Help about adjustable power supply

[preebi4]

Hello,
I want to make adjustable power supply to 24 volts 30 amps. If there is a way to with current adjustment will be great.

I've try to make this scheme 1.5 to 37 Volt 30 Amp Power Supply
http://www.zen22142.zen.co.uk/Circuits/Power/3730.htm
All works fine on idle with no load. When i try to load ( put 2 light bulbs 12 v) - 3 amps, power transistors start to heat a lot although i put much bigger heatsink from described one ".. used the heatsinks from the computer power supply for the power transistors.. ". Do you know what might be the problem ?
Is there a way to equip this scheme with current limit adjustment ?


Thank you

[KhronX]

What value is your input voltage? And what's the output voltage set to?

[preebi4]

Input voltage is 25 V DC and output was set to 12 V DC. I try the scheme with 2 car lightbulbs. For 4-5 min the whole heatsink was hot. Is it possible to be from some fake transistors ?

[DaJMasta]

What's the resistance of your light bulbs?


It actually sounds pretty reasonable to me.  The supply you've made is a linear one, so across the transistors you should expect to see about the voltage drop times the current delivered watts of heat, so 13 times... maybe a few amps if those lights are headlights and are in parallel.


A lower input voltage, something closer to your supply, would mean less heat dissipation.  Less current delivered would have a similar effect.

[preebi4]

i am comparatively new at this so sorry if i sound stupid.  Didnt check resistance on bulbs ( dont have them right now will check later) but ammeter on exit showed 3 amps when loaded. so 25 V - 12 V is 13 V drop.  13*3 = 39 watts, divided to 6 transistors.... they are 90 watts each

You think if i try to load more will do better ?

[not1xor1]

Hello,
I want to make adjustable power supply to 24 volts 30 amps. If there is a way to with current adjustment will be great.

I've try to make this scheme 1.5 to 37 Volt 30 Amp Power Supply
http://www.zen22142.zen.co.uk/Circuits/Power/3730.htm
All works fine on idle with no load. When i try to load ( put 2 light bulbs 12 v) - 3 amps, power transistors start to heat a lot although i put much bigger heatsink from described one ".. used the heatsinks from the computer power supply for the power transistors.. ". Do you know what might be the problem ?
Is there a way to equip this scheme with current limit adjustment ?


Thank you

  30A is a huge lot of current. What do you need that for?

In any case the article you linked is pure bullshit. 
1) to get 37V-30A DC you would need a 2000W transformer, as transformer AC current should be at least 1.5 times or more than the maximum output DC one and a 2000W transformer is huge and heavy
2) you would never be able to get up to 37V because the input voltage would easily exceed the absolute maximum ratings of the LM317 as input voltage would vary greatly depending on AC line fluctuations and on load
3) even if you had a constant reliable 40V DC input, each 2955 transistor would have to waste as heat about 180W (worst case 1.25V-30A output)
I seriously doubt that the author ever tested all his PSU claimed features.

Now about your circuit, where are you getting the DC input voltage from? A transformer? An AC-DC adapter? A 24V battery pack?
Please give more informations.

In any case, supposing a 24VDC input voltage, 12V 3A output means that both your load and the power devices dissipates 36W as heat.

But if you want to get 24VDC as maximum output you need at least 27VDC or more as input because of the LM317+BJTs dropout voltage at full load. And if you intend to use a transformer you have to take into account AC line fluctuations and output voltage vs. load so you would need an even higher input voltage.

At maximum output current and lowest output voltage you would need to dissipate about 8-900W as heat.
No way you can do that with just 6 TIP2955.

You would need more powerful devices and at least 10-12 of them and a 100W heatsink for each (e.g. CPU heatpiped heatsink with fan).

[preebi4]

I guess i wont need 30 A, just thought i can use this scheme. Suppose 20 A will be enough - i need this to power up tools with damaged batteries. I have few with different work voltage - 12, 14.6, 18, and 24 volts. But most of it i make this for hobby and for fun.

So i guess i used wrong scheme again. Can you suggest scheme that will work for what i want ? I want to have current adjustment as well so can use it for charger.

I have transformer wich gives 20 V AC and 28.5 V DC after bridge.

[not1xor1]

I guess i wont need 30 A, just thought i can use this scheme. Suppose 20 A will be enough - i need this to power up tools with damaged batteries. I have few with different work voltage - 12, 14.6, 18, and 24 volts. But most of it i make this for hobby and for fun.

So i guess i used wrong scheme again. Can you suggest scheme that will work for what i want ? I want to have current adjustment as well so can use it for charger.

I have transformer wich gives 20 V AC and 28.5 V DC after bridge.

Most DC motors need high current only on start-up or heavy load.
I think the most appropriate choice would be an AC-DC switching module + a current/voltage adjustable buck regulator module.
That would easily allow you to get an overall efficiency greater than 80% while the higher noise (vs. linear PSU) would be irrelevant for driving DC motors or charging batteries.
There are plenty of such modules for sale on ebay/amazon/aliexpress/bangood/etc.

Since I do not have any of them I can't recommend any particular one.
Just remember that most of those chinese modules are grossly overrated.

In any case you can search for more informations here in the forum. I think there are various threads about those modules.
Some of them have a nice panel and a graphic display. I remember Dave tested and repaired one of them.

[T3sl4co1l]

Get an off the shelf switching supply.  They're cheap, efficient and capable.

Not usually adjustable so you may need a few to replace different batteries.  But hey, the lifetime of that linear supply will be measured in milliseconds if you actually try to draw 30A from it.

Tim

[T3sl4co1l]

Most DC motors need high current only on start-up or heavy load.

I recently put an ammeter on a cheapshit (Horrible Freight special) cordless drill I have.  Powered by an adjustable high current switching supply, it's capable of hitting the 40A current limit for short overloads!

I'm impressed that they used a motor with armature resistance that low, and that the speed controller can even deliver that much current without bursting into flames.

I've been running it from my bench supply (a beefy 20V 10A linear), no wonder it can bog it down.

Tim

[DaJMasta]

i am comparatively new at this so sorry if i sound stupid.  Didnt check resistance on bulbs ( dont have them right now will check later) but ammeter on exit showed 3 amps when loaded. so 25 V - 12 V is 13 V drop.  13*3 = 39 watts, divided to 6 transistors.... they are 90 watts each

You think if i try to load more will do better ?

To answer this specifically, while the transistors may be rated to far more, they will only physically stand up to that if adequately cooled.  In this instance, the issue isn't that the transistors aren't up to the job, it's that under the stated operating conditions, they are generating way more heat than can be dissipated down to a cool temperature by the heatsink.  With a bigger heatsink or maybe some airflow forced over it, they will still operate fine, but the limitation of linear supply designs is this inefficiency.  Many commercial/lab adjustable supplies use several transformer taps at different voltages, so if, for example, you only wanted 3.3V out, you wouldn't have 20.7V of voltage drop and absolutely massive heat dissipation, you could use a 6-7V transformer tap and only dissipate the heat from a 2.7-3.7V drop.  There are also switching power supplies which can be MUCH more efficient, sometimes in the realm of only 5% energy as waste heat (whereas your original post has about 50% of the energy used as waste heat), but involve a more complex design and have more inherent noise on the output, though there are switching supplies which can be very clean.

At least in this example, noise from a switching supply, even a cheap one, would be entirely unnoticeable.

[preebi4]

I am not sure i can make switching power supply but will look for some schematics. I choose linear because the are much simpler i have powerful transformer.   Saw this scheme i think to try it because already have most of the components.

[prasimix]

If your initial requirement is not changed (variable 24 V, 30 A) then this circuit cannot deliver that just as first circuit that you presented. Few people already informed you why is that and what is a reasonable path to follow (a SMPS topology) but since information is not a knowledge now you can spend some time and money to build this circuit to gain needed knowledge that you can compare with information that you already got .

EDIT: for linear 30 A you don't need just a powerful transformer: you need powerfull output stage (bjt or mosfet), powerful pre-regulation that comes before output stage. Pre-regulation could be before main trafo primary, after main trafo secondary as SMPS or perhaps mag-amp, etc.

[preebi4]

I am sure that switching supply is the better solution as people said, just dont know am I capable of making it.  Can you offer some scheme that is not so complicated and can be build from someone like me with not much knowledge and practice ?
Thank you

[DaJMasta]

There are switching supplies in a lot of devices - while making an efficient-across-many-ranges sort of jack of all trades power supply would be quite challenging, salvaging a couple of units with different voltage outputs could give you a range of voltages with fairly configurable outputs.


Take a computer power supply, for example.  You'd get a lot of power on 12V and 5V rails, so you'd be able to supply a 12V line, a 5V line, or with the addition of the power supply you're already using (at least the regulator and transistor part), you can get something like 10.5 to 5V off the 12V rail (and a maximum voltage drop of 7V, so much less heat generation), and you can get like 3.5V and down from the 5V line using that same regulator setup.  For higher voltage options, you'd need a higher voltage input, but as you scale up, expect to have difficulties.  12V at 30A is 360W of power, no small amount to deliver, and at double the voltage, that would be double the output power before you factor in any losses from regulation or other things.  Even if you do it right, it's going to generate a bit of heat.

Now you can probably see why variable lab power supplies capable of high currents and voltages tend to be expensive   They are tricky beasts to get right, and getting reasonable efficiency to minimize heat generation and noise is actually a pretty tall order.

[T3sl4co1l]

Buy an off the shelf SMPS.

They are usually adjustable, over a small range of voltage.  Buy a few to cover your needs.  One is cheaper than building a linear supply, and even if you need half a dozen separate supplies to cover your needs with this strategy -- you're still coming in way less than the cost of learning how to make them in the first place!

As for learning, if that is a desired goal -- take your time with that, using smaller examples.  These things are easily scaled up, so you don't need to jump right into a ~1kW size project to get direct experience and understanding with it.  It will take years to become fully fluent in their design -- take your time.

Tim

[John Heath]

You can always weeny out and buy off the shelf. Check out this sweety. A nice addition to your bench and all the grunt work is done so you can return to designing interesting circuits not no brainier power supplies. Give yourself a treat.

[preebi4]

Thank you for your help and advices. Will see what mess will create