| Author | Message | ||
     Reepicheep |
Oldog asked for a discrete component voltage indicator (high and low, go / no go). I am using a $15 wal-mart trolling motor battery voltage tester, which has 4 LED's and an LCD voltage read out. It works great, but you have to get pretty creative and hack the case up (and reseal it) and find a place to mount it. I used hot melt glue to seal it up, and put it upside down with the LED's pointing out between my tach and my flyscreen, and I can read the volts LCD through my translucid blue flyscreen. A nice "borg" effect.  FtBastard also did a great write up of using the single LED PIC controlled voltage monitor, that one little LED can tell you a LOT (multi colors, multi blink states) and can be mounted in one of the existing blanks inside the instrument cluster, another very cool solution. I think that device is like $30 or something. But because I am a sucker for any welding or small analog circuit excercise, I designed one up for Jim on principal. Here it is:  It should come on at a fixed voltage, which you set with the 5k potentiometer R3. If you want it more accurate, us a potentiometer to get the what you want, and stuff in fixed resistors in its place. I think Jim was going to build two of these, one for high voltage, one for low. It is sized for a safe input of up to 15 volts. Above that, and the LED may have a short but exciting life. If you want a bigger margin, put in a bigger R1, but you will loose brightness at normal voltages. It is also sized for a red or green LED, which will have about a 2v drop. Blue / White LED's have like a 4V drop, so you will want a smaller R1 in there to give them the current they need if you want them running at full brightness (which may be WAY too bright). Careful, I designed this, but did not test it. As a general rule, I don't trust my designs until I see them work. That being said, I suspect I nailed this one, its fairly basic stuff and has lots of room for error. The PN2222A transistor comes in big blister packs at RadioShack, always have a pile around. As does the 5.1V zeener. So parts ought to be easy to get. I release it under the perl artistic license. Do what you want with it, but give me design credit if you republish. And if anyone tries to build it, let me know if it actually works Maybe I will build it with my son, just for fun, and teach him to solder. | ||
| Oldog |
Thanks Bill! I will put this together, soon maybe this weekend as its supposed to be raining. | ||
| Bombardier |
Ah! The old 'Zener Diode in the Circuit trick Hey?' I have no clue but it looks good Reep! | ||
| Jlnance |
and find a place to mount it. I have considered making something that mounts in a cig lighter plug. Then I could just plug the entire assembly into the accessory outlet on the dash. | ||
| Reepicheep |
Sorry it can't easily be calculated for fixed resistor replacements for the pot to aim for a particular voltage... It will depend on the transistor base-emitter forward voltage drop, and the zeener reverse breakdown voltage at that low current, and be a fussy little calculation. If you cant get a variable supply voltage to just set the pot manually, measure the voltage at the pointy end of the zeener with the LED turned on and tell me what that says, and I can probably make a good guess from there. | ||
| Reepicheep |
Oh, and I spelled license wrong. Thats... err... the Hungarian variation of the term, I got into the habit of spelling it that way when I studied under Oppenheimer.  | ||
| Court |
Why am I seeing Badweb sued because of a run on trolling motors. . . . that, after legendary success sends Wal-Mart into a commercial death spiral . . . . | ||
| Reepicheep |
I couldn't help myself, and went ahead and calculated a fixed resistor setup that I *think* will go on at 13 volts. That should mean it is on when the bike is charging, off when the bike is drawing, but still give you sufficient warning to get yourself to a place you can deal with it if your stator goes south. In that diagram above, drop R2, it's gone. Replace R3 (the potentiometer) with R3a above the zeener diode, and R3b from the zeener input to ground. Pick a 1K resistor for R3a. It can't be too big as it would then starve the base of the transistor for the current it needs, and can't be too small or it would burn up the transistor and resistors at high input voltages. 1k is a good choice, but not the only one. This involves some guessing at the voltage drop across the zeener at these really low currents. I am guessing it is a total of 5.7 volts (5.1 for the zeener, and .6 for the base emitter junction). If that is correct, the equation for the trigger voltage would be very close to: 5.7 Volts / R3b Ohms * R3a + 5.7 That would give the following (with R3a = 1k ohms) R3b = 700 ohms means the light comes on at 13.8V R3b = 800 ohms means the light comes on at 12.8V R3b = 900 ohms means the light comes on at 12.0V I'll try and redraw and post the schematic. (Message edited by reepicheep on August 29, 2008) (Message edited by reepicheep on August 29, 2008) | ||
| Midknyte |
 http://gadgets.boingboing.net/2008/08/28/basic-ele ctronics-sy.html  | ||
| Reepicheep |
OK, here is the update. As usual, if anyone builds it, I'd love to know if it actually works  | ||
| Reepicheep |
Update... Jim was playing around with a quad comparator that is also cheap at radioshack, and got a good result. I think I could use that same chip and get a 4 position voltage monitor with just one zeener, and maybe 4 more resistors, and no transistor. So from a parts count standpoint, it would probably make more sense to build that then this. But it lacks that cool old school analog feel I think my son and I are going to try building the one above now and see if it works... Jim had a bad result, but I think he might have baked his zeener... | ||
| Court |
Bill: If you build one, I'll buy it. Court | ||
| Reepicheep |
You'd be getting cheated And besides, then Id be violating sponsorship rules. Ben (my oldest) and I built it, and it works OK, it needs something more in line with what Jim was proposing. With about a 280 ohm resistor for R3, a green LED is *barely* glowing with my 9sx (good battery, fully charged) not running. When you start the motor and it idles, the LED goes moderately on. Revving the motor lights it up brighter. Total VR failure and overvoltage would probably make it glow bright green. So it actually ends being a narrowly tuned LED analog guage. Which is entertaining, but probably not optimal. The reason is clear (as it always is in hindsight ). The Zeener has a "fuzzy" voltage range at which it "knees over". So it starts leaking current, and slowly turns on the transistor within (I'm guessing) a .3 to .6 volt range. So you could try and salvage my circuit above by making the transistor switch *really* touchy and put it in a darlington configuration, which means it is a primitive amplifier with a gain of like 1000 or more. That would probably help, but it could open the door to other problems. But they make special operational amplifiers that have one purpose in life... instead of making a little signal a big signal, they take two inputs. So long as input A is lower voltage then input B, it outputs nothing. As soon as input A is even a little bit higher voltage then input B, it goes straight to full voltage. They call these amplifiers comparators, and you can buy an IC that has 4 of them in a chip for under $2 at radio shack (LM339). (http://www.national.com/mpf/LM/LM339.html) Looks like they can output 16 mA, which is a bit on the low side, but should probably be able to make an LED light up, so long as it isn't one of the "ultra bright" ones. So thats the better route to go, though it is scope creep. It is up to $8 to $10 in retail parts (probably under $5 wholesale), and remember the $15 wallmart trolling motor battery monitor has 4 LED's *and* an LCD voltage readout. Though this is clearly a lot more fun... And the wal-mart solution didn't teach my 9 year old to read resistor color codes and use a breadboard, so from that perspective, the "better" solution is the worst deal ever made... | ||
| Reepicheep |
Here is what the basic schematic would look like, ignore the resistor values, I didn't calculate any of them out yet. Also, I couldn't find the schematic symbol for an LED, so I just used a normal diode, so ignore the part numbers on that. Also ignore the zeener voltage listed, I didn't fix that either, though until you calculate out the resistors it doesn't really matter. So anyway, here is the footprint of what it would look like. Also, the comparator might be able to sink more current then it can source, so the whole thing should probably be turned around (V+ to diodes into op amp, and the + and - inputs to the comparators reversed. That would probably get you a brighter LED without baking the 339. If there is a burning interest, I could finish the calculations. Maybe I'll make another saturday project with my sun and grind out this one. If anyone has a burning interest, it's probably less then 30 minutes to polish off the rest of the design. As a heads up, that circuit isn't rocket science, but there are a lot of opportunities to wire things up wrong... it's not a trivial excercise to solder it up and package it. But it would be cool...  | ||
| Oldog |
Here's the diagram I sent Bill, I have done the Low volts side ( lights on low volts ) Unless I missed some thing the high side ( light on over volts ) should work I will build it in a few days as time permits.  the parts came from radio shack you may be able to use just the 5v reference, I wanted to try the 12V ref too the chip has 4 comparators built in this uses 2, I plan on building the circuit and placing it in a film container and running leads out for the lights and power connections | ||
| Jlnance |
I love circuits. a green LED is *barely* glowing with my 9sx (good battery, fully charged) not running. Careful, this circuit draws power. You don't want it wired in in a way that would cause it to be connected when the bike is switched off. | ||
| Jlnance |
The Zeener has a "fuzzy" voltage range at which it "knees over". So it starts leaking current, and slowly turns on the transistor within (I'm guessing) a .3 to .6 volt range. Plunk a 100k resistor between the base and emitter and it will probably help that. Currently, any leakage through the diode will cause collector current to flow, and as you note, there is a bit before it really brakes down. | ||
| Jlnance |
See, now you've got my head on a roll, it's 4:30 am, and I can't sleep cause I'm thinking about your circuit. If you move the Zener into the emitter circuit, and reduce the value of r1 accordingly, the Zener will see not only the base current but the collector current as well. This should make the knee much sharper. It will also greatly increase the base impedance, so you can up r2 & r3 by 10 or 100 times if you want, which will decrease the current this circuit draws when it is off. | ||
| Reepicheep |
Good thought! I bet that would work really well. Wonder if the zeener can sink enough current to run the LED. I should clarify that I meant ignition on, but bike not running, not really bike switched off. The meter would be put on a switched feed. Fun, isn't it! | ||
| Jlnance |
Wonder if the zeener can sink enough current to run the LED. They are 1W zeners, so they can sink almost 200ma, which should be a gracious plenty (that's a technical term .) Fun, isn't it! I was going to spend my life doing this stuff before I got sucked into computer programming. | ||
| Court |
I understand "green" and "red" . . . when you yahoos get it to that point . . . call me. | ||
| Reepicheep |
Same here... I was in my senior year EET, and looked at the help wanted adds. 10 programming jobs to every EE job. That was as much as I needed to know... (and frankly, though analog design is a lot of fun, I was a better computer guy anyway). I have the parts lying around, I'll try it. And I got most of the 4 light comparator thing calculated out this weekend as well... that might be fun to build. | ||
| Reepicheep |
dammit Jim, now you made me work through lunch using your "moved zeener" and Courts "I want a red light also". Now I am trying to figure out how to do an over and under voltage indicator (as described by Court, the light is red or green for go or no go). Probably getting complicated enough to go to the comparator, but I love the "zen" of pure analog components... | ||
| Jlnance |
How about just turning the light back off for over voltage? Add a second transistor that shunts the existing transistor's base current to ground. Drive it with a 15V zener connected to vcc (and appropriate current limiting resistor.) (Message edited by jlnance on September 03, 2008) | ||
| Jlnance |
BTW, do you remember when we were talking about redesigning the voltage regulator for these bikes? I came up with a cool little scheme that used the inductance of the stator to make a boost converter. That would let you lower that stator voltage and still charge at low RPM. And by adjusting the switching, you could morph it back into a shunt regulator at high RPM if you needed to. I wrote a computer program to simulate it, using made up numbers for the stator impedance. I never designed the control circuit, which of course is the difficult part. Was cool to play around with though. | ||
| Reepicheep |
Can't let it drop... The resistor values (R1 through R5) are wrong. The targets we want are 14.4V for the first light (over charging), 14V for normal happy, 11.9 volts for an OK battery that is not being charged, and a 11.0 volts for "get ready to push".  Gotta calculate the resistors... thought I did it with Maxima, but the math isn't working. Gotta solve 8 equations for 8 unknowns... and was too lazy to do it by hand. I'll try it again tomorrow. | ||
| Reepicheep |
New and improved. I built this one far enough to verify it works on a breadboard (built two of the four indicators). I have to get some better matched LED's and I will solder up a prototype. It will actually be pretty cool. It will be tunable, and when the bike is turned off but properly charged, one green LED will be on. If the bike is running and properly charging, two green LEDs will be on. If the bike is overcharging, two green and one red LEDs will be on. If the bike has drained 50% of the battery capacity, one red LED will be on. Anybody know how good those little simple wiper pots hold a setting over time? I can seal them from the elements, but they will still see a lot of temperature cycles. Will they hold up?  Oh, and of course on the LM339, pin 3 goes to bike positive voltage, pin 12 goes to ground. (Message edited by reepicheep on September 25, 2008) | ||
| Bikerjim99 |
"Simple" is it? Sure is a relative term. | ||
| Reepicheep |
Scope creep. It gets less simple by the moment. I blame Oldog. | ||
| Court |
This is very cool . . . . I have the donor test bike . . . with electrical accessories mounting by the day . . . ready. We'll sort this babe out, pot 20,000 of them and be the next Microsoft! |
