Topology of 3kW 24V to 350V step up converter

Thread Starter

wkaiyu

Joined Oct 18, 2021
3
Hi all,

I try to design a prototype step-up converter from 24V to 350V. The power range is about 3kW. It will be transformer-isolated.

Currently, we don't need to meet certain efficiency spec, so soft switching is not strictly required. What will be the best topology to implment this considering control simplicity and component size and numbers?

I am considering phase-shifted full bridge, but I am open to any other topoolgies.
 

Papabravo

Joined Feb 24, 2006
21,159
I'm not clear on where transformer isolation comes into play. First things first. A step up ratio of 350 / 24 ≈ 14.58 is impractical for a boost converter.
Second thing. 3kW on the output, assuming an efficiency of 80% requires 3.75 kW on the input or a current of 3.75 kW / 24V ≈156.25 Amperes. These power levels are unsuitable for a boost converter.

This suggests that you might require a transformer topology. What do you know about potential transformer topologies in DC-DC converters? What do you know about fabricating a custom transformer?
 

Thread Starter

wkaiyu

Joined Oct 18, 2021
3
Hi,

The transformer isolation in my original post means the DCDC converter with a transformer between primary side and secondary side. It is the same thing as the transformer topology you mentioned in your post.

I'm not clear on where transformer isolation comes into play. First things first. A step up ratio of 350 / 24 ≈ 14.58 is impractical for a boost converter.
Second thing. 3kW on the output, assuming an efficiency of 80% requires 3.75 kW on the input or a current of 3.75 kW / 24V ≈156.25 Amperes. These power levels are unsuitable for a boost converter.

This suggests that you might require a transformer topology. What do you know about potential transformer topologies in DC-DC converters? What do you know about fabricating a custom transformer?
 

ronsimpson

Joined Oct 7, 2019
2,989
If this was my job I would look at forward PWM current mode like below. This transformer has 5V output but could be made for 350V output. (change the error amp). Right now is has no isolation but a TL431 & isolator and that is fixed. This type does well with low voltage and high current in the primary side. There will be 10A in the output diodes.
1634600703166.png
 

Papabravo

Joined Feb 24, 2006
21,159
Modifying this design for the TS purposes looks like a pretty heavy lift.
After providing the TL431 and the opto for feedback. is the isolation still achieved if the Vcc for U1 on the primary side is derived via L3 on the secondary side?
 

ronsimpson

Joined Oct 7, 2019
2,989
is the isolation still achieved if the Vcc for U1
I have used this type for off line work. "Vcc" could come from that winding or it could come from the 24V supply. Vin 16 "VIN" could connect directly to +24 and it makes a 7.5V on "VCC" internally. The part will run a little hotter that way but it is simple.
 

Papabravo

Joined Feb 24, 2006
21,159
I have used this type for off line work. "Vcc" could come from that winding or it could come from the 24V supply. Vin 16 "VIN" could connect directly to +24 and it makes a 7.5V on "VCC" internally. The part will run a little hotter that way but it is simple.
OK, I'll grant that you have a number of options for generating and isolating the Vcc voltage, but there are still issues involved in making the output 350 V @ 8.57 Amperes. The most critical of which appears to be the specification and fabrication of the transformer.
 

ronsimpson

Joined Oct 7, 2019
2,989
fabrication of the transformer.
The 9A secondary is a problem but the 100A primary is a real problem.
To make the transformer simple, forget post #4 and go with a half or full bridge circuit so there is 1/2 as much primary.
I once tried making a 3 turn winding out of copper sheets. The interwinding capacitance was way too much. I think one turn of copper will work if the switching frequency is high.

On a different subject. This type of transformer is a possibility. I used one turn per layer and in high current windings used parallel windings. We need 15:1 or 20:1 turn ratio.
1634649850281.png.
 

Thread Starter

wkaiyu

Joined Oct 18, 2021
3
Thank you folks for the input. Could we use 3~4 modules and interleaving them? We will use TI DSP for the control loop, so phase shifting is easy to implement.


The 9A secondary is a problem but the 100A primary is a real problem.
To make the transformer simple, forget post #4 and go with a half or full bridge circuit so there is 1/2 as much primary.
I once tried making a 3 turn winding out of copper sheets. The interwinding capacitance was way too much. I think one turn of copper will work if the switching frequency is high.

On a different subject. This type of transformer is a possibility. I used one turn per layer and in high current windings used parallel windings. We need 15:1 or 20:1 turn ratio.










.
 

Ian0

Joined Aug 7, 2020
9,671
Hi all,

I try to design a prototype step-up converter from 24V to 350V. The power range is about 3kW. It will be transformer-isolated.

Currently, we don't need to meet certain efficiency spec, so soft switching is not strictly required. What will be the best topology to implment this considering control simplicity and component size and numbers?

I am considering phase-shifted full bridge, but I am open to any other topoolgies.
What frequency?
A 50Hz transformer will have an efficiency well above 90%, and the transistors will have negligible switching loss, but the transformer will weigh 17kg.
 
Top