflon wire to handle 10A though. Don't bother with copper magnet wire, it will have more losses.
On my small converter (18W) I needed maximum efficiency possible because its portable, and I ran into a problem where I could not source enough current, so I have to use 2 identical windings and rectify them center-tap grounded full wave.
If you do a small-scale experiment and find you cannot source the current you need reliably, you might also need to double-up each secondary for more current.
Re: SMPS: What topology for 3kW 36 V to 360 V converter?
hen918, Sun Jan 01 2017, 12:09PM
I have done almost exactly that. (mine had a 400 V output voltage and a 30 V input voltage)
I used a phase shifted full bridge with one large ETD-59 transformer.
The UCC28950 provided my control, with the synchronous rectifier outputs not connected.
SiC diodes were used for the rectification, and IRFB3077 MOSFETs (two in parallel) for the primary drive.
I haven't tested it at above 3 kW but it has been over-designed to perform well at very high powers.
The only issues are the parasitics in the transformer, made worse by having four parallel primary windings.
Push-pull is impossible with that specification. The energy stored in the transformer's leakage inductance (in the order of 1 uH) will be dissipated in the 'FETs on every cycle.
hen918, Sun Jan 01 2017, 12:09PM
I have done almost exactly that. (mine had a 400 V output voltage and a 30 V input voltage)
I used a phase shifted full bridge with one large ETD-59 transformer.
The UCC28950 provided my control, with the synchronous rectifier outputs not connected.
SiC diodes were used for the rectification, and IRFB3077 MOSFETs (two in parallel) for the primary drive.
I haven't tested it at above 3 kW but it has been over-designed to perform well at very high powers.
The only issues are the parasitics in the transformer, made worse by having four parallel primary windings.
Push-pull is impossible with that specification. The energy stored in the transformer's leakage inductance (in the order of 1 uH) will be dissipated in the 'FETs on every cycle.
Print this page