Be nice to your Vref....

[Bruce Bowling]

A note on wiring up the harness on your vehicle....

The wire labeled "Vref" is the +5V voltage reference for the MAP sensor and the TPS sensor. This wire only connects to these components, so be very careful to not connect this directly to ground or (worse) +12V battery. Bad things will occur if you mis-wire this one - double check your work (what is that old saying ..... "measure once, cut twice" ....).

When wiring up a vehicle, I recommend performing this in discrete and deliberate stages - this is how I approach this:

1) I always make sure the dangling wires on the MicroSquirt harness are not touching each other or any metal (ground return path). Next, I wire up all of the sensors (TPS, O2, coolant, intake air temp, MAP) and the signal return grounds (use the sensor ground return wire for these). I then hook up the main ground wires to the engine block and the switched +12V for the +12V lead. I again check to make sure the remaining dangling wires (which includes the outputs - you have not hooked them up yet) are insulated from each other and anything else - tape these if need be.

2) Hook up the laptop and run MegaTune and establish communications - remember that with the latest MS-II embedded code (V2.881 and later) you need to set the "processor type" to MicroSquirt the first time you run (click the "burn to ECU" button for this and any other change to stick).

3) I then like to let MegaTune and the MicroSquirt just run for a while and talk to each other, at least for several minutes. If there are any communications burps/hiccups, this is the time to find them out and correct (its a nice, clean electrical environment right now). Do not worry about the MicroSquirt sucking down the vehicle battery, its current draw is around 40 milliamps so it is a puny load - the dome light above your head draws many times more current. This is also a good time to get familiar with MegaTune - set up the number of cylinders, REQ_FUEL, ignition trigger type, etc. - and, maybe save off a MSQ or two, just to get the feel for how it all works. You ahould also check the sensor readback readings, calibrate the TPS, set the MAP sensor type, set the O2 sensor type etc. If you have everything coming back A-OK then you are in great shape, and you are potentially minutes away from a running engine.

4) Also, in this state, with no ignition or injectors hooked up (and no fuel pump active), I like to turn over the starter motor with the MicroSquirt turned on and MegaTune running. This is a big current draw and if there are grounding issues you will see the sensor readings jump all over the place - if so then check your engine-to-battery-to-chassis grounds (see the sticky on this). Less of an influence but perhaps more fun is to turn on the windshield wipers, heater fan, headlights, and even blow the horn - each of these can produce noise on the +12V feed and it is better to find this out now so you can rectify instead of tracking down, say, a random tach glitch that is really caused by battery feed noise - it happens....

5) Next, I hook up the tach input source, and I check for tach signal with MegaTune with the starter cranking. If you are using EDIS, GM DIS, or HEI module it is often better to test this along with the outputs (next step).

6) Finally, I hook up the fuel pump relay, injectors, and ignition. And then start it up (they always start right up, or at least do something), tweak the tune, and drive away.

A lot of the above is available at the http://www.megamanual.com site, tons good info here and an absolute must-read before installing your MicroSquirt!

- Bruce

[geoffct]

What is the true limit of Vref? I have seen 20mA, I know I have exceeded that, and will have to in the future.

My system uses 11mA for a cam sensor, 8mA a piece for a couple of map sensors, 1mA for a 5k tps, and whatever the microsquirt itself pulls. Thats 28mA right there. I don't think this is too extreme.

Also I do not have a serial port on my primary laptop and would like to use a bluetooth to serial converter which pulls upto 55mA.

Unless I am missing something, the circuit of the microsquirt is the same as the V3 which is claimed on the bluetooth page to be able to drive the 200mA Socket bluetooth adapter.

Thanks
Geoff

[Bernard Fife]

geoffct,

Bruce can obviously comment better, but I believe the Vref can source a few hundred milliamps. Vref comes essentially straight from the 1 Amp LM2940-5.0 voltage regulator, so as long as the on board requirements (< 200 milliamps, and usually about 100 milliamps) the remainder is available for external use. (It's possible the regulator has been changed to a different spec - Bruce will know.)

Perhaps the 20 milliamp limit you mention is the port pin driver current limit, which is a different thing entirely?

Lance.

[geoffct]

I am referencing the introduction to uS page which states at the very bottom:
"Maximum permissible source from Vref: 20 milliamps"

Geoff

[Bernard Fife]

Geoff,

Bruce can comment, but I suspect that is a typo.

Lance.

[Bruce Bowling]

What is the true limit of Vref? I have seen 20mA, I know I have exceeded that, and will have to in the future.

My system uses 11mA for a cam sensor, 8mA a piece for a couple of map sensors, 1mA for a 5k tps, and whatever the microsquirt itself pulls. Thats 28mA right there. I don't think this is too extreme.

Also I do not have a serial port on my primary laptop and would like to use a bluetooth to serial converter which pulls upto 55mA.

Unless I am missing something, the circuit of the microsquirt is the same as the V3 which is claimed on the bluetooth page to be able to drive the 200mA Socket bluetooth adapter.

The regulator on the MicroSquirt is a surface-mount version, and the heat sink is a copper area on the board plus the ground place (there are several "vias" stitched around the regulator for this).

The MicroSquirt, "unloaded", meaning just the input sensors (with MPX4250 MAP) and serial, no outputs, has an overall current draw of roughly 55 milliamps. The temperature rise of the regulator case is about 8 degrees above ambient temperature (deg F) with a 13V input power supply.

The power dissipation of the regulator is simply the voltage drop across the regulator multiplied by the current. In the case above, the voltage drop is 13 volts minus 5 volts (Vout - Vin) = 8 volts. The power in watts is 8 * 0.055 = 0.44 watts. This is the amount of heat generated, and in this case the PCB has to dissipate about 1/2 of watt of heat, which is it quite capable of doing.

Now, lets take an extreme case - say the same 13V power supply, but now try to draw 1 amp of current. The math is easy = 8 volts * 1 amp = 8 watts. Think about the MicroSquirt PCB having to dissipate 8 watts! The board dimensions are 2.5 x 3.5 inches - not a lot of surface area to radiate heat, for sure.

The 20 milliamp number I put in for the Vref spec came from temperature measurements of the PCB while adding load to the Vref line. I have to go back to my notes, but if I recall, the 20 ma additional load resulted in a PCB temp rise of 20 degrees above ambient. I could pull more current but then the board will rise in temperature. Think of a relly hot day where the ambient temp is high, then add another 20 degrees on top of this for the regulator.

On the PCB itself, there are two mounting holes right next to the regulator. They are there for those people who want to pull a little more current from Vref - simply add a heatsink setup to these mounting holes. How much more current you can obtain with this just depends - use an infrared thermometer and monitor the temp rise under the added load with the heatsink. The heatsink could also be bolted to the case as well.

I know people have asked why not split the Vref signal from the other 5V logic supply with another regulator on the uS board? Sounds good, however the same amount of heat is being dumped in on the PCB even with two regulators (we learned this lesson the hard way with the sequencer - hence the move to a switchmode setup there). Yes, both will contribute less, however the the board will again equalize at the same higher temperature. For those who have run the uS with the PCB out of the case you will notice the board get warm, and for the most part the heat travels across the entire board. This is because of the thermal sinking to the ground plane, which covers the majority of the board. If we had a really large PCB splitting the power up may work, but there really is not a lot of surface area to be dumping a lot of watts into, even if they (heat sources) are distributed.

Finally, remember this current rating is for the Vref +5V line. It does not affect the current capability of the drivers - these are referenced to the +12V battery and only share the common return ground path with the Vref. The relay drivers also are control +12V lines, only the gate drive for these drivers are using the +5V from the board, and this is already accounted for in the ~55ma number.

- Bruce

[Bernard Fife]

Bruce,

Thanks for that, I stand corrected! If anyone needs more than 20 milliamps, implementing a separate 5 Volt regulator on a bit of proto board ought to be easy enough - start with something like a LM7805 or LM2940-5.0 in a TO-220 package (their data sheets should provide all the info anyone might need).

Lance.

[geoffct]

Thanks guys,

Lance, wouldn't adding a heatsink be tremendously easier and better than a separate 5V regulator?

I will be adding a block aluminum heatsink to the pad unless you recommend otherwise. When I am "done" working on the guts I expect to open the back side of the case and add some thermal grease between the ignition coil driver heatsink and my new heatsink and the case. It should keep things cool.

Thanks again,
Geoff

[Bernard Fife]

Geoff,

Yes, probably. But it may not provide all the current every user wants/needs. It also requires some monitoring.

When that isn't sufficient (i.e., it gets too hot at the required current level) then a separate regulator (which is cheap and easy to wire) may be the next step.

Lance.

[geoffct]

I made a heatsink. I measured the pad as best I could, but if you could post the dimensions of the pad and holes that might help people in the future. I measured it as 18mmx5mm, with a the two #4x40 holes along the center line spaced 3.25mm off then back face and 10.25mm between the holes.

Also I noticed the VB921s heatsink is out from the case, so I will be replacing that heatsink with 1/4" thick material.

Geoff