Wat wo waggie aka How NOT to build a tattoo power supply

[AE7OO]

Greetings,

This thing was brought to me by one of the daughters former boyfriends who did tattoos, with the complaint it would not turn on.  I said I would take a look at it.

The only labeling on is on the front:

Code: [Select]

INPUT AC 60V-250V 50Hz-60Hz
-- display here --
OUTPUT DC 1.5V-16V 2A
No manufacture, no serial, no FCC, no CE, etc.
It is a simple design, power switch on the back, front jacks for foot peddle and gun leads, voltage display and voltage control. 

Code: [Select]

And just for reference, the 3 active devices in this thing are:
THX203H - PWM switching controller
AMS 817C - Optocoupler
YD317 - Voltage Regulator
Pictures 7, 8 and 9 show the devices.  Picture 8 also shows off that snazzy "filter cap".  Please note, none of the filter caps are agency rated, they are generic 1KV ceramic caps.
I'm still trying to figure out why almost all tattoo power supplies out there take mains instead of using a wall-wart.   19V wall warts are not a rare item. In this case, as shown in the 2nd picture, it takes a simple 2 conductor mains plug, which told me right off the bat I was not going to like what I found.  No 3 wire, no earth/safety ground. 

First off I took the front panel off, the attached parts are shown in the 3rd picture.  The only surprise was find a 2W rheostat as the control for the voltage.  Also, for whatever reason, the maker of the voltage display sanded the IC.     These things are just about a commodity item, not a super secret design.

After taking the back panel off, as shown in the 5th picture, I found more surprises.  Now some of this stuff maybe me nit-picking too much(meaning some of it is up to the designer), but my spidey senses were getting a workout.  To a small degree it is based on the reference design from the datasheet, but to be truthful I'm not sure it was an engineer(of ANY type) who designed this thing. 
It looks like what they do is set the 317 output voltage to ~16V and use the rheostat to control output.  Is there an inherit reason to set the output this way instead of varying the adjustment voltage?  If I was worried about the output being uncontrolled, for safety reasons, I would have used a 16V or 18V zener in a crowbar with a fuse.  If that rheostat shorts, for whatever reason (how about an atypical liquid spill), you'll get that 17V output until something fries. 
Also, while most manufacturers show current output as greater than 1.5A, it seems hard to find a legitimate version rated for 2A output.  I guess they are depending on the intermittent usage of a tattoo gun.   I did not fire up the analyzer, but I've got no doubt this thing spews garbage everywhere.  How could it not?  There are no filters anywhere, which would be bad enough when using a SMPS, then when you take into account what the output waveform looks like, makes you want to cry.  A coil gun would be the worst, but even a rotary is going to be nasty.

This thing was a contradiction from the get go.  The 6th picture shows good soldering and static creepage distances.  But then there are the WTF items or lack thereof.
First off, outside of the transformer, there was not an inductor in sight.  Next, the majority of the caps were small in one way or another;22uF instead of 33uF for the input cap, using a 50V cap in the HV area, etc. Then we have the thin hookup wire, small diodes(Yes, they are 1N4007's, but still...), a glass fuse, a total of 940uF for the combination of smoothing/regulator input, and a whopping 10uF as the 317 output cap(which has to get hammered, with two coils on the common tattoo gun).  And here is a nice one, the 317's tab is output, in this case they did not use an insulator so the heat sink is live at base output voltage.  There is also no protection against a short circuit except either thermal shutdown or rheostat failure.

And now to the part that made me glade that this thing had failed when it did.  Take a look at picture 10 please.  As I could not find any of rulers, I used a common .1 header to show scale.
You also need to know this board slides into the case on rails that are .1 deep with almost zero clearance side to side.  Then throw in the fact the clearance from board bottom to case is .11, not counting any projecting pins.  North American 115V would be bad enough, I don't want to think about 250V usage.  Correct me if I'm wrong, but the whole thing sounds like a disaster waiting to happen.

After seeing the how this thing was designed and constructed I told him that it was trashed and not worth repairing.  While the design obviously works, there are just too many corners cut for me to feel comfortable doing the repair. Then when you take into account that they cost peanuts to buy, it just was not worth my time and effort to go any further.

Instead what I did was whip up a 317 design with variable voltage via the adj pin, with over voltage/short circuit protection via 18V zener and PTCs.  The input voltage was supplied by an old Dell 19V/4A laptop power supply(it had a blown auth chip and so would not charge a battery).  I sprayed various ferrite beads at obvious places for the hell of it. It has 1410uF/50V(470uF x 3) on the input and the same on the output.  I reused the on-off switch, voltage display(it was close enough), the two jacks and the case.  Everything is set back 1/4", even though my max voltage is 19V. Oh, the heat sink is no longer live(insulator and grease).  I could have added a below 0V reference(via charge pump or reversed boost) to allow for 0V - 16V output, but why(He asked me about it, go figure. Foot peddle anyone?)? Plus I'm lazy...  The entire thing, including creating a PCB took me maybe 2 hours.

It can not compeate with the Chinese specials, but I'm willing to bet it's much safer and will last almost forever.  Used Dell power supplies are cheap(especially if they have a blown auth chip..)

To me this thing is scary.  And I can not imagine that they are actually legal for sale.  With no filters, that power cord is going to act like an antenna, to say the least.

[tooki]

If you ask me, most tattoo power supplies are sketchy as hell. There are some good vendors, but most look surprisingly DIY and old-school.

I suspect this is a legacy of tattoo equipment originally being all DIY by artists, few of whom would have had any real understanding of electronics. And of course, once the "standard" in an industry is established, it tends to stick around...

This example is, of course, Chinese crap. It probably isn't legal for sale, which of course doesn't stop them from making their way to USA anyway.

[Mechatrommer]

I sprayed various ferrite beads at obvious places for the hell of it.

for what?

It has 1410uF/50V(470uF x 3) on the input and the same on the output.

for what?

if somebody is going to use your LM317 psu at max rated current 1-2A, i suggest be ready with some LM317 stock in case if they come back. dont you want to see whats inside your Dell psu?

[floobydust]

None of us could make unapproved mains-powered equipment and sell it in another country. Surprised there hasn't been a crackdown on this, Trump could really cause more trade drama in a good way banning this junk.
Tatoos are pretty much taboo in china so they don't actually use these power supplies and machines.
 
The tattoo gun machine is like a telephone ringer, a couple solenoids as a mechanical oscillator but extra resonance electrolytic 22-47uF. The contacts arc a fair bit, so I hope the LM317 has the three protection diodes. The big output capacitance is needed.

[Mr. Scram]

None of us could make unapproved mains-powered equipment and sell it in another country. Surprised there hasn't been a crackdown on this, Trump could really cause more trade drama in a good way banning this junk.
Tatoos are pretty much taboo in china so they don't actually use these power supplies and machines.
 
The tattoo gun machine is like a telephone ringer, a couple solenoids as a mechanical oscillator but extra resonance electrolytic 22-47uF. The contacts arc a fair bit, so I hope the LM317 has the three protection diodes. The big output capacitance is needed.

Pay crap, buy crap. Most Chinese manufacturers cater to the market. If there's a buyer, it will be built.

[AE7OO]

Sorry about the late reply, I've been busy.

Spray was really the wrong word and they are not the little beads one would think.

The entire supply was made from parts from the junk bin.  On the 19VDC in, there are two threaded ferrites(picture #2) (plus a choke & cap combo(from some SMPS)) to keep the hash OUT of the DC power supply, and the scope confirms it(Unless it's throwing junk that my 60Mhz analog scope can't see, which I doubt).   Then for the two jacks I used ferrite barrels(picture #1).  The reason for using them was to cut down EMI on the cords as much as I could.  I realize that the gun is going to act almost like a broadband jammer, but there is not much I can do about that.  The caps are there to act as a reserve in and out.
I treated the gun jack as an inductor with a 1N4007(I had bought a 50 pack years ago) diode to keep any possible reverse current spikes out of the 317.

I ran it into a dummy load (an old radio one that I had modified to also take DC in addition to the N connector already on it, 2 big 50 Ohm wire wound resistors in a paint can filled with mineral oil) and adjusted it for 1.5A max.  After 10 min the 317 was hotish(but not HOT) so I was happy with the way it was heatsinked.   Oh, I guess I should mention that it was the HV version of the 317.

Sorry I had not gone in to further detail on how it was put together in the original post.

NOTE: The pictures are from other projects, but they are representative of what went into that power supply.  The kids have since broken up and he and it are long gone...