Maker Pro
Maker Pro

help with designing a circuit

undescriptive

Apr 14, 2013
11
Joined
Apr 14, 2013
Messages
11
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.


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

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!
 

HellasTechn

Apr 14, 2013
1,579
Joined
Apr 14, 2013
Messages
1,579
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.

TIA!

setp 1. understood.
step 2. what do you mean ? divert the dc power to one output at a time? like a wsitch ?
step 4 in DC current the capacitor is open circuit when fully charged.
then again how will you attach it onto the load.......

I think what you need is a microcontroller circuit...
 

undescriptive

Apr 14, 2013
11
Joined
Apr 14, 2013
Messages
11
Hi,

Thanks for the reply.

2 - yes, two capacitors being charged and discharged at opposite times:
i.e

Cap1 - Charging
Cap2 - Discharging

then

Cap1 - Discharging
Cap2 - Charging

So, yes, this is just like a double pole switch for each capacitor and the switches are changed together at the same time.

There must be a way to do this without using a micro-controller - Ideally this should just used passive components and transistors...

you could do this with a mechanical disk and wiper contacts....

Thanks!
 

undescriptive

Apr 14, 2013
11
Joined
Apr 14, 2013
Messages
11
something like that, but I'm trying to stay away from a relay as it will burn out quickly (constant switching)
 

sirch

Dec 6, 2012
109
Joined
Dec 6, 2012
Messages
109
What are the capacitors driving? What frequency do you want to run the change over at - i.e. how often do you want it to flip from C1 to C2?
 

undescriptive

Apr 14, 2013
11
Joined
Apr 14, 2013
Messages
11
after looking at the specs for relays, I was a bit daft in discounting them....

So, relays are back on the table again!
 

undescriptive

Apr 14, 2013
11
Joined
Apr 14, 2013
Messages
11
The flip time should be when the capacitors reach "maximum" charge or half discharge time - I'm quite flexible on this, it doesn't matter about the frequency.

The capacitors could be driving anything, it doesn't really matter...
I have some 1F capacitors in mind so they should be able to cope with virtually anything that I could throw at them!
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Does the capacitor need to be completely disconnected from the drive circuit when the load is connected? Like in a switched capacitor power isolator? Or can one terminal of each capacitor stay connected to the ground rail of the driving circuit?

If there's no need for full isolation, you can use transistors or MOSFETs. Although relays might last long enough, the click-clack-click-clack-click-clack will drive you out of your cotton-pickin' mind! Unless you use a little one that you can hardly hear.

Give us some more details, such as the expected charge current and the voltages you'll be using.
 

undescriptive

Apr 14, 2013
11
Joined
Apr 14, 2013
Messages
11
I think totally isolated will be best...

Voltage will be no more than 15V
Current could be up to 30A (eventually) - I want to try is smaller though to demo - so only about 1A to begin with...

Noise isn't so much of an issue, I won't be in the same location!
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Noise isn't so much of an issue, I won't be in the same location!
Ah, but you will be when you test it!
You'll probably be alright with small relays, but those big ones... after ten or fifteen minutes... Arrrrggggggh!
 

undescriptive

Apr 14, 2013
11
Joined
Apr 14, 2013
Messages
11
I quite like the sound of clicking things! :)

but if you have any diagrams on how to rig this up, that would be awesome!

Thanks!
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Yeah, I thought that too. "Clicky clicky noise, sounds cool!" For the first ten minutes or so.

You'll need one DPDT (also called DPCO) relay for each capacitor.
Here are some examples of DPDT relays. They all have 12V coils, but other coil voltages are available.
2A contact rating: http://www.digikey.com/product-detail/en/TX2-12V-TH/TX2-12V-TH-ND/2709466
10A contact rating: http://www.digikey.com/product-detail/en/LY2-DC12/Z790-ND/126870
30A contact rating: http://www.digikey.com/product-detail/en/LY2N-DC12/Z2603-ND/369492

Connect the capacitor between the changeover terminals.

A good option for driving the relay coils is a 555 timer. It is as common as dirt - just Google "555 oscillator" for examples. It needs a power supply between 4.5V and 15V. Its output can drive up to 200 mA in each direction, so you connect one relay coil between the output and the positive supply, and the other coil between the output and the negative supply, and they will switch alternately.

Make sure you connect a diode in reverse across each relay coil. Relay coils are inductive, and when the driving voltage is removed, they generate a "back EMF" which will damage the 555 if the diode is not present.
 

undescriptive

Apr 14, 2013
11
Joined
Apr 14, 2013
Messages
11
So... I could get away with a DPDT relay designed for a car/vehicle?
They can usually handle a large current and will probably be cheap and easily obtainable...

So, set the 555 into an astable multivibrator (or 2 transistors)
then use the activation time to activate the relay, 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....

That sound right?
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
That's all exactly right.

You will want some way to limit the charging current flowing into the capacitor, especially if you plan to use a 1 farad capacitor. A discharged capacitor will try to draw an extremely high current (limited only by the resistance in the circuit, which includes the effective series resistance or ESR of the capacitor, which is very low for a large 1 farad capacitor). A typical voltage-regulated power supply would not be able to supply enough current, and its output voltage would drop at the start of each charge cycle. This would be bad for every component in the system, and it would also wreak havoc with the oscillator if the oscillator was powered from the same power source.

The simplest current-limiting arrangement is a series resistor. This causes the capacitor voltage to rise over a period of time with a characteristic curve. If you want to charge the capacitor as quickly as possible, a current regulator is better; it allows you to charge the capacitor steadily at a maximum rate. It would also allow you to disable the charger briefly during the relay switching, so the relay contacts (on the charging side, at least) will not have to switch a heavy current.

Whichever option you use, the series component (resistor or current regulator) will dissipate significant power, as heat. This could require some heavy heatsinking, depending on how much power you need to transfer into the capacitor. You can use a switching power supply for better efficiency; a 15V 30A power supply of any kind will be pretty big and expensive though.

It would be interesting, and could be helpful, to know what you intend to use this circuit for.
 
Last edited:

undescriptive

Apr 14, 2013
11
Joined
Apr 14, 2013
Messages
11
this is basically a science project!
me being curious I suppose...

to help limit the charging, could I use an incandescent bulb to help limit the power going into the capacitor? (just me musing here)

though I assume that the bulb would increase in resistance as it glows (making it the opposite of what I want)

30A is probably way overkill - so if we could limit this sensibly to 5A out, it would take approximately 5 seconds to charge a 1F cap from nothing (with no limiting resistor or such)

my thinking is that if I have a current source that can supply many Amps, it would be better to have multiple smaller circuits in parallel to provide larger currents rather than a single giant system.

Sorry for being so mysterious - I do want to build something though!
So, how would I current limit the capacitor charging system to say 1A(or at most 5A)?

Thanks :)
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Right, an incandescent bulb is the opposite of what you want.

The formula for charging a capacitor is:

dV / dT = I / C

so if I = 5A and C = 1F, the capacitor voltage will rise at a rate of five volts per second. So it would take three seconds to charge from 0V to 15V.

I don't think it would be better to have multiple relays and multiple capacitors. If you need to transfer a large amount of power to the load, use one big capacitor and one big relay, unless several smaller ones are needed for price or feasibility reasons.

A current limiter or current regulator is also called a current source or current sink, depending on whether it's connected in the positive or negative side of the load (the capacitor, switched by the relay contacts). Start at http://en.wikipedia.org/wiki/Current_source#Zener_diode_current_source. That shows a simple current sink using an NPN transistor, which would connect between the relay and ground (with the other side of the relay connected to V+).

This circuit is fairly simple and not very accurate, but it will be fine for your application. The current that it sinks (into the collector of the transistor) can be calculated using Ohm's Law, I = V / R, where I is current in amps, V is the voltage across the emitter resistor (called R2 on the Wikipedia page) and R is the resistance of R2. The voltage across R2 is roughly the base voltage (which is set with a zener diode, or an LED, or some diodes in series) minus the transistor's base-emitter voltage, which is about 0.8~1.0V for a typical NPN power transistor running at significant current.

In this case you might be best to use a Darlington transistor, which requires much less base current, but has a higher base-emitter voltage. Suitable devices would be:
BD679 (4A, 40W, TO-126): http://www.digikey.com/product-detail/en/BD679AS/BD679AS-ND/976566
BDX53C (8A, 60W, TO-220): http://www.digikey.com/product-detail/en/BDX53C/BDX53C-ND/965855
TIP142G (10A, 125W, TO-247): http://www.digikey.com/product-detail/en/TIP142/497-2541-5-ND/603566
2N6284 (20A, 160W, TO-3): http://www.digikey.com/product-detail/en/2N6284/497-2563-5-ND/603588
There are a few bigger ones as well.
 

undescriptive

Apr 14, 2013
11
Joined
Apr 14, 2013
Messages
11
I have another thought/musing....

If I were to attach a capacitor across the circuit at the front (a smoothing capacitor if you will) of sufficient size to cope with any kind of power drop (protected with a diode to prevent it being sucked out to the 1F cap) to be able to run the circuit for lets say 10 seconds (shouldn't be too hard as the circuit is pretty low current draw

would this be possible and allow the cap to charge at it's maximum rate - providing that the current source is massive? or am I totally barking up the wrong tree?

Thanks!
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
I'm not sure what you're suggesting. You can (and should) have a smoothing capacitor across the 555 oscillator part of the circuit, to make sure it's not affected by power supply variations caused by the heavy charge current for the 1F capacitor (assuming the control circuitry is powered from the same supply). But there's no point adding a smoothing capacitor to help with the current delivered to the 1F capacitor. You would need a smoothing capacitor much bigger than 1F to get any useful effect, and then that smoothing capacitor would need a long time to charge up.

If you only want the circuit to run for a short time, and only periodically, then a smoothing capacitor might help. In that case I would use a sealed lead-acid battery, like a car battery but probably a bit smaller, and trickle-charge it. This is a good way of getting extremely high currents for short periods of time (think of the starter motor - these draw hundreds of amps). But the current regulator, and the relay, would need to be huge!

I think you might be best to build up something that's not too extreme, and play around with it to see if it will do what you want. You don't seem to know how much power you need to transfer to your load. If you can't calculate it, and you don't want to explain what the project is for so someone else can estimate it, trial and error is your next best bet I think.
 

undescriptive

Apr 14, 2013
11
Joined
Apr 14, 2013
Messages
11
Hi,

I wasn't very clear...

yes adding a "large" smoothing capacitor for the 555 (or other) oscillator to protect against large drops is what I meant.

I am not sure how big the load will be, which is why I'm being extra cautious in saying what the final output will be!

ok... lets get hypothetical for a minute :)

say the load will not exceed 10A and normally the load would be say... 5A
The current source can supply much more than the 10A maximum

would those values help at all? (they are guestimations, but I think they would be typical of the final output of the project)

I totally agree about not building a monster one at first as I have no idea if this will even be successful!
As you may have noticed, it's all a bit of a learning experience for me...
 
Top