It's been some time since theory days. What I am wondering is if I have say 28V at nearly .5 Watts and I use a regulator to drop the voltage down to say 6V. What would be the result of power I have available. I realize I need a dropping resistor and I'm sure I can calculate the power lost accross the resistor, but what can I safely get away with. Is there a formula I can use? Thanks
Posted on 2004-09-10 14:56:44 by mrgone
The maximum you'll get is 0.107 W (if .5 means 0.5W). You can't get more current than supplied unless you use (another) transformer or some trick with capacitors.
Posted on 2004-09-10 15:39:11 by Ultrano
So essentially it's the same result as a voltage divider only with tight voltage lock. Well that's no good. I need to maintain the wattage but indeed not looking to use Xformer. Hmmm what to do now? 6 X .5A = I need 3 watts.
Posted on 2004-09-10 15:48:24 by mrgone
What about devising a circuit that uses the trick, described in the attached image? The link-up will have to be triggerred at a specific frequency that depends on the power necessary, capacitors value and output voltage. With relays it'll be fairly easy to make, but who uses relays these days :) not to mention their power consumption... So there'll have to be made up some way to use bipolar transistors in switching state. Haven't seen such a circuit, maybe it's not possible with transistors :/ .
Posted on 2004-09-10 16:12:36 by Ultrano
I'm already doing some switching with opto-isolator. I getting about 28V at atleast .5Watts. I'm timing the output pin 11 off the AC cycles using port change interrupt as the AC input. Are you saying I need to take it to a higher freq.?
Posted on 2004-09-10 17:09:53 by mrgone
The AC is 110/220V , right? The bitmap has been resized and most text cannot be read - can't get the values of elements. Can't you change some of the elements so that the output is 6V, and meanwhile use powerful transistor and optocouple to get more current?
Posted on 2004-09-11 07:37:02 by Ultrano
Yeah it's 110 AC here in US where are you? But using half wave rectifier so really 55 or 60V then smooth it, then run it into PNP Xsistor and switch it with opto-isolator. At the output (far right) is actually in this circuit, tied to a relay coil who's power consumption is 430 mW almost a half watt at a voltage of approx. 33V. The problem is if I use the switching mechanism to lower the voltage any further than I will lose power also. What you don't see in schematic is the first power supply that powers the processor. It is a bridge rectifier capacitively coupled through a 2.2uf chip non-polarized cap. (no Xformer). So what I did was couple the full wave pulsating DC signal to the PIC's port change interrupt so I would know when the power is passing threshold in a positive direction. I suppose I have some room to play with the timing here but I need atleast 3 watts at 6V for what I want to do.
Posted on 2004-09-11 10:03:43 by mrgone
It's been some time since theory days. What I am wondering is if I have say 28V at nearly .5 Watts and I use a regulator to drop the voltage down to say 6V. What would be the result of power I have available.


Power in is the max power out. So if you have .5W in, the best you can get out is .5W

You can't quite get that out, as there is always an efficiency factor. And to get this much out you would need some sort of a switching regulator; that isn't as scary as it once was, as custom IC based designs abound. Maxim and Nat Semi would be my first choice for a look-see.

If you go for a linear regulator (and a resistor counts as such), your max current out is just 0.5W/28V = 17.8mA. At the 6V level you ask, this is just about 100 milliwatts. Yikes!


(Capacitor based switchers like Utrano speculated on exist, but mostly for UP converters. The cap is charged at the input, then switched to charge a cap from the input plus the cap voltage, thus it is a voltage doubler. Not too terrible efficient but excellent for low power apps

(Converely, the DOWN converter design has some problems right off the start, as the caps will charge unevenly due to mis-matched capacitance (tollerance effects). Then when placed in parallel, the highest voltage caps will then work to charge the lowest voltage charged ones. Lots of trouble when a single switched inductor will bypass all these effects)
Posted on 2004-09-11 14:42:09 by Ernie
IC "pun" the current remains the same. Well that would mean I would drop 400 mW across a resistor and the regulator. What a horrible waste of power. I haven't done much with switching power supplies. I ve done some and come to the conclusion that buck-boost converters are the smallest circuits. It seems you always get strapped with some sort of inductor or small Xformer. I'm trying to avoid that although I am quite familiar with Amidon's product line. Guess I was looking for a quick fix so I could get started on the circuit it's self I wish to power. Always something :roll:
Posted on 2004-09-11 16:16:21 by mrgone
I would use one of National's Simple Switchers, such as LM2574. You just need an inductor (which can be off-the-shelf, not really critical, datasheet gives you selection curves, no need to be scared of it), a couple of capacitors, and possibly a couple of resistors to set the output voltage. These parts do not need anything else to run properly other than careful layout, but the datasheet gives you enough info on how to do it.

With a switching regulator like that, you can get about 77% efficiency at full load (0.5A out), so at the input you need about 4W.

You can also use the 5V version of the part, just add one resistor in series with the sense input to increase the voltage to 6V (nice thing about National: they give you the values of the internal resistors used in the divider, so you can easily calculate the resistor you need to add).
Posted on 2004-09-13 11:29:58 by VVV
I'm gonna have a look at that one :alright:
Posted on 2004-09-13 11:37:26 by mrgone
Multi-package optoisolators. I had forgot about them but just found some quads. Two would handle the 8 bit buss. Need to check smd size. By the way what happened to our shematics?
Posted on 2004-09-21 22:03:42 by mrgone