You buggers beating me to it again...
My vote for high power fets, TO220 easy to mount vertically or horizontally and plenty of devices to chose from.
That's no good for ignition on board and a terrible waste (2mA with a to220?? lol) for driving external ignitors.
Hmmm, should I title one of the general outputs as fan? Kind of like I did with the fuel.
If you are installing GP FETs put them on the PWM pins as that is what will be top of the list of things to code for.
My post :
jharvey wrote:- LED resistors are now 3k so much less loading, a 5v signal should produce 1.6ma
It should be 12v for a bunch of reasons. Number one, keep the load AND noise off the regulated lines. Number two, reduce the reverse potential applied to the LED by kick back. Number 3, 1.6mA is pushing it ;-) 3k is a good value for 15v, hook em up just like the load.
Fred did you mean I should have a R for pins 9 and 11?
No, I meant across the points P45 and P46 as an input impedance that the VR signal must drive.
- Moved the injector LED such that it lights when current flows.
To 12v right?
- Changed AN's per Fred's request, but don't plan on them there, I expect to let PCB layout rule what should go where. I don't know of physical reasons why they should be that way, so if I can avoid crossing over wires at the PCB, I'll likely change them around. I will try to keep them as noted above.
There is more to where they go than just making the PCB simplest. Firstly I have and have had for 6 months a setup on my desk using them in this way which will be a pain to change. I don't want to waste time (significant time) reworking that for a small benefit. Secondly, we need to have them in some consistent and sensible order for clarity of instructions. If you show me a good reason to change them at some point, I'll consider, but if you do, thats more time I'm NOT developing the firmware that we need to make this thing happen. Having a couple of traces crossing is way way way less important than me wasting half a day screwing around changing hardware IMO. Obviously if it makes the board a mess, I'll change, but if it's minor the traces should cross.
- ignition current sensing... I wonder if that is what the 4th wire in my wife's Jetta is doing. I'll add this to the TODO 2.0 list. Don't tell Fred, but I'm kind of tempted to move the injector current sensing to 2.0 as well.
Shhhhhhhhhhhhhhhhhhhhhh
- FET on the thermistor is required for my application, and is a good idea on a normal setup. I don't have a heat sink for the thermistor, and it will go up in flames with out the FET duty cycle keeping the overall power draw down. I plan on default wiring it as shorted, then I'll have to cut a trace to make it work Most people will simply ignore three little holes.
This I care about a lot less than the injector ones, esp with the trace bridged. I have to ask though, what the hell kind of sensor are you using that will "burst into flames" let alone distort the reading enough to care about with a 2k ish resistor feeding it? Hell, you could feed 5v straight into most thermistors and they would be OK. Were you just speaking colourfully?
- about "an protect" and "digi protect", they show extra parts, most will probably be jumpered thru anyhow. For now I'm leaving in the pads so we can add parts if required. I expect after a proto of 1.0 is done, many parts can go away, like the 0R in the injector schematic.
but but but, the 8 EXTRA ADC channels are extra. IE, they are for functions that code won't be written for for months. As I see it, the board should be focussed on "what we will need now/soon" rather than "what we might need one day". This goes for all aspects. We need injector drive, but we want low Z drive. We need separate grounds and powers up front, we want an RTC on board. We need the 6 core sensors, we want the other 8 adc inputs. etc.
- about the clean and dirty supply, is there a dirty supply on the PCB? Injectors, fuel, ignition, ect will be directly wired through relay's to the battery, then grounded via PCB. I know the injectors show a 12V lead, but that really shouldn't be connected right. Ignition also drives external power IGBT's right, so dirty is even farther away there.
We want a 12v dirty in the case so that people that need 12v to expand the system don't steal it from places that would screw up the system. We shouldn't be putting extra stuff on the board, but we should be planning to plug in another card with xyz functions on it.
- I think we want some LED's from the CPU. One min to tell us CPU is running, and other(s) for potential feedback, but mostly for places to solder wires if we want to connect something to them. If we have something else that goes there, I'm happy to remove LED's.
Well, those are supposed to be the staged injector channels. If you just want to verify the thing is on and running, put a single one on the same pin as the "user led" is on the TA board, one of the PWM pins. Past that LEDs are a total waste of space/time/effort when you have serial comms functioning properly. The only reason I wanted any LEDs on the board is to visually verify that outputs we are actually using are being switched as they should be. Give the fuel pump drive a led, the injectors and coils a led each and past that, nothing IMO, though you can put one on the pwm pin as an "i'm alive" indicator. That is useful. Each LED should show that something real is happening, not be there on it's own as some signal. If you are going to put any on, you should use a 2k resistor to keep heat out of the CPU. We shouldn't be encouraging people to "connect something to them" at all. Everything should be buffered. People that show us non-buffered outputs should be kicked firmly in the goulies and told to fix their setup before being helped :-)
Just looking at A13... the purpose of on board LEDs is to verify at the end of the signal path that stuff is happening. For the outputs, it should be on the output outside everything else. For the inputs it should be after absolutely everything else, ie, right on the CPU input or just outside the 1.6k resistor or something, not at the input from the world. We need to verify that the CPU pin itself is seeing signal, and for outputs we need to verify that the device, or at least connector is seeing signal. Do you see what I mean?
Hmmm, I don't think the LED in the ignition schematic works this way. Also Fred is this the way you objected to? Should I put it back the way it was, or should I put it differently.
I don't know which are ignition as none are labeled as such. All the LEDs appear to be connected to 12v with a 3k though, so thats all good :-)
- about the zener in the reg, I know it's not a crow bar, but it does help with OV. Should it stay? Don't have to populate it if you don't want to.
The thing is, if there is an extended fault it will burn and may take the board with it. Caps will "just" explode if exposed to higher than spec voltage. I think it should go as in it's current form it is not useful. If you want to force it low then Deltas idea is a good one, but we decided not to do either I thought.
Oh also still looking for footprints for the ignition driver FET/IGBT. Should I just use a TO220? I'm guessing it's going to be a much smaller package.
I thought we decided to use IGBT on board as the default configuration? We definitely decided to use XOR chips as the buffers/drivers for them anyway. The actual IGBTs are to220 devices, so if we are going with on board ign then that is what you should use. Some options are :
BIP373 :
http://www.megamanual.com/ms2/bip373_datasheet.pdf - new default MS ignition driver with protection (copy url and paste in new tab)
ISL9V5036P3 :
http://www.fairchildsemi.com/ds/IS/ISL9V5036P3.pdf - people are using this with success
irgs14c40l :
http://www.irf.com/product-info/datashe ... 14c40l.pdf - people are using this with success
I think they all have the same pin out, but double check that.
Right now we are making a proto, that will have extra stuff in it. Like test points for inserting in line amp meters, and extra holes for parts we may or may not want. Rev's will remain A.XX or (heaven forbid A.XXX) until the proto is ordered.
Dude, it's an EMS which in it's basic form is a dead simple circuit, what exactly would we want to measure with an ammeter? If you put in test points like that, they should be hard wired by default like the thermistor stuff. I think we already know what we do and don't want on the board, it's more a matter of figuring out how much of it can fit reasonably.
Good work, I still haven't figured out the correct pins for a few things, so I need a little more time for that.
Fred.
