I'm currently building this:


And I want to include an LED inline that will light when the switch is in the on position. Where is the best place to attach the LED leads?
I know its probably obvious, but this is my first foray into soldering, and wiring of this type.
I'm more used to romex and junction boxes.


Thanks!

That is a sucky schematic...It isn't at all clear which position on the switch is 'on', or whether the 12V is the supply voltage, or an output voltage. Can you give us more background on this, ie where the circuit came from, and what it's used for? It'd probably help lots.

For starters though, it is generally a good idea to tie that 'not used' trim pot connection to the middle (ground) connection. The theory behind that generalisation is that when trim pots fail, they usually fail because the middle connection fails. As the circuit stands when that happens the trimpot in effect becomes an open circuit, with the left leg of the regulator floating, ie unpredictable. With the 'not used' connection tied to the middle, when the pot fails it will fail to it's maximum resistance. That still may not be a good thing, but tends to be a more graceful degradation. Again, more background would help to clarify that.

I agree with genixias comments.

I assume the +12V line top right is input fed from car battery - ie anything from 9-15V from when you crank the starter to when the alternator is charging fully...

So that means that you should have a regulator with lower output than 12V, and that you can adjust its output voltage somewhat by using the pot.

A SPDT switch doesn't have one "on" position - it has two. In one position the center connector is connected to one of the other connectors and in the other position the center connector is attached to the other of the "other" connectors. Note that "center" refers to logic, not neccessary physical position.

So going strictly by how the switch is drawn you should have either the "from original harness" or the adjusted voltage from the regulator goint to the common "MAP" wire depending on switch position.

Depending on the type of signal that is on the "from original harness" it may be impossible to distinguish switch position easily. Simplest way to get a switch position indicator LED going is probably to change the switch to a DPDT switch instead. In practice it is 2 parallell, commonly activated but electrically isolated SPDT switches - ie has 2x3 connectors.

Use one of the sets to switch the regulator output/from original harness/MAP wires and use the other set to switch power to the LED. A typical LED will in car use work OK with a 1kOhm series resistor. The generic formula for series resistor is (V_supply - V_ledforward) / I_ledforward.
ie the required voltage drop over the resistor (supply voltage minus the forward voltage drop over the LED) divided by the LEDs current draw at the voltage in question. Supply voltage is a bit tricky in a car - as long as you don't use the max parameters for the LED, 13V is a useful average considering engine off or on. Also check power dissipation in the resistor - P=(V_supply - V_ledforward) * I_ledforward. Here it's a good idea to use 14-14.4V as supply voltage (worst case...). Then get suitably sized resistor.

/Michael

Whoah you guys, thats a lot of information for my tiny brain.

Heres some background info.
What I'm trying to build is an Adjustable Map (Manifold Air Pressure) Sensor for my Jeep. The map sensor calibrates the fuel delivery to the engine. I have some engine mods, and am running a little lean. So the solution is to change the voltage slightly thats supplied to the stock map sensor, this in turn witll supply more fuel to the engine.
The idea (I think) is to install this module inline to the stock map sensor, and then you can switch it to active or inactive. That way if it fails you wouldnt get stranded.

I could buy one for a little over a hundred bucks, but somebody pointed me to some websites that (sort of) illustrate how to build one myself. Parts were about $20.
Figured it might be kind of fun, and I dont really have much experience with this kind of work, so it might be good practice (or a complete disaster).
Here are the links to the websites:

This is the original site

And this site sort of expands on the original.
This is where I got that diagram.

I already bought all the parts.
This is what I got from my friendly neighborhood Radio Shack:

Multi-turn Cermet Potentionmeter 1000 ohm
LM317T Adjustable Voltage Regulator
220 ohm metal film resistor
Circuit Board
Heatsink for Regulator
Housing for Project
SPDT toggle or rocker switch
SubMini LED Lamp

Does that fill in any gaps?
Thanks for all the info, and any other bits you think might help.

Have I sufficiently illustrated how clueless I am yet?

OK - makes sense.

Get the DTDP switch I suggested.

Run the wiring in your original diagram to one set of contacts and then wire the LED on the other set. Hook the LED up to the center and same side pin as the input from the voltage regulator is on the other set. r=1kOhm is probably good for most normal LEDs (non-highintensity or low-power)



As to connecting the unused pin on the pot to the grounded pin depends on what happens worstcase if the slider disconnects from the resistive path. Is it worse if there's no connection to ground or worse if there's a high resistance path? Probably depends on what the voltage regulator does if you disconnect ground - I don't know but guess you'd get 0V out. If you did the extra connection you will get a higher than wanted voltage (don't feel like looking up the LM317 spec sheet and calculate just how high the voltage would go, but probably 12+ V)

Edit:

Looked at the original site you linked.

Voltage out is Vout = 1.25V*(1+R1/R2) where R1 is 220 Ohm
and R2 0-1000 Ohm.

If you included the extra connection, the failure of the slider to connect would give 1.25*(1+220/1000) which is 1.525V.
Leaving it out would probably lead to 0V out. Compared to the nominal 5V the original circuit had, both might be effectively just as bad... but that's just speculation.

/Michael

Hmm.

Since;
Vout = 1.25V*(1+R1/R2) where R1 is 220 Ohm,
and we want Vout ~6V, let's look at the value of R2;

R2 = R1 / ( (Vo/1.25) -1)
= 58 ohms.

So why are we using a 1Kohm pot? It doesn't provide much resolution for adjustment either.

Let's assume that we're going to tie the unconnected end of the pot to the middle leg to make it fail safe at Rmax. When that happens, we presumably want Vout=5V (to match stock voltage - mildly lean, not dangerous).
That gives R2=73 ohms. So if we use a 75ohm pot then all should be dandy - good resolution, and good fail-safe.

Here's the bummer, Ratshack only sell 10K or 1K pots. And since Fusto's already bought a 1K pot, looking for a 75ohm pot is going to be a pain. Furthermore, we might start hitting power dissippation issues with those low values.

So let's keep the 1K pot, and change R1 to make the pot failsafe and gain resolution;

For failsafe, Vout=5V, R2=1Kohm.
R1=( (Vo/1.25) - 1) * R2.
R1=( (5/1.25) -1) * R2.
R1= 3Kohm.

That then gives R2 for Vout=6;
R2= R1 / ( (Vo/1.25) -1)
R2= 3K / 3.8
R2= ~790 ohms. Perfect.

Now your next challenge is finding a 3K resistor. Ratshack don't sell them.
You could just use a 3.3Kohm resistor instead.
This would give a failsafe Vo= 5.375ohm, ie a little richer than stock. This will probably be perfect, assuming that your ideal mixture is indeed around the 6V mark. However, if for some reason you need 5.2V for an ideal mixture, be aware that using the 3.3Kohm resistor would preclude this. I think that it's worth pulling the gamble handle though. You can always change the resistance afterwards.

So to recap;
Use Mike's circuit. Tie the unconnected and middle connections of the 1Kohm pot together, and use 3.3Kohm resistor instead of a 220ohm.

Good catch!

Another thing:

Do you (fusto) know how much current the input this is connected to draws?
Max current rating shouldn't be a problem, but the minimum (for good regulation) might be...
Worst case requires 10mA (typical 3.5mA) To be safe you might want to put a resistor between the output and ground, so that ~5 mA always flow (ie 1kOhm resistor). To cover the worst case, use a 520 Ohm resistor instead.

To increase stability and transient response (don't know how sensitive the MAP input is to small fast changes) you might also want to add a 0.1uF cap on the input and a 10uF on the output (both to ground).

Spec sheet and application notes for the LM317.

Edit: In your parts listing you say subminiture LED lamp. Is that a LED with its own resistor, so that you can drive it, as bought, direct from 12V? If so, skip the 1kOhm resistor for the LED in the drawing - it's only required if you had a raw LED.


/Michael

OK...
So, I'm using Mike's revised schematic, I'm upping the 220ohm resistor to a 3.3Kohm, and I'm tying the unused leg on the pot to its middle leg.

To be even more safe I could also:
put a 520 ohm resistor between the output and ground.
And add a 0.1uF capacitor on the input and a 10uF capacitor on the output (both to ground).

Have I got that right?
Do I even know what that means?

Do you (fusto) know how much current the input this is connected to draws?

I'm not sure what current the input draws.
I did get an LED which has a resistor built into it.

Heres a revised diagram. New changes are in red.
Am I on the right track?
Have I got a clue?

Oops, mistake in my initial reply...

The caps aren't needed on the common MAP wire - you don't want them involved when you're in "stock" mode on the switch. Same for the 520 Ohm resistor, they should bot attach between the LM317 and the switch.

The caps don't go "inline" on the wires.

One end of the 10uF cap attaches to the point where the 3300 Ohm resistor connects to teh LM317 and the other to ground. The 520 Ohm resistor connects to the exact same points.

One end of the 0.1uF cap attaches to the +12 line, close to the LM317, the other end to ground.

/Michael

Revised diagram

To cover the worst case, use a 520 Ohm resistor instead.

Just went to Radio Shack and they didnt have any 520 ohm resistors. They only had 470 or 560.
Could I use either of those instead?

Sure, 560 should be OK, go with the 470 if you want to cover that last 0.001% or so...

Hmmm, I could have sworn 52[0...] was one of the standard values, but obviously I was wrong...

/Michael

What would be a good way to test this?
I mean other than hacking it into the wire harness for my car and seeing what blows up.

I have a fairly good multimeter. Dont know what most of the settings are for. I mostly use it for car/motorcycle and house stuff.
Can I use it to test my end product?

So far so good with the construction. Although my soldering leaves a lot to be desired.
Although considering I've never done it before, (except for plumbing) I think Ive done pretty good so far.

Heres the backside.

Set the pot somewhere in the middle of its range

Apply 12V to input (and ground) - battery or AC/DC power adaptor.

If nothing goes up in smoke it's a fairly good sign

Put multimeter on the closest higher DC Volt setting (might be shown as V=) than the voltage you intend to measure. In this case a setting of like 20VDC (ranges vary betwen meters, but you get the idea) is good.

If you measure between the the output and adjust pins on the regulator (or any thing directly connected to them) you should have 1.25V (the internal adjustment voltage used by the regulator)

If you measure between ground and the MAP wire (with switch in appropriate position) you should get a reading somewhere around 5V. Adjust the pot and see if/how the reading changes. Adjust until you have a reading you think your car will be happy/at least functional with...

Temporarily hook up another resistor (100-200 Ohm or so) in parallell with the 520 Ohm one
(adds a small load) and check that the voltage remains the same (the regulator regulates properly).

Flip the switch and you should get 0V (unless you have the "from original wiring" wire hooked up to something)

For the record heres how I wired it.

I broke the connections down into 4 groups. I then soldered the wires into these groups on the circuit board.

This is the group breakdown:

Group 1:
.1uf Cap
10UF cap
470 res
mid leg pot
ground out

Group 2:
last leg pot
3.3k res
left leg VR

Group 3:
10UF cap
mid leg VR
3.3k res
last leg switch side A
470 res

Group 4:
.1uf cap
Right leg VR
12V in
last leg switch side B

Other connections are:
Red LED wire to mid leg switch side B
Black LED to Ground
12V to Last leg switch side B
Out to map to second leg switch side A
In from map to third leg switch side A
first leg connected to mid leg on pot

I think thats it.
Does that sound correct so far?

(made some corrections)
corrected LED hookups

Hmmm, the LED hookup sound spurious - red should be centre pin side b, and black should be connected to ground. +12V should be wired to last pin side b on switch

Otherwise I can't (after a very quick look over) see anything off...

-------------
ooops ignore the first edit - misinterpreted what you said

/Michael

Yes, you can use your multimeter...but firstly;

At least one of your capacitors is an electrolytic capacitor. Electrolytics are polarised, ie it matters which way around they are. I hope (and think from what I can see in the photos) that you noticed the grey stripe with the '-' sign on it. The lead nearest this stripe needs to be at the ground connection. If this is reversed, then it will probably explode - probably wouldn't damage anything besides itself, but it would make a nice bang and stink. One other thing to check on the electrolytic - they are voltage rated too. As long as it's 16V or above you're fine. It's probable that this is the case, but check anyway. Typically the same result if it's wrong. I think that the second cap is unpolarised, although the photo isn't clear. If it is polarised, then it should be clearly marked like the electrolytic is.

After that's been checked;
Verify that the regulator's pins are connected correctly. There should have been a pinout on the package. Note that the tab is not electrically isolated, and is connected to Vout (middle pin), and not ground. That means that you have to ensure that it is eventually isolated from the car chassis somehow. For the time being, just make sure it doesn't touch anything during testing.

Set the multimeter to measure resistance, measure across the pot connections and roughly set the pot to 870ohms.

Look to see if your multimeter has a continuity test. This is usually shown by a buzzer symbol.Test the mode by touching the probe ends together - you should hear a buzz.
1) Move the switch to the 'stock' position. Test that the MAP connection and the 'stock wire' connection are connected together (buzz)
2).. and that neither are connected to Vout (the middle leg of the regulator) (no buzz).
3) Move the switch to the 'tuned' position, and now verify that the MAP connection is electrically connected to Vout,
4) and that neither are connected to the 'stock wire' connection.
5) (Still 'tuned') Now check the LED is connected to the 12V connection.
6) Move to 'stock'. Check the LED is not connected to the 12V connection.
7) Change the multimeter to resistance mode again. Measure the resistance between Vout (middle leg) and Ground. It should be ~490ohms. If it's out by more than 5% then verify the pot resistance again.

Assuming all is good up to now;
You ideally need to find a 12V source somewhere to test this. Ideally it will be protected by a fast acting .5A fuse, but we don't live in a land of ideals.
I'd use an old PSU from an 'AT' type PC. Actually, I lie, I bought a new 'AT' PSU as a 12V supply! PC power supplies are useful as cheap power supplies for us geeks because they can typically supply large currents at 5V and 12V, both of which are very useful. An 'AT' type is more useful than ATX because it doesn't require any hacking to get it to turn on (ATX expect a motherboard to be there).
Actually, as it stands, you should be able to test it with a 9V battery, which should be nice and safe. Note that we are at the limit of the regulator here - it really wants >3V difference between Vin and Vout, but as long as the current draw is low (which it should be) then it should work.
Anyhoo, find a 9-12V test source, and connect the 12V and ground wires appropriately.

8) If there is a sudden burning smell, or you notice anything getting really hot, or melting, or going bang, disconnect immediately. (Goes without saying really )
9) Set the multimeter to measure DC Volts. Measure Vout (with respect the Ground). It should be very close to 6V. Verify that tweaking the pot does indeed change the voltage, and then set it to be whatever the Jeep enthusiasts deem to be ideal. ( ~6V IIRC..)
10) Verify the LED works If it doesn't, but all the previous tests passed, then the LED is probably wired in reverse.
11) Disconnect the 12V wire. Change the multimeter to measure DC Current (mA). You'll possibly need to move the probe to a different hole in the multimeter as well. Now connect the multimeter between the 9/12V source and the 12V wire, ie;
9V/12V -> MM -> Circuit -> Ground.
Measure the current drawn with the switch in 'stock' mode. It should be about 12mA, give or take.
12) Measure with the switch in 'tuned' mode. It should be about 30mA, give or take. (This depends heavily on the LED specs. As long as we're not over 40mA, don't worry. Otherwise post the specs, so we can confirm that it isn't a circuit fault).

If you've got this far, then it looks like it's all good. It's probably a good idea to use an ignition-switched 12V supply to avoid that 10-40mA drain on the battery when the car is off, but not absolutely essential unless you leave your car for weeks at a time. Remember to ensure that the regulator tab remains isolated from the car chassis. Reliability of homebuilt circuits really relies on the final finishing touches - it will last much longer if put in a plastic case and components are suitably affixed to prevent vibration damage.

hmm I think your right.
I think I do have the LED hooked up incorrectly.
Let me try that again...

**A little later on...**
I did have the LED hooked up wrong. I dont know what I was thinking...
I now have the black going to ground, and the red going to mid terminal B side of the switch.
The last terminal on the B side is hooked up to 12V.

Whoo...
Ill give all that a try in the AM.

I did notice that the capacitor that I got was polarized. I wasnt sure what to do until I noticed that mike was nice enough to put + and - for the cap in his revised diagram.
I probably would have figured it out eventually, and hooked the - side up to the ground. Maybe after I blew one up first though.
The packaging for the eloctrylitic cap doesnt say specifically about voltage. Unless this is it:
35WVDC maximum
On the back it says:
Op temp: -40 to +85 C
tolerance +-20%
Lead spacing 2.0mm

Thats pretty much it though.
Its Rat shack part number 272-1025

35WVDC maximum...it's fine.

God i love this BBS. This isn't even my project, but just reading through all the help you guys are giving is GREAT stuff. You all realize that given our collective knowledge, we could take over the world right?

Heck, I'd be happy with just my little corner of it

Actually, it really bothers me that this circuit would even work to begin with. The concept that a vehicles A/F mixture is reliant on a sensor that is so dependant on its supply voltage is abhorent. The environment is depending on that sensor. Its output should be only dependant on the air pressure, and independant of small supply voltage changes, as that is the only way that we can have absolute confidence in its readings.

But that doesn't detract from the enjoyment of a challenge, and at least Fusto can take advantage of it.

The concept that a vehicles A/F mixture is reliant on a sensor that is so dependant on its supply voltage is abhorent.

Welcome to the wonderful world of Chrysler engineering Hey Fusto, want to build me one of these for my Jeep once you get this all worked out

-Mike

Sure, why not.
Its kind-a fun.

And the more I practice my soldering the closer I am to doing those damn tiny LED's.

OK, everything checks out so far in terms of connections. The only steps I ran into problems with were these...

Set the multimeter to measure resistance, measure across the pot connections and roughly set the pot to 870ohms.

The highest I could get the pot was 763. The screw reached the end of its travel and just clicked...

If there is a sudden burning smell, or you notice anything getting really hot, or melting, or going bang, disconnect immediately.

Nothing blew up and nothing smells funny, but the heatsink I connected to the regulator got really hot. Like burn your fingers when you touch it hot. So I disconnected. I had it hooked up to a car battery charger that was outputting about 13.98V. I also tried it with a 9V battery, and it got kinda hot that way too, but not as much.
Maybe I've got the regulator hooked up incorrectly?
See edit in next post

The switch works fine, and the LED works fine.
This is the potentiometer I got and this is the regulator.

Maybe I've got the regulator hooked up incorrectly?

Actually, in looking at the pinout on the package (which is a terrible diagram by the way) I think I may have done just that...

(edit)
Yeah that was it.

But still the highest voltage out I can get with the pot at its limit is 4.9. I can decrease it to 0V by adjusting it down, but it wont go any higher.
Wrong pot?

Change the multimeter to resistance mode again. Measure the resistance between Vout (middle leg) and Ground. It should be ~490ohms.

And this test produces the same ohms as checking the resistance between the pot legs.
??

Ok, you appear to have the pot and the 3.3k resistor connected in parallel, and not in series. That would explain why you see the same readings, and is borne out by the math; (3300*1000)/(3300+1000) = 767.

I think that you've probably connected the non-regulator end of the 3k3 resistor to Ground, ie, to the wrong end of the pot.

hmm. Not sure I follow you.

I've got the last leg of the pot (furthest from the screw adjuster) attached with the left leg of the regulator (ADJ in the diagram), and one end of the 3.3K resistor.

The other end of the 3.3K res is attached to the pos side of the 10UF cap, the mid leg of the VR, the last leg switch side A , and one side of the 470 res.

The other 2 legs of the pot (mid and front) are connected to one side of the .1uf Cap, the neg leg of the 10UF cap, the other side of the 470 res, and the ground out wire.

Should I switch the wires as they are attached to the pot?

You all realize that given our collective knowledge, we could take over the world right?

I still have dibs on New Zealand. Oh, and I want one of those big frickin lasers. (I don't ask for much.)

What? You don't want sharks (or mutated sea bass) with lasers on their heads then?

- Trevor

Nah, just one big frickin laser to play with. That'd be enough.

Nah, just one big frickin laser to play with. That'd be enough.

I was about to make fun of the double-post thing that I saw in two other places on the BBS tonight, and now it just happened to me. I swear I didn't submit the message twice, the BBS is messing up on us tonight.

Well I switched the wires as they were connected to the pot and it didnt help any. In fact I think I may have fried something. I no longer get 763 ohms but about 650 max when I test it.
Maybe Ill hit Radio shack tomorrow and get a new batch of parts that I can ruin.

After all is said and done, I probably could have just bought the pre-made unit and saved myself a lot of trouble!
But this is a learning experience, right? And what $ value can be placed on learning?

OK, so I took the whole damn thing apart, and started over again (using the same components though, except for the .1uf cap and the circuit board).
Apparently my soldering skills have actually degraded since the first time I put this together, so now it looks even more like crap than before.

Now when I do the resistance test on the pot (with it at its max adjustment), the max ohms I get is around 783. And when I test between the mid leg on the VR and ground I get about 420.
So at least its different. Still not right, but at least different.

I think next I'm going to swap out some of the components, and if that fails, Im just going to order this stupid thing pre-made.

Swap out the regulator first.

??

Unfortunately I was unable to get it working with the last rendition of the schematic.

My sister came down to the Cape from Boston the other day (shes an electrical engineer out of MIT) and she helped me with a few things (while also mocking my inept soldering skills).

She pointed out that we had R1 and R2 reversed in the equation we were using. According to the diagram on the voltage regulator package it shoud be Vout = 1.25V*(1+R2/R1). The original website where I found out about this whole thing had it wrong too, so maybe thats where we got it from.
I changed a few things around and starting off with the original basic diagram, eventually got it working.

I changed the 3.3K resistor to a 150 and used a 10K pot instead of a 1K. The 1K only went up to 750 ohms and I just couldnt get it to work with that. I still used the DPDT switch as opposed to the SPDT and I wired in the LED as recommended.
So with this configuration it works perfectly. Wether it should or not I have no idea, but it does.

So then I tried modifying what I had, to the latest version of the schematic by putting in the capacitors and resistors that were reccomended but as soon as I would do that the voltage would drop to almost nothing.
So now Im not sure what to do.
Ive got it working, but with no safeguards.

Maybe I bought the wrong kind of capacitors?

This is what I got to work...

as soon as I would do that the voltage would drop to almost nothing.
... Ive got it working, but with no safeguards.
Maybe I bought the wrong kind of capacitors?

hmmm, caps on the input to ground and output to ground should not affect voltage, just stability, unless they're shorted internally... just about independent on type.

/Michael