Solve Power Supply Questions: Electrical Engineering Forum Help!

[Shadrack]

Wow! An Electrical Engineering Forum! :!)

Please help!

I'm a newbie in the field, just got my degree last May (2005, BSEE). I focussed on computer engineering and can program the crap out of a micro-controller. Now I'm working on some automated test systems and have to work on a lot of other people's poorly documented work. :yuck:

My electronics know how is very rusty, and my power management know how is non-existent. Ok, enough with my intro...onto a few questions:

1.

What do I do with multiple power supplies in a system? The system that I am currently having a lot of problems with atm has 3 power supplies throughout the entire system . They are all DC: 1x 5V, 2x 12V. What should I do with ground? Should I just connect all the grounds together? I measure about a 1mV difference between the different ground planes... connecting them together causes a small amount of current, should I be concerned? What is the best way to inner-connect multiple power-supplies throughout a system? The components are not optically isolated from each other.

For instance, the output of one module is a 5V DC signal, and the module is being powered by the 5V power supply. The signal goes into a series of Inverter units that is powered b a 12V power supply to bounce the signal around between 0V and 12V. ? this seems very problematic to me unless the inverter units are optically isolated which they don't appear to be.

Wouldn't it be better to use 1 DC power supply at the highest voltage rating that is needed in the system and could deliver enough current to everything, then use DC-DC converted and voltage regulators to take care of the small stuff? How do you manage your power?

Any good references on system power management? I'm just using basic P=I*V and keeping track of my currents going between modules.

2.

Is there anything wrong with using a voltage divider to bring 24V down to ~5V to feed a 3.3V regulator? What "level" of resistors should i use? High resistors will lower power usage, so it seems to me I should use those.

3.

Why is there a small amount of resistance between the + and - terminals of my power supplies. Only a couple hundred of ohm (not connected to anything else). Intuition tells me there should be a HIGH amount of resistance between the terminals. Then again "low output impedance, high input impedance" is ringing in my head. But for a power supply?? Wouldn't there be a lot of internal current flowing between those two potentials with such a low resistance between the two? What is going on here?

4.

O.T.: can anyone recommend a good IDC connector crimp tool?

Thank you for your input!

BTW: love all the emoticons on this forum...very appropriate.

-Shadrack

 

[pallidin]

Just some thoughts:
A common ground plane for the power supplies is acceptable if, and only if, ground potential is of no concern.
That is, a common ground from three separate power supplies can certainly alter the ground potential for a specific power supply.
I would isolate.

 

[NoTime]

1. Not enough info.

2. Generally it would be a problem to use divider resistors to the input of a regulator. They don't like high impedance inputs and can go into oscillation.
Either heftier cooling on the existing regulator (If V max in > 24) or setup a two stage regulator, the first stage dropping to the 5v.
Depending on the load it might be possible to use a divider with some filtering.

3. Shutdown discharge safety feature?

 

[Shadrack]

DC-DC converters

I just got some 5V regulators in today that will take 24V, so I will just step it down that way. The 3.3V only will only regulate with a max input of 6V...after that you start getting screwy outputs.

My plan is to find 1 power supply that will adequately power the whole system and then use DC-DC converters to knock the voltage down for the various components.

What is the difference between a DC-DC converter, and a power transformer? Is there any? DC-DC converter are expensive, and I'm open for a cheaper solution.

 

[berkeman]

What is the difference between a DC-DC converter, and a power transformer? Is there any? DC-DC converter are expensive, and I'm open for a cheaper solution.

Welcome to PF, Shadrack. It's a very helpful place. A power transformer is just that, a transformer. it will transform AC voltages up or down, based on the turns ratio. There are issues with input line voltage variation and series resistance and leakage inductance that affect the output AC voltage that you get. But basically you use power transformers to change one AC voltage (like to 120Vrms) to another (like 24Vrms). You would usually follow a power transformer with a full-wave rectifying bridge, a large DC storage capacitor, and then linear regulators to make your final output voltages. Keep in mind that there are UL (Underwriters Laboratories) regulations that govern how you connect and deal with AC Mains voltages. You need to follow those design guidelines/regulations on the hot side of your power transformer, and get the final design UL approved if it will be use out-of-house. If you are just designing and building in-house test fixtures that work with AC Mains, you should still design to UL guidelines/regulations, but you won't generally submit to UL.

A DC-DC converter is more efficient than the linear regulator topology (the transformer + linear regulator design described above), but it is more complex and noisier. If you buy one off-the-shelf, then it will meet the published specs for Vin and Vout. Keep in mind that you will get more ripple on the output voltages compared to the linear approach, and you will generally get some external magnetic field noise from the switching magnetics inside the DC-DC "brick" that you can buy.

You can design and build your own DC-DC converter circuits, and there are some fairly easy to design versions (like National Semiconductor's "Simple Switcher" series). But you still have to think about and deal with output ripple issues and external magnetic field interference issues.

Plus, it sounds like you have multiple output voltage issues in your ATM system. You can get bitten by latchup and ground loop issues if you don't do things right. Probably it's easiest (especially on my fingers) if we just talk on the phone. PM me and I'll give you my phone contacts if you want to talk about cleaning up the architecture of the ATM system whose design you inhereted.

-Mike-

 

[Shadrack]

Thanks for the offer berkeman!

I actually have an electronics goto guy available, I have just been hitting him with a lot of questions lately and wanted to expand my question and answer base to a forum like I have done in the past with programming and digital design. Also, everyone can bennefit from answers in a forum... web crawlers can index these answers and future searchers of the same problem have a better change of arriving at a solution.

The noise that is generated from a DC-DC converter has made it clear to me that I do not want one of these next to any instrumentation. Fortuently, this is all a relay driving circuit and my instrumentation signals are in separate housing.

-Shad