ashmason

Hans, would you spend the extra to put diodes and fuses if it were going in your car? I am on the fence, the more parts there are, the more chance of things going wrong, but on the other hand I would rather have the extra protection if you feel its justified. If its over-engineering for a 1% chance then I vote no, but if its more probable then I vote yes spend the extra.

As far as the fuses go, same thing really. I don't mind pulling the unit once in a blue moon to change a fuse and would rather the fuse pops than damaging components.

finally!

Vote yes for the $7.20 safety net

928er

I would say do what is right to make this as problem free, and follows the plan to make this a true upgrade.

hans14914

Borland,

In theory, yes, I could have the heater valve solenoid energized for a delay period, but since the valve itself is fail open, it won't do anything without the vacuum from the running engine.



I am on the fence to regarding overengineering. The TVS array is just one clear example. There are other components that could be trimmed. The difference between the current planned design and a trimmed down "bean counter" version could translate into $20 or more in costs... Which does not reflect the true retail increase because of labor in assembly and other intangibles.

I will sleep on it and see how it looks in the morning. Leaning towards designing most of the protection in, and making it in a way that they can be bypassed or simply left off if need be.

Thanks
Hans

Jadz928

Hans,
Good move to sleep on it. I'm sure you have consulted folks with a more vested interest in the product launch.
It's really up to you guys to know your market and what they want. Only you know what your "sweet spot" cost targer really is.
I could say you should design the most robust system, because I would be willing to pay for it no matter what. But then, my voice may not be speaking from your prime market. Then again, it might.
Your call, bud! GL.

BTW, the developments read awesome, although I only understood 1/10th of it.

Dictys

As above, really. I would prefer to have the protection built in despite the extra cost.

borland

Hans,

Actually the car has a vacuum reservoir, so if the system is acting normally (no gross leaks), there should be plenty of reserve to handle such intervals.



BTY, those solenoids in the above schematic are shown as two-way valves, but they are actually 3-way. Thanks for giving it consideration.

blown 87

Plus one.

And I also want to say plus one to do not listen to us, you are the designer, you need to do what you feel is best.

sendarius

I vote for including the protection.

The original was pretty robust - crude as only dated electronics designs can be, but robust.
Having a more elegant and contemporary, but less robust, replacement design kind of defeats the purpose.

EspritS4s

Hans,
IMHO, the fuse to protect the AC relay seems like overkill, particularly if it offers more protection than the original system. As for the transient voltage suppressors, it's hard to advise on this without knowing the MCU that you have chosen as well as your power supply design. Does the manufacturer offer any app notes on using the MCU in an automotive application? The 12V supply in a car is incredibly noisy, but there are multiple ways to deal with this. I wouldn't include TVS on the voltage supplies for extra protection unless really needed. Are you using switching or linear regulators for the 5V supplies? I'm assuming that you are using a linear for the low current 5V, but how about the main 5V? This is probably not a decision that should be made just based on cost or people's preference about over engineering. Again, just my opinion as a (former) board designer. Personally, from a cost perspective, I say go for it. However, as a designer, I say over protection is an unnecessary complication.

blown 87

Smart folks put fuses on the original system to protect that relay.

hans14914

Gary,

Thanks for the feedback. I am using an Arduino Nano for the computing core. The reason being is that it has the FTDI serial/usb bridge on board, and that is an expensive chip. I can get the Arduino Nano assembled for less than I can source all the individual components and assemble them here... so its a no brainer for me. Plus it allows me to socket the assembly, so it can be field replaced quickly if its ever necessary.

Regarding the power supply, I am using linear LDO regs for both 5v supplies, and the 3.3v for the LCD. Both of the 5v regulators are automotive grade, with internal load-dump protection and other automotive design features. I am using low ESR tantulum caps on the power supply, in accordance with manufactures specs, so the voltage should be pretty clean.

I am using an ISOFACE driver for the solenoids and external drives. I have measured that the soloenoids pull upto a constant 500ma with the standard inductive flyback. The ISOFACE is rated at 1.2a/pin and uses internal galvonic isolation. Of course its not true isolation in this application, as we are using a common ground. The chip may seem overkill, but its cheaper than using relays, and reduces assembly time significantly over several discrete driver circuits.

There is kind of an odd discrete digital pot as part of the design, as I was not able to find any common component to replace the standard pot. I have a shift-register driving the pins of a standard ULN2803 type octal NPN Darlington relay driver array. The first 5 pins are connected to small signal relays rated at 2A which bypass a 0.5w resistor. The, values are 100, 200, 400, 800, and 1600ohm, which gives all 20 discrete 100degree increment positions for the 20 degrees the system can resolve. by engaging the relay, it provides a bypass to its paired resistor, allowing it to be switched in or out... but it will default to hot under system failure, as in the original design. The 6 and 7 outputs drive relays that short to ground through a resistor or pull up to +12v as the factory microswitches do. The last driver is used to engage the AC relay. I am using a socketed DIP style DPDT relay with the two poles summed to increase the power handling. If the AC relay does burn out, a new one can be plugged in with ease.

The AC functions have been broken out into discrete sections, so that they can be addressed individually in the code. I will make the code and the schematics open source to purchasers of the system to help foster community improvement. I am not the best electronics designer on this board, and certainly not good at software, but I hope providing these "tools" will encourage people who are good at these things to improve the original design.

Back the the TVS array, I was going to put them under the socket for the Arduino, and to just blanket protect all the pins. I do have one voltage divider which has limited isolation. It has a zener and reverse current protection diode. It is used to monitor the voltage on the illumination input to set the dimming for the local LEDS. Most of the other components are reasonably isolated, but its always tough to try and guess what is going to happen in an automotive environment.


Borland-

If you think the internal vacuum will stay active that long, I can probably include something in there for you. All the hardware supports addressing individual elements, so if its not included in the first software revision, it can be added relatively quickly thereafter. Update is via a simple USB connection to a laptop.

Thanks
Hans

blown 87

Wow, something I finally know about in this thread.
If the system is in good working order it will hold for a very long time.

hans14914

OK, back on track after trying to get some things ready for Sharktoberfest. During the trip, I was bumped on the inbound flight, so I had a captive ~4hours to tweak the schematics without interruption. This was a great opportunity in disguise, and got me motivated to finish it up when I got back.

Schematic capture and package drawings for the "Final" climate layout are all complete now. Everything has been transitioned over to SMD in hopes that I can get everything on a small board which will backpack behind the display in special custom applications.

There have been some slight changes in the feature set, after calling and discussing with owners and 928 professionals in various climate zones throughout the country. The end result being that there will be complete rear AC integration into the new controls. There is a full resolution discrete 10K pot with 100ohm steps allowing for fine control of the rear system. Additionally, there is support for both using the stock resistor back for fanspeed control, as well as the possibility of using PWM to allow complete control (and linking the front and rear fans so they have the same relative speed). That said, adding these features does increase the component cost, but they have been laid out in a way that a board can be assembled with or without the rear AC components. Similarly, the rear AC an be outfitted with the less expensive control for stock resistance blocks over the PWM module.

Also included is the ability to either retain or replace the stock AC switch. This will give customers more flexibility with custom console layouts. For those who which to delete the AC button panel, I have also added a dedicated output for operating the door locks as well. This is a simple integration, and will just simulate pressing the button. There is no visual feedback from the central-locking system on the control unit. Nor is there support for the central warning light/reset. However it is possible for users to include these features... i just dont want to be responsible for the timing belt warning system from a liability standpoint. Additionally, I think it is slightly difficult to aesthetically and ergonomically include dissimilar functions, such as the locks, warning system, odometer reset into the climate head in an elegant manner. The infrastructure is there on the board if someone wants to do it.

The factory climate card-edge connector is now on it own board. This board will house all the current intensive devices, such as the relay for the AC, the speed control relays for the rear AC, and the blower motor control. There will be a dedicated fuse for the AC power system, as well as both the +12v constant and switched power buses. This board will have thick copper traces, minimum of 3oz, and will connect to the standard 1oz SMD board via a 40pin ribbon connector.

I dont know exactly how much time its going to take to do the board layouts, but I am hoping to submit files to my build-house this weekend. We should be in the physical testing stages of the "final" revision in about 10days or so.

I will post board layouts as I get them completed.

Thanks
Hans

Jadz928

Keep on rockin' the free world, Hans.