Isolated DC:DC converter

Go To Last Post
19 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm looking to build some low cost Voltage meters to do some production testing / monitoring.

Each device will be low power (Approx 0.5W 5V, 100mA). Running a MCU with ADC + Some seven segment displays.

Plan is to use a 24V DC bus and have each node run a isolated DC:DC step-down to power the device. I realise you can use a intergrated module to do this but I would like to know if there is a simple discrete solution out there anyone can recommend?

oddbudman

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Isolation is a bit tricky., The standard is a transformer, I guess. The challenge is getting feedback.

Some schemes use an opto-isolator and others measure voltages and currents in the primary. I've used some Linear Tech switchers that do the latter with middling results. I've tried some that use the opto feedback with even less success.

Another challenge is finding the right transformer. A few switcher manufacturers specify particular transformers and sometimes they are available through distribution. Those that are tend to be expensive.

There has to be circuits that work.

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You can regulate a flyback converter from the primary side and probably get good enough regulation for your needs.

Tom Pappano
Tulsa, Oklahoma

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Come to think of it, I happen to have some flyback transformers left over from one of my devices that I no longer manufacture. They are designed for 24 vdc in 50khz, and +12,-12v, +5v, and isolated 5v out. If you would like a deal you can't refuse, let me know 8-)

Tom Pappano
Tulsa, Oklahoma

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks for the offer Tom, but I should be right here- I live in Dongguan China (plenty of transformer factories in my neighbourhood ;)). Plan is to make 1200 of these things, so I'll be able to spec the trx I want to suit this application.

Your right in that I'm not so worried about regulation, I plan to put the mcu behind a linear reg on the secondary side of the dc:dc.

I like the looks of the circuit on page 20 in this Appnote.

http://www.irf.com/technical-info/appnotes/an-937.pdf

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you want to learn, then get a commercial dc-dc converter similar to the specs you require and crack it open. If encased in epoxy, heat it to 100C and most epoxy will go like rubber. Use a small screwdriver to pick it out - carefully. Then sketch the circuit. Decide if it is easier to buy or to build. You can bet each and every component is there for a reason, so if you think it can be done simpler, then there will have to be a tradeoff.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I ran across a cute IC used on the Atmel VFD demo board. It is the IR2153 "self oscillating gate driver", and appears to be a nice low parts count solution for simple power supplies.

Tom Pappano
Tulsa, Oklahoma

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Isolated DC-to-DC but can generate more power than you need:
http://www.powerint.com/en/produ...
China link is at the top of the page.

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The Fairchild FAN6754 has an example of the whole design for running off the mains with isolation and feedback control

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm not really interested in running with mains voltages.

I'm quite familiar with these intergrated switches (ST, Fairchild and Power intergrations etc), but for this application i'm trying to avoid using one.

I only need around 100mA at 7V. Enough to power a MCU and some seven segment displays hiding behind a 5V regulator.

So right now I'm tossing up between a 555 or a logic gate to generate the waveform that i will apply to a small ferrite transformer to isolate the supply.

Anyway, atttached is the current schematic, it might help give you a better idea of what the plan is. Any extra eyes / feedback appreciated. The method to get a negative voltage for the amplifier may raise a few eyebrows, i'm still thinking about this approach.

For this circuit DCIN will be 12VDC.

Please don't put me on a stake for using a Holtek ;) It's a case of the shoe fitting for this app.

oddbudman

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

oddbudman wrote:
I'm not really interested in running with mains voltages.

I'm quite familiar with these intergrated switches (ST, Fairchild and Power intergrations etc), but for this application i'm trying to avoid using one.

I only need around 100mA at 7V. Enough to power a MCU and some seven segment displays hiding behind a 5V regulator.

So right now I'm tossing up between a 555 or a logic gate to generate the waveform that i will apply to a small ferrite transformer to isolate the supply.

Anyway, atttached is the current schematic, it might help give you a better idea of what the plan is. Any extra eyes / feedback appreciated. The method to get a negative voltage for the amplifier may raise a few eyebrows, i'm still thinking about this approach.

For this circuit DCIN will be 12VDC.

Please don't put me on a stake for using a Holtek ;) It's a case of the shoe fitting for this app.

oddbudman


Out of curiosity - why the LC filter on the input? (L2+C1). Additionally, isn't your DC/DC just a flyback converter? Why the full bridge rectifier? Additionally, what purpose do R5 and R6 serve?

I think your method of getting a negative voltage is perfectly fine if you don't mind the loss in efficiency. However, I think it'd be wise to put more filtering on it.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

R7 and R9 seem to be or no use. Those gates are unused and you tied the inputs low so why use those resistors on the outputs to ground? No harm but maybe waste a little power? Just curious

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

For a flyback converter, the phases of the transformer are wrong, the phase of the primary should be the other way around, and you would need only one diode on the secondary.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've changed to the 500khz version from the appnote, this way i can save some pcb space and I can put the unused logic to use.

R5 and R6 are used to load the converter a little so the output voltage won't overshoot too much if load after the regulator drops (MCU down etc).

This is a forward converter, not flyback. See the application note I mentioned earlier on page 20-21 for more information.

The LC filter on the input is to reduce conducted emissions back to the main psu. I plan on running a daisy chain of these devices so i'm just trying to avoid too much noise finding its way onto the power supply line.

Updated schematic attached. Thanks for the help ppl :)

oddbudman

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Before committing to a PCB I'd prototype the converter first :) It's not a standard topology, and the appnote says it becomes a flyback at loads below 5mA. Just oogle Gmages for 'forward converter' and you'll see a few variations, but not this one :)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Ok, just an update. Pics of the beast in action below.

Not too many speedbumps along the way. Its been a struggle to get the right ferrite core, still waiting for its arrival. In the meantime I've used an EMC core found at the parts market.

oddbudman

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

A quick question.Did this topology work?As i see in the picture has 1:2 windings,a round ferrite core with no airgap,so it must be no flyback.
In the accuracy of readings the offset of the opamps must be taken care,especially in current measurement since it is offset(mV) X 69.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yeah, it does work. I was running this one off a 6V battery, hence the winding turn configuration. Thanks for the heads up on the offset voltage. I'll try and post some sort of load regulation graph as well as efficiency specs when I get round to it.

oddbudman

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

What about using a TL431 shunt regulator in secondary with a PC817 optocoupler for output regulation.