How It's Made: Egr Emulators

[FlossyTheSmart]

Hey everybody, I was playing around with lasers and acid and stuff tonight, and decided to take some pictures. I thought this might be sort of interesting to some of you. Let me back up a bit. I made a batch of EGR emulators a while back for someone, using plain old perfboard (this stuff) and through hole components (Like these). While these worked, they were big, bulky, and took a good amount of time to make. All that in mind, I decided to make my own circuit boards, so that I can use little surface mount components (Like these!) I started by grabbing the latest iteration of the EGR emulator circuit by smartdriver (from here)

This is all well and good, but isn't particularly usable in it's current state. With that in mind, I headed over to Digikey and found suitable components to be used in the circuit and ordered them. At this point, I knew the size of the components being used. With that info, I used a circuit layout tool called EAGLE to make the circuit below:

Then I hid some of the layers to make this, which is just the bits that we want to actually show up on the circuit board:

Now, this is where things start getting fun I grabbed some copper clad board from my stash:

Then I cut a 2" strip off of one end (Since this was really just an experiment). I then spraypainted the copper side of the board (Black, although it doesn't really matter). Once the spraypaint was dry, I chucked the now black copper board into my laser cutter/etcher (A 40W CO2 laser, in case you were curious). I used it to etch the black and white inverted image of my circuit from above into the spraypaint, leaving spraypaint only where I wanted my copper lines to be:

(Action shot!)

Once that was done, I was left with a strip of copper board, with five spraypaint circuits on it. I mentioned acid before, right? This is the part where we start playing with extra dangerous stuff

I pulled out my container of Ferric Chloride (Attainable at your nearest quality electronics shop), and got the bubbler going in it (Not just for show!)

As it happens, ferric chloride eats (etches) away at copper, but not spray paint (Which makes it an "etch resist"). The bubbler helps this process by keeping the liquid moving, and getting fresh acid in contact with the copper. I threw the copper board into the jug o' acid, and waited. and waited. and waited. The process should only really take 20-30 minutes, but my acid isn't particularly fresh, and has been used a bunch of times before this, so it's slower. Also, it wasn't particularly warm in the basement where I was doing the etching (Heat speeds the process). In the end, it took about 1hr, 10 minutes to etch the exposed copper off of the boards. Once that was done, a quick rubdown of the boards with acetone removed the spray paint, and I was left with this:

For scale, each of those etched out circuits is about 1" x 1". The components that go on the board are absolutely miniscule. That's really as far as I got tonight, but it seems like good progress. All that needs to be done after this is to drill a few holes for the wires to connect to the EGR harness, and soldering on the components. That's a job for tomorrow with a magnifier! Hope it was interesting!

[KurtMan]

Dude, this was a really cool post to read!

[smartdriver]

Great writeup!

Clever use of new technologies.

I am interested to see how well the circuit functions without the 100 Ohm 2W resistor.

Just a suggestion if you would like to make 5 boards at once.

Using SOLDER PASTE (NOT solder flux paste) fasten all the components to the boards and put them into a toaster oven set to broil for a few minutes.

Example:

http://www.zeph.com/zephpaste.htm

http://www.freetronics.com/pages/surface-m...en#.U8ZC33Jo8ps

The solder paste contains eutectic solder particles in a paste flux which holds the components to the board.

A tiny puff of smoke will indicate when the solder has flowed and the board may be removed..

Edited July 16, 2014 by smartdriver

[FlossyTheSmart]

A good suggestion! I've read about solder paste before, but only in the "tub of solder" format, where you have to use a stencil and squeegee the paste through the holes onto the pads. Using a syringe looks like it would be dead simple! I think I'll still solder the first one up "manually", but assuming it ends up working, I think I will end up popping out tonight and grabbing some solder paste and a toaster oven from Sally Ann.

[stickman007]

Awesome write up!I sold another 1000ohm 5v setup to somebody last week. He said he'll report back if it works well or not.

Edited July 16, 2014 by stickman007

[FlossyTheSmart]

D'oh, just realized I didn't order 1/4W resistors for R6. Looks like I'm placing another order from Digikey. Getting solder paste along with the resistors.

[smartdriver]

D'oh, just realized I didn't order 1/4W resistors for R6. Looks like I'm placing another order from Digikey. Getting solder paste along with the resistors.

Power = E^2/RThe max voltage across R6 is 5V.Max power dissipated by R6 = 25/1000 = .025W which means that likely that same 1000 Ohm resistor for R4 may be used for R6.

[FlossyTheSmart]

Ah, perfect. I was concerned as you had mentioned in the post where you made the updated schematic to use a 1/4W 1K ohm resistor for R6. The power rating on the resitor I was using is 0.1W, so sounds like I should be safe.

[Francesco]

Awesome write-up, and great tips about toaster oven reflow work!

[dalplex]

should D1 be reversed?

[FlossyTheSmart]

Nope, my circuit matches the original drawing. You might be getting mixed up by the "crosshair" in the middle of the component drawing?

[smartdriver]

The ECU wants to see 0.8V at idle (0% duty cycle) and 3.3V at 70% duty cycle.

http://vimeo.com/1810724

The 0.8V is provided by the voltage drop across the diode plus the voltage drop across R2 determined by the ratio of R2 to R4 across 5V minus the diode drop.

Idle Output Voltage = 0.6 + 220 / (220 + 1000) X (5-0.6) - 0.6 = 0.8V

The diode offsets the Base-Emitter drop of the output transistor Q2 and also provides temperature compensation to ensure that the output voltage-duty cycle transfer curve doesn't change over wide variations in temperature.

Edited July 16, 2014 by smartdriver

[FlossyTheSmart]

And this is why you are the designer, and I'm just manufacturing them

[smartdriver]

May I make a suggestion for your next iteration?By moving R6 and adding a jumper it becomes possible to bring the output pins right to the edge of the board.With the proper pin spacing this would make it possible to use short pieces of heavy gauge copper wire as pins which would plug directly into the EGR cable connector and obviate the need for an external connector.I have also added an extra pin for the unused 12V line which might eliminate plug in errors.The drawing isn't to scale. The red wire is the jumper.Every smartcar owner could carry around 2 of them as cufflinks.

Edited July 16, 2014 by smartdriver

[FlossyTheSmart]

I like the idea of using heavy gauge wiring (Standard household wiring would probably be suitable) as pins to plug straight into the connector, however, I'm NOT a fan of the jumper. My primary goal with this particular layout was actually explicitly to avoid jumpers, and to go entirely SMD. Of course, the transistors are still through-hole, so I wasn't entirely successful. I don't really know what I'm looking for in terms of transistors, would it be possible to find a 2N3904 equivalent in an SMD form factor?I could still do the built in pins without having the connections at the edge. They could just be through holes, with the pins ending up perpendicular to the actual board.

[smartdriver]

Advertised on EBAY U.K.

Originating from China.

The part number may be useful to you.

http://www.ebay.co.uk/itm/100PCS-MMBT3904-...r-/140936228098

Edited July 17, 2014 by smartdriver

[FlossyTheSmart]

That is unbelievably cheap, and with free shipping. 100 from Digikey would be $10, plus $8 shipping (Although I'd get it the next day). I think I'll order from that guy and when I get around to another batch of these, switch to using the SMD ones. I assume they'll take a really long time to get to me.

[FlossyTheSmart]

Got my solder paste last night, but first I finished hand soldering my test board. I have used household wiring as "pins", and we'll see how those work. I'll be throwing this one in my car and dogfooding it for the next little while.

[stickman007]

Hey Flossy,I just called MB today to order a few more OEM connectors with the harness and found out that the price has been updated. It was $29 and now its $40. Webber Motor in Edmonton gave me the old price, but if you want to collect a few more connectors, time to buy now!P/N is still C0019185V001000000

[FlossyTheSmart]

Finally got a chance to test out my first emulator from this new batch this morning - Failure. Fortunately, upon doing the autopsy just now, it seems that a few traces lifted then cracked around the "pins" (small segments of household wiring in through holes in the board). THis is partly because the holes that I had to drill left very little copper around the pin, so there was some strain focused on just the trace leading up to it when I first plugged it in. I've just applied some extra solder, and hot glue to seal/support, and I'm about to head out for another test.(BTW, the household wiring works wonderfully as pins, after using a screwdriver to gently ease the receptacles apart in the car. Everything holds quite nicely.)EDIT: Still no go, and for some reason I was getting a traction control light...

[FlossyTheSmart]

Update 9000:I started fresh, and used the solder paste to do a whole new board. I just used wires as I normally did, instead of "pins", and I just had a successful test run. The pins are good in theory, but I'll need to do some further work to make it feasible. In the meantime, I'll keep making them with wires and solder paste.(solder paste is so much easier and cleaner!

[FlossyTheSmart]

4 Emulators made now, have to etch a whole bunch more boards

[smartdriver]

Great news!I had not tested the latest 1K Ohm 5V update and I am happy to see that everything seems OK.With surface mount transistors you could likely build a circuit small enough to put on a postage stamp.

[FlossyTheSmart]

Another update, after my "recall" on the emulators I sent out over the last little while.

The 4 wire emulator design seems to prevent limp mode (most of the time), but still lights the CEL. Obviously no good. I had to order some 100 Ohm, 2W resistors (surface mount of course), and do some manual tweaks to the circuit boards to make it work. First I cut the trace between R6 and the blue wire, then I replaced R6 with the 100 ohm resistor, then soldered a fifth wire on after R6.

In addition to those tweaks, it seems I burnt some of the resistors on the boards I had cooked up before. I've had to replace 1 or 2 resistors on most of the boards (Generally from R1, R3, and R4). Presumably this means I cooked the boards too long in the toaster oven, or I fried them while correcting tombstoning after cooking the solder paste. Some time with a multimetre and soldering iron, and all is good now!

I installed one of these refurbed emulators into Kurtman's car earlier this evening, and I quickly threw one in the cabrio while I was at it.

I also got the SMD transistors that I had ordered, and they look good. I tweaked the designs to use them, and switched back to the 5 wire design, and was able to reduce the size of the emulators by about half. It looks like I can fit about 30 on a 6" by 6" copper clad board.

[smartdriver]

That would work for a while, but the ECU requires a feedback signal from the EGR valve.

Even though the valve is blocked by a plate and prevented from inputting exhaust gas to the engine, it will still move and it can still eventually become clogged and throw an error code or the feedback potentiometer will still wear and also could throw an error code.

Just adding a blocking plate doesn't really buy you anything in either the short or long run and I see very little difference between a disconnected EGR valve (which requires an emulator) and a blocking plate.

Some remaps electronically disconnect the EGR valve so that it is effectively not in use and do not require an emulator.

Edited October 31, 2015 by smartdriver