Not getting 5V to CPU socket! Help!

[robd]

Hi, I've assembled my MS2 V3 DIY kit and am now trying to get it working. I've connected it to the car and done some testing with a mulimeter and I'm getting all the correct readings it seems to U5, so I've got 10V(should this be 12V? Dodgy ground maybe?) to the pin closest to DB9, then under 1 Ohm to middle pin and then 5V to third pin furthest from DB9.

However I'm not getting 5V to the CPU socket, all the pins that should get 5V on the CPU socket only seem to get about 0.3V so I'm guessing there is a short somewhere. I've gone through all the power supply assembly checks and all the components are the correct way around where they need to be and there are no solder bridges that I can see. I've jumpered S12C to JS9 and I'm getting 10V(again should this be 12V?) to CPU socket pin16.
Also not sure if this helps, but I'm also getting about 1.9V to pin27 on the CPU socket!!

So my question is, does this sound like a short on the 5V circuit and if so, do I just need to look for anything that could be shorting on the 5V circuit, or could other components be affecting it as well? I've been looking at the below diagram which shows the 5V circuit, should I just be working off this? Also if I have anything connected to the 5V on the proto area, will I need to rule this out as well?

I've cleaned the board with isopropyl and rinsed it and I can't see any visable solder joints except for possibly one but its very hard to tell, this is Q2.

If anyone can point out anything obvious would be great as I really dont want to have to pull up things one by one as it's a real pain as I can't really be in the car with power and working on the board at the same time!

Thanks in advance.

[elaw]

Here are a couple of things to check. First, measure the voltage on both sides of F1. If you have 5V on one side and 0V on the other, there could be a short. If you read 5V on both sides, the problem could be an open circuit.

But another thing people have trouble with is the CPU pin numbering... are you sure you're understanding it correctly?
If you look at the board like this:

...pin 1 is at the lower left. The numbering goes up the left side, so the upper-left pin is pin 20. Then it goes to the right-hand side, so the upper-right pin is 21, and then down the right side so the lower-right pin is number 40. So the CPU is "upside-down" compared to some of the other chips on the board.

[Matt Cramer]

Do you have 5 volts on the U5 regulator itself?

[robd]

Here are a couple of things to check. First, measure the voltage on both sides of F1. If you have 5V on one side and 0V on the other, there could be a short. If you read 5V on both sides, the problem could be an open circuit.

But another thing people have trouble with is the CPU pin numbering... are you sure you're understanding it correctly?
If you look at the board like this:

...pin 1 is at the lower left. The numbering goes up the left side, so the upper-left pin is pin 20. Then it goes to the right-hand side, so the upper-right pin is 21, and then down the right side so the lower-right pin is number 40. So the CPU is "upside-down" compared to some of the other chips on the board.

Thanks, I'll do that test on F1 and see what the voltage is, anything else I could test at the same time? Also am I thinking correctly in that I mainly need to look at things on the 5V circuit? Could I literally test every component in the circuit from U5 on to see if I had 5V on one side but not the other to see where the fault is, or doesn't it quite work like that?! Sorry I'm a bit pants when it comes to electronics! Doh!

I thought it was pin1 at the bottom right? I'm sure that's what the megamanual said! But I did test every CPU socket pin and there wasn't 5V anywhere, I also tried it by grounding to the car so even if I got numbers wrong I know I had a ground. Pin16 did have 10V though which seems correct as I've jumpered S12C to JS9!

Do you have 5 volts on the U5 regulator itself?

Yeah I've got 5V at U5. Tests at U5 all came back ok, it's just 5V at the CPU socket I dont seem to have, but I did have 10V as said above and also a random 1.9V at pin27(or pin13 if I'm looking at it wrong way!)

Thanks very much for the help so far, keep it coming, I really hope it's something simple as I can't wait to get it running.

[Matt Cramer]

Check if you have 5 volts on each side of the inductors (L1 and L2).

[robd]

Ok, just been outside in the car and tested the ecu again with the above tests. I'm getting 5V on one side of F1 but not the other and not getting 5V on either L1 or L2.
The side I got 5V on F1 was the side closest to the heat sink if that helps?!

So does this sound like I've got a short somewhere? If so as I said before, would it only be something on that 5V circuit I'm looking for or could any shorts anywhere on the board affect it? The only one from looking at it there's a small possibility of a short is Q2!
Also I have the Joe P mods on my board so am also utilising both 5V holes on the proto area, I guess I would need to rule these out as well for shorts? If so, could I just de-solder the wires to the 5V on the proto area to rule anything in the proto area out?

Hope someone can help and give suggestions on what to try next and I can do some more testing tomorrow then.
Cheers
Rob

[elaw]

Well... you're in a bit of a pickle because there are a large number of things in the circuit that are connected to +5V.

On the not-5-volts side of F1, what voltage are you getting? Zero or very close (like < .2 volts) would indicate a short. For a short, I'd start with a really close visual inspection to make sure there aren't any solder bridges anywhere. Also as you said, those wires in the proto area, try disconnecting them. The MAP sensor runs on 5 volts too - check its connections carefully.

If the voltage is a little greater, like 0.4 volts or more, it may not actually be a short but a component installed improperly - probably backwards. Possible parts include D19, C22, U6, and the CPU daughterboard. I can't see Q2 causing this problem.

[robd]

Thanks very much for the reply, I do hope I don't have to go through the process of pulling up everything, but I suppose it's my own fault for not testing it earlier before I finished the assembly.

On the not 5V side of F1, I was getting just over 0.3V so there was a reading there, just not a lot! I'll visually re-check all those components you mentioned again although I have checked them all already a few times, so unless I'm missing something everything should be the correct way around.

I'll also try disconnecting the wires to 5V in the proto area and check around the map sensor. I may just remove Q2 as well just to check.

Thanks again.

Edit: Also don't know if this helps, but I was getting about 0.3V on L1 and L2 too, and when I test the CPU socket pins where there should be 5V I was getting about 0.3V there as well!!

[SymTech Laboratories]

We recommend carefully lifting one side of D19, then re-checking the voltages. D19 is a very common failure point; it's usually the first to go if 12V is accidentally applied to the 5V rail.

[robd]

Just checked the board again, and everything is definitely orientated the correct way so I'm 99.9% sure it's not that unless I'm being completely stupid and missing something!

In work at mo so can't try lifting D19 but will try that tonight.

Another question on this, will the 5V be stopping at F1 then, or could I test each component in the 5V circuit to see which one specifically is causing the issue? I assume if there is a short or bad component after F1, that's breaking the circuit, so they would all read the same voltage so I would have to just pull each one up and see if I then get 5V on both sides of F1 so I would only be able to narrow it down by lifting one component at a time!
I guess I've just answered my own question there!

[robd]

Ok, just done some testing and got some interesting results! To save time, I've just connected a car battery to my MS, so I've used my harness adapter that would normally plug into car loom, and just connected 12V from battery to 12V on loom adapter, then connected ground on battery to chassis ground pins 18&19 on loom adapter.

CPU was unplugged before I did any testing.

Now when I power up the MS, both LED 14 and LED 16 light up and stay on, I must mention neither of these lit up last time! Not sure what this means!!?!

Also I tested the 5V circuit, and with just the 12V and ground to MS, I am now getting 5V to both sides of F1, also to L1 and L2 and I'm also getting 5V to the correct pins on the CPU socket. However I'm also getting 5V to some other pins on the cpu socket that I'm not sure are correct?!

I'm getting 5V to the following cpu socket pins, 1, 12, 13, 14, 15, 17, 20, 21, 22, 24, 25, 31.
I also get 12V(well 11.5V) to pin16 which should be correct as I've jumpered S12C to JS9 as previously mentioned.
I also get 1.8V to pin23 and 2V to pin 27!

So at first it was looking like maybe I've done something wrong on my wiring harness adapter, or maybe I need to disconnect something from the stock loom that I don't need, like my TPS or Fuel Pump maybe?!

So my questions now are, what does both LED 14 and 16 lit up mean with no processor installed and are those above voltages correct. If it helps I'm running the following mods on my board which use 5V and my car is a 1990 Mazda MX5 1.6.

This is the assembly guide for the components I need I followed, I'm using the Standalone setup with a harness adapter..
Assembly Guide

This is the mods I've had to do, all barring the yellow changes I've been told..

And also this fan control relay mod(replace Q7 with Q4, run wire from top of D15 to 1AC1B)...the black lines are the Joe P mod as well, but is wired up differently for some reason!


And this is the Joe P mod, now I've done my wires from D16 to IAC2B and D14 to IGN as decribed on his page, not as shown above to swap them around.
Joe P Spark Output Mod

Also another idea just popped into my head, I've made my harness adapter for a VTPS, but I haven't got that side setup yet. Now you mentioned before about an open circuit. If my stock On/Off TPS is still connected, but my harness adapter is wired for a VTPS with the +5V, not having the VTPS connected, could this be causing an open circuit and that's why I'm now getting 5V on the MS with just power and ground to my MS?
Here is how I've done my harness adapter..


Ignore the 12V for the fuel pump though, I'm doing it a different way so that wont be wired up.

Hope this all helps and what I've tested so far now is good news?!

Any further help would be greatly appreciated and let me know if I need to test anything else!
Thanks very much,
Rob

[SymTech Laboratories]

I'm getting 5V to the following cpu socket pins, 1, 12, 13, 14, 15, 17, 20, 21, 22, 24, 25, 31.
I also get 12V(well 11.5V) to pin16 which should be correct as I've jumpered S12C to JS9 as previously mentioned.
I also get 1.8V to pin23 and 2V to pin 27!

Those are all reasonable voltages you measured.

So my questions now are, what does both LED 14 and 16 lit up mean with no processor installed and are those above voltages correct. If it helps I'm running the following mods on my board which use 5V and my car is a 1990 Mazda MX5 1.6.

Q6 and Q8 are either "on" somehow, or their collectors and emitters are shorted. This could be due to the ignition output modifications you did.

Also another idea just popped into my head, I've made my harness adapter for a VTPS, but I haven't got that side setup yet. Now you mentioned before about an open circuit. If my stock On/Off TPS is still connected, but my harness adapter is wired for a VTPS with the +5V, not having the VTPS connected, could this be causing an open circuit and that's why I'm now getting 5V on the MS with just power and ground to my MS?

An open circuit on the TPS input won't bring down the whole ECU. It shouldn't have an effect on the internal voltages at all, unless maybe you accidentally connected pin 26 (TPS Vref) to ground.

[robd]

I'm getting 5V to the following cpu socket pins, 1, 12, 13, 14, 15, 17, 20, 21, 22, 24, 25, 31.
I also get 12V(well 11.5V) to pin16 which should be correct as I've jumpered S12C to JS9 as previously mentioned.
I also get 1.8V to pin23 and 2V to pin 27!

Those are all reasonable voltages you measured.

Awesome, well that's good news at least as I was worried I shorted my cpu socket...phew!

So my questions now are, what does both LED 14 and 16 lit up mean with no processor installed and are those above voltages correct. If it helps I'm running the following mods on my board which use 5V and my car is a 1990 Mazda MX5 1.6.

Q6 and Q8 are either "on" somehow, or their collectors and emitters are shorted. This could be due to the ignition output modifications you did.

Right ok, do you think that's likely to be that I've wired something up wrong on the mod, or maybe the mod just does that as it's using those two LED's? I know part of the mod is so the coils won't fire until the processor does, so maybe that's done something so the LED's are on when processor is out! I'll try putting the processor back in now I know the voltages to cpu socket are ok and I'll see if the LED's go out.
On a normal setup, should they light up if you've just got ignition on and the car isn't started(with processor in)?

Also another idea just popped into my head, I've made my harness adapter for a VTPS, but I haven't got that side setup yet. Now you mentioned before about an open circuit. If my stock On/Off TPS is still connected, but my harness adapter is wired for a VTPS with the +5V, not having the VTPS connected, could this be causing an open circuit and that's why I'm now getting 5V on the MS with just power and ground to my MS?

An open circuit on the TPS input won't bring down the whole ECU. It shouldn't have an effect on the internal voltages at all, unless maybe you accidentally connected pin 26 (TPS Vref) to ground.

Ok thanks, well I assume I've done something wrong as I got 5V with just car battery connected and powering ecu, but when I use the harness adapter I don't! So I'll check all my connections again and make sure I haven't wired them up wrong. Does it sound reasonable that if I've wired something up wrong on my harness adapter that it would then affect the voltages on the ecu? Or should this not happen at all?

Thanks very much for all the help so far, really appreciate it. I wish I'd bought all my stuff to work with me today now, itching to test it some more!!

[robd]

Just another question on this before I try it, with the above voltages measured at the CPU socket, does it sound safe for me to put the processor back in and test it? Just want to make sure I'm not risking killing the processor.

[elaw]

It looks like the CPU would be safe but CPU or not I'd strongly advise one thing: make sure there's a fuse in the line powering the Megasquirt! The fuse should be no larger than 5 amps.

As far as the two LEDs being on, if I understand the mod you did that's to be expected. If you take a 2N2222, conect its emitter to ground and its base to the cathode (pin 1) of one of the LEDs, the LED will light up. That's because the base-emitter circuit looks like a diode - when supplied with more than about 0.5 volts, it conducts. So the LED cathode is essentially being grounded by the base of the transistor you added, which causes the LED to light.

[robd]

It looks like the CPU would be safe but CPU or not I'd strongly advise one thing: make sure there's a fuse in the line powering the Megasquirt! The fuse should be no larger than 5 amps.

As far as the two LEDs being on, if I understand the mod you did that's to be expected. If you take a 2N2222, conect its emitter to ground and its base to the cathode (pin 1) of one of the LEDs, the LED will light up. That's because the base-emitter circuit looks like a diode - when supplied with more than about 0.5 volts, it conducts. So the LED cathode is essentially being grounded by the base of the transistor you added, which causes the LED to light.

Thanks very much, well hopefully that explains the LED's then and will clear that up.

On the fuse, is that because I'm powering it off the car battery without all the normal car fuses attached? If so and I was to add a 5amp fuse in between my DB37 connector and harness adapter connector, am I ok to leave that there when I then connect it back up to the car? Would it then just be one extra fuse on the 12v wire? Also if I was to just connect it back to the car now, is it still advisable to have a 5amp fuse on the 12v wire or is the standard car fuse setup fine to protect the MS?

Sorry if that's a stupid question, I'm literally learning as I go here and electrics has never really been my strong point as you can tell!!

I'm really wondering now if it's my TPS causing the issue here. From what I understand so far, my standard TPS has 3 wires but is just an on/off switch type TPS. I've wired my MS up for a vTPS using pin 22 and 26 and then I assume it would use one of the normal sensor grounds. So if the MS wiring differs from my stock TPS 3 wire setup, that could be causing a short yeah? So if this is the case, would simply disconnecting the sensor(connector) from the TPS on the car in the engine bay resolve the issue and prove if it's this causing the prolem? Hope that makes sense!

[elaw]

The fuse thing is just a basic principle... the Megasquirt should always be protected by a suitable-sized fuse. If the fuse in the car is the right amperage (3A preferred but a little hard to find, otherwise 5A) then you're fine... when operating it in the car. If you're running it directly off a battery, you definitely should have a fuse in the line - a battery can put out 100x the amount of current needed to destroy the Megasquirt so the MS needs to be protected.

Re having two fuses, it's not likely to cause a problem... unless one blows then you have the headache of figuring out which one it is!

The TPS being just a switch could definitely cause what you're seeing if it's wired wrong. If you're stuck with that TPS, there's really no point connecting it to the Megasquirt at all as the MS can't use a switch as a TPS input. So you could just ground pin 23 on the DB-37 instead. But... is installing a potentiometer-type TPS an option? You're likely to have much better luck with the Megasquirt if you have a proper TPS.

[robd]

The fuse thing is just a basic principle... the Megasquirt should always be protected by a suitable-sized fuse. If the fuse in the car is the right amperage (3A preferred but a little hard to find, otherwise 5A) then you're fine... when operating it in the car. If you're running it directly off a battery, you definitely should have a fuse in the line - a battery can put out 100x the amount of current needed to destroy the Megasquirt so the MS needs to be protected.

Re having two fuses, it's not likely to cause a problem... unless one blows then you have the headache of figuring out which one it is!

The TPS being just a switch could definitely cause what you're seeing if it's wired wrong. If you're stuck with that TPS, there's really no point connecting it to the Megasquirt at all as the MS can't use a switch as a TPS input. So you could just ground pin 23 on the DB-37 instead. But... is installing a potentiometer-type TPS an option? You're likely to have much better luck with the Megasquirt if you have a proper TPS.

Ok great, I'll check what amp the fuse is in the car then first. I may well fit a fuse to the 12v line then anyway even if the car side is all good, my ecu will be pretty easy to get to in the car anyway so checking the fuses isn't a problem, and I'll definitely fit a fuse if I run it off the car battery again then. Thanks for that.

I'm getting more and more confident then that it's my TPS causing the issue. I will be switching to a vTPS as I'm going ITB's, but I don't have the ITB's on the car yet until I'd tested the MS to make sure that was all ok! lol The original reason I wanted to connect it to the car before fitting the ITB's, was just to make sure I'd done it all correctly and update the firmware so I could start playing with a map to get some form of base map going. I know I need to create a new map on the car, so was planning on just connecting the MS to the car, temporarily attaching my vTPS sensor to the MS/TPS wiring, obviously the correct way around, so I could then create a new base map which I think should be ok to do, so I can hopefully get it running even if it's not great, just to get me the 5 mins down the road to work with no WOT action or anything like that, just limping it along!

[robd]

Yipeeeeeeeeeee
Unplugged the TPS from the car and tested the MS and it worked, all the voltage readings came back ok and the LED's lit up just as they did when I tested it off the car battery.
So I turned it all off, unplugged my coils etc and put the processor in, fired it up, connected first time and it now has updated MSExtra firmware and I was able to connect it and create a new project in TS, so a very happy bunny today now!

A massive thanks to everyone for their help, really apreciated.