New guy to the forums needing some help making a 0-18VDC regulated variable power sup

[Mike86H]

Ok, so the more amps you put out through a power supply and the lower the voltage, to more heat produced, correct?

So if I run an LM350 and supply it with 2 amps, I need to keep the voltage from the transformer as close to my operating ranges as I can? If I want to run between 1.25 and 18 Vout, then I need no more than a, roughly, 19.5V output transformer? Anything more will result in more heat?

 

[KeepItSimpleStupid]

Sort of.

The more AMPS you draw, the more heat dissipated.

For the LM350, 1.414*(secondary) voltage has to be 3 volts greater than the regulated output.

That's why that last design you showed had the ability to use 1/2 of the transformer secondary.

Lets say you used a 24 center tapped transformer. You would get either 17 or 33 VDC of raw output. So, you could conceiveably use it for <14 VDC reguated or 15 to 30 V regulated with the same power disipation (@17V and 33 V) depending on the "range switch".

Power =( V^2)/2, so reducing the drop across the regulator yields a decent improvement.

So at say 14 VDC @ 1A you have (17-14) or 3 W or (30-3)/1 or 27 W; Those 3 W or 27 W are wasted as heat.

An Aside:
Some transformers can be purcased with "dual secondaries" for parallel or series operation. So, if you got a 2 idential windings of 12 V, 1 A, you could wire it such that you can get 12 VAC @ 2 A or 24 VAC @ 1A. Nothe the available power V*I remains the same. 24 W and 24 W.

An analog meter might be a nice touch.

I'm trying to stay somewhere between simple and complex, but it's becoming harder to do so.
It's no longer as simple as it was when we started. We know now of the peak loads.

I would put the two protection diodes in place which ronv said not to. I'd also put a TVS or 3 on the AC line depending if your going to use a 2 conductor or 3 conductor cord.

 

[Mike86H]

Yeah I was hoping to keep this really simple and just make a really robust power supply, but its seemingly harder to keep it really simple. I see things that I would like to incorporate into my design from other power supplies, such as the 3-prong receptacle with a built in fuse for the power cord and the internal fan.

My big problem is wanting to try and blend old school and new school technology. When I set out on this course, I had hoped that I would be able to come up with something better than what was out now, but with the move to digital everything, its incredibly hard to do and maintain a cost effective point, because at some point this could easily become a project that was just cheaper to buy pre-made from the beginning.

I asked though about the heat dissipation and such because I found an LM338 based power supply that is at 5A/1.25-30V that runs a pretty small heat sink with a fan on top, but I figured that would put off alot of heat and would require a bigger heatsink than what it came with. Im just really concerned now with the heat dissipation and keeping the whole thing cool so it does cut off in the middle of use.

 

[ronv]

Ok, so the more amps you put out through a power supply and the lower the voltage, to more heat produced, correct?

Right

So if I run an LM350 and supply it with 2 amps, I need to keep the voltage from the transformer as close to my operating ranges as I can? If I want to run between 1.25 and 18 Vout, then I need no more than a, roughly, 19.5V output transformer? Anything more will result in more heat?

Now you got it. Right again.

Like Kiss said, it's a little more complicated about the transformer. Transformers are speced in RMS but once thru the bridge it will charge the caps to peak voltage which is 1.414 times higher than RMS. I think the last one we used was a 16 volt one which would charge the caps to 22.6 volts. We will loose about 1.2 volts thru the bridge so 21 volts into the regulator. But wait, there's more! The caps only get charged when the sine wave from the transformer is higher than the voltage on the caps. So in between peaks the voltage will sag. This is called ripple. The higher the current draw for a given capacitor the higher the ripple. This is also why the transformer needs a higher current rating than the load. And finally the regulator needs some voltage higher than it's output to regulate. This is often in the spec sheets as drop out voltage and also depends on how much current you draw. The advantage of the switch that Kiss is talking about is that the output of the transformer can be cut in half. This is nice for power, but you would have to stop and switch then readjust the voltage.

 

[ronv]

Fans make a really big difference in the size and temperature of heat sinks. If you could use a fan that would help.

Almost anything you buy on line won't have the ten turn pot and transformer which is where most of the cost is.

 

[KeepItSimpleStupid]

@ronv; OOps, forgot about the 1.2 V drop on the diodes. Now spec it for 95VAC for the input too, and add the account for the transformer regulation. The transformer voltage will be higher when it doesn't operate at the full load capacity. Fun stuff. Right!

 

[Mike86H]

Well what I was thinking of is running one, maybe two fans, in the case with some slots cut for ventilation as well. That one box I posted earlier only had the fan, which I dont understand because theres no place for the air to flow out. But yeah, Im thinking at least a fan or two, I just dont really know how to plug those into the schematics.

As for the transformer, the one that was posted here from Mouser would probably be the one I end up going with. I've noticed that they have several transformers in stock in the 20V range, but I figure the one posted in this thread would be the best bet. And I think Mouser also sells the 10 turn pots, although really, if the Vout doesnt go up past 20, I could probably just use a regular pot instead.

So if at all possible, could yall help me with the values of components so that I can then go over to Mouser's site and make a price list and such?

 

[ronv]

Send me a private message with your e-mail and I'll make you a bill of material on Mousers web site tomorrow. Then you will have prices and part numbers. What parts do you want to buy somewhere else because you like their looks or something?

 

[Mike86H]

Really as far as internals, I dont have too big of a preference as far as any other retailers. Mouser seems to be alot cheaper than most of the other places and since I can go pick them up and save on shipping, that makes them my preferred choice. The only request I do have is for a similar power switch to the one on the box I posted, but with an LED built in to the switch. So for power, a round rocker with an LED. Thats my only preference.

And I just sent you a PM with my email.

 

[ronv]

I sent the link for the Mouser order. I'll make a schematic of the whole thing in a bit.
Just click on the part and you can see the data sheet. Check the switch and the 10 turn dial to make sure they are what you have in mind.

 

[Mike86H]

Thanks Ron. I really appreciate it. The list is great!

 

[ronv]

Here is the schematic that I think matches the parts. Keep me honest.

 

[Mike86H]

Thanks for all the help on this Ron, and KISS. It is greatly appreciated indeed.

I do have a few more questions regarding this project before I order the parts though.

For the heatsink, Im looking at Mouser, and the majority of the TO-220 heatsinks seem really small. Being that I will be ordering the rest of the parts from Mouser, if I can get the heatsink there as well that would be great.

My question is, if I run a fan, possibly 2, and have ventilation holes in the box, can I use a smaller heatsink? This may be completely off the wall as well, but what if I used an extruded aluminum case? If I mounted the IC to the case, would it work as a heat sink as well? I've just never done much with heat sinks, so Im not too sure what will work and what wont.

 

[KeepItSimpleStupid]

@ronv
Can you throw in a few surge supressors and put the protection diodes in?
Label them - optional.

Remember the extruded heat sink that ronv suggested earlier. That extrusion can be placed on a metal box. The paint, however is somewhat of a thermal barrier.

You can look through this search https://www.google.com/search?site=&tbm=isch&source=hp&biw=994&bih=605&q=audio+amplifiers&oq=audio+amplifiers&gs_l=img.3..0j0i5l9.1628.5627.0.5840.16.10.0.6.6.0.105.930.9j1.10.0....0...1ac.1.25.img..0.16.979.wERewK56wcg#bav=on.2,or.r_qf.&fp=6676649ff4ca754f&q=audio+amplifiers+heat+sink&sa=1&tbm=isch to get an idea

on some heat sinks.

An Audio Amplifier is a good example.

I really don't think you want a fan unless it doubles as a pink noise generator.

 

[Mike86H]

What about a heat sink this size: https://www.mouser.com/ProductDetai...q9R1p1wF%2bpkzxkk34TmCTnV6etLpPOYxVsthvDwbQ==

 

[Mike86H]

Ok, so I've been doing some reading about heatsinks and size needed and all the math and what not. Im going to kind of think out loud here that way if Im totally wrong, I can be corrected.

The LM350T has a thermal resistance, junction to case, typically of 1.2C/W. So that means that, if mounted to a heatsink, and let say that Im running the power supply at 8V, then 10V*2A would be 20W dissipated as heat and the IC would heat 24C above ambient? If ambient temperature is 27C, then the IC would reach 51C under those conditions?

Im still not super clear on where the thermal resistance of the heat sink comes in, but the one I posted from Mouser has a resistance of like 1.3C/W. From what I understand thats how you base what temperature the heatsink will rise above ambient based on watts to the IC. So basically the heatsink would mirror the IC in terms of heat at that specific condition.

If operating temperature on the LM350 is 0~125C, then it sitting at 51C would seem to be fine? At least in my mind. Coupled with a fan and ventilation, I would think that I would be capable of staying well within the operating range of the IC?

And as for the fan, how would adding it to the circuit make it into a pink noise generator? (I'll even be honest, Im not too sure what a pink noise generator even is.)

 

[KeepItSimpleStupid]

Yea, for simple calcs like:
Tcase = 95°
Tambient = 25°
Pd = 100 W
(95-25)/100 or your heat sink needs Rθ to be < 0.5 °C/W

The thing is, that when you do the FANCY calcs, every interface has a thermal resistance.
So, you start out with the Max Junction temp and Tambient

You account for each interface and change things based on the emmisivity (color) of the heatsink, the type of insulating washer used etc.

So, yoour quick and diry calc looks about right. The Junction temp of a silicon based IC is quite high.

OK, I was adding a bit of humor which I didn;t expect you to get.

White noise is the statc you hear when you detune an analog Radio and hear "Static"

"Pink noise" is supposed to be soothing because it consists of all frequencies at the same amplitude.

There is our friend Wikipedia: https://en.wikipedia.org/wiki/Pink_noise

I once built a "surf synthesizer" project for someone. It was the PE article here: https://www.preservationsound.com/?p=1864 Here is a better one: https://www.google.com/url?sa=t&rct...=k94-8LcCDBtEzGo-6a7VCw&bvm=bv.50723672,d.aWc

It worked quite well.

A fan, might well make a noise that is soothing. The joke was basically, I don;t think you need the fan, but the noise it makes may be soothing to the patient, so you get two things for the price of one: pink noise+cooling

 

[KeepItSimpleStupid]

See here: https://www.wakefield-vette.com/products/natural-convection.aspx

What you should be looking for is a heat sink that would have it's fins vertical when mounted on the outside of the chassis.

 

[ronv]

The LM350T has a thermal resistance, junction to case, typically of 1.2C/W.

This is true if you could keep the case at 27C. But alas the heatsink heats up. So the temperature junction to case is one part.

So that means that, if mounted to a heatsink, and let say that Im running the power supply at 8V, then 10V*2A would be 20W dissipated as heat and the IC would heat 24C above ambient? If ambient temperature is 27C, then the IC would reach 51C under those conditions?

Not quite. Remember back a ways we talked about the cap charging up to the peak voltage? What that means is when (if) your line voltage coming in is 10% higher than nominal the input voltage to the regulator could be around 28 volts. So it would drop 20 volts X 2 amps or 40 watts. So we would have junction to case of 1.2X40 = 48 + the heat sink 1.3X40=52, + ambient = 27 for 127. There is a little more lost in the interface between the transistor and the heat sink but not much in this case because it needs no insulation from the heat sink. This is kind of a worst case number as your line voltage would have to be high while your voltage setting was low and you were drawing maximum current. My bet is you would be ok, but for sure you would be ok with the one from amazon that I posted. That one would make it without the fan if it was outside the box.

Im still not super clear on where the thermal resistance of the heat sink comes in, but the one I posted from Mouser has a resistance of like 1.3C/W. From what I understand thats how you base what temperature the heatsink will rise above ambient based on watts to the IC. So basically the heatsink would mirror the IC in terms of heat at that specific condition.

This is the other major part of the equation and needs to be added to the first part. Please also note that the heat sink manufactures are kind of sneaky - the 1.3C/Watt is with air. Without air that heat sink is 2.85C/watt.

Do you want to put it inside the box?

[/QUOTE]

 

[HiTech]

Mike86H

Before you go off and try to build a PSU which will cost you as much if not more than buying a ready-made one, consider getting a used one online like off of eBay. The Hewlett Packard power supplies are built like brick $hit houses and are true lab-grade quality, designed to last for tens of years operating at 100% duty cycle 24/7. The two models I list below are small enough to set off to the side on a work table. They are short-circuit protected, have coarse and fine controls for both voltage and current, are fully regulated to tolerances better than hobby-grade circuits and they have high quality analog meters on them to indicate your exact settings. The voltage and current can be set to a very near 0 output and the fine adjust controls really do yield incremental adjustments. Best part is they are listing on eBay for well under $50 (plus S&H). If you are operating a professional shop, you should be relying on one of these units and not something fabricated into a project box. Besides, your customers will be more impressed with a cool looking electronic device operating the tattoo gun. Just enter HP 6284A into the search box on eBay and you should see several results. You can narrow your search by selecting BUSINESS & INDUSTRIAL, followed by TEST EQUIPMENT and then enter the HP model numbers.

Hewlett Packard 6284A 0-20vDC 0-3 amps
Hewlett Packard 6201B 0-20vDC 0-1.5 amps

There are other equal quality brands of similar power supplies on fleaBay if you search for names like POWER DESIGNS or KEPCO, or SORENSEN. Some of those brands offer dual and triple outputs all individually adjustable w/ individual meters. That would allow you to connect multiple tatt guns to just one power supply unit!

Here's a nice Sorensen XT-20-3 with digital readout listed for $30 with no bids yet 0-20v at 0-3 amps
**broken link removed**