Need galvanic isolation

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

I have to build a device that can measure current among other things. For current measurement I need galvanic isolation. The complete device shall be supplied by USB. For communication I want to use a FTDI chip. For current measurment I want to use an ATtiny261V. I did some worst case current consumption calculation and came to the conclusion that the circuit never exeeds 2mA.

I don't want to use any other/additional power supply than the USB power supply. So I need to add a galvanic isolation circuit. The attached schematic explains my first idea with optocoupler. But the CNY74-2 optocoupler can't switch enough current to the ATtiny261V circuit. With a LED current of 20mA the opto-circuit can only pump 1.5mA to the ATtiny261V side.

The 2nd thought was to use a simple voltage doubler circuit on the input side of the galvanic isolation circuit, so the voltage drop accross the circuit can be higher. Then the circuit would be able to pump 2mA but no reserve left if the optos age.

The next thought was to discard the circuit and try a transformer - something like this one. It has a ratio of 1:3. But I don't have this transformer at hand and no experience with it.

if I use a low-drop/low-quiescent-current regulator on the ATtiny261V side, is it possible to gain stable 3.3V DC @ 2mA? I can pump up to 35mA into the transformer (to stay below the 100mA USB limit).

Or can you recommend better optocouplers or even a better circuit?

Any idea would be appreciated.

Regards
Sebastian

Attachment(s): 

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

Why not use some of the nice AVAGO optomosfet-optorelay like ASSR-1511?

Guillem.
"Common sense is the least common of the senses" Anonymous.

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

Simple answer: I didn't know about it and I don't know where to get it. Can you recommend a seller?
Many thanks!

Regards
sebastian

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

I checked all by AVAGO recommend distributers. It seems no one has these nice parts in stock. :(

Regards
Sebastian

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

You can try ADI ADUM5241.

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

They look perfect with the exception that they consume a lot current. More than 100mA are needed, to output some mA and I have only 35mA left for my application.

Regards
Sebastian

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

Use Hall sensor for current measurement. It's simple and fault resistant solution.

Limit for USB isn't 100 but 500 mA.

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

I think with respect to "safety" your opto-circuit
would not provide galavanic isolation between USB and
Tiny-supply.
I would use a small transfomer based solution.

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

Sebastian,
I'll add a thought or two:

Isolation typically means there is no current path from one side of the circuit to the other. A transformer has current going in and out one side (winding), and a current induced in the other winding. Ignoring the finer points, no current flows in a loop from the primary to the secondary side.

An opto coupler provides isolation, also, You feed your signal in the LED side, and the photodiode turns on and off. Light is used to couple the two halfs. Again, no current flows from the LED side to the output side.

I believe, has you have drawn you circuit, there is no isolation, as ossi pointed out. Think of the optoisolators as signal couplers, not power transfer units.

An option for you might be a MURATA Power Solutions DC-DC Converter. Mouser, and others, sell these. They come in an a large selection of voltages and current ratings. They are about the size of a DIP chip. You feed in +5 / Gnd from your USB source, and you get out +5/Gnd, or +/-5, or +12, or +/-12, etc...
The output power is truely electricallly isolated from the input to a rating of 3,000 V isolation. This means you can put 3,000 V across the input and output sides of the device before it breaks down and current flows.

I'd condiser measuring the current, either with a Hall as mentioned above, or with a series resistor, etc. Convert it to a digital number. Feed it digitally through an optocoupler back to your main device.

JC

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

Many thanks for all your great help!
A DC-DC Converter seems to be the way to go. Expensive but might avoid a lot of headache. However I missed details about the input current when not needing the max. output current. I want to stay below a total current consumption of 100mA of the complete circuit because in this case no self powered USB hub is needed. Can somebody share his/her experience about power consumption of DC-DC Converters? Maybe they consume a lot of current even if the output is not loaded.
Many thanks!

Regards
Sebastian

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

Pick one that has the voltage and current specs you need for the remainder of the project, then pull the data sheet. It will provide this info, or info on its efficiency.

A few mAs for the uP, and 10 mA for an opto coupler should not be a problem.

Note that if you are drawing 100 mA at 5V from the USB you have 0.5 Watts total power to play with.

JC

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

Thanks JC. I have selected the smallest available type (1W). In the datasheet I only find information about the efficiency at the max. output current (69%). I think the device has a min. input current consumption, even if the output current is zero. The min input current might be too high for my application - but I don't know it.
I have send a question to the manufacturer. If I get any reasonable response, I will quote it here.

Regards
Sebastian

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

You may wish to google "DC-DC Converters" for some other units, perhaps 0.5 W devices, etc.

I mentioned the above only because I have used them previously.

Good luck with the project!

JC

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

I have checked some german distributors. It seems like I can't get 0.5W devices. If I don't get a response by the manufacturer, I'll order a device and check it by myself.

Regards
sebastian

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

IIRC Tracopower is distributed by Farnell, and sells pretty small DC-DC converters. I use one of them, 5V in, 5V out, 1W, to supply an isolated RS-485 interface. Can't remember the reference, thought.

Anyway, if you can buy to Farnell, RS, Mouser or similar, they should have a wide range of small DC-DC converters.

Guillem.
"Common sense is the least common of the senses" Anonymous.

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

Hi!

For little power consumptions the suggested RS-485 isolator is perfect.
The IC You can use is the MAX253 which is available from Farnell. Order Nos: 9725393, 9724680, 1379887.

The transformers those may be used with this IC are also available at Farnell. Order Nos:
1362404, 1362405, 1362408, 1362407, 1139146, 1139144, 1139145, 1215120.
These are from the Murata 78253 series.

Regards,

Peter

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

Don't forgret to read the data sheets for any DC:DC converter, while they may quote 1-3kv isolation, they are usually designed for ELV-ELV isolation.

They usually wont meet EN60950 or equivalent so be carefull if you are trying to get isolation from a mains supply to a microcontroller.

Ron.

 

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

Yesterday I received a NME0505DC. Sweeeeet!

Consumes less than 17mA with no load.
Very nice part to isolate USB power supply. I only need the isolation to make some rough current measurements of a 12V automotive ECU. So no big safety requirements for my application.
Many thanks for all your valuable help!

Regards
Sebastian