PDA

View Full Version : DC to DC board, to supply propeller projects.



rwgast_logicdesign
03-25-2012, 11:24 PM
Ok so I have taken everyone's advice and and put my main PCB on the back burner for a little bit. Im working on one piece at a time. Last night a built a power regulator prototype on a breadboard. I am using the lm2940ct 5.0v and the lm2937-3.3v for my regulators, they are the two regulators sold in the parallax store. The data sheets for the 5v reg calls for a .47cap parallel with the input pin and a 22uf cap parallel with the output pin. The 3.3v regulator calls for a .1uf cap and a 10uf cap respectively. I had to improvise on the 5 volt regulators input and use two .22uf metal film caps in parallel to make a .44uf since I didnt have any .47 caps. I am using a ceramic .1uf on the 3.3v input and the 22uf and 10uf output caps are both aluminium. I wired the regulators in series with the 5v supplying the 3v. If anyone feels they need a schematic I can make one and post a pic of the breadboard.

Now Im having a problem, well im not sure if its a problem or normal. When my circuit is complete and I use a meter to test it the 3.3v reg output 3.27v and jumps to 3.28v every minute or so for a split second. The 5v regulator is 4.96 to 4.97 except this one fluctuates alot, the 3.3v also fluctuated like this when it was the only regulator on the board. So Im worried that my voltage are up to .04 off and that they fluctuate, I was under the assumption when the circuit was set up with the right caps I would get a totally steady supply when there was nothing on the regulators load. Is it regular for the outputs to be a little low?

Leon
03-25-2012, 11:30 PM
Metal film capacitors aren't suitable for use with regulators. Most modern devices are designed to use ceramics.

rwgast_logicdesign
03-25-2012, 11:36 PM
well like I said when I have the same problem when i use the 3.3v reg by itself, and it has a .1uf ceramic and 10uf electro. So Whatever is causing these these fluctions and lowish voltage isnt the metal film, or all to blame on the metal film becuase the same thing happens when i dont use them. Why are metal films not suitable. Im just trying to work with what I have and I dont have anything close to a .47uf nor can I find one on junk boards.

I would definately be using all ceramics if I had them, but at this point I dont and i see lots of people use the aluminum cans for the higher value caps on regulator circuits.

http://www.parallax.com/Store/Components/Capacitors/tabid/150/CategoryID/28/List/0/SortField/0/Level/a/ProductID/161/Default.aspx those are the two caps im using the ones above them that look the same say poly, im not sure if these are acually metal film, just know there not ceramic.

Leon
03-25-2012, 11:47 PM
Details of the capacitors are in the data sheets. Note the ESR requirements.

KP
03-25-2012, 11:47 PM
My guess is that you have a digital voltmeter and the voltage from the 3.3 V regulator is about 3.275V

rwgast_logicdesign
03-25-2012, 11:57 PM
yes I do have a cheap *** digital meter right now, Id wondered if that was part of my problem. So this means my circuit is working like it should? The regulator not putting out 3.3v isnt an issue? Can you reccomend what qualities to look for in a meter?

Cluso99
03-26-2012, 12:09 AM
What is your input to the voltage regulators (where is it coming from)?
The input and output capacitors you describe must be "at" the input and output pins on the regulators. Other capacitors are required at each IC.

However, the big issue is likely that your input power supply is not steady and you have no bulk capacitor to keep it steady. If you are supplying it with a rectified ac then it will need a large bulk capacitor like 1000uF electrolytic with a voltage at least 2x the input voltage. Not all power packs (wallwarts) are the same. Most are unregulated but rectified with 1000uF capacitors. Their voltage varies wildly above the rated voltage when the current is lower than the quoted value. This is why we often say a 6V supply is better than a 9V supply.

Again, you need to consult the internet to research power supplies. A good start is to read the App Notes on using regulator ICs from the IC manufacturers. Also look at power supply design. There is a big difference between designing electronics and just understanding the basics behind them. Google and the internet is your friend :)

Leon
03-26-2012, 12:20 AM
Many years ago I designed a power supply (5V and 12V) for a 5 1/4" disk drive, for a TRS-80 Model I. A friend of mine copied the design, but left off the capacitors on the outputs of the regulators, and then wondered why he wasn't getting any output. They were oscillating at several MHz! Soldering capacitors between the output pins and ground solved the problem.

Peter Jakacki
03-26-2012, 12:20 AM
First off the term "DC to DC" implies a conversion of DC back into DC and this is true of isolated power modules etc which convert DC to AC and back to DC. If you are simply regulating without converting then it's a DC converter or just a regulator.
All things digital will suffer from "quantization error" where an analog value does not quite fit into an exact digital value. As Leon mentioned it may be halfway in between a reading and the meter is simply put, doing the best it can.

I'm a bit surprised that you actually thought a 5V regulator would be "precisely" 5V and if you checked with an analog meter it would measure 5V. You can get devices with tighter tolerances but that's $$$ and the LM2937 datasheet which you have read of course :) says that the output at 5ma test load could be between 4.75V and 5.25V. Anyway by the time it's connected to a load there may be some voltage drop along the way etc. Use tantalums instead of electrolytics (in general) on the output as the former have lower ESR which most regulators require. Also, most cheap meters etc aren't too bad but aren't very precise , which requires precision which requires $$$ :)

rwgast_logicdesign
03-26-2012, 12:57 AM
Yes I realize that the either a ceramic or tant is better than the elctros on the output pins, At the moment I dont have any, i will be building something better later on, but right now i just need something that works and will provide stable voltage to plug into my prop board. Im not sure what the esr on the metal film caps are, i looked through the data **** did not see it, seems like only a few caps acually have a good data sheet with info like this. Anyways they seem to be operating fine, i could throw 5 .1uf ceramics in parrallel I guess but wont this add the esr for each cap?

Where would I want to add the bigger cap at? I am using a power supply from wal mart that lets me select usb or 1.5 - 12 volts. Im using it mostly at 7.5 volts that seems to be about the sweet spot for the 5v regulator 6 volts is a little two low. I have tested at 12 volts everything works fine, the regulator gets a bit warm depending on what its hooked up too, probably add a heatsink to both, which will help the esr problems too if I understand right?

Things would be easier if we still had a radio shack its a headache to find components or wait for them to ship, or have to pay shipping for a 15 cent item

Leon
03-26-2012, 01:29 AM
ESR is a property of capacitors, and has nothing to do with heatsinking!

evanh
03-26-2012, 01:31 AM
Those are perfectly normal voltages. Linear regulators like that are pretty forgiving of capacitor selection.

For future piece of mind, I recommend spending money on an oscilloscope. This gives you the visibility you currently don't have. Preferably a digital one with decent storage, these are becoming very reasonable prices now - http://www.jaycar.co.nz/productView.asp?ID=QC1932&keywords=oscilloscope&form=KEYWORD

PS: That scope pricing is in my local NZD and, although it's not listed, the capture memory depth is 2 Mpoints (Probably 1 M per channel).
PPS: I don't have one of those as I purchased when such cost effect tools were not available - and spent 15 times that price 10 years ago (Just to get deep memory capture).

rwgast_logicdesign
03-26-2012, 01:50 AM
I understand that but according to the data sheet the cooler the regulator the more forgiving the ESR reqirements tend to be. Anyways Ive fixed the problem but Im trying to design as top notch as I can with what I currently have on hand, now i have a tone of these http://www.parallax.com/Store/Components/Capacitors/tabid/150/CategoryID/28/List/0/SortField/0/Level/a/ProductID/160/Default.aspx .1uf ceramics, There is no data sheet link on that page. I searched the part number and looked at the face with a magnifier and came up with 104 e1m on one side and 12v rhu on the other. Is there anyway to find out what the esr of these caps are? From what I read Im going to need some more advanced tools than I own... Im hoping someone knows where a data sheet for these are or a way I can measure esr with a meter.

evanh
03-26-2012, 02:00 AM
Yep, one of those 0.1uF ceramics on either side of the regulator is perfect. Any extra storage added depends on consistency of supply, ie: If it's from a mains transformer then the rectified supply vanishes every half cycle so you need a decent sized cap for continuous supply, usually an electrolytic.

Peter Jakacki
03-26-2012, 03:17 AM
Some linear regulators are forgiving of output capacitor selection but this is to a point and dependent upon dynamic load etc. However most LDO regulators are slow to respond to load transients and normally have a larger output capacitor from 1uf to 10uf with low ESR as the capacitor has to do the work of supplying the load until the regulator "responds" (trying to keep the language simple here). If you do not have a large enough capacitor with the required ESR (equivalent series resistance) then it will not do it's job and you may find that the regulator ends up responding when it's too late and then it tries to compensate so that you end up with ringing or just outright continuous oscillation. A large capacitor is insufficient, it must be able to supply enough current on demand which requires a low internal resistance usually expressed as the ESR of the capacitor. By paralleling capacitors you increase capacity and decrease the ESR, so this is another way of achieving a result.

The LM2937 datasheet says:
Required for stability. Cout must be at least 10 μF (over the full expected operating temperature range) and located as close as possible to the regulator. The
equivalent series resistance, ESR, of this capacitor may be as high as 3Ω.
91010

kwinn
03-26-2012, 08:45 AM
A variation of 0.01V is nothing to be concerned about. It is known as "last digit bobble". Your output voltage is very close to the point where the least significant bit of the ADC switches from 0 to 1 and vice versa. Almost anything will cause that bit to toggle back and forth.