my first project :) - need a little help

[SnakeBite]

Hello,

After I finished second year at Uni, I decided to start a new project. It's not something new just to improve my project design experience (with you help of course…).
 I want to build adjustable linear power supply 1.2-20v@3A. This PS will have a seven segment display that you can change the voltage and the current (current limit).
The main demand from the PS that it will be as reliable, efficient and stable as possible.

I found the PERFECT voltage regulator - LT3083.
I will use 10,000uf 50v cap as a filtering cap (probably a Rubycon brand(.

I did my homework  and it's seems that I need a 230VAC TO 25VAC @ 3A transformer but i have a few question since i'm newbie :

1. Which transformer should I choose  "Laminated steel cores" transformers or toroid transformer (since i need more efficiency( ?
2. Since the range of the voltage is 1.2 to 20 volts ,  i don't think that the voltage regulator should take all that HEAT (ABOUT 20v*3A = almost 60W!!!) I think i should use a transformer with a multiple taps to minimize the problem. I decided the voltage regulator shouldn't dissipate more than 9w, so I thought to use a few taps that i can connect in series when needed to get the required voltage. here is the diagram of the transformer taps I thought to use:

0v             9v
                 4v
                 4v
                 4v
230v         4v



i thing the easiest way to put my hands on this transformer is wind one myself you have more ideas ?is it hard to wind a transformer?
3. What is the best way to switch the transformers taps? I thought using a relay - what do you think?


And that's it for now
Thanks A LOT for you help
Ido ,Israel

[IanB]

You can likely find a 12-0-12 V transformer off the shelf that you could use in a 0-12-24 V configuration.

If you don't heat sink the regulator it will limit the output current to avoid overheating. Plan on mounting the regulator on a big effective heatsink if you really want 3 A out of the supply over a range of voltages.

[Simon]

why not use a SMPS ? at least to drop the voltage from the high to low range. You could put a linear reg after it if you like, you will claw back heaps in efficiency that way

[SnakeBite]

The main demand from the PS that it will be as reliable, efficient and stable as possible.

SMPS are don't be that stable as linear PS .

[Simon]

The main demand from the PS that it will be as reliable, efficient and stable as possible.

SMPS are don't be that stable as linear PS .

well it depends on what you mean by stable. if your talking output noise then yes linear is better which is why you could consider a first SMPS stage to drop the main of the voltage and then a linear reg.

if your talking reliability, well i don't know, I thought SMPS was reliable

[SnakeBite]

simon do you mean to use some kind of PWM for the main voltage drop ? it won't add noise to the voltage?
yes i meant low noise when i said "stable".

[Simon]

you can use a 2 stage regulation, say you want 1.2V, you use a SMPS to drop the 20 to 3.5V and then use a linear reg after it to drop to 1.2V, the result is that the SMPS drops the main of the power efficiently and the remainder is done by the linear reg with also cleans up the power and produces low noise.

[SnakeBite]

can you explain more about "2 stage SMPS" ? what stages you mean exactly? i know  a little of  SMPS theory .

Ido

[Simon]

use a standard SMPS to drop your transformer voltage and then put a linear regulator after it so that you feed the linear reg say 12v instead of 20V and save o heat dissipation

[SnakeBite]

i don't understand what is "standard SMPS " take computer PS and connect it to mine? how ?

[DrGeoff]

If HF noise is an issue, then avoid the SMPS route. You could use a multi-tap transformer and use a range switch to select voltage ranges (or auto tap selection if you are adventurous). Otherwise I'd suggest using several T03 BJT's mounted across an appropriate heatsink to distribute the power dissipation, rather than have it all come from the one device. 60W is easily manageable on a reasonable heatsink, you'll need to do the thermal calcs to choose teh correct device and temperature rise.

[Simon]

how about we define what the exact requirements are ? as this is a first project the OP may be trying to go over the top causing unnecessary heart ache. Do you really need near "0" noise from the supply ? what will the power supply the realistically used for ?

[SnakeBite]

how about we define what the exact requirements are ? as this is a first project the OP may be trying to go over the top causing unnecessary heart ache. Do you really need near "0" noise from the supply ? what will the power supply the realistically used for ?

there is no specific use for this PS i just want an stable PS if i build one.

If HF noise is an issue, then avoid the SMPS route. You could use a multi-tap transformer and use a range switch to select voltage ranges (or auto tap selection if you are adventurous). Otherwise I'd suggest using several T03 BJT's mounted across an appropriate heatsink to distribute the power dissipation, rather than have it all come from the one device. 60W is easily manageable on a reasonable heatsink, you'll need to do the thermal calcs to choose teh correct device and temperature rise.

how can i make auto tap selection  ? with a relay?

thanks
Ido

[SnakeBite]

i want to buy toroide transformer from china . will it be as reliable as other none Chinese transfomer? i know that transformer are mostly very reliable devices but still...

Ido

[Simon]

should be ok, a visual inspection should tell if it is really bad, if it is claim your money back

[SnakeBite]

can you tell??



Ido

[Simon]

not without handling it and taking it out of the wrapping, no

[img]

And while you are in the process of power supply design, it might be a good idea to think how you're going to test whether you designed what you think you designed.
For example, since you make an emphasis on stability, what experiment(s) have you got planned to gauge that?

[SnakeBite]

I'm planning to test the ripple with high-end analog scope with no load and under full load and tweak the rails with capacitors accordingly (of course I'll not wait til then to do that i will design it like that that I'll need a little tweeking).

Ido

[SnakeBite]

i have a few question about the ADC and the counts of the display.

i want to have good resolution of 5 digit i mean (20.000 volts) so thought i need 4.5 digit ADC but the most significant digit can be only 1 (4.5 digit = 19999) so i have a little problem there . i tried to find 4.75 digit ADC but i couldn't find any .

can you help me out?

[IanB]

i have a few question about the ADC and the counts of the display.

i want to have good resolution of 5 digit i mean (20.000 volts) so thought i need 4.5 digit ADC but the most significant digit can be only 1 (4.5 digit = 19999) so i have a little problem there . i tried to find 4.75 digit ADC but i couldn't find any .

can you help me out?

19999 would give a reading of 19.999 V = 20.000 V, which is the resolution you are looking for. Can you clarify the question a bit more?

[SnakeBite]

from what Dave's said in the VBlog on counts i understand (correct me if i wrong) that with 4.5 digit resolution the MSD can only be "1" maximum so it will never can show 20.000 v only 19.999 v.
plus, the 20v i'm limitting the PS is under full load (20V@3A) and under no load it should be more than that like 20.700 or something...

do i understand Dave correctly?

Ido

[IanB]

For practical purposes a display range of 0 to 19999 counts give a meter range of 0 - 20 V at full resolution. If you want to read above 20 V then you need a display with 39999 counts. But with the 19999 count display you could show a reading of 20.70 V, which might be accurate enough for practical purposes? Also, a properly regulated power supply will be 20.00 V at no load and 20.00 V at full load. If you are expecting the power supply to be unregulated, then I am not sure why you want to indicate the voltage with such precision?

[SnakeBite]

i have a PS and it limited to 40v and when it's not connected to a load it shows 40.7v on screen i thoguht it's normal - guess not...

i want it to be regulate PS but i want it to be more than 19.999v , in case that something happent  , to user will know that the PS is overvoltage...

thanks for clearing it out...

so you say that 4.5 digit ADC will be good enough to show on screen more than 19.999v? what is the maximum limit of 4.5 digit ADC ? what is the maximum voltage that it can show?

thanks
Ido

[IanB]

I think you are confusing terms here. An ADC is an electronic device or component that converts an analog signal to a binary digital signal. The resolution of an ADC is usually expressed in bits, e.g. 8-bit or 12-bit. (For instance an 8-bit ADC could output values in the range 0-255 counts.) You then have to process this digital value in a computer or microprocessor of some kind.

What I think you are looking for instead is a digital panel meter to read volts. These come in a package so you just connect an appropriate voltage to their input and they display the reading on the display. Many such meters do appear to come in 1999 count ranges, or maybe 19999. This is unfortunate, but if you want to display above 20 V then you will have to find a 3999 count meter or use a 19999 count meter without the 1. (If you ignore the 1 you can pretend it is a 9999 count meter, but you will lose some accuracy.)

[SnakeBite]

thanks for the clearification but i know what i meant.
there are ADC's that convert the voltage (analog signal) to digital Binary signal and they also have a seven segment driver in one package what i need is a way to convert the analog voltage in a resolusion of 4.5 digit but with a little more than 20000 counts and then drive the five seven segment that i have .

i almost sure that i can do it with PIC or AVR microcontroller but i'm in a learnning process of how i program and use microcontrollers .... so if the ONLY way to make it happent is with a micro i will have to speed up my learnning speed...

Ido 

[SnakeBite]

i found a way to make voltmeter with 40000 counts using MAX1499 ADC and MAXQ610 MCU but i don't have real experience with MCU yet and also the package of the two chips it's defiantly not for beginners (32 pins TQFP)

you can see all the info here
http://www.eetasia.com/STATIC/PDF/201101/EEOL_2011JAN31_EMS_POW_AN_01.pdf?SOURCES=DOWNLOAD

do you think i need to buy the starter kit and other accessories of this chip and get familiar with it OR there is more easy way let's say with pic micro+MAX1499 ? which pic do i need if the MAXQ610 is 16 bit micro?

what is better the pickit2 or the pickit3 ? do they fix the bugs of pickit3?

thanks
Idoaricha

[IanB]

I think the answer to your question depends on your objectives. If your intention is to learn about electronics, then you can go ahead and make it as complicated as you like.

But if your intention is to construct a practical power supply, then I do not understand why you are getting so hung up over the 40000 counts? What good will it do you to display 20.700 volts instead of 20.70 volts? Unless you are planning to make a super accurate, highly regulated, highly stable laboratory grade power supply, then the extra digit will be a waste of time.

[Bored@Work]

Before you think about stuff like which microcontroller to use, do the sensible thing, run the numbers.

If I got my math right then 1 count (aka the last digit) in 40000 is 25 ppm. Can you construct a resistive divider that good?

For example, lets assume one of your resistors is perfect, and the other is just 0.05% off, and the meter is perfect, too. Then you end up with a display of 39.990 V for 40.000 V input (or 40.010 V - i.e. overflow).

Now you might be lucky. You might get two resistors from the same lot which happen to have a similar error in the same direction, canceling each other's error out in the resistive divider. But do you feel lucky?

And if you aren't lucky, do you have the instruments to adjust the divider, or do a correction in software?

[Zero999]

The problem with some cheap transformers is they don't have enough iron and copper in so run hot and will catch fire if they don't have a thermal fuse and there may not be enough insulation between the primary and secondary which can be a shock hazard.

Automatic tap selection can be done with a relay or transistors, attached is a basic tap changer which uses transistors.

[SnakeBite]

The problem with some cheap transformers is they don't have enough iron and copper in so run hot and will catch fire if they don't have a thermal fuse and there may not be enough insulation between the primary and secondary which can be a shock hazard.

Automatic tap selection can be done with a relay or transistors, attached is a basic tap changer which uses transistors.

i can't find match transformer in sane price. and  even at anyprice i can't find something with the same rating output voltage is it really that critical ?

BoredAtWork and IanB i thought about it this Saturday and i came to your same conclusion -  that i don't need that much resolution and i'll lower it to 3.75 digit (4000 counts) 20.70 is enough for me.

thanks
Ido

[Zero999]

Oh sorry, I forgot to post the schematic of the tap changer.

[Zero999]

i can't find match transformer in sane price. and  even at anyprice i can't find something with the same rating output voltage is it really that critical ?

You won't find an off the shelf transformer with all of the specified taps. To get the exact specification you desire, you'll either need to wind it yourself or get it custom made. You could use two 6-0-6V transformers in series to get 0-6-12-24V but it'll be more expensive and have poorer characteristics (efficiency & regulation) than a 12-0-12V transformer with the deisred current rating.

[SnakeBite]

i do found one but you said that we can't trust it because it's chinese transformer.

my question is  : Is it really that critical to buy made in USA transformer after all it's just a wires and iron core.

 i started to read about transformer theory but it seems a big hassle to wind it yourself to many variables to consider and a lot of formulas of the Power loss , magnetic flux also types of cores etc...

Ido

[Zero999]

Of course it doesn't matter where the transformer is made, as long as it's good quality. I'm surer there are plenty of good Chinese made transformers but buying a non-name part is often a gamble.

[SnakeBite]

thanks . i contact the seller and he was honest enough to tell me that this transformer is NOT high quality one.
so I'll for more...

i have another question :

what is the difference in protection between EMI suppression capacitor and ferrite bead  ? do they make the same thing or one is different then the other ? i did a little research and the ferrite bead is generally a low pass filter that blocks RF and EMI from going through a cable and the EMI capacitor is do the same .

if they are the same which one is better to put in a power supply near the input transformer?

thanks
Ido
 

[Zero999]

A ferrite bead acts like a lossy inductor which is a high impedance to RF and is used to block it. A capacitor has a low impedance to RF and is used to short circuit it to earth.

EDIT:
A mains transformer will be very good at blocking RF because the eddy losses will be high at high frequencies. It's possible for some RF to get through the transformer to the secondary, via the interwinding capacitance. This can be minimised by using a screened transformer or by connecting both sides of the secondary to earth via Y capacitors.

[SnakeBite]

which one you think will be better to use at the input of PS ?