http://opalkelly.com/

Author Topic: DC/DC 48V/60V to 12V-10A  (Read 309 times)

0 Members and 1 Guest are viewing this topic.

Offline sangseu

  • Contributor
  • Posts: 7
  • Country: vn
DC/DC 48V/60V to 12V-10A
« on: June 19, 2017, 09:58:34 PM »
Hello everybody. I'm study DC/DC converter. I've teardown a "DC/DC 48V/60V to 12V-10A" product and wana copy that.
I've draw schematic


and trouble with a Trace has mark in picture


Is That Trace a Shunt or some things else ? Could anybody explain it for me ? Please.
 

Offline fourtytwo42

  • Regular Contributor
  • *
  • Posts: 104
  • Country: gb
  • Interested in all things green/ECO
Re: DC/DC 48V/60V to 12V-10A
« Reply #1 on: June 19, 2017, 10:23:53 PM »
I see several problems with the schematic, I assume Q1 is N-channel ? well if it is the integral diode is drawn the wrong way around. The trace you ask about is a current sensing shunt, if you want to copy this pcb then make sure you copy that traces dimensions exactly (weight, width & length).
 
The following users thanked this post: sangseu

Offline sangseu

  • Contributor
  • Posts: 7
  • Country: vn
Re: DC/DC 48V/60V to 12V-10A
« Reply #2 on: June 19, 2017, 10:42:07 PM »
Yes, MOSFET is N-Chanel. If I can measure resistor of that Trace, I can replace it by a Shunt Resistor, don't I ?
 

Offline MagicSmoker

  • Frequent Contributor
  • **
  • Posts: 745
  • Country: us
Re: DC/DC 48V/60V to 12V-10A
« Reply #3 on: June 20, 2017, 01:44:20 AM »
The MOSFET is drawn correctly, but this circuit won't work for a number of reasons.

The first is that you need some kind of floating or charge-pumped supply for driving an n-channel MOSFET in a buck converter, as the gate needs to be brought 10V higher than the source to turn it on, but the source will be at the same potential as the drain at that point... (there are alternative approaches which I am omitting in the interest of brevity and simplicity).

Next is that the current sense signal needs to be obtained with a differential amplifier or some other high-side sensing scheme because the Isense input on the PWM controller IC expects to see ~1V with respect to ground.

Using a trace for the current sense resistor is a time-honored hack for doing things very cheaply while also getting a degree of temperature-dependent overload protection for free - the resistance of copper goes up 0.4%/C. But, as I said already, you need more circuitry to do this in a buck converter in the first place.

 
The following users thanked this post: sangseu

Offline bktemp

  • Super Contributor
  • ***
  • Posts: 1606
  • Country: de
Re: DC/DC 48V/60V to 12V-10A
« Reply #4 on: June 20, 2017, 02:16:57 AM »
The MOSFET is drawn correctly, but this circuit won't work for a number of reasons.

The first is that you need some kind of floating or charge-pumped supply for driving an n-channel MOSFET in a buck converter, as the gate needs to be brought 10V higher than the source to turn it on, but the source will be at the same potential as the drain at that point... (there are alternative approaches which I am omitting in the interest of brevity and simplicity).
The circuit is correct and will work. The UC384x operates as a floating highside switch driver:
D2 is the bootstrap supply using the 12V output voltage. R4 is the startup resistor.
The current sense is also correct, since UC384x GND is connected to the switching node.
 
The following users thanked this post: sangseu

Offline MagicSmoker

  • Frequent Contributor
  • **
  • Posts: 745
  • Country: us
Re: DC/DC 48V/60V to 12V-10A
« Reply #5 on: June 20, 2017, 02:26:27 AM »
The circuit is correct and will work. The UC384x operates as a floating highside switch driver:
D2 is the bootstrap supply using the 12V output voltage. R4 is the startup resistor.
The current sense is also correct, since UC384x GND is connected to the switching node.

You are correct - I should have noticed that the node labeled Gnd wasn't connected to the 0V common line between input and output.

Carry on!
 
The following users thanked this post: sangseu

Offline fourtytwo42

  • Regular Contributor
  • *
  • Posts: 104
  • Country: gb
  • Interested in all things green/ECO
Re: DC/DC 48V/60V to 12V-10A
« Reply #6 on: June 20, 2017, 05:12:42 AM »
The MOSFET is drawn correctly
Hmm the picture is so poor I had thought the diode cathode was connected to the source but perhaps its the drain after all :)
 
The following users thanked this post: sangseu

Offline fourtytwo42

  • Regular Contributor
  • *
  • Posts: 104
  • Country: gb
  • Interested in all things green/ECO
Re: DC/DC 48V/60V to 12V-10A
« Reply #7 on: June 20, 2017, 05:14:35 AM »
Yes, MOSFET is N-Chanel. If I can measure resistor of that Trace, I can replace it by a Shunt Resistor, don't I ?
You should not have to measure it, according to the schematic it's 0R2 or is that your estimate ?
 
The following users thanked this post: sangseu

Offline sangseu

  • Contributor
  • Posts: 7
  • Country: vn
Re: DC/DC 48V/60V to 12V-10A
« Reply #8 on: June 20, 2017, 11:13:36 AM »
You should not have to measure it, according to the schematic it's 0R2 or is that your estimate ?

It's my estimated. take some math, V_ref = 5V, I_sense about 1V, resistor devider is R2---R3-R_Shunt so U_Shunt ~ 0,24V.
I_in ~ P_out/U_in = 1,7A so R_Shunt ~ 0,24/1,7 = 0,14 Ohm.

labeled Gnd wasn't connected to the 0V common line between input and output

Right, this schematic may work follow this INVENTION
https://www.google.com/patents/US6313616


copy that traces dimensions exactly (weight, width & length)

I can do that by import picture of that PCB to KiCad and make the same Trace Shunt. But, different PCB quality causce different value of Trace Shunt.

Please give me a suggestion for copy that designed.
« Last Edit: June 20, 2017, 11:27:03 AM by sangseu »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf