hi

Is there a simple way to isolate grounds from a single DC supply, in order to provide isolated grounds to different circuits? I've seen this asked elsewhere, but have not seen a good solution.

Note, I'm obviously not an EE, just a self-taught experimenter. i don't even know enough to be dangerous.


Separate supplies:

In this example, i'm charging 2 caps from separate supplies, yet i can still keep the cap connected in series. The charging remains isolated:

Common supply:

But in this scenario, with a common supply, the bottom cap does not charge-- it's return path takes a shortcut back to ground through the series wire connecting the 2 caps:

How can i isolate the return ground of the bottom cap, without losing the hardwired series connection between the caps?


Mosfet Driver (MAX628):

This could be my answer, if i can understand it. This circuit, outputs 2 separate, isolated Vsupplies from a single source! using a MAX628 dual mosfet drivers.

The outputs are not fixed V-- the V's vary with the input V, as i want.

"Simple Circuit Generates Multiple Isolated Supply Voltages"
https://www.maximintegrated.com/en/app-notes/index.mvp/id/1932
https://datasheets.maximintegrated.com/en/ds/MAX626-TSC428.pdf

i gotta figure out if this means i can chain more mosfet drivers to drive more outputs:

"Additional diode-capacitor networks can be added as shown to create a multitude of isolated voltages, limited mainly by the maximum total load current of all outputs and the maximum drive available from U1"

Also need to decipher this:

"This circuit is suitable for systems with fixed DC potentials, but is less suitable for those in which the potential between low- and high-voltage sections is varying."

What is "the potential between low- and high-voltage sections"?

Balancer:

This balancer circuit seems to have isolated grounds, correct? Those mosfets again. Hoping to simplify. I'm not trying to balance cells, so i can remove the balancing parts, and just keep the isolation parts. If i could only figure out where they are

http://www.nxp.com/docs/en/application-note/AN4428.pdf
(fig 2)

It would be awesome if someone could share a quick analysis of this circuit. Here's my weak attempt:

- What's the inductor for? Some sort of delay, which helps isolate the grounds?

- What do the caps do? i think the caps at the control pins filter out DC offset, correct? Is that just to help condition the the control V's, or is it related to isolation?

- Why N/P mosfets, instead of just N or just P? I'm guessing that's to allow bidirectional current, so current can pass either direction between the cells. But i don't think that's related to isolation.

- The diodes seem to me the most likely related to isolation. The doc says there's a relationship between the schottkies and the diodes inside the mosfets. That's beyond my understanding, but makes me think the solution might involve diodes and mosfets.


Other options:

Also looked at galvanic isolators, but cannot say i understand them.

This article mentions, "distributed power supply technology and the power bus technology. A DC voltage bus and an AC quasi-square wave voltage bus are built by relevant converters". "distributed power supply" sounds promising, but complex. i could not find further info on that stuff.
http://ieeexplore.ieee.org/document/7037948/

i think it could also be done by putting a DC-DC converter / regulator on each circuit, but i believe that will output a fixed V. I want the output V's to vary as the source supply varies.

i also considered transformers, but seems bulky. Maybe a surface-mount transfomer?
http://www.ebay.com/itm/ED8-SMD-Audio-Nickel-Steel-Transformer-600-600-1-1-/272443169653

Hoping the solution, if there is a very simple one, might be accomplished with just diodes and/or mosfets. I experimented a bit with diodes, but could not figure it out.

cheers!

 

There are simple DC-DC converters out there that you could use for that purpose. If you're willing to pay the price. One I've been using recently is the CME0512S3C 5 to 12 VDC converter. It's perfect for driving standard mosftets from a 5VDC supply.

 

There are simple DC-DC converters out there that you could use for that purpose.

thx for tip!
but, i believe that will output a fixed V. I want the output V's to vary as the source supply varies.

cheers

 

What are you trying to do? A "source" is a very general term. Every detail makes a big difference.

 

by "source" i mean voltage source ie the power supply

thx!

 

There are variable output dc-dc converters out there. I suggest you do a little searching, and then get back here if you have any doubts.

 

There are variable output dc-dc converters out there. I suggest you do a little searching, and then get back here if you have any doubts.

Thx for reply, but I did not explain well.

I'm not looking for a variable supply. I'm already using one!

I want to get multiple isolated grounds from it.

 

Why does your source supply vary? Do you want the source to vary? Why? What are you trying to do?

 

Do you want the source to vary?

Yes, I want the source to vary. I'm only asking how to generate isolated grounds.

I understand you'd like more info about my whole project, but sorry, I can only say that isolated grounds would be very helpful! I promise i'll share the details if i get everything working!

No one has commented on the several possible methods I mentioned in the OP. Surely there's a diode ninja out there

 

Wow, the Mosfet Driver (MAX628) method in my OP sure is looking good.
I think that might be the answer.
Any opinions?

 

Maybe you are over-thinking this problem.

Problem: You want to charge batteries connected in series using a single battery charger.

Solution: Connect the positive and negative leads of the charger to the positive and negative terminals of the battery, respectively, charging one battery at a time.

You can do this by manually connecting the two leads to the battery.
You can also do this with mechanical selector switches or with mechanical relays.

 

Connect the positive and negative leads of the charger to the positive and negative terminals of the battery, respectively, charging one battery at a time.

You can do this by manually connecting the two leads to the battery.
You can also do this with mechanical selector switches or with mechanical relays.

Exactly, MrChips. I'm using switches.

In your scenario, i'd need 2 switches per cell, to connect/disconnect both charger leads (hot and ground) from each cell in turn. Right?

I'm trying to reduce parts count, by using just 1 switch per cell. That would be half as many switches. Yes i know i can use a double-switch, but that still more wiring and soldering.

To use just 1 switch/cell, i'd have to keep the ground-lead of each cell connected to ground of the charger permanently. Instead of moving the ground from cell to cell as i charge, I would only switch out the hot lead.

According to my tests, keeping the charger ground connected to all cells at all times requires isolated grounds.

But, getting isolated grounds may end up being more parts than simply using 2 switches per cell. I'm comparing! i did not know before i started exploring, which option would be lowest-parts-count.

Please share if you know another way to do it with just 1 switch per cell, without isolated grounds!

Even if i decide to go with 2-switches/cell, instead of isolated grounds, i think isolated grounds is a very neat trick, might come in handy sometime. So i'm glad i found a method (mosfet-driver in my OP).

(btw, when you say "the battery", you mean "cell", right? i think "battery" = "pack".)

cheers!

 

I cannot see any way of doing this with just one switch per cell.

 

I cannot see any way of doing this with just one switch per cell.

--except the isolated grounds method, right?

 

To get your "isolated grounds" you would have to have isolated power supplies.

 

To get your "isolated grounds" you would have to have isolated power supplies.

Yep! I believe that's achieved by the mosfet driver solution in my op.

But, was hoping it could be done with just a couple diodes.

 

Pardon me. I must be missing something. A power supply cost less, and is more efficient, and is smaller...... than a extra switch?

 

A power supply cost less, and is more efficient, and is smaller...... than a extra switch?

- I need each voltage to track the common supply, meaning vary with the common supply (ie charger).

- plz share a supply that's cheaper, takes up less room, and fewer parts than a switch!

Thx

 

Sorry I couldn't help......good luck.

 

Sorry I couldn't help......good luck.

i misread your message. You're right, a switch is simpler.

As i mentioned above, i did not know when i started looking at this, if i could achieve isolated grounds simply with one or 2 parts. If so, then it would be a viable alternative to a switch.