We need to lay out a basic board

From DIY contraptions to sophisticated FreeEMS-specific designs! Plus general hardware development!
User avatar
sry_not4sale
LQFP144 - On Top Of The Game
Posts: 568
Joined: Mon Mar 31, 2008 12:47 am
Location: New Zealand, land of the long white burnout
Contact:

Re: We need to lay out a basic board

Post by sry_not4sale »

I want to run my 12a on it, and plan on helping with the firmware where I can.

Only need 2 injectors for now, can live without ignition (but later, 2 plugs per rotor, split)
How many people actually think they might want a prototype board set, where
the express purpose of these boards is as a short-term solution to allow more
efficienct and effective coding, not necessarily a perfected board that is 'good' in
the sense of weatherproof, automotive temp grade components, etc... ?
+1

Want want want!
Owner / Builder: 1983 Mazda Cosmo 12at (1200cc 2-rotor turbo) coupe [SPASTK]
165hp @ 6psi standard - fastest production car in japan Oct 82
User avatar
jharvey
1N4001 - Signed up
Posts: 1607
Joined: Tue Jun 10, 2008 5:17 pm

Re: We need to lay out a basic board

Post by jharvey »

510rob wrote:Just to get an idea of the scale of things here:
  • How many people are planning on doing firmware coding?
  • How many people plan to implement a raw prototype system on an engine, and
    of how many cylinders?
  • How many people actually think they might want a prototype board set, where
    the express purpose of these boards is as a short-term solution to allow more
    efficienct and effective coding, not necessarily a perfected board that is 'good' in
    the sense of weatherproof, automotive temp grade components, etc... ?
  • Where are we at with respect to the general consensus of the ideal 'initial config
    for development'?
If I had working firmware, there's a good chance I would help with enhancing firmware, but I'm honestly not very horny about learning S12X. I just don't see that as a handy chip for me to know in my normal life.

I'm certainly willing and capable of implementing a raw proto. It's very likely I would go for it as long as the costs are reasonable. Cold weather is just around the corner, that might add some delay, as my work space is much more limited when I have to play inside.

I guess I'm tempted to encourage we take a stab at the pin out, add a bunch of vias to the layout I've drawn, such that we can cut traces and add jumpers easily enough. That way if we are successful, great. If we have to move some pins, or we find a sub circuit is bad, we can cut it up and respin sections as required.
gearhead
LQFP112 - Up with the play
Posts: 120
Joined: Sun Feb 03, 2008 9:30 pm
Location: Chicago, USA

Re: We need to lay out a basic board

Post by gearhead »

Guys, I have been really busy. Broken car, lots of work, etc.

Jharvey, I have not reviewed your schematic recently. If I remember from looking at it previously, there were hall sensors on drivers and such which are all good, but add components and complexity that most of us will not want/need. Most of this is not 'developing the circuits', but developing the firmware to run cars. The rest of the circuits are well understood and well tested. I vote we creatively re-purpose simple circuits that are used elsewhere in the field. I have nothing against an alternate to the LM1815, but suggest its use as it is well known and has been used for years on VRs with no issues in the Saab ECUs I am familiar with.

Low current drives should all be 2n2222 or equivalent and should be able to be used as I or O. Most should have a readily available 5V pullup on B and C. I am not against using fets here, either, but I have experience with 2n2222 and BS170 (fets). Also if a 12V pullup is needed, it should be able to be done with a jumper wire. I am not against soldering this across as it will be only on a few outputs (tach, drivers for external control devices) that will need 12V vs 5V. All hall sensors must be used with 5V reference. Using a noisy 12V reference for these is just not a good idea. An alternate tested suggestion would be to use a chip for this which provided 4 outputs at a time. One is a CA3236 which is used in Saab trionic ECUs to trigger the ignition, 4 outputs. I have not found a data sheet on this one, though, and it appears hard to find. A multi I/O chip with similar to 2n2222 or BS170 capability would be a plus as board consumption would be reduced.

The one thing that has been discussed but not fully agreed upon is the injector drivers. I know that there are many MS units running vnd5n07 directly off the micro pin (without fet driver). Based on this experience, I am willing to go with this for initial boards. Personally, I feel it is 'better' to have fet drivers and that is why I am running them on my install. I have not verified any of this, but can look closely at my Saab ECU and see how it is driven. It uses MTD3055s for injector control. http://www.datasheetcatalog.org/datashe ... 55VL-D.PDF. I will check tonight to see if these are driven by FET drivers, but they are also logic level FETS and avalanche rated, so they may also be driven directly by the micro. If so, I would endorse this direction as it minimizes parts and these are dead reliable.

Lets start a requirements list and make sure we hit these or strike them from the list.

Requirements:
1) Power supply
2) 6 analog inputs (CLT, MAT, TPS, MAP, TP, Batt)
3) 6 injector outputs
4) 6 low current ign outputs *no high level coil drivers on board, these should be separate*
5) Separate ground plane for digital, analog and high current
6) Separate battery sense 12V line which is not used to power any of the ECU components
7) Grounding to be done by low impedance cable (large size) to the block for each ground plane. (These can be jumpered together at the board at your own peril.)
8) No LowZ injector drivers on the board. If you want to put in a header to drive Jean's boards, I am fine with this as there may be people trying to run FreeEMS to develop it with LowZ injectors
9) Extra high and low current drivers which can be used as either I or O, have 5V digital nearby for pull up. 12V pullup does not need to be as easy to get, but could be used in the case of certain I or O.
10) no current sensing needed on board as part of layout. KISS

Any more Requirements?

Gearhead
tpsretard
QFP80 - Contributor
Posts: 99
Joined: Thu Mar 19, 2009 3:05 am

Re: We need to lay out a basic board

Post by tpsretard »

just a question, why not make the low curent outputs out of something like a vnd5n07.
they are cheep enough and can sink enough current to be useful, can drive a realy, a solenoid or even a light. then you also do not need to worry about flyback etc.

Just suggesting since we are talking about keeping it simple.
it will also help with board space.
User avatar
jharvey
1N4001 - Signed up
Posts: 1607
Joined: Tue Jun 10, 2008 5:17 pm

Re: We need to lay out a basic board

Post by jharvey »

If I remember from looking at it previously, there were hall sensors on drivers and such which are all good, but add components and complexity that most of us will not want/need.
Correct, I have some extra potential stuff in there, and in the current version of my layout, this is jumper-ed out by default. Such that the normal install won't have it. It does have some vias and pads that could be used for it, if so chosen. If one wants to try it, they will have to cut a trace to make it possible.

Low current drive? I'm guessing you mean some sort of buffer drive for the power drive. The MOSFET drives can be directly driven by the ECU, they have a very high impedance. The Darlington style devices can not be driven directly by the ECU. I recommend against the Darlington style, and get rid of the need for that low current drive altogether.

The vnd5n07 clamps to 70V, I see the VNP7N04 handles 7 amps instead of 5 amps, and is clamped to 40V instead of 70V. The lower voltage clamp might be handy. Also the VNB28N04 also clamps to 40V, but allows up to 28 amps. The VNB28N04 has a lower Rds than the VNP7N04, so it will dissipate a bit less heat when saturated. However the VNP7N04 will only dissipate around 1/8 watt, so it's not a huge difference really.

These VNXXXXX chips can be driven directly, and are likely to not have a problem. The concern is isolation voltage. IGBT's are a duel stage device that jack the isolation voltage way up. The MOSFET offer a much lower isolation voltage. This means that ESD or other transient spikes have an increased chance to propagate back to the ECU causing it to get smacked. One way to increase the isolation voltage is to have a lower current drive, like the 2n2222. However we can get the same isolation voltage with a XOR gate, and XOR has the additional benefits of lower latency, and flippable polarity.

I suggest the VNP7N04 (or VNB28N04) with XOR gate.

The board layout I've drawn up also includes digi protection circuits, but those don't do anything unless you exceed 5v or -.5v relative to gnd. So they aren't really required for a proto. Wouldn't hurt to have them, but are not required for a proto.
gearhead
LQFP112 - Up with the play
Posts: 120
Joined: Sun Feb 03, 2008 9:30 pm
Location: Chicago, USA

Re: We need to lay out a basic board

Post by gearhead »

I am of the opinion that at this point there can be too much protection especially at this phase of the project. If it takes board space and requires more components, we need to discuss it. For example, I do not think that digi protection is required if it burns board space. I have built multiple MS units and know others with less skill that have done same and only a very few burned inputs that I have read about...

As for the VND*** being able to be driven by a micro pin directly. Yes, It can. Yes, it is probably fine as there are many Microsquirts and others out there that do this. My concern is more that of consistency. I am ready and willing to be proven wrong and can even back down on this for progress to be made. I have made this point before. If you read the specs on the ST Omnifets. They are all 'logic level'. If you look at the specs, though, there are many ratings with the gate at 10V. Automotive frequencies are slow in comparison to the full bandwidth of these devices, true, but if we want reliable, snappy turn ons and offs, we may want to use a fet driver for important outputs especially for those sensitive type of timings such as ignition and fuel. I looked at the Saab ECU again this morning and it appears that they are using FET drivers for their injector drivers as well. I was unable to find the chip specs for this to confirm it via Google, but it appears to be an ST L9824D on the inputs of the injector drivers (logic level FETs I listed yesterday). I think this is a FET driver of some sort. (aside: There are a lot of 74xxxx chips which I think are buffers as well for protection of the 332 chip on this ECU as well). B&G as well as the guys running Jean's sequential board and Microsquirts would prove that driving these directly off the pin 'works'. I think Bruce claims it is merely time shifted and the extra microseconds are not important. I cannot comment on this as I have not scoped it all, but I have tried to run some omnifets at much higher frequencies and *absolutely* needed a fet driver to get them to work without overheating. For other outputs, using these instead of 2n2222s to drive coils, I am fine with as well. It just may burn board space *and* in teh spirit of Fred, these are even more SMDs and 2n2222s can be through hole...

Gearhead
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: We need to lay out a basic board

Post by Fred »

As I see it :

* Snappy turn on is more about heat than injector control accuracy
* Protection circuits are for a final production GP board that normal people will use

In the case of high imp injectors I doubt we need a driver, for low impedance with some form of PWM I doubt we can get away without it.

I doubt the isolation voltage is a practical issue at all.

I have more things to post about earlier posts, these just cover the last two posts by Jared and GearHead.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
510rob
TO220 - Visibile
Posts: 8
Joined: Sat Feb 16, 2008 1:32 am

Re: We need to lay out a basic board

Post by 510rob »

Hello world,

I've been playing with PC layouts in between my required philosophy readings & psychology papers. I know you guys disagree with P&H drivers, but I left them on there anyway. Working on PC layout stuff seems to be a good mental exercise to clear my head of scholastic & epistemic baffle-gab.

I increased the boards size to allow room for:
  • 6 injector outputs
  • 6 medium current outputs
    • 6 ignition outs
    • 1 idle control output
    • 1 boost control output)
Well, fire away! That's what we are here for - to openly discuss opinions about pathways, without getting wound up or non-objective about it.
Attachments
6-inj_6-ign_2-gp_output_board.jpg
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: We need to lay out a basic board

Post by Fred »

510rob wrote:epistemic baffle-gab.
I had to look that bugger up! :-)
I increased the boards size to allow room for:
  • 6 injector outputs
  • 6 medium current outputs
    • 6 ignition outs
    • 1 idle control output
    • 1 boost control output)
Can you ensure that the high current fuel drivers carry their own ground and power planes if and when required? I ran this on the Volvo with great success. I figured I had to do it properly right from the start ;-) Some other systems are still doing it wrong 10 years later!

Image
Well, fire away! That's what we are here for - to openly discuss opinions about pathways, without getting wound up or non-objective about it.
Amen, good call! :-)

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
510rob
TO220 - Visibile
Posts: 8
Joined: Sat Feb 16, 2008 1:32 am

Re: We need to lay out a basic board

Post by 510rob »

The board has three separate grounds - one for low power, one for high power injectors, one for high power ignition + PWM outs...
Post Reply