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

[Mike86H]

I don't want to get into a big argument here, and I mean no disrespect, but the thing is, power supplies for tattoo equipment aren't exactly the most elaborate things. At least not the vintage ones. Sure, now you can buy fully digital units with an external transformer/rectifier set-up, but from what I've seen in the industry, those tend to crap out after a couple of years. However, there are power supplies that are from the late 60's and 70's still in use today, its just that the younger guys dont want old school technology. Me personally, Im a big fan of the old school technology, and the power supply I have envisioned would be perfect with the one small problem of both the guys who designed it and put it into production are dead now, so buying one isnt going to be exactly easy and if anyone new tries to mass produce it or keep production up, I can almost guarantee that component quality is going to slip drastically.

I'm guessing now, after reading that, that there probably wasnt a voltage regulator included in the original design for a reason. But Im still confused on all of this. I've seen alot of power supplies (i.e. lab power supplies) running an LM317 and a rather large heat sink and doing just fine. Granted Im not sure that lab power supplies are used for extended periods of time (1-6 hours) in one sitting over pretty much all day, but I do know that alot of artists are switching over to laboratory power supplies and modifying them to make use of our foot switches and machine cords.

The transformer that I would probably use is a Radio Shack transformer rated at 25.2V/2A output. My machines run at around 7-10V max. The wont even come on at 2V. So, if roughly 2V are lost through the circuit, the 25.2 would drop to 23.2, but we'll call it 23 for the ease of math. Now, lets say that I run an LM317 regulator, which I think rates amperage at 1.5A max. So if Im using your math, which Im sure is the standard formula for this equation, at max, I would be dissipating about 24W of energy. Would a large heat sink and a fan not be enough to keep the circuit cool?

Im not terribly concerned about heating up the box because all power supplies get warm. As long as it doesnt get hot enough to burn skin, I can live with that.

The biggest thing I can think of that would make this different than other applications would be that, computer power supplies need to provide a constant stream of power throughout the length of time that a computer is plugged into it. However, with a tattoo machine, we are running them in short intervals over the course of several hours. There is never powered supplied to the machine for a period longer then roughly 2 minutes. At which point we let off the foot switch and our machines arent drawing the current. This power supply isnt something that is going to be plugged in, hooked up, and left to run for a multitude of hours at a time. Efficiency in electrical current and heat dissipation isnt of primary concern in my business, however robust power from a unit that is less apt to break than something made in a Chinese sweat shop is. Believe me, power supplies going out in the middle of a half sleeve tattoo isnt a good thing.

If you would like, I can link to the YouTube video showing the power supply in question so that you can get a general idea of what Im looking at putting together and maybe you would have some ideas on how to tweak or improve upon it from there?

 

[Mike86H]

The guys are right the switcher is better, but they are more difficult to build and make work.

I get the impression you want a full sized knob to turn not a screwdriver slot.

Yeah, when we tune our machines, we need access to a full size knob, usually a 10 turn pot, in order to fine tune the machine speed because its all done on the fly based on needle size that we are driving at the time. So the full size pots with a knob thats easily accessible is a big priority.

As far as efficiency goes, I know that there are always trade offs. The older power supplies with less components were far from efficient but were built like old battleships. They have nicknames such as "tank", "brick", and "lunchbox" (think of the old 40's steel lunch boxes) just because they are so rugged. The more modern ones that use more circuitry and more complex circuits are great as far as electrical efficiency, but Im assuming that based on the amount of components they tend to fail alot easier and more frequently. So, for me at least, something that is more rugged and robust, and potentially less likely to fail when I need it, is a fair trade off for efficiency.

Here is a link though to the power supply design in question: https://www.youtube.com/watch?v=r-ZmZbuvL8M

 

[KeepItSimpleStupid]

OK.

Why don't you design the power supply for 7-10 V then. We can make the adjustment work that way. Then the transformer can be sized for say 15 Volts of Raw DC. You can use 1.4 * Transformer RMS voltage. Capacitor can be sized for 1000 uf/amp. There should also be two small caps per the datasheet close to the regulator.

A 10 turn screwdriver adjustable pot could also be used.

It's also very simple to ground the ADJ terminal which will set the output to 1.25 V, the equivalent to OFF. If you want to get fancy an external MOSFET high side driver can be used to get that output to zero volts.

To round out the power supply, A couple of surge suppressors could be added in a few places. i.e The line or the input to the regulator.

And, if you want to learn something maybe even a bi-color LED to show it's turned on when green and red when the foot switch is activated.

Again, since I like to discuss alternatives, there are medical rated power supplies such as these: https://www.trcelectronics.com/Meanwell/medical-power-supply.shtml They do have a limited adjustment range. At $10 they are a bargain. Probably even cheaper than a transformer. Medical supplies pay special attention to leakage currents and isolation. e.g. Use around defibrillators.

You can follow them with a linear regulator if you wish too.

The LM317/LM350 supplies that we have been talking about have a fixed current limit.

One might have to be careful displaying a home-brew power supply in front of a perspective customer, so do your best to make it look VERY professional.

You are correct, with no load there will be little dissipation. You can also use a metal case for a heat sink. A plastic case would be double insulated, but then you have to add vents or make the heatsink external.

The case, the power supply and "real estate (the PCB) generally contribute the most to he cost of a project. Little things tend to get you. Power cord, power cord socket, punch for a hole, standoffs whatever.

Go ahead link.

FWIW. I tend to when discussing a project to suggest things that might be absurd or off the wall. I also tend to try to think about "What if I had all the money and all the resources in the world, what would I like this thing to do"? From here, I design/program the essentials, but leave "hooks" for enhancements. This has actually led to computer programs running for 16 years without any modifications. In one case, I must have "anticipated" the invention of spreadsheets, BUT ran out of memory.

So let's say, you:
1) might want a power indicator
2) you might want surge supression
3) you might want a voltmeter
4) you might want to change to a switching supply

It might be helpful to know the characteristics of your footswitch, so that can be incorporated now. The other thing to get an answer for is whether you really want 0 volts to be output when the footswitch is off. It's BETTER to turn off or reduce the output of the supply to 1.25V than to turn the AC on and off because of transients.

 

[KeepItSimpleStupid]

Since mine came after ronv's post, the voltmeter might be a good addition and voltage and current even better. You my find that having the current meter is more beneficial than the voltmeter. You also mentioned the 10 T pot being panel mounted which is OK.

Current, more than voltage probably defines the operating point of your tatoo equipment. The current is dependent on the skin resistance.

What I don't undertand, is if the equipment is nothing more than a foot controlled power supply and a "needle"? In which case, I would SERIOUSLY consider a CV/CC supply (Constant Voltage/Constant current).

That video is Yuk. Don't like the fact that the AC is switched with the foot pedal. There are no small caps like a 0.1 uf on the regulator and no surge suppression. And it looks UGLY.

This is the first link for a high side driver: https://www.electro-tech-online.com/custompdfs/2013/08/1154fb.pdf
Yes, it makes it more complex and there are definately others, but it gets you zero volts out.

One of the better ways of doing the footswitch in my opinion would be to use an optocoupler on the ADJ terminal and to force the current through the footswitch to be at least 10 mA. This usually is tthe minimum about of current required to de-oxidize the contacts of the footswitch.

 

[JonSea]

Man, after watching that video, it's no wonder you are confused! Almost nothing he says is accurate or correct!

But if you want to duplicate that, it would be pretty simple.... A transformer, bridge rectifier, filter caps and a rheostat and your set. The rheostat is nothing like the variable resistors you can get at RadioShack – they are actually wire-wound and can handle amps of current. Digikey should have them....not exactly sure of the value you'd need...but it would be in the tens to a hundred range.

 

[KeepItSimpleStupid]

I want to pick apart the power supply design for reliability:

1. Total exposure to line transients just plugged in and on. I recommend an ISOBAR surge suppressor.

2. The lack of the small caps makes the regulator "prone" to wanting to oscillate. This is not good and increases power dissipation.

3 Line and load transients because the line voltage is switched. Switching the output would be better, providing the foot switch is rated for DC.

 

[Mike86H]

Those are all good points, and I will do my best to elaborate more thoroughly and describe what all is truly needed. Im going to try to touch on all of these in order KeepItSimple, so I apologize in advance if this is a long post.

Traditional tattoo machines run on electromagnet coils, that pull on a spring loaded bar to drive the needles forward and into the skin. At low voltages, the electromagnets dont have the force to pull on that bar. Typically we dont see the machine start to run smoothly until around 7+V. The amount of voltage needed depends on the size grouping of needles that we use at that time. However, there are also rotary style machines which run off a standard DC motor to spin a yoke and drive the needle forward that way. Those machines run off less voltage than a standard coil machine. But they are widely used now, so being able to adjust below 7V would be a big plus just in case I ever wanted to use one.

Most of our standard power supplies are adjustable from 0 (or the equivalent, my current one starts at 1.25) up to around 18V max. Anything over 18V is typically considered overkill as we dont normally run machines that high, but I have seen some instances where a machine needed to run up around 12V and some as low as 5V depending on the type of machine. So limiting the power supply to a small window, limits the size of needles we can use and overall the size of the piece we can work on. When using traditional machines, most efficiency of the power supplies goes out the window simply because we need such a wide array of voltages to be able to choose from, because as well as the needle grouping sizes, different machines run on different voltages due to how they were built. Most machines have a different "sweet spot" that they run best at, voltage wise.

A screw driver controlled pot wouldn't really be the best, simply because varying the voltage needs to be done on the fly while tuning the machine to the needle grouping, etc., as we set up for a tattoo. The pot really needs to be mounted in a way that a knob can be accessed without taking the power supply apart. A single pot to adjust the voltage on the outside of the enclosure also helps keep down chances of cross contamination of blood borne pathogens. So it really needs a pot accessible from the outside.

The minimum value of voltage doesn't have to be 0. It can be 1.25 and be just fine. As I said above, my current power supply starts out at 1.25. I only know this because it has a digital volt meter, but I never really look at it.

Im not entirely sure what you mean by a surge suppressor. When I see or hear that terminology, I think of the surge strip with like 10 outlets that plugs into the wall outlet.

For LED's, I am planning on using an LED rocker on the main line coming in to the transformer so that it illuminates when it is on. I had though about putting an LED that would come on when the foot switch is pressed, just because that would look cool. To add to the LED foot switch light, I have thought about having two 1/4" female jacks that could be switchable from one to the other via a small toggle for the output to my machines. But that isnt going to be a requirement.

My concerns with cheap electronics is the quality of components. Given that a power supply is absolutely critical to a good tattoo, I really dont want to go the route of buying cheaper pre-made items just because its easier if it means that it could crap out on me when I need it most.

When talking about the "fixed current limit" of the IC, whether it be the 317 or the 350, I again am not sure entirely what that means, but I assume that it limits the amperage?

On the note of the power supply being displayed in front of customers, most of my customers are pretty strange individuals. Alot of us are into the macabre side of life. I've seen power supplies built into things like skulls and even wooden boxes decorated with a steam punk style. I have a background in metal work, fabrication, and welding. I figure that I would build a chassis out of sheet metal and would distress it a bit for a vintage look, but then rivet the majority of it together for a real "industrial" look. But I also know that it being made of steel that I will have to take great care to keep from shorting the power supply on the chassis or electrifying the whole thing and shocking the crap out of myself at some point.

For a heat sink, since Im going for the industrial look, I figure that getting a decent size aluminum heat sink and mounting it off the back would not only help keep the circuit cool, but would also aid in the overall appearance of the entire unit.

Cost wise, I fully intend to over build this so hopefully it will last me for a good long while. With grommets, heat sinks, standoffs, etc, I know that it wont be cheap by any sense of the word, but I figure for around $100 I can build something that will last for quite a while, and probably even less than $100.

I appreciate the "off the wall" ideas and suggestions, because, well, I'm a bit off the wall too. Random and strange ideas are usually the ones that appeal to me the most.

To reference points numbered 1-4; I really dont need a power indicator other than perhaps a light for on/off and possibly the LED for the foot switch. And I also really dont need a voltmeter anymore as I normally just tune my machines by sound and feel.

The footswitch is just a simple SPST momentary switch mounted in a case, that when stepped on, completes the circuit and sends power to my tattoo machine. It is connected to the power supply by a 1/4" male jack.

 

[KeepItSimpleStupid]

Yikes! Now we know A LOT More and all of that makes a BIG difference.

So, based on what you said, what is the max and minimum activation time on the solenoid? And anything else you can say about use.

Why don't we switch gears 180 degrees. e.g. Use a solenoid driver IC. e.g. https://www.ichaus.de/iC-GE

No, I'm not saying that one. These IC have come into existance since car fuel injection.

I designed a couple of solenoid drivers and they were long tern ON type of drivers.

In one case, I ad LOTS of cooling available and it was my first one. It was designed to drive a high inertial load connected to a rotary solenoid. A shutter.

In the second case, I had a cycle time of about 1-2 minutes. It was also a shutter, but it was located next to a 1 m thick $1000 piece of glass. For this, I chose I different type of activation (slow, but with a small kick). I essentially used a Pulse Width Modulated (PWM) motor controller PCB, about $20 kit.

With these designs, you also help the actuator. It sees less power.

So, what might 30 seconds of tatooing look like in terms of on/off cycles.

BTW KISS is OK to use.

So, we might be able to build a MUCH better mousetrap starting from specifications. What if
one could adjust the "kick time" and the "hold voltage"?

So, give some idea of the cycle time.

PS:
Current limit isn;t as important as the ability to handle a brief surge.
The solenoids should be protected with a reverse diode, hopefully at the solenoid.
Surges can come from a variety of places: The power line being the major source.

 

[Mike86H]

Ok, most of that went entirely over my head. As far as solenoid drivers, all I really know of solenoids is when I replace them in my car. Haha.

When talking about the maximum and minimum time on the solenoid, are you referring to the time it takes for the needle bar to move forward? If so, I honestly couldnt tell you in exact terms. I can tell you that the cycles per second, meaning one full stroke of pushing a needle into the skin and retracting it, is anywhere between 100-140 cycles per second, but that number isnt a definitive number either as it can be anywhere in between with +/-10 on either end of the spectrum. So on the low end we are somewhere around 2700 CPS and 4500 CPS on the high end.

 

[ronv]

Wow. That is old. The boys are right not all he talks about is accurate. Having said that I think the LM350 is for you. They are virtually bullet proof. If they get to hot they shut off, if you short them out they protect themselves. Kinda like that toaster for this century. I see now why you kept asking where the potentiometer went. In the toaster case it was in series with the output of the supply, but in our case it controls the 350. The regulator he showed is not adjustable it just sets the voltage to 12 volts.
In any case if you don't need the 18 volts things would be cooler and thus more reliable for the lower voltage. For example you could get more amps from the same size transformer so it would not have to work so hard. Same with the LM350.
A couple of words of caution. The heat sink I showed needs a fan to be safe at low voltage and high current so get a big one from E-bay or someplace (real expensive new). I would also use the Mouser transformer not the Radio Shack one. Not that I have anything against them but to supply 2 amps to your machine requires a transformer greater than two amps because of the bridge setup and the filter cap.
10 turn pots and dials that read them out are available. A little pricey but you will probably still be under $100. So nail down the voltage and if the heat sink will have a fan. When we are all done I can make you a Bill of material for Mouser. Sometimes it's easier because I get the catalog.

 

[KeepItSimpleStupid]

Up the size of the filter cap because of the peak currents.

I was going on the assumption that it was one push of the foot switch is one solenoid activation.

@ronv
Aside: I wonder whatever happened with the coloidal silver guy?

 

[Mike86H]

Ok so now that you guys have seen what was stumping me, it leads me to more questions. I know the one I linked to in the video is pretty much ancient and totally antiquated, but that is an unregulated power supply correct? Being unregulated means that the sine wave still has a bit of a dip in it instead of being completely flat correct? Where as a regulated power supply would produce a completely flat sine wave? So by adding the VR circuit, it would smooth the sine wave completely correct?

If I went with the modified circuitry and put the LM350, would a heat sink like the one Im linking to work?
https://www.amazon.com/Aluminum-Hea...UTF8&qid=1376101974&sr=8-34&keywords=heatsink

 

[KeepItSimpleStupid]

Nope. the LM350 is a TO-3 package, I believe so a heatsink similar to this: **broken link removed** could work.

Well, I checked, you would probably want the TO-3 package and not the TO-220. In both cases, the heatsink would not work UNLESS you had access to a mill to remove some of the fin material for the screw. The TO-220 package could be mounted on the inside of a metal case, with thermal goo.

This, https://www.digikey.com/product-detail/en/S606B-30/1168-1380-ND/3042209 is heat sink grease. Radio Shack sold a very small tube of white grease. Don't get it on your hands. it's difficult to remove. A very thin film is required.

SOME insulators don't require it.

The TO-3 package needs a mounting kit such as this one https://www.digikey.com/product-detail/en/4725/4725K-ND/27407 ou also need a ring terminal to attach a wire to the case using this kit.

Here is a TO-3 socket https://www.digikey.com/product-detail/en/4606/4606K-ND/27393 but it doesn;t come with the insulating washer.

In any case, USUALLY the reguator has to be electrically isolated from the heatsink.

Unregulated, means that there will be 120 Hz ripple, assuming a 60 Hz power line frequency that depends on your load. DC means flat. there will ALWAYS be some ripple.

There is no such thing as a "flat sine wave". Flattened or clipped, maybe.

Some search terms:
thermal grease
to-3 mounting kit/socket
t0-220 mounting kit
to-220/To-3 thermal insulator
to-220/to-3 heatsink

The slash meaning different searches "to-3 mounting kit" or "to-220 mounting kit"

Digikey has a free schematic drawing thingy which I can't find. You can also use www.webpcb.com as well which references digikey part numbers.

 

[Mike86H]

I made up a schematic that seemed to make sense to me, using the power source from the old "toaster" power supply and then adding on RonV's circuit on the end. Im going to post it up and see what yall think.

Pardon the fact that there are no real values for components, Im still trying to figure those outs.

**broken link removed**

 

[KeepItSimpleStupid]

That's kinda funny. use a rectangle for the regulator, In, Out, adj.
I'd remove the parallel combo of R3 and the pot and replace with a resistor A + pot + resistor B + ground

We can then figure out the aprox values to give you a min/max range.

Protection diodes should go like fig 24, here: https://www.google.com/url?sa=t&rct...DDQ9WiO3yBOVPLA&bvm=bv.50500085,d.aWc&cad=rja

Do, you want the gizmo to use 1.25 (off) and Vout(on)? We can put an optocoupler there, a few resistors and a current source, e.g. https://www.digikey.com/product-detail/en/4N27/4N27VS-ND/1738518 and this https://www.digikey.com/product-detail/en/LM334Z/NOPB/LM334ZNS/NOPB-ND/6215

You can add it, and then decide later what version to use. Switched AC or switched DC.

You can add attachments to the forum using "go advanced", "manage attachments"

 

[ronv]

This heat sink would be better.

https://www.amazon.com/Large-Rectan...ew/dp/B00BNR4122/ref=pd_sim_sbs_petsupplies_1

It will still get quite hot at 7 volts at 2 amps, but if you bolt it to your metal box that will also help.

Lets also change to the 16 volt transformer (PN in schematic). This will drop the power dissipation a bit which will cool thing down. There may be a little ripple get thru the regulator with this one when the voltage and current are at maximum.

Kiss, the regulator we were looking at is in the TO-220 package so it doesn't need all the TO-3 stuff.
The values for the adjustment resistors in the schematic should give a range of 1.25 to 18 volts. The parallel resistor keeps you from setting the voltage higher than 18 volts with a 2K pot.

The simulation shows the "ripple" and the DC output from the regulator at a nominal load.

 

[KeepItSimpleStupid]

OK, the 2A makes the TO220 package OK in my book. I was also looking at the other way of doing a range. 6 in one and a half dozen in the other.

What's your take on the foot petal? 1.25/Vout or AC switched?

 

[ronv]

I think the foot pedal is just a switch to turn the pen on and off. Not sure if the smarts are in there or not.

I lost track of the silver guy. Sad story.

 

[Mike86H]

Hey gentleman, I wanted to say thanks again for all of yall's help. I havent really been able to keep up with this over the weekend due to my sons birthday.

As far as the footswitch, yeah its just simply a means for forwarding the power on to my machines. I actually found a couple of pictures this evening of the internals of some cheaper power supply units that honestly seem to just be the power section followed by a voltage regulator. They seem a bit poorly designed, and thats probably why they fall apart so fast, but it seems similar to some of the schematics yall have posted.

I still havent started buying any parts yet as Im still trying to wrap my head around all of this cause I really dont want to have to build a half dozen or so before I perfect it. But if yall would like, I could post the pictures of premade internals, including the power supply I have sitting here, and maybe attempt to go off those, or yall can tell me how close to the schematics posted the PS units are already.

 

[ronv]

Can't hurt. It would be good to make sure we have the right current at the same time.