Power Supply

Hi,

Well there are loads of current limiting adjustable regulator IC's out there, but using one of them is useless if you want to actually learn something. With that in mind, I set out making my own power supply using the (Not so) trusty simulator Livewire.

So here it is. The values are not worked out or anything, and it's not been simulated properly because Livewire seems to like throwing a fit whenever possible.

Features:
*Adjustable from about 0.7V right upto around 26V
*Max current is around 4A according to Livewire, before the voltage drops too much. Oh course, Livewire could be totally wrong.
*Totally incorrect input fuse, because I don't know the max output !
*Adjustable (At build time) "V Droop" via R18. I.e. The output voltage is slightly raised when more current is is drawn, to counter the "droop" when a high load is applied.
*Built in Ammeter via the connection "Current Adjust", 1V = 1A.
*Separate "V Sense" conenction for connecting to the output terminal, to counter the voltage drop due to internal wiring and the current-sense resistors.
*Built in current limiter which (Should) latch the output low and light an LED when the current limit is exceded. Can be reset using the push button "Reset Current Limit".
*Posible to add a current limit meter by adding a Voltmeter to measure the output of VR1 - "Current"", 1V = 1A.
*Probably some other things that I've forgotten.

Anyway, I've attached a low quality picture of it. (Best I could get due to the dimensions.)
I've also attached a .zip file which contains a high res .xps.

.xps (XML Paper Specification) is some kind of Microsoft image format, kind of like a PDF. As far as I know it comes with Windows Vista so anyone with Vista should be able to see that. Link about .xps.


So yea, basically I'd like suggestions, improvements and comments about it. If anyone has a lot of free time and is willing to test it, please let me know how it goes !

Thanks,
-Andrew.

I have some experience with designing and building power supplys. I have found if the voltage sense circuitry senses the output terminals that overcomes any internal voltage drop caused by internal components. I have used 3 terminal regulators for current limiting but find they have some limitations. One being the internal circuitry monitors the in/out differential and the power dissipation and it can shut down the regulator prematurely.
Most current sense circuitry monitors the voltage drop across a resistor, this voltgage drop can go to the input of an OP amp or in many cases is the base emitter voltage of a transistor. Making the sense resistor variable makes the adjustable current sense.
In a battery charger that I designed, the current sense is actually a bridge circuit. Half the bridge monitors the charge voltage, and is on pick off point for one of the current sense points, the other half of the bridge is the other pick of point.

If you want to learn something then go for an LM723 - it's a regulator IC alright but it provides more scope for learning than the LM317.

k7elp60 said:

I have some experience with designing and building power supplys. I have found if the voltage sense circuitry senses the output terminals that overcomes any internal voltage drop caused by internal components. I have used 3 terminal regulators for current limiting but find they have some limitations. One being the internal circuitry monitors the in/out differential and the power dissipation and it can shut down the regulator prematurely.
Most current sense circuitry monitors the voltage drop across a resistor, this voltgage drop can go to the input of an OP amp or in many cases is the base emitter voltage of a transistor. Making the sense resistor variable makes the adjustable current sense.
In a battery charger that I designed, the current sense is actually a bridge circuit. Half the bridge monitors the charge voltage, and is on pick off point for one of the current sense points, the other half of the bridge is the other pick of point.

Click to expand...

Can you explain how you'd use a bridge to measure current ? I can't think how you'd do it.

Hero999 said:

If you want to learn something then go for an LM723 - it's a regulator IC alright but it provides more scope for learning than the LM317.

Click to expand...

I will look into using the LM723 when my switched mode book arrives, so I can design a switching PSU using that IC. In addition to that, I can get the LM723 from my local Maplin which will be great when I want to build a prototype.

However right now the idea was to just use analogue electronics and no fancy IC's - So I can learn analogue electronics better.


So any error's I've made with my design ? What else could be added and/or improved ?

Thanks.

yngndrw said:

I will look into using the LM723 when my switched mode book arrives, so I can design a switching PSU using that IC. In addition to that, I can get the LM723 from my local Maplin which will be great when I want to build a prototype.

However right now the idea was to just use analogue electronics and no fancy IC's - So I can learn analogue electronics better.

Click to expand...

The 723 is an analogue IC, a VERY, VERY old analogue IC for building linear stabilised PSU's - but it's never really been bettered?, and is still a current chip.

You can't use it to build a switching PSU.

You can use it to build a switching regulator, have you checked the datasheet?

I wouldn't recommend it if you want good efficiency; I'd suggest you use a real switch mode IC.

Hero999 said:

You can use it to build a switching regulator, have you checked the datasheet?

Click to expand...

Haven't checked it for 30 years or so?

OK then, a 'really bad choice for a switchmode PSU?'

The datasheet is where I got the idea from. Well I'll use some IC to make a switched mode PSU when I get my book, but untill then comments about my analogue design ? :c

EEK! A very fuzzy JPG file type schematic. I can't read the parts numbers. If it was saved as a GIF or PNG file type then it would be very clear.

Your opamps have only a 15V supply. So their max output voltage is about 13V and then the max output voltage from your circuit will be only about 10V.

Your output transistors won't share the current without each one having an emitter resistor. The one with the highest gain will hog the load current then burn out. Then the other one will burn out.

When I get back home I'll take the .xps and make a very high res version of it. Sorry about the current .jpg, but it's the only way I could export it currently.

The opamps have a 30V supply, via two cascaded 78L15's - Or atleast that's what I intended. :x

What value should I use for the emitter resistors ? I'll add them also when I get home.

Thanks.

yngndrw said:

The datasheet is where I got the idea from. Well I'll use some IC to make a switched mode PSU when I get my book, but untill then comments about my analogue design ? :c

Click to expand...

I think you're missing the point, the LM723 is better suited to an analogue linear regulator than a switching regulator so you're better off using it as a linear regulator.

I see two very fuzzy cascaded 78L15 regulators in your schematic.
The schematic wouldn't be the size of a city block if it is cropped properly.
The JPG is too fuzzy for me to crop it.

Hero999 said:

I think you're missing the point, the LM723 is better suited to an analogue linear regulator than a switching regulator so you're better off using it as a linear regulator.

Click to expand...

I ment that I would use a different IC than the LM723, I'd use one soley designed for switching power supplies.

audioguru said:

I see two very fuzzy cascaded 78L15 regulators in your schematic.
The schematic wouldn't be the size of a city block if it is cropped properly.
The JPG is too fuzzy for me to crop it.

Click to expand...

Yea that's how Livewire exports it. Had to print it to a document then screenshot it, and due to it's width it wouldn't let me take a decent screenshot in one piece. Tomorrow I'll use photoshop and make a new image out of small sections of the .xps, each zoomed in. The .xps file is a vector image and can be zoomed in upto 6000x, but will only work on Vista.

I cropped part of your fuzzy schematic and made it big enough to try to see it. I tried making a few words bigger but they are too fuzzy.

The one you tried to enlarge said +V (Un Regulated) and was just to pass the voltage down to the control circuit below.

yngndrw said:

Can you explain how you'd use a bridge to measure current ? I can't think how you'd do it.
Thanks.

Click to expand...

12V LEAD ACID BATTERY CHARGER
In the schematic the junction of R2 and R4 is the pick off point for two things, 1 the output voltage sense the goes to pin 2 of the OP amp and 2, the pick of point for the current sense that goes to pin 5 of the OP
amp. R4 adjusts the output voltage of the charger. The junction of R1 and R3 is the other pick off point or the current sense. This point goes to pin 6 of the OP amp thru resistor R11. R3 adjusts the point at which IC2B’s output goes high and turns on LED2. In this circuit when the charge current drops to about 75 mA, LED2 illuminates. So the voltage drop accross R8 is sensed by the bridge.

Ah I see, thanks you for the diagram and explination. Can you explain to me what the advantage of this is over a current sense resistor and a standard differential amp is ? Also why does adjusting R4, the voltage adjustment, not affect the current sense ?

It seems I have an aweful lot to learn. Thanks again.

PS: Still not back home yet, later tonight I'll get that high-res image done though.

Okay then I've made a nice big image for you. .jpg set to full detail. It's 5453x768, so it should fit on your screen full height, then you just have to scroll sideways to see the rest of it. It's very detailed so I'm sure you'll be happy.

Clicky for image
(628KB, 5453x768 .jpg)

I've added:
* Emitor resistors to the power transistors.
* Vdroop adjustment, VR3.
* Voltmeter to display the current limit, "Current Limit", 1V = 1A.

Enjoy.

It is the biggest schematic that uses the most amount of memory that I have ever seen!

I don't like your cascaded regulators. Opamps don't need a regulated supply. Use an amplified voltage reference or an amplified low temperature co-efficient zener diode instead. (about 5.6V).

audioguru said:

It is the biggest schematic that uses the most amount of memory that I have ever seen!

Click to expand...

But hey, it's detailed. Took a while to piece that together from screenshots of the .xps.

audioguru said:

I don't like your cascaded regulators. Opamps don't need a regulated supply. Use an amplified voltage reference or an amplified low temperature co-efficient zener diode instead. (about 5.6V).

Click to expand...

So you're suggesting that I run the references off a lower voltage, then adjust the opamps so that the lower input is still correct and run the opamps off the unregulated supply ? What if it peeks too high ? I mean the supply will be around 35V - Isn't this too high ?

I also worry that the lower voltage for the references would make them less accurate - Is this true ?

Thanks.