Help with designing an odd capacitor system...

Thread Starter

undescriptive

Joined Apr 14, 2013
9
Hi everyone,

I am trying to design a circuit that does the following:


1. Takes a DC input (currently 6-15V)

2. cycles the DC input between two (or more) outputs

3. the output charges a capacitor (when the capacitor is charging, the load must not be attached.)

4. once the capacitor is no longer charging, it is "disconnected" from the rest of the circuit and attached to a load.

(sort of like 2x DPT switches that are linked together but wired opposite)

I was thinking about doing this with an astable multivibrator (transistor - not op-amp based), but I get lost from step 3 onwards

I have an empty breadboard, some bc-337's and a bit of time to play with!

If anyone could help, I would be very grateful

TIA!
 

#12

Joined Nov 30, 2010
18,224
At first glance, this seems to be a job for relays.
Any crude attempt at a drawing would be a lot of help.
The part about, "two or more loads" really throws a wrench in the works.
I'll have to ask you to get more specific in defining that.
 

Thread Starter

undescriptive

Joined Apr 14, 2013
9
attached is a probably massively wrong diagram of one half of the cycle

I'm fairly certain you can imagine what the other half looks like!

I was considering the astable as it already has the timed switching capability...

I hope this makes sense!

The astable I can build, it's the next bit of charging a capacitor, then "disconnecting" it from everything except the ground, then using that to power the load.

Basically the load should never see that is connected to the power source, only the capacitors...

Thanks!

p.s. oops looks like I missed the ground connection on the load!
 

Attachments

Thread Starter

undescriptive

Joined Apr 14, 2013
9
interesting opamp....

however, I'm not sure it will do what I want (though I'm not an expert in these things at all!)

the capacitors both need to be charged to positive DC bias, as in charged to +12V (not +ve and -ve as the data sheet suggests)

I could go the mechanical route for this too...

imagine a disc with 4 conductive tracks around half the outside.
these tracks are setup to be 180 degrees from each other.
there are brushed contacts on each of the tracks
ground is common
as you rotate the disk, one half of the disk is dedicated to charging a single capacitor and the other is discharging - and vice versa on the other half of the rotation.

This links both "sides" together, but makes them mutually exclusive - if you catch my drift?

It's tricky to explain in text and even for me to get my head around this at almost midnight!

Thanks for all the input though!
 

#12

Joined Nov 30, 2010
18,224
Yes, I get the picture, and that is what the 7660 chip does, but it only does that for about 10 milliamps. Do you want more than 10 milliamps?
 

Thread Starter

undescriptive

Joined Apr 14, 2013
9
Hi, yeah, more than 10mA...

I would like to be able to handle upto 100A in the end...
but for this experiment around 1A would be good....

Thanks!
 

Thread Starter

undescriptive

Joined Apr 14, 2013
9
Looking at the specs of Relays again...

This might be better suited to that - they could last a very long time without issue...

So, lets look at a relay (or 2) to charge and discharge the capacitors in an alternating fashion....

The capacitors will be big chunky 1F types...
 

Thread Starter

undescriptive

Joined Apr 14, 2013
9
@ #12

Yes! that's it!

Just that is a single cycle, so there would need to be another capacitor/relay combo which would be in the "ON" state while the other is off....
 

#12

Joined Nov 30, 2010
18,224
If you're going to make a 2 transistor oscillator you can run one relay from each collector, but I think a 555 chip would be a good choice if the timing is a few seconds or more.

ps, I have to go to sleep now. Other people might help while I'm gone.
 

richard.cs

Joined Mar 3, 2012
162
So you want to use a flying capacitor to generate an isolated supply? That sounded reasonable until you mentioned 100 amps. I assume you want a continuous current by swinging the capacitor back and forth, 100A doesn't seem feasible with any sensible combination of capactor size and switching frequency.

What's your application? There has to be a better way of doing it. I'm imagining generating an isolated supply by switching at 50 kHz or so and driving a transformer.
 

Thread Starter

undescriptive

Joined Apr 14, 2013
9
well, perhaps 100A is overkill (could always run more of these in parallel to get more current)

so lets go for 15A max... but for this demo... lets stick with 1A :)

Switching speed can be as slow as required to charge the capacitor (or discharge it)
I know that is load dependent, but several seconds per cycle is fine....
 

richard.cs

Joined Mar 3, 2012
162
So you want to deliver pulses of 1 amp to your load several times a second, rather than switch rapidly to try and simulate a continuous current source?
 

Thread Starter

undescriptive

Joined Apr 14, 2013
9
not quite...

I have been having a think... see if this makes sense...

So, set a555 into an astable multivibrator (or 2 transistors)
then use the activation time to activate a relay (DPST), which then charges the cap directly from source voltage, when the system then flips, the charged cap is isolated and able to discharge into the load while the other cycle is charging.

so, this has 2x Capacitors and 2x DPST relays

So the load will essentially see a constant but isolated DC source
 

richard.cs

Joined Mar 3, 2012
162
So you do want to supply a continuous current to the load then. As a minimum such a system would need to have three flying capacitors, or considerable capacitive storage on the load side. Otherwise there will inevitably be breaks in the current flow when the relays switch.

If you do this with mechanical relays you should expect to need huge capacitors, switch very high peak currents, and generally produce a system that is highly inefficient and has short service life, if it works at all.

When this is done with capacitors it usually uses capacitors of a microfarard or two switched at many kilohertz by mosfets, and supplies at most a few 10s of mA. A transformer based approach would be much more appropriate for what you want to do here.

Something akin to this http://www.powerguru.org/wordpress/...07/Isolated-AC-DC-Converter-Block-Diagram.jpg would work well, and if you don't need good regulation on the output then the feedback path isn't needed. With suitable component selection you could transfer 10s of amps across the isolation barrier with a transformer the size of a deck of cards and a few mosfets and diodes.
 

shortbus

Joined Sep 30, 2009
10,045
I'm doing something similar to this in an EDM machine circuit. Two capacitor 'banks' with a charge and discharge mosfet for each bank. While the one bank is charging the other is discharging. If this is the type circuit you need I'll draw up a quick sketchfor you and post it.

My circuit also includes PWM for the discharge, voltage measurement on both charge and discharge also. But the main part of the charge/discharge can be done with a D flip-flop and inverter chips, plus a 555.

Let me know if you think it will serve your needs.
 

Thread Starter

undescriptive

Joined Apr 14, 2013
9
shortbus - that sounds really good!

I'm trying to keep this as simple as possible (I'm pretty simple!)
So any kind of diagram would be excellent!

Thanks for all the help guys!
 

#12

Joined Nov 30, 2010
18,224
As a matter of opinion, you haven't said why it is necessary to use a switched capacitor device to isolate a load. Lacking that information, a switched capacitor circuit seems way more complicated than just using a DC to DC converter which isolates the load perfectly and is cheap to buy.
 

shortbus

Joined Sep 30, 2009
10,045
shortbus - that sounds really good!

I'm trying to keep this as simple as possible (I'm pretty simple!)
So any kind of diagram would be excellent!

Thanks for all the help guys!
I'm pretty simple myself.:p
OK, don't know if this is something that will help you or not. I left out the part to introduce PWM to the circuit, since you didn't say you needed it.

This is a basic circuit that will need more to finish it out. Didn't want to spend too much time with my hitech drawing hardware (pencil, paper and ruler) if it doesn't look like you can use it. It will need a few more parts for the mosfets, that I can show if it will work for you.
 

Attachments

Last edited:
Top