LM317 voltage regulator breakout kit

[reagle]

I have a simple LM317 through-hole voltage regulator kit for sale on tindie. IT's for those cases when you get tired of building it once again on a perfboard and need a bit more permanent solution

https://tindie.com/kuzyatech/lm317-power-supply-kit/

Kit contents:


Assembled look

[reagle]

Three left !

[reagle]

Actually that was by design. I've seen too many variable things like that fry downstream circuits.
You could also just solder two single pin female headers in and make it a socket to change resistors in, I suppose

You could have used a pot for R2. Particularly if the intent is versatility and reuseability when breadboarding low power stuff.

[Zero999]

How can you mount it? There are no holes on the PCB for screws.

Have you considered doing a pre-assembled version with SMT for everything except for the LM317 and LED? Both values for R2 could be fitted but optionally bypassed by jumpers giving the choice between 1.25V, 3.3V, 5V and 8.3V.

[reagle]

Added to the wish list for the next version - thanks

[ptricks]

Next design I would use a design like this, covers all the bases.

http://www.ladyada.net/library/equipt/diypsupp.html

[Zero999]

he didn't want to use a potentiometer because they're a pain to use.

Another option could be to use a DIP switch sorting out several resistors to give different voltages.

[reagle]

Well, it's only been like 8 months, but I finally got around to updating this design 

Thanks for the suggestions, most of them made their way in:
http://kuzyatech.com/wp-content/uploads/2013/06/SCH_-201306_LM317ADJ_Reg.pdf

The kit is up on Tindie at
https://www.tindie.com/products/kuzyatech/new-version-adjustable-lm317-regulator-kit/

Changes:
1. Add mounting holes- a biggie! There are now three mounting holes sized for 4-40 screws (3.2mm diameter)
2. Move LM317 regulator to board edge, so it can be mounted on a heatsink
3. Add ability to adjust voltage: you now get a choice to either use an included 5K potentiometer for adjustable voltage or install a resistor instead to get fixed output. The values for 3.3V, 5V and 8V are still included
4. Add protection diode to handle shorts on the input.
5. Add cap to the feedback pin (together with a protection diode) to lower output ripple if desired as per appnotes.

[mariush]

I can't say that I like it. It looks too bulky and too... like you threw the components on the board, visually it doesn't look nice to me.

A cheap heatsink is maybe 20 cents, screw and washer is another 5-10 cents. I would add them to the package. See
http://uk.farnell.com/aavid-thermalloy/507302b00000g/heatsink-to-220-24-c-w/dp/1611415
http://uk.farnell.com/multicomp/mc33278/heatsink-to220-x-2-7-6-c-w/dp/1710623

I'd add some reverse voltage protection - you can't rely on people not connecting the input the wrong way... i'd consider using a p channel mosfet so that there won't be such a big drop at input. see    though p channel fets in through hole variety are a bit expensive

The output pins are in the center of the board - if I'd want to put it on a breadboard it would basically use up space both above the breadboard, and on the side of the breadboard... more clutter on the desk. Just picture it.

The green led is ugly as hell. Couldn't you find a rectangular or smaller led that would not take so much space? Is that led even lighting up when you set the output to 1.5-1.8v, considering it's a 2-2.2v forward voltage led (green usually is around that value) and you use a 3900 ohm resistor? That's what... 0.5 mA?
I'd use something like this http://uk.farnell.com/avago-technologies/hlmp-s201/led-rect-he-red/dp/1003236 .. 2.54mm pitch, small space on board, 1.9 forward voltage so would work even at low voltages..
It would be nice to have a USB jack on the board in a corner, jacks that have through hole pins are less than half a dollar, even less in quantity - a usb jack would allow people to power the board from a usb charger/pc etc so no extra investment in separate supply. Though since the Lm317 is used, there's little point in because of the voltage drop on it ... maybe if you use a lm1117 instead or something with even less voltage drop... I don't know.

[reagle]

Thanks for taking a look, I appreciate it, though I can't say I agree with you - it looks pretty to me
Considering it took me this long to update the design, if I wait for perfection it will never ship.
Sure, there is a bunch of other things one can add, but you gotta stop at some point. If you make the thing un-killable for example, some kids will never learn to pay attention or what the magic smoke smells like

[c4757p]

I agree, I don't think it needs reverse protection. Reverse protection is for entire devices, not components - it's the job of the guy assembling it to connect it right. What's next, electrolytic caps with embedded protection diodes?

Can I make one suggestion - one that should be an easy fix? Technically, the 240 ohm resistor they give in the datasheet is wrong - the LM317 theoretically has a minimum load of 10 mA, and 240 ohms across the 1.25V reference doesn't quite meet that. (240 is actually chosen for the LM117 with half the load requirement.) It's usually fine, but if you get a marginal batch of LM317s you may find some don't regulate very well if they're not otherwise loaded. (And they do the worst possible thing when that happens - the output goes high.) I'd halve both resistors if I were you.

[reagle]

Can I make one suggestion - one that should be an easy fix? Technically, the 240 ohm resistor they give in the datasheet is wrong - the LM317 theoretically has a minimum load of 10 mA, and 240 ohms across the 1.25V reference doesn't quite meet that. (240 is actually chosen for the LM117 with half the load requirement.) It's usually fine, but if you get a marginal batch of LM317s you may find some don't regulate very well if they're not otherwise loaded. (And they do the worst possible thing when that happens - the output goes high.) I'd halve both resistors if I were you.

Not a bad idea.  I could  also bump up LED current. Though the difficulty there is- without knowing what the end user will set output to, you don't know if the poor beast will be under or overdriven 

[c4757p]

With one extra transistor you can drive the LED with a constant current, regardless of output voltage. Hold on, I'll throw a schematic together.

[c4757p]

The trace is the LED current, and output voltage is on the horizontal axis. You can see that once the output hits around 4.5V, the LED current peaks at about 8mA and stays there no matter how high the output voltage goes. R3 sets the current - roughly (1.25V - 0.65V) / R3 = I(D1).

It's an emitter follower. It takes the 1.25V reference voltage of the LM317 and puts it across R3 (less the 0.65V threshold of the transistor). That makes (1.25V - 0.65V)/R3 flow through the resistor (Ohm's law), and the only place for it to go is through the LED. There will be a small brightness variation in real life, because as the output voltage rises the transistor will get warm and its characteristics will change, but this will be minimal.

By the way - in this example the LED current plummets to zero around 3V, but that's just because the LED model I chose in the simulator has a very high threshold voltage.

[mariush]

There are led drivers like this one, which limits led to 10 mA no matter what voltage and don't take up as much space as a transistor (not to mention you won't have problems with hFe/beta and differences between transistors) but it's not through hole:

http://uk.farnell.com/on-semiconductor/nsi50010yt1g/ic-led-driver-50v-0-01a-sod123/dp/1794979RL

[c4757p]

One obvious advantage of that is that the PCB layout doesn't have to be rearranged at all, it just directly replaces the resistor. I'm still a big fan of dirt-cheap jelly bean parts like transistors, though, and I don't think those problems really are going to matter when it's just driving an LED. But I do like it.

[reagle]

The other benefit, if we consider this to be a learning kit, is that we illustrate a few concepts like using transistor in this configuration.
If we use a CC chip, it just works,  but does not really teach much to a newbie

[edavid]

Look what you can get from China for $7:

http://www.ebay.com/itm/PX-1011-LM317-Adjustable-Power-Supply-Module-with-Voltage-Meter-/220965444339

It has a built in DVM!

[Bored@Work]

It has a built in DVM!

If I am not mistaken an autoranging DVM.

[mariush]

It's possible to do it in 4-5$, but in volume... the biggest cost reagle has is probably the 3-5$ pcb , probably made by laen at http://oshpark.com/.

If I can order  50 panels of double sided 300x250mm for 1350 euro at eurocircuits.. that ebay pcb is 70x55 so there's about 18 boards per panel, 900 pcbs in total.. so about 1.5 euro / 2$ per pcb (21% vat included, but as a company I wouldn't pay vat so take off about 250 euro and we're down to about 1.2 euro per pcb) ... and this is not cheap china pcb manufacturer.... they can probably get it down to less than $0.5

The seven segment display is some overstock/extra from a custom order (it's not regular seven segment with dots, it looks like it's from some type of clock display) so it was super cheap, i saw some of those on digikey from time to time for under 1-2$ ... the potentiometer is probably 0.5$ at 1000pcs or something like that... the ic is a plain microcontroller probably not even programmable (programmed in factory) that uses some voltage dividers to get the voltage... resistors and caps are a dime a dozen...

[reagle]

Yep, these are very low volume, using oshpark made boards. There is no way I can compete with China. I could probably knock a few bucks off by going to itead, but I'd rather stay with US fab. Plus their quality is much nicer