Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations GregLocock on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

power supply questions

Status
Not open for further replies.

pshah72

Electrical
Jun 25, 2013
1
Hi All,
I just joined this forum.
I am planning to work on power supply design very soon. I wanted to make sure my ideas going to work or not. :)

One of our products running on both AC (85-265VAC) and DC (20-28VDC) power supply. We have those as separate supply circuit at this point.

I am planning to design common power supply circuit which can take both AC and DC power supply. Is it possible? If yes, I am looking for help on the same.
The output of power supply should be 24V and 5V in both situations.

I will come up with all other parameters like load, current requirements etc later on.
Before I move ahead I want to know is it possible to design such a circuit that can sense AC or DC and behaves according that?

Or I can explain same thing in little different way: - "How to distinguishes/sense AC(85-265) voltage and DC(20-28) voltage signal?" Once I know the method, I can use it before Transformer stage and after that transformer behaves according to incoming voltage that will sync up with secondary side of transformer.

Thank you for looking in it!!!!!!
pshah72

 
Replies continue below

Recommended for you

I don't think sensing is the issue, per se. As I look through the Vicor or Ametek websites, I'm wondering why there are no product offerings as you describe. I think the biggest issue is that designing a such a front end requires either:
> DC input components capable of handling AC, which means bigger components and more of them
> AC sensing that switches to DC if no AC present; still need something that can convert both AC and DC, but possibly at lower power

If the customer's intent is to eventually use only one input or the other, then the supply as described incurs a weight, power, and reliability penalty for carrying all those extra components that aren't actually going to be used.

TTFN
faq731-376
7ofakss

Need help writing a question or understanding a reply? forum1529
 
The power cords are going to be different anyway, so the advantages of using a single input seems minimal.

For the AC source, you also have to think about the myriad of standards and approvals that would be required. Many power supply bricks are simply covered with logos of all the approval agencies.

That being said, it would seem that a universal bulk power input (rectifier, filter) followed by a *very wide range* (20 to 375 volts) DC-DC converter stage is *basically* all that would be required (ignoring all the rules and regulations for the moment).

Detecting and switching is probably fraught with unexpected failure modes and dangers.

 
Is it possible? Yes, it's always possible. However is it practical or cost-effective? Probably not.

One thing you have not indicated is the power level involved. If this is a real product and not a one-off, then with the AC input you're going to need UL or CE approvals for most countries. These regulatory requirements may be in conflict with your additional DC input requirement.
 
A company I used to work for came up with a novel approach for some small modules with that input range. They used a constant current LED driver chip feeding a zener. It wastes a little power and is not line isolated but 24-280V AC DC is a neat trick.
 
Given the disparity between the voltage ranges, I'd accept the complications of a detect and switch approach. A triple-throw switch would default to the detect circuit, which would be a low current resistive divider with a diode off the appropriate voltage tap. The diode would feed an RC circuit tuned to give you an analog indication of the input voltage. >85Vac would produce well above some voltage threshold while 28Vdc would produce well below. A comparator would then convert that to a binary signal. Once a power-up timer circuit expires, the binary signal would be latched and sent to the triple-pole switch, which would then switch to the appropriate power supply input.

Then you only need to spent the other 90% of the design dealing with edge-case scenarios caused by power supply fluctuations!
 
pshah, Not knowing what the restrictions are for your product but If your AC circuit is a typical bridge followed by a PFC boost, you could always run this circuit with 28VDC input right into the bridge. The PFC will act as though the "sine wave" input is stuck at 3 or 4 degrees but it will work nonetheless. It will just have a very inefficiently high duty cycle and the weight and size would end up being very non-ideal for either AC or DC operation. If the weight of the product isn't a big concern, this may be the most inexpensive approach though. Again, not knowing what the restrictions are, what market you're in etc, it's hard to advise on this.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor