Build your own switchback - turn/running flasher controllers

[aquillen]

PLEASE REVIEW POST 44 IN THIS THREAD IF YOU PLAN TO BUILD THIS YOURSELF FROM MY DRAWINGS.

I build a lot of things instead of buying - within reason, and then sometimes without reason. I just like doing a lot of different things in a project. It's why I'm building an 818 - to do instead of buying "done".

These days a lot of people who would never get into electronics, are doing so, with just a little info to get them going. This is not really a complex circuit by any means. If you have just a bit of skill, or an 'electronics type' friend who wants to help, or even a high-school kid who wants to learn and try it - I say go for it.

I haven't done a PCB board for this yet but will, and will post up a drawing for that. It can be done just on a piece of perf board, etc. for that matter.

I did the breadboard setup and tests, with several of the LED lights listed below, from Amazon. Also verified it works with the standard Subaru 3 terminal flasher module, with LED's. (in the olden days flasher modules often needed the "load" of incandescent lamps to do the flashing - hence I tested it). This would get in trouble running incandescent bulbs, or LED's drawing much over a total of 300mA (Q1 would get HOT at least if not fail).

Other transistors (must be reasonably equivalent in characteristics to these) can be used, these are the ones I had in my junk boxes.

http://res.cloudinary.com/aq007/image/upload/v1500499683/818S-SwitchBack_nrtf6b.jpg

How it works:

With "turn signal off", Q2 is off, allowing Q1 to turn on. This lights the white LED's connected to Q1. When turn signal is energized, the Yellow LED's light by direct wiring through the flasher 12V. Also this turn signal 12V almost instantly charges capacitor C1 via diode D1 and low ohms resistor R1. The 12V at C1 and the gate of Q2 turns on Q2. Q2 turns off Q1 so the white LED's are turned off instantly.

Once the turn signal is off, the high resistance in R2 begins to discharge C1. The resistance/capacitance time constant of R2, C1 takes about 2 seconds to discharge enough that Q2 can turn off. By this time a blinking turn signal has recharged C1 repeatedly so Q2 remains turned on. After the blink condition ends, C1 will finally be able to discharge until Q2 turns off, allowing Q1 to turn back on and resume the running white LED mode.

With 6 each of some 50 milli-ampere white LED's connected to Q1 it can get slightly warm, so packaging of the circuit (pot in epoxy for example), should be such that it can dissipate some heat, about 1/4 to 1/2 watt is typical. Also the Q1 transistor case must be insulated from ground, etc.

The LED's I tried were pointed out in another post here, I bought them and like them. Yes tiny wires, but that is all they need for the current demand in play.

One way to package this might be in an Altoids box, or a cigar tube, then pot (filled) the container with some JB Weld. I believe JB Weld has metal filler but not to the point of being conductive enough to affect this circuit. The Q1 should be close to the surface of the container so it can pass heat out, but not touching. JB will move the heat out fairly well. If you get to fancy it finally reaches the point where buying a couple commercial modules would make more sense. But then you didn't build it yourself (818 wink).

https://www.amazon.com/gp/product/B00X5NI9HU/ref=oh_aui_search_detailpage?ie=UTF8&psc=1

If the link to the site goes bad, here is what they were described in the ad:

Everbright 8-Pack White / Amber Switchback 12V 9W Eagle Eye Lamp Led Light Bulbs For Car Tail Car Motor Backup Light Fog Light Daytime Running Lights(DRL)

[STiPWRD]

Art, this is perfect, thank you for the schematic and the explanation behind how it works. Time to go scrounge in the scrap bin!

[frankc5r]

I build a lot of things instead of buying - within reason, and then sometimes without reason. I just like doing a lot of different things in a project. It's why I'm building an 818 - to do instead of buying "done".

These days a lot of people who would never get into electronics, are doing so, with just a little info to get them going. This is not really a complex circuit by any means. If you have just a bit of skill, or an 'electronics type' friend who wants to help, or even a high-school kid who wants to learn and try it - I say go for it.

I haven't done a PCB board for this yet but will, and will post up a drawing for that. It can be done just on a piece of perf board, etc. for that matter.

I did the breadboard setup and tests, with several of the LED lights listed below, from Amazon. Also verified it works with the standard Subaru 3 terminal flasher module, with LED's. (in the olden days flasher modules often needed the "load" of incandescent lamps to do the flashing - hence I tested it). This would get in trouble running incandescent bulbs, or LED's drawing much over a total of 300mA (Q1 would get HOT at least if not fail).

Other transistors (must be reasonably equivalent in characteristics to these) can be used, these are the ones I had in my junk boxes.

http://i1078.photobucket.com/albums/w487/aquillen1/818_wiring/818S-SwitchBack_zpso0b0iiq8.jpg (http://s1078.photobucket.com/user/aquillen1/media/818_wiring/818S-SwitchBack_zpso0b0iiq8.jpg.html)

How it works:

With "turn signal off", Q2 is off, allowing Q1 to turn on. This lights the white LED's connected to Q1. When turn signal is energized, the Yellow LED's light by direct wiring through the flasher 12V. Also this turn signal 12V almost instantly charges capacitor C1 via diode D1 and low ohms resistor R1. The 12V at C1 and the gate of Q2 turns on Q2. Q2 turns off Q1 so the white LED's are turned off instantly.

Once the turn signal is off, the high resistance in R2 begins to discharge C1. The resistance/capacitance time constant of R2, C1 takes about 2 seconds to discharge enough that Q2 can turn off. By this time a blinking turn signal has recharged C1 repeatedly so Q2 remains turned on. After the blink condition ends, C1 will finally be able to discharge until Q2 turns off, allowing Q1 to turn back on and resume the running white LED mode.

With 6 each of some 50 milli-ampere white LED's connected to Q1 it can get slightly warm, so packaging of the circuit (pot in epoxy for example), should be such that it can dissipate some heat, about 1/4 to 1/2 watt is typical. Also the Q1 transistor case must be insulated from ground, etc.

The LED's I tried were pointed out in another post here, I bought them and like them. Yes tiny wires, but that is all they need for the current demand in play.

One way to package this might be in an Altoids box, or a cigar tube, then pot (filled) the container with some JB Weld. I believe JB Weld has metal filler but not to the point of being conductive enough to affect this circuit. The Q1 should be close to the surface of the container so it can pass heat out, but not touching. JB will move the heat out fairly well. If you get to fancy it finally reaches the point where buying a couple commercial modules would make more sense. But then you didn't build it yourself (818 wink).

https://www.amazon.com/gp/product/B00X5NI9HU/ref=oh_aui_search_detailpage?ie=UTF8&psc=1

If the link to the site goes bad, here is what they were described in the ad:

Everbright 8-Pack White / Amber Switchback 12V 9W Eagle Eye Lamp Led Light Bulbs For Car Tail Car Motor Backup Light Fog Light Daytime Running Lights(DRL)

I am in the middle of building a similar circuit using p chan mosfet.
I bought and tested the Brightview leds u show and at 13.8 vdc they only
draw about 60 ma ea. Going to use them anyway since hard to find any others.
They are about 0.9 watt ea not the advertised 9 w. To give them some credit they
.may mean like a old regular bulb. I also have adj delay. Like the circuit.

[aquillen]

I am in the middle of building a similar circuit using p chan mosfet.
I bought and tested the Brightview leds u show and at 13.8 vdc they only
draw about 60 ma ea. Going to use them anyway since hard to find any others.
They are about 0.9 watt ea not the advertised 9 w. To give them some credit they
.may mean like a old regular bulb. I also have adj delay. Like the circuit.


The comment about 9 Watt - I suspect that was for all the lamps in the package as sold. Or more likely 0.9 Watt typo'd into 9 Watt. Definitely about 60mA at 13.8Vdc though and that is of course what we need to know.

[aquillen]

I revised the value of R1 to improve immunity to voltage spikes if they ever should happen (was 100 ohm, now 10K ohm). Either value would work but 10K more prudent. Circuit could be made much more robust against voltage spikes, but probably overkill to go there. (See my original post, not followups that link the original drawing with 100 ohm for R1). -art

[Frank818]

I am in the middle of building a similar circuit using p chan mosfet.
I bought and tested the Brightview leds u show and at 13.8 vdc they only
draw about 60 ma ea. Going to use them anyway since hard to find any others.
They are about 0.9 watt ea not the advertised 9 w. To give them some credit they
.may mean like a old regular bulb. I also have adj delay. Like the circuit.

0.9watt each... that explains why a tiny 22gauge wire can easily handle 3 of them...
But they are VERY bright for just 1watt! Very efficient LEDs, their lumens per watt must be very high.

[frankc5r]

The comment about 9 Watt - I suspect that was for all the lamps in the package as sold. Or more likely 0.9 Watt typo'd into 9 Watt. Definitely about 60mA at 13.8Vdc though and that is of course what we need to know.

How about replacing Q1 with a 2N3055 power npn transistor? Takes up a little more space but with significant more power capacity.

[aquillen]

How about replacing Q1 with a 2N3055 power npn transistor? Takes up a little more space but with significant more power capacity.

I'll toss one on the breadboard and see how it behaves.

[AZPete]

I share the thrill of trying new projects and I need my LED turn signal lights to work right, but I'm an electronics dummy. When the rest of you are digging into your scrap bin for parts, I had to look up PCB and perf board in Google! I have found all the parts in Amazon but can you show me how to put them together on a board? Can I fill the Altoids box with epoxy or will JB Weld dissipate heat better?

[frankc5r]

Both Alex and I have circuits. I trust his more but
will be testing both. Alex can speak about a pic board
He had mentioned designing one. I can build 2 for you if you
wish. I am at my vacation home w/o most my tools and board making supplies I should be testing and building when rest of my parts arrive

[Frank818]

Frank, what difference in trust are you thinking about between your circuit and Alex's?
You don't think yours will explode, don't you? lolll

[frankc5r]

Frank, what difference in trust are you thinking about between your circuit and Alex's?
You don't think yours will explode, don't you? lolll

We will see. I am going to build it tomorrow and test it with multiple displays, measure temp rise, current draw etc. Alex is an experienced designer. I know enough to be dangerous!!

[aquillen]

We will see. I am going to build it tomorrow and test it with multiple displays, measure temp rise, current draw etc. Alex is an experienced designer. I know enough to be dangerous!!

I did a PCB layout and put one of mine together this evening. Ran 8 lamp sets on it, 400ma load. It runs about 1/4 watt out of the Q1 darlington transistor. Then potted it in JB weld. Will play again tomorrow evening.

In a good heatsink it would handle a couple amps - ie. regular lamps, but that is not what my project is about. Driving the LED's is enough and it will do that skating along.

Regarding using 2N3055. It needs a darlington transistor in my design, so that is not a fit, but not needed for LED's either.

[frankc5r]

Thanks for the info. Going to try it!

[aquillen]

So here is my first assembly. Uses the schematic I posted above. Before potting it I ran 8 of the cat-eye lamps above. I think the new front lamps use 4 so that's 2x the load. Q1 just barely got warm at 13.8V supply, as expected. I believe the packaging I use here is suitable. A bit unconventional way to do this, but shows how you can toss a couple together with the simplest of containers - aluminum foil shaped over a 1/2 x 1/4 piece of wood. (I used oven foil - a bit heavier). I will mount these directly against a frame member so they can pass the marginal heat on out to the frame. The JB needs overnight to fully harden so I'll post up final results later.

Also it is made "surface mount style" but using leaded components that would normally be mounted through drilled holes. I do some surface mount now and then right here at home but only when doing a hundred or more of something. The extra setup is worth it only when the volume gets up a bit.

About potting with epoxy. In 1990 I designed and built about 300 battery testers for an R/C motor/battery business. Drained individual NiCd's at 10amps and timed the run down. The guy wanted the circuit potted in epoxy so it would be hard to reverse engineer. I mixed in some alcohol to help the epoxy flow more like paint. The units developed problems due to conductivity in the epoxy - but it was brought on by the alcohol. Set in the Sun for a week and they baked out and worked. But I wouldn't mix anything into the JB Weld - just warm it up and it will flow well enough. (but don't wait long after you heat it - it will start to gel).

Top view circuit, compared to a 9V battery

http://res.cloudinary.com/aq007/image/upload/v1500499995/SWITCHBACK_01_ju5l8p.jpg

I want the transistor to have at least some heatsink on the bottom, and it needs to be insulated, along with the rest of the circuit board, from any container. The more traditional "correct" heatsink is a screwed down, or clamped affair of some sort. At maybe 1/4 watt the method I'm using here will work fine, in my humble opinion. That is - a bit of the JB Weld on the bottom of the transistor, then press in a regular mica transistor insulator, then a bit more JB on the bottom of the transistor and smeared all along the circuit board.
http://res.cloudinary.com/aq007/image/upload/v1500500066/SWITCHBACK_02_sr9y1u.jpg

http://res.cloudinary.com/aq007/image/upload/v1500500127/SWITCHBACK_03_sfldb9.jpg

http://res.cloudinary.com/aq007/image/upload/v1500500191/SWITCHBACK_04_ydj1pz.jpg

http://res.cloudinary.com/aq007/image/upload/v1500500246/SWITCHBACK_05_sap4at.jpg

Could be made smaller but really?

http://res.cloudinary.com/aq007/image/upload/v1500500318/SWITCHBACK_06_ap0h7d.jpg

[aquillen]

AZ Pete shoot me a pm. Might have a freebie for you

[frankc5r]

625686256962570

Tested the circuit I used today. Did a video but too large for site. Current draw for 3 lights was 0.17A and mosfet stayed at room temp
for the 10 min I ran it. When t/s stopped it took to long, about 8 sec, to return to white but that is easy resistor change.
I like Alex's circuit-very clean . Looking forward to see his results and time delay he has.

[aquillen]

So I ran my prototype detailed above most of today. It will run about 10F above ambient with 8 lights, using 1/2 Watt, and that would never be an issue. With 4 lamps it would be 1/4 Watt which is negligible.

I realized the schematic I posted called for 1/4 Watt resistors, but I did my PCB for 1/8 Watt. So I decided to do a layout for 1/4 Watt and post that. Also to simplify the packaging and build process.

During testing at one point I had the 8 cat-eye lights act up, seemed to be related to bumping the module (but it was setting on the bench with the lamps). I did destructive testing on the unit to see if anything had happened inside the epoxy. Nothing found I'm happy to say. Eventually I found that one of the cat-eye's was shorting between the yellow and white wires - inside. This will keep the switchback from working correctly of course. These cat-eyes are very bright and otherwise seem a great light to use, but I'll show you what to look out for - you may get one or more that are not reliable - you can fix or just test and get extras depending on your skills.

First the cat-eye problem, then how to build a switchback using the larger 1/4 Watt resistors (more likely to find them around unless you buy from a electronics supply like Mouser or Digikey, etc.)...

You can test the cat-eyes by power one up with 12Vdc - black = gnd, red - white lamps, yellow - yellow lamps. With one or the other lamps on, avoid both for this test, gently tug and push in on the wire at the back of the unit. See if lights flicker, the other light comes on when it should not. Don't pull too much or you'll have it broken - it's not a sturdy seal in the back. If you pass this test, it probably will be a keeper. Examine the silicone seal closely and if you find a "pin-hole" or unbonded area, work in some silicone seal or dab in some Plio-Bond rubber cement (hardware store item - great stuff to keep around), to build it up a bit.

The wire cord going into the back of the cat-eye is marginally bonded with a little dab of silicone seal right at the back. Not much strength to prevent the cable from moving around inside, and it is held in by the small wires soldered to the back of the lamp pcb. The pcb is only held in by the tension of the cable, and is trapped from the front by the lens. It can wiggle if the wires move. It has lumpy solder connections to the wires. Some of the lumps are big enough that when the pcb moves a bit, it may touch the casing of the cat-eye (aluminum). The result in this case was the red and white wires shorting together when the cat-eye was bumped or rattled. You can gently pry out the lens from the front by lifting the edge with a hobby knife. Glue seems better in some spots than others, so try different spots to find one that will let it come loose.

Push the wire in from the back and the pcb will come out with the wires. Then you can attempt repairing the solder connections - wick off the blob and resolder it. I cut a small washer from electrical tape and worked it in there too. The clearance in the aluminum hole is so close it is almost a wonder most of them don't short out. But I still like the style and they can dazzle you - bright.
http://res.cloudinary.com/aq007/image/upload/v1500500407/Cat-Eye_guts_nce3my.jpg

[aquillen]

625686256962570

Tested the circuit I used today. Did a video but too large for site. Current draw for 3 lights was 0.17A and mosfet stayed at room temp
for the 10 min I ran it. When t/s stopped it took to long, about 8 sec, to return to white but that is easy resistor change.
I like Alex's circuit-very clean . Looking forward to see his results and time delay he has.

Looking good. Hope you'll share too. Guys can build whichever they want. Did a 2nd prototype today using 1/4 watt instead of 1/8 watt resistors - just to follow my schematic above, and make mine a bit easier to assemble for beginners. - see below -art.

[aquillen]

So my schematic at the start of this thread called for 1/4 watt resistors. Then I prototyped using 1/8 watt parts because I have bins with about 1000 of each, and as usual was trying to get the board really small - a habit I have.

I ran the board all day, packaged up in the epoxy I show above and except for the goofy problem with the cat-eye lights, the board worked great. In the meantime I layed out a board for 1/4 watt resistors simply because they are more likely to turn up at a Radio Shack, etc., and that was what I put in the original drawing. A little easier to assemble. I also wanted to do a package that was easy to do, and had a mounting tab built in.

Here is the assembly layout. The PCB is 3" long and has a foil block under the power transistor. Hidden under the wires at the right end of Q1, the transistor power tab is soldered to the pcb foil (not a lot of solder, just some in the "hole". Not absolutely necessary but soldering the transistor's large power tab to the foil increases heat dissipation. If you do that step and don't have practice doing this - a warning - use a larger soldering iron, sand, scrape or file the tab on the transistor to help the solder flow and bond to the tab, and "get in and out" quickly so you don't cook the transistor with your soldering step.

The wires here don't really need to be heavy guage - I'm using #26 which is fine for the LED current levels in this circuit, because it is big enough and I have about a zillion feet of it left from another project. The only good reason for using larger wire would be it seems a poor fit up against the wiring it connects too, which you would likely do near the lamps themselves (but you decide location). But #18 would be as big as I'd go just to keep the bulk from getting difficult to manage.

If you are building these on perforated "perf" board and not using a printed circuit board, I'd say just mount the Q1 power transistor so it is close to the surface of your container, with an insulator as I show in the first build above. 1/4 Watt is really very low heat to dissipate - consider those little resistors can do that in open air.

If you go with a pcb layout like shown here, the layout view is "from the top looking down" with the intent of soldering the parts on the top, not using the through-hole design. Keep that in mind when making your pcb.

JP1, 2 = Yellow wires (turn/blinkers)
JP3, 4 = White wires (running lights)
JP5, 6 = Black wires (ground 0V)

http://res.cloudinary.com/aq007/image/upload/v1500500487/Switchback_top_view_1-4_Watt_hvos30.jpg

The above board was "painted" with JB Weld using a small screwdriver blade and then heated slightly with a hot-air gun to help it flow into crannies and under the wires. A few minutes later it was cool and already hardened enough to handle. The bottom of the board is not coated. The mount hole takes care of how to securely mount it if bolt-in is preferred.

http://res.cloudinary.com/aq007/image/upload/v1500500546/Switchback_1-4-Watt_done_eulxue.jpg

Board pattern:

http://res.cloudinary.com/aq007/image/upload/v1500500597/Switchback_Foil_1-4-Watt_miqca0.jpg

Components:

http://res.cloudinary.com/aq007/image/upload/v1500500665/Switchback_Layout_1-4-Watt_Photo_e7w1h2.jpg

[aquillen]

Oh - and the delay on my circuit is about 2 seconds after the last blink on time.

[frankc5r]

I tested ea of mine too but did not yank on the back very hard. I worry the vibration of car will cause failure. I thought we might be able to use clear silicone
on the back of the pcb to pot them . Also if I used some 3 to 1 heat shrink to bond the rear aluminum case to the exiting wire to prevent their silicone rubber
movement and resultant failure of unit. Thanks for all your work.

[aquillen]

Yup. I was sitting here thinking about these after I wrote the bit above. I think having one of the lamps go out after getting the entire headlights finally assembled would be a real Debbie Downer. They are nice enough to put up with having to do a bit of product re-engineering so we can trust them.

They are so easy to take apart (at least the 1st one was). I'm going to see about taking them apart, and putting something in there to solid bond them. I'm never quite sure which silicone RTV to use - many of the different types will outgas acetic acid as it cures and so corrode the parts. Plio-bond cement is a reasonable material to use, dries pretty fast and has been used for some electronics assembly, and other stuff since at least the early 70's.

Cyanoacrylic - super glue - at first thought - decent choice. But would need to give it 24 hours to ensure it doesn't release that white vapor-coating that would spoil the back of the clear lens.

[Frank818]

Super interesting, plz continue! :)

[aquillen]

I guess one more go-around on the Amazon cat-eyes. They are nice enough and haven't seen another yet to use. So if you get these, to ensure they are reliable you should know about this. I took all 8 of mine apart. Pry out the front lens with hobby knife. Some take persistence, some nearly fall out. Of the 8, the wire/cable was held in ranging from fully filled with silicone to a rubber grommet to peice of foam to almost no silicone. The wiring at the back of the pcb was either very neat ranging to poor solder/bare wire begging to short out.

So to have these be reliable, and not have to tear into your lighting later on, you may want to inspect and tune them.

Some of them I rewired by soldering off the leads, removing excess solder from the circuit board, and soldering back on after trimming the wires. This is on the smaller end of a soldering job, so be ready with magnifier, small tip iron and a bit of skill.

The clearance between the solder points and the inside of the aluminum housing is almost zero - very likely to have contact on some of these and may even be relying on the anodized aluminum to be an insulator. I happened to have some fiber washers of the right size, but you could cut washers from thin cardboard - .475" OD x .2" ID, slit the washer so it will slip over the wire if you don't remove the wires, and solve the problem. Not a fun hour and a half but probably worth it to keep these lights.

here is the clearance problem:

http://res.cloudinary.com/aq007/image/upload/v1500500744/Cat-Eye_guts2_kdomo3.jpg

[frankc5r]

Art and others,
Here is what mine look like. Lens came off easy. At Art's suggestion, I am going to use some pliobond on back of pcb to vibration proof it and then use
3/1 heat shrink tubing on back of housing to lock in wire exiting reat to housing and prevent movement.6263062631

[frankc5r]

https://youtu.be/znpa4_pbjGUis
Here is video of testing a p channel mosfet switchback driver with me acting as a turn signal flasher at various speeds.

Hope this works. I used Utube to host.

[Frank818]

How about using silicone like the one we used on the alu panels all around the car to create a clearance between PCB and anodized housing?

And for those without heatshrink to prevent the wires from moving on the back of the housing, maybe some electrical tape rolled tight around the housing and wire will work. When done right it prevents movement.

[frankc5r]

How about using silicone like the one we used on the alu panels all around the car to create a clearance between PCB and anodized housing?

And for those without heatshrink to prevent the wires from moving on the back of the housing, maybe some electrical tape rolled tight around the housing and wire will work. When done right it prevents movement.
Art mentioned silicone outgases acetic acid which can corrode the connection.
Tape can come loose. Heat shrink will stay tight and prevent movement. I ordered
3/1 ratio heat shrink so it will shrink tightly around both housing and wire

[Frank818]

I ordered 4:1 heat shrinks, as the housing is 9.7mm and the wire is 3.5 and there are no 10mm heat shrink I could find. I have another solution if I don't get those ebay heat shrinks (75% of my orders since Sept 29 have never arrived).

If all silicones use that acetic acid then how do you glue the round PCB back on if using fiber washers? Or maybe you don't need to, you just push it back in?

What adhesive do you use to glue the lense back on?

[aquillen]

I ordered 4:1 heat shrinks, as the housing is 9.7mm and the wire is 3.5 and there are no 10mm heat shrink I could find. I have another solution if I don't get those ebay heat shrinks (75% of my orders since Sept 29 have never arrived).

If all silicones use that acetic acid then how do you glue the round PCB back on if using fiber washers? Or maybe you don't need to, you just push it back in?

What adhesive do you use to glue the lense back on?

Not all silicones release the acid. I THINK you can go by smell - if pungent then probably does. If no odor or nothing burns your nose - probably ok. I used Pliobond for lens.

I'd put at least a thick piece of paper behind the board if you have it out. Because just putting silicone or other glue, doesn't keep it from possibly touching some part of the metal case, unless maybe you'd put a skin and let it dry over night. Even a bit of paper then silicone or glue would help prevent a short. Paper spaces it away, the sealant keeps it from moving and touching later.

[frankc5r]

Frank818,
When u buy 4/1 heatshrink u do not need to buy 10mm for a 9.7 hsn. It shrinks a lot so buy bigger so it goes on easier. I bought an assortment from Amazon cause I am here in Fl for winter away from my shop which has lots. Here is the assortment including 1/2 in which fits the 3/8 hsn easy.

https://www.amazon.com/gp/product/B01M4MF8U6/ref=oh_aui_detailpage_o01_s00?ie=UTF8&psc=1

[Frank818]

Ok we'll see if I can slide them in at 12mm, should work the housing is 10mm.

I will cut cardboard washers and slice them in. Not sure yet what I'll use to glue everything back on. Pliobond is freakin expensive cuz only found in USA so shipping is high and I can't really make it under 30 bucks for a small amount of it, if I recall. Will look for a replacement. Maybe krazyglue.

[aquillen]

Something reasonably close to pliobond is some of the rubber cements for office use - rather similar actually, and the rubber cements for tire patch prep - if it has some filler so it is not just totally runny.

[frankc5r]

I took Art's (aqillen) picture of his pcb layout and saved it, then printed it out at 43% reduction to get a proper size 5/8 x 3 in pcb layout. Did not have
a laser printer here so took to Staples and had them print onto glossy photo laser paper. When ironed onto a copper pcb board, the fused toner resists
etching by ferric chloride and you can make your own pcb board cheap. Pictures attached.62753
Arts layout
627546275562756627576275862759
Black coating is the toner. After ironing, wash off paper in water, etch copper where there is not any toner. wipe toner off etched board yields
finished pcb board.
Have all parts to make circuit, then will test-expect great results.

[UnhipPopano]

Given that there are not that many electronics parts used, making a circuit board may be a bit of overkill. I would think that not many people would be aware of an old but good method of making of wiring up a circuit called Wire-Wrapping. Here is a link on how it is done [http://makezine.com/2009/07/27/lost-knowledge-wire-wrapping/]. You can buy a hand tool for $15 to $20 and wire is cheep.

[frankc5r]

Given that there are not that many electronics parts used, making a circuit board may be a bit of overkill. I would think that not many people would be aware of an old but good method of making of wiring up a circuit called Wire-Wrapping. Here is a link on how it is done [http://makezine.com/2009/07/27/lost-knowledge-wire-wrapping/]. You can buy a hand tool for $15 to $20 and wire is cheep.
Wire wrap is a good technique and pcb board is a little overkill but it took very little time to do. Plus the design called for the circuit to be bolted to a frame member
for a heatsink. To do that requires a bare bottom which is done here by wiring the board as a surface mount rather than the regular wire through board. Put the wire wrap board in a box or insulate the wrap wrap pins poking thru board and that works too.. Many ways to do it.

[frankc5r]

Did a test coating of Pliobond on back of dual color eagle eye leds being used as turn sig and drl lights. There is some concern about the back of pcb or wires shorting to anodized aluminum housing. This is about a drop. A second drop after first dries should give a pretty good insulation and vibration solution especially
if coupled with paper washer.

[aquillen]

I expect only a few will have interest in making their own, but would like the PCB. This is an 818 "love project" so what I'll do is this (I've no interest in vendoring here and won't). PM me if you want a couple PCB's. I'll wait a couple weeks and see how many I should make. Takes me an hour to make one or 30 of them, costs just a few bucks at most, and all stuff I have already on hand. In a couple weeks I'll make as many as ask and mail them out in envelopes. Cost is so small I don't care to worry about it.

If you want a kit with all the parts, say so, you can see the cost of the mail when you get the package and send that amount back to me. I should hope that would not be considered a vendor activity. I don't want or expect a penny over the shipping.

After the good will runs its course I may post up a wish list of 4 or 5 emissions pieces that I suspect some are not using. Maybe that will cover the pay it forward.

Don't forget your mailing address. A couple guys have already got parts on the way from mid last week - same deal.

[Frank818]

Wow this thread moves quickly. I'm still at the "pliobond replacement" phase. lolll

Anyway, instead of paying over 40 US to get pliobond into Canada (the one you get for 17 bucks), Art's alternatives made me look into a few things. The tire repair rubber glue and other rubber cement are pretty easy and cheap to find. I wanted to see if there wasn't anything I already had at home.

I think this copper gasket silicone may work, it doesn't say there is acid and it BARELY smells, if at all. It's non-conductive too.

627666276762768



Other than that, some various Contact Cement would probably work too, I guess? Not sure they are rubber, but do they need to be...

http://www.canadiantire.ca/en/pdp/permatex-contact-cement-0382518p.html#srp
http://www.canadiantire.ca/en/pdp/lepage-pres-tite-contact-cement-0670310p.html#srp

[frankc5r]

You should check out McMaster-Carr. It is a US
company but I checked on couple web sites who members
live in canada and they got shipments w/o any
problems. They have EVERYTHING!!! Give them a call to check.
330-995-5500. Also look at web site.

[Frank818]

I've asked McMaster twice in the past 3 years and they don't deliver to residential addresses outside USA. Only commercials. It's not clear on their website which is why I asked twice, just in case they changed mind.
But that's ok, I have still some tire repair rubber cement, I'll use that.

Will start dismentling the Eagle Eyes later this week or next w-e (as soon as I get my new soldering iron). I expect 6 to 10h of work on these.

[aquillen]

If you make your own PCB using toner transfer process, don't forget to use paint or some app to mirror the image before you print your toner paper. ...

[aquillen]

If you are assembling from my pictures and drawings - please take note. Q2 the BS170 transistor comes in different packages. Some have a flat face, round back side (TO-92 case style). Others have a larger flat face and a smaller flat back instead of round. The drawing that shows parts layouts, shows the top looking down and the rounded side of Q2 faces away from Q1 - this is the correct orientation.

With the round back it is easy to identify the front/back. And the numbers are printed on the flat side. With the large flat and smaller flat back, the numbers may be printed on the back side, which can lead you to install the transistor backward.

The photograph of my boards shown above are built with the confusing transistor style, with the printing on what would normally be the rounded side the TO92 style case. So my transistor as shown in the photo has the back facing up (with the numbers visible).

If you have the TO92 style BS170 with a flat on one side only, then you need to mount the transistor with the flat facing down, where you will NOT see the printed info on the part.

A forum member found this out for us the hard way... I'm so used to these parts it didn't cross my mind to cover this before.

[Frank818]

The clearance between the solder points and the inside of the aluminum housing is almost zero

Mines aren't made of aluminum, they are full plastic, including the inside. I guess I got a different version and that I am immunized against possible shorts.

Plus, the wire that comes out of the housing is glued with some kind of silicone or black glue that makes the wire very tight in place and won't really move. I know yours do.

Also the lenses were very glued in place and only way to remove them was with a lot of heat and quite some force with a small flat screwdriver, before it would cool down.

[arieshton]

Hi

I just saw your switch back circuit and Im planning to build one, but theres a modification I would like to make. My LED have this built-in controller that when voltage is applied it has this flowing effect, I tried to connect it to my flasher lights but I encounter this problem - it is not syncing.

So I came up with this idea to hook it up to a constant 12v but the problem is it stays ON. I came up to your design and it looks the perfect circuit for my project but your circuit turns OFF the constant 12v when the flasher is ON.

Heres what Im planning to change:
1. Once the flasher turns on the circuit will be activated and it will turn on the LED supplying it constant 12v and vise versa.
2. I will replace the NPN transistor with PNP (TIP107 transistor)
3. Move R3 to the ground.

Heres the circuit diagram
102077

Please let me know what your thoughts. Im just new to electronics, Im consulting to avoid blowing up something :)

Thank you and looking to hear from you soon

Aries

[aquillen]

If I understand, you need to apply steady 12V to your lamp module whenever the turn signal is on = flashing. When the turn signal is not on you want 12V turned off. Use this circuit for such function:

https://res.cloudinary.com/aq007/image/upload/v1549939337/818S-SwitchBack_-_reversed_function_version_efxpiy.jpg

[arieshton]

Thanks so much

[arieshton]

If I understand, you need to apply steady 12V to your lamp module whenever the turn signal is on = flashing. When the turn signal is not on you want 12V turned off. Use this circuit for such function:

https://res.cloudinary.com/aq007/image/upload/v1549939337/818S-SwitchBack_-_reversed_function_version_efxpiy.jpg

HI me again. I just realise we cannot apply the MOSFET switch in the Negative (-) of the circuit, because it is also connected to the white DRL, which means if the yellow is OFF, the white is also cause the negative is cut by the MOSFET. So I made this circuit, please let me know if this is correct, instead I will use a P channel MOSFET and put it in the constant positive (+) 12V.

102151

[arieshton]

HI me again. I just realise we cannot apply the MOSFET switch in the Negative (-) of the circuit, because it is also connected to the white DRL, which means if the yellow is OFF, the white is also cause the negative is cut by the MOSFET. So I made this circuit, please let me know if this is correct, instead I will use a P channel MOSFET and put it in the constant positive (+) 12V.

102151

Can we use a relay instead of MOSFET?

[lance corsi]

This is so far beyond my ability to comprehend, my head is spinning. I read the entire post. You electrical guys are amazing! Thank you for the info, even tho I cannot take advantage of it. Isn't it great that we all have different fields of expertise?
Thank you Aquillen, Frank818, and Frank c5r. Your intelligence is exceeded only by your generosity.

[aquillen]

Lance it is just what a person learns, thankfully we all learn different things. Sharing is the key to a wonderful world.

About the relay - yes a relay can be inserted, but you need that mosfet transistor because the gate on it does not consume current. We want a transistor that does NOT use current to switch on or off (mosfets transistors are voltage controlled whereas bipolar or "regular transistors" do use current).

The two resistors and capacitor form a timer circuit so that that the circuit quickly switches state when the turn signal comes on, but holds it's state for a couple seconds after turn signal is turned off (that way blinking state of turn signal does not fool the circuit into changing state until the turn signal is really done blinking - i.e. completely off).

The mosfet transistor (BS170 or IRF20, etc.) being voltage controlled, does not interfere with the current charging and discharging the capacitor. So the time constant of the R-C (resistor - capacitor) circuit is not modified by the transistor. A relay by itself won't work. You could fiddle with a diode charging a capacitor and turning on a relay with discharge characteristics setup by the capacitor value and relay coil resistance. But you would have a relatively high initial charge current to get the relay on, you need a pretty good size capacitor microfarad wise to keep it on between flash pulses, etc. If you change the relay for any reason and it is a little different coil resistance, your setup changes. So pain in the butt to setup.

The mosfet is a tough transistor these days, I've used them for many years now in small to complex circuits and up to about 2000 ampere control setups. The only rule is to avoid exceeding voltage spec's on them and of course max current rating. This circuit is smaller than a relay and should work for most people.

Use this circuit to control a relay directly if you need double pole contacts like a relay can give you. Add protection diode across the relay coil to protect the mosfet.

[arieshton]

Lance it is just what a person learns, thankfully we all learn different things. Sharing is the key to a wonderful world.

About the relay - yes a relay can be inserted, but you need that mosfet transistor because the gate on it does not consume current. We want a transistor that does NOT use current to switch on or off (mosfets transistors are voltage controlled whereas bipolar or "regular transistors" do use current).

The two resistors and capacitor form a timer circuit so that that the circuit quickly switches state when the turn signal comes on, but holds it's state for a couple seconds after turn signal is turned off (that way blinking state of turn signal does not fool the circuit into changing state until the turn signal is really done blinking - i.e. completely off).

The mosfet transistor (BS170 or IRF20, etc.) being voltage controlled, does not interfere with the current charging and discharging the capacitor. So the time constant of the R-C (resistor - capacitor) circuit is not modified by the transistor. A relay by itself won't work. You could fiddle with a diode charging a capacitor and turning on a relay with discharge characteristics setup by the capacitor value and relay coil resistance. But you would have a relatively high initial charge current to get the relay on, you need a pretty good size capacitor microfarad wise to keep it on between flash pulses, etc. If you change the relay for any reason and it is a little different coil resistance, your setup changes. So pain in the butt to setup.

The mosfet is a tough transistor these days, I've used them for many years now in small to complex circuits and up to about 2000 ampere control setups. The only rule is to avoid exceeding voltage spec's on them and of course max current rating. This circuit is smaller than a relay and should work for most people.

Use this circuit to control a relay directly if you need double pole contacts like a relay can give you. Add protection diode across the relay coil to protect the mosfet.


Thanks so much. Based on your comment I created this circuit. Did I got it right?

102162

Will this relay sufficient?
102163

[aquillen]

Both should be fine.

[brittrix]

PLEASE REVIEW POST 44 IN THIS THREAD IF YOU PLAN TO BUILD THIS YOURSELF FROM MY DRAWINGS.

I build a lot of things instead of buying - within reason, and then sometimes without reason. I just like doing a lot of different things in a project. It's why I'm building an 818 - to do instead of buying "done".

These days a lot of people who would never get into electronics, are doing so, with just a little info to get them going. This is not really a complex circuit by any means. If you have just a bit of skill, or an 'electronics type' friend who wants to help, or even a high-school kid who wants to learn and try it - I say go for it.

I haven't done a PCB board for this yet but will, and will post up a drawing for that. It can be done just on a piece of perf board, etc. for that matter.

I did the breadboard setup and tests, with several of the LED lights listed below, from Amazon. Also verified it works with the standard Subaru 3 terminal flasher module, with LED's. (in the olden days flasher modules often needed the "load" of incandescent lamps to do the flashing - hence I tested it). This would get in trouble running incandescent bulbs, or LED's drawing much over a total of 300mA (Q1 would get HOT at least if not fail).

Other transistors (must be reasonably equivalent in characteristics to these) can be used, these are the ones I had in my junk boxes.

http://res.cloudinary.com/aq007/image/upload/v1500499683/818S-SwitchBack_nrtf6b.jpg

How it works:

With "turn signal off", Q2 is off, allowing Q1 to turn on. This lights the white LED's connected to Q1. When turn signal is energized, the Yellow LED's light by direct wiring through the flasher 12V. Also this turn signal 12V almost instantly charges capacitor C1 via diode D1 and low ohms resistor R1. The 12V at C1 and the gate of Q2 turns on Q2. Q2 turns off Q1 so the white LED's are turned off instantly.

Once the turn signal is off, the high resistance in R2 begins to discharge C1. The resistance/capacitance time constant of R2, C1 takes about 2 seconds to discharge enough that Q2 can turn off. By this time a blinking turn signal has recharged C1 repeatedly so Q2 remains turned on. After the blink condition ends, C1 will finally be able to discharge until Q2 turns off, allowing Q1 to turn back on and resume the running white LED mode.

With 6 each of some 50 milli-ampere white LED's connected to Q1 it can get slightly warm, so packaging of the circuit (pot in epoxy for example), should be such that it can dissipate some heat, about 1/4 to 1/2 watt is typical. Also the Q1 transistor case must be insulated from ground, etc.

The LED's I tried were pointed out in another post here, I bought them and like them. Yes tiny wires, but that is all they need for the current demand in play.

One way to package this might be in an Altoids box, or a cigar tube, then pot (filled) the container with some JB Weld. I believe JB Weld has metal filler but not to the point of being conductive enough to affect this circuit. The Q1 should be close to the surface of the container so it can pass heat out, but not touching. JB will move the heat out fairly well. If you get to fancy it finally reaches the point where buying a couple commercial modules would make more sense. But then you didn't build it yourself (818 wink).

https://www.amazon.com/gp/product/B00X5NI9HU/ref=oh_aui_search_detailpage?ie=UTF8&psc=1

If the link to the site goes bad, here is what they were described in the ad:

Everbright 8-Pack White / Amber Switchback 12V 9W Eagle Eye Lamp Led Light Bulbs For Car Tail Car Motor Backup Light Fog Light Daytime Running Lights(DRL)


This is exactly the type of circuit I'm seeking. However, being interested in old British sports cars, I have a requirement to make this work in both positive and negative ground cars. Having built the circuit with bridge rectifiers ahead of the switchback circuit, it fails to delay the return to marker, so the sequence is white-amber-white-amber instead of white-amber-amber-amber-white. Is it possible to make this in a polarity neutral version so that I don't have to make two different versions (I have several friends who want one as well, but there is a mix of positive and negative ground cars).
Thanks,
Lew Palmer

[chmhasy]

For the people that don't want to deal with the chemicals and drilling of PCB's

I have used the free version of eagle cad https://www.autodesk.com/products/eagle/overview design the pcbs and have used https://oshpark.com/ to produce some PCB's with great results.
and the prices are very reasonable.

[aquillen]

Is it possible to make this in a polarity neutral version so that I don't have to make two different versions (I have several friends who want one as well, but there is a mix of positive and negative ground cars).
Thanks,
Lew Palmer

Retracted my first reply after thinking instead of just writing. Also - would those pos gnd cars be 12V or 6V? Overall it would have to be redesigned though. I'll think about it and get back.

[brittrix]

Retracted my first reply after thinking instead of just writing. Also - would those pos gnd cars be 12V or 6V? Overall it would have to be redesigned though. I'll think about it and get back.

All 12 volt.

Thanks
Lew

[aquillen]

What vehicle is this?

[brittrix]

What vehicle is this?

1950 MG TD and 1934 MG PA and 1936 MG PB

Lew

[aquillen]

1950 MG TD and 1934 MG PA and 1936 MG PB

Lew

FYI looking into it. I'm assuming you would like to use the single cat eye style white/orange-yellow LED modules we started this thread for. Bit tricky working out a simple circuit for this because those modules are common cathode LED devices. Can be done but I'm trying to come up with minimal parts/simple circuit that is reliable of course. Let you know.

[aquillen]

Lew, I haven't dreamed up a simple way to make a universal/polarity neutral circuit. Probably could with relays and transistors for example. With a lot of time spent I would probably get a hit, but after an hour or so I have one drawn using only transistors which should do specifically for positive ground. I need to test it and if good will post.

[brittrix]

I appreciate anything you can offer, Art.

Lew

[aquillen]

We have gone somewhat off topic here doing positive ground designs. BUT, fyi I'm just waiting on a part I designed with this week so I can verify a circuit I want to post. Should have it next week and will update.

[aquillen]

We are definitely off the 818 grid now for sure. I guess it was the challenge so here is a low cost solid state switchback circuit for positive grounds. It's so close to a universal circuit I should spend more time and get it there but... I leave you to layout the circuit board since I don't need one of these... I benched it but did not test with temperature extremes or way low voltage, etc. It works fine on my bench on 12V to 14V, should be good in a typical car.

https://res.cloudinary.com/aq007/image/upload/v1572488576/Switchback_for_positive_grounds_oxbewc.jpg

How it works:

The dual cat-eye LED's are going to usually be common cathode which makes switching the anode sides of the LED pair a requirement. This addresses that problem with about as few components (low cost to boot) that I could think of. Setting up a time delay also makes using a normal coil based relay an extra complexity. Using a solid state opto-coupled mosfet switch (literally a solid state relay) helps simplify the timer circuit.

With normal "lights on" battery negative power via R1 biases transistor Q1 to turn on which then powers the white LED.

When the turn signal/flasher is on, battery negative is to be applied to both diode D1 and R2. Resistor R2 biases transistor Q2 to turn on and this powers the yellow LED at each flash pulse. D1 feeds the negative to the opto-isolator switch device IC1. Current flows through IC1 and R3 to the battery positive. Internally IC1 has an LED that sends light into an MOSFET switch that essentially functions as a relay contact. When the LED is on, the switch is closed. The closed switch in IC1 "shorts" the Q1 emitter-base so that Q1 turns off and the white LED is now off.

During the signal/flasher on time, D1 also supplies charging current to timing capacitor C1 (this current is limited by resistor R4 to prevent excessive current in the diode D1). When the flasher goes into "off" time, C1 discharges slowly through the IC1 and resistor R3 path. C1 and R3 create a time constant so that current flows through IC1 for about 1-2 seconds and keeps IC1 "turned on". This keeps Q1 turned off during the off part of flashing, but only as long as the flashing continues so that C1 can keep getting charged.

When flashing is off for a couple seconds or more, C1 finally discharges enough that IC1 is no longer kept turned on. Switch IC1 is off and Q1 again turns on lighting the white LED. A nice feature of this particular IC device is it has built in logic so that it switches off suddenly instead of fading from on to off as the C1 current supply "fades away", this means the white LED comes on suddenly instead of fading from off to on.

IC1's pins 3 and 4 are interchangeable - when laying out a PCB you could swap them to simply the board layout if desired.

If a separate connection to the cathode of the LED's from IGNITION on, etc., is a problem, then use a pair of diodes, one from the lights on and the other from blink on, to supply the BAT(-) to the common cathode of the LED module.

enjoy!

[k2mpw]

Is there any chance you have a gerber file to share? I'd consider getting a sheet made and make the boards available at cost. I tried my hand at CAD PCB design and just about lost my mind. I've made a dual channel version deadbug style in a small enclosure that works fine.. I would like to add a pair of active reflectors to the front and rear of my Honda Goldwing. Making single circuits wrapped in heat shrink would work much better than what I have right now.

qsl.net/k2mpw/honda/gl1800 is my personal farkle page for the Goldwing.

Thanks
Michael
k2mpw at that Ya Who dot com site.

[aquillen]

Post 20 has pcb board image you can copy and produce from (after you scale it properly). I use a program called Boardworks from over 20 years back. It produces screwed up gerber files - had a professional board shop try to use that data once and they probably cussed me the whole time. If I ever make something that I need a bunch of I'll have to break down and use something from modern ages I guess.