power supply circuit design

Go To Last Post
93 posts / 0 new

Pages

Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I was hoping some of the more experienced folks here would take a look and see if I made any glaringly obvious mistakes in designing the power supply portion of my circuit.

Some quick notes:

 - The "12V" input supply is actually an automotive electrical system.

 - D1 and D2 may not be totally necessary, but I feel better protecting from someone connecting the external lines backwards. 

 - D1 and D2 are rated for 9A forward current.

 - The MCU side of the circuit also has a 100nF decoupling capacitor at each power pin.

 - The two TVS diode have a breakdown voltage of 15.5V and a clamping voltage of 19V.

 - Voltage regulator input capacitor is ceramic.

 - Voltage regulator output capacitor is aluminum electrolytic with an ESR of 45mΩ.

 - The ATMega328P-AN MCU is only running three inputs (two capacitive touch keys and one reed switch) and one output (the MOSFET you see here).

 - MOSFET data sheet: http://www.mouser.com/ds/2/115/ds31738-765013.pdf

  - Regulator is ON Semiconductor part #NCP1117DT18RKG, the data sheet is here: http://www.mouser.com/ds/2/308/NCP1117-D-81326.pdf

 

Unless anyone see anything glaringly obvious that is wrong I'm going to build it and give it a shot!

Last Edited: Thu. Aug 17, 2017 - 06:43 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You forgot something?

I see links to 2 datasheets, but no schematic.

 

10nF caps seem a bit small for decoupling. 100nF is more "standard", but as far as I know anybody who is not int to much voodoo ciruits (Real HF of, processors with GHz clocks etc) just slaps on a 100nF cap on each Vcc/gnd combo with short leads and it's "sufficient".

 

Personally I always put some inductor or ferrite core in the power leads. There are too many nasty transients in the outside world to make a uC circuit run reliably without them.

 

PCB layout / wire length is important.

Have you read some application notes about power supply decoupling?

 

Eidt:

LsD wrote:
supply is actually an automotive

Automotive is a quite harsh environment.

One of the troubles is "load dump". You have TVS's, probably in the right place.

From the top of my head: Load dumps can be a temprary incrase (maybe also negative) in voltage up to 60V or so from a source which is able to deliver multiple amps (10A or more?) into your circuit during serveral hundreds of ms. Can your TVS's survive that? You might want to add something to limit the current during those events.

Paul van der Hoeven.
Bunch of old projects with AVR's:
http://www.hoevendesign.com

Last Edited: Thu. Aug 17, 2017 - 06:44 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well... bullocks... I was sure I attached the .png of the schematic.
It's on my work computer and I'm home now, I'll try to upload it tomorrow.
I studied the MCU and voltage regulator data sheets pretty closely, but this is my first rodeo.
Any tips or suggestions are welcome, but that will probably require that I insert the missing image...
I'll also correct the original post, the caps at the MCU are 100nF.

What type of inductor would you suggest and where? (A response to that will probably also come after I add the missing image.)

Last Edited: Thu. Aug 17, 2017 - 06:45 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

After "browsing" and "selecting" you also have to press "upload" (Blue botton to the right of "browse").

Paul van der Hoeven.
Bunch of old projects with AVR's:
http://www.hoevendesign.com

Last Edited: Thu. Aug 17, 2017 - 11:23 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Up-loading sequence.

 

I'm eager to see .

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

@LsD

 

I am lazy and simply click the file name, Ctrl-C and then click in the draft post and use Ctrl-V. Pasted.

 

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

What protects the TVS diodes? What protects the mosfet? If the switched output is going to the real world, you might want to use a protected mosfet and a high side one at that.

Realise that the automotive environment is rather harsh - vibration,temperature,electrical and very high ESD. Also there is high currents involved, ensure your wiring is adequate for the currents expected and fused accordingly.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Be sure to watch out for load dump--it can cause big trouble

 

There are parts (diodes) designed for this specific condition. 

 

When in the dark remember-the future looks brighter than ever.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Now I know why the attachment was missing. The security protocols at my office blocked the upload.

 

I think it should be accessible via this link:

https://www.circuitlab.com/circu...

 

I know that's rather a pain for people trying to help me, sorry about that :(

 

If anyone wants more info, here are the data sheets:

12V relay: http://www.mouser.com/ds/2/164/f... (FTR-P6GN012WA)

flyback diode: http://www.mouser.com/ds/2/427/r... (RS1A-E3/5AT)

fuse: http://www.mouser.com/ds/2/643/2... (2JQ 2-R)

D1 and D2 protective diodes: http://www.mouser.com/ds/2/196/I... (IDD09E60)

TVS diodes: http://www.mouser.com/ds/2/308/S... (SL12T1G)

voltage regulator input capacitor: http://www.mouser.com/ds/2/281/c... (GRM31CC8YA106KA12L)

voltage regulator: http://www.mouser.com/ds/2/308/N... (NCP1117DT18RKG)

voltage regulator output capacitor: http://www.mouser.com/ds/2/445/8... (875105240001)

And, just because it was mentioned, the decoupling caps that I put at the MCU: http://www.mouser.com/ds/2/445/8... (865230640001)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:

What protects the TVS diodes? What protects the mosfet? If the switched output is going to the real world, you might want to use a protected mosfet and a high side one at that.

Realise that the automotive environment is rather harsh - vibration,temperature,electrical and very high ESD. Also there is high currents involved, ensure your wiring is adequate for the currents expected and fused accordingly.

 

As far as what protects the TVS diode and MOSFET, you just lost me? Forgive my newbiness...

The protective high side MOSFET comment also went over my head.

 

I do understand the wire sizing and fusing requirements for the external circuitry. Automotive electrical is by no means new to me, but electronics very much is and I appreciate all of the input, tips, and help!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:

Be sure to watch out for load dump--it can cause big trouble

 

There are parts (diodes) designed for this specific condition. 

 

 

I guess my inexperience is glaring here, I though that was what the TVS diodes were doing?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

LsD wrote:
I was sure I attached the .png of the schematic.

Rather than attaching, it's usually better to embed the image - so that we can actually see it in the post.

 

Full instructions here: http://www.avrfreaks.net/comment...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

awneil wrote:

LsD wrote:
I was sure I attached the .png of the schematic.

Rather than attaching, it's usually better to embed the image - so that we can actually see it in the post.

 

Full instructions here: http://www.avrfreaks.net/comment...

 

I tried that as well.

Unfortunately we are connected to a government server and our security protocols are pretty tight, although seemingly random.

Some sites I can't even get to from here. Others I can get to and even post images.

I have no idea what the selection process is for what is blocked and what isn't. 

Last Edited: Fri. Aug 18, 2017 - 11:45 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Here you go:

 

 

EDIT

 

from the https://www.circuitlab.com/circuit/635cq7g72m6f/screenshot/1024x768/ link posted earlier

 

EDIT 2

 

Clickable link from the foot of the image: https://www.circuitlab.com/c635cq7g72m6f

Last Edited: Fri. Aug 18, 2017 - 12:07 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Awesome, thank you!!!

 

EDIT: on a side not - it's kind of annoying that I have to use that clunky software to do the initial drawing and then re-draw them in my PCB design software when I get home. I don't know why they don't just put the design software on my work computer or let me work on my personal laptop here. LOL (obviously kidding!) :)  

Last Edited: Fri. Aug 18, 2017 - 12:25 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Ok, you've got a fuse that hopefully will protect the transzorbs - although we have no idea of what they are. You only should need one unidirectional transzorb, as a significant reverse voltage will kill the regulator and most unidirectional tranzorbs will be a normal diode in reverse conduction and a zener in forward conduction. Besides, you have a series diode as well. In the case of a load dump, will the transzorb survive long enough until the fuse blows?

No fuse on the relay switch circuit - what protects the wire?

Why only 1V8? Most automotive electronics try to stay with 5V logic for noise immunity.

LsD wrote:
As far as what protects the TVS diode and MOSFET, you just lost me? Forgive my newbiness... The protective high side MOSFET comment also went over my head.  

If you get a load dump, then the transzorb conducts and gets hot. Eventually it will fail and melt off the board. Net result you device has failed. You want your device to tolerate a load dump without failing. Similarly with the mosfet - what happens if the load (relay in this instance) draws too much current? If it's on the circuit board, you're probably safe, but if it is off the board then the wiring may short. Net result a smoking mosfet.

Most circuits in a car are 'high side' ie: the 12V gets switched, not the 0V. Companies like infineon have protected high side mosfets that have overvoltage and current limiting. You pay more for them, but they are much more likely to survive abuse.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:

Ok, you've got a fuse that hopefully will protect the transzorbs - although we have no idea of what they are.

Sorry... no idea what what are?

 

You only should need one unidirectional transzorb, as a significant reverse voltage will kill the regulator and most unidirectional tranzorbs will be a normal diode in reverse conduction and a zener in forward conduction.

If I understand correctly, you're saying I should remove the TVS on the left in the diagram?

 

Besides, you have a series diode as well. In the case of a load dump, will the transzorb survive long enough until the fuse blows?

My ignorance is crippling me on this question. The fuse is quick blow, I'm not sure exactly what I'm looking for on the TVS data sheet.

 

No fuse on the relay switch circuit - what protects the wire?

There will be a fuse on the main power wire, at the source. All of that is external to this device.

 

Why only 1V8? Most automotive electronics try to stay with 5V logic for noise immunity.

I was trying to keep current draw as low as possible, that is why I went with 1.8V. This device will be connected a vehicle battery power source, not ignition switched. I didn't want the parasitic draw to be problematic.

 

LsD wrote:
As far as what protects the TVS diode and MOSFET, you just lost me? Forgive my newbiness... The protective high side MOSFET comment also went over my head.  

If you get a load dump, then the transzorb conducts and gets hot. Eventually it will fail and melt off the board. Net result your device has failed. You want your device to tolerate a load dump without failing. Similarly with the mosfet - what happens if the load (relay in this instance) draws too much current? If it's on the circuit board, you're probably safe, but if it is off the board then the wiring may short. Net result a smoking mosfet.

The relay is surface mount and is on the PCB. I believe this is what you mean?

 

Most circuits in a car are 'high side' ie: the 12V gets switched, not the 0V. Companies like infineon have protected high side mosfets that have overvoltage and current limiting. You pay more for them, but they are much more likely to survive abuse.

That is interesting, I will look into that. I originally had the MCU switching a solid state relay that controlled the magnetic relay, but was advised that that was unnecessary and over complicating things and switched to the MOSFET. So, the use of a MOSFET in this application is new to me.

 

Thank you for all of your input!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Using a single diode: The most basic and crude form of power supply design is the one which uses a single diode and a capacitor. Since a single diode will rectify only one-half cycle of the AC signal, this type of configuration requires a large output filter capacitor for compensating the above limitation.
A filter capacitor makes sure that after rectification, at the falling or decreasing sections of the resultant DC pattern, where the voltage tends to dip, these sections are filled and topped by the stored energy inside the capacitor.
The above compensation act done by the capacitors stored energy helps to maintain a clean and ripple free DC output which wouldn't be possible just by the diodes alone.

For a single diode power supply design, the transformer's secondary winding just needs to have a single winding with two ends.

However, the above configuration cannot be considered an efficient power supply design due to its crude half wave rectification and limited output conditioning capabilities.

 

* Careful! Moderator is watching you closely now. *

Last Edited: Fri. Aug 18, 2017 - 02:35 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

What?!

This is a DC source powering a DC device...

Maybe "power supply" was a poor choice of terms?

 

 

EDIT: I just looked at all of that user's posts and "spammer" seems like a strong possibility

Last Edited: Fri. Aug 18, 2017 - 01:57 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

A bit about load dump:

https://duckduckgo.com/html?q=au...

 

brings you pretty fast to an application note of TI with a nice example schematic:

TI Application Note snva681a

http://www.ti.com/lit/an/snva681...

 

Note that the IRF840 in the schematic is a pretty beefy fet, with a Vds of 500 Volt. (Which seems a bit exessive they must have found one in the parts bin ?)

http://www.vishay.com/docs/91070...

 

The STF1342 however only has a Vds of -60V which seems too low for this application. Weird.

http://www.onsemi.com/pub/Collat...

Paul van der Hoeven.
Bunch of old projects with AVR's:
http://www.hoevendesign.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thank you.

 

Looks like I have a LOT more research to do before I start prototyping :(

Last Edited: Fri. Aug 18, 2017 - 02:57 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm a bit confused about the 1.8V.  We saw that mention in last week's Tiny45 thread as well.  What is the motivation for running at a voltage level not suitable for the internal BOD?  And apparently you have near-infinite automotive supply source.

 

From the other thread, it sounds like a simple app.  If your response to above is "let's try to use as little power as practical", then in nearly all cases a lower total consumption is had with going "fast" periodically, then sleeping.

 

So if you want to set the BOD at the lowest level, then e.g. ~2.2V is 10% over the nominal trip level.

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm sure that I'm the one who is confused. The device needs to be operational when the vehicle is parked and will be drawing from the battery. It was my understanding that running the MCU at 1.8V would draw substantially less current than a higher voltage. I may very well be confused.

I am listening to advice I get here and taking it very seriously. That is why I ditched the idea of using a Tiny and replaced it in the design with the Mega328.

Your other comments about BOD just went completely over my head.

I'm sorry if I'm asking questions that are too low level or just don't understand what is being said. I have been doing research on my own and working on the design for months and thought I had a decent grasp, but maybe I am way off.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

When taking a step back an looking at the overal design I get the idea you are an (almost absolute) beginner with electronics trying to design the "perfect" circuit from only theoretical knowledge found in books / internet.

 

Do you realise there is a good reason it takes years of studiying to get an engineering degree (in electronics / software / ... ) ?

So we have situation here where you are trying to learn, and we are trying to help but because of the very limited capabilities of text on a forum like this it is a very inefficient way of communicating.

Because of the limits of communicating through this medium we have to make lots of assumptions.

Yes, it is true that a uC on 1.8V draws less current than on 5V, but that also easily becomes irrelevant if your 12V relay draws 100 times more current than your uC. Heck, even a single led draws several orders of magnitude more current than a uC in sleep mode.

 

When re-reading your other posts (which took me more then 10 minutes) I get the idea that you want to wire some software without having hardware, then order some components and program an uC, then put it in a circuit and expect it to work. I'll predict it's not going to happen that way.

My advise to you is to roll up your sleeves and get your hands dirty. But a m328 breakout board, usbASP (or arduino board, whatever (I don't like arduino's)) a breadboard and some small stuff (leds, resistors, capacitors wires etc) and just start building stuff. Because you've said you want to build more projects you'll need all that stuff anyway (and a DMM, LA, ...).

Start small, have some fun on the way while building things and then extend your knowledge.

It does not make sense to bang your head till it hurts on trying to understand datasheets without much prior knowledge of electronics.

 

When dealing with cars, have you read (and understood) the load dump challenge which has been mentioned several times here?

I haven't seen any feedback from you on that (pretty important subject) yet.

When something goes wrong with your circuit the whole thing can catch fire (which is not so likely because of the fuse) or you have a flat battery when you need your car to go to work on a monday morning.

Also trying to debug a circuit lying upside down under the dashboard of your car is no fun.

 

If you power your uC from 2 or 3 AA batteries you circumvent a whole lot of potential problems, you can easily take the whole thing out of your car if something has gone haywire and you will get a good idea of how much power your ciruit consumes. 3AA batteries would easily last a month, an depending on the power your led takes it can be multiple years or the same time als the shelf life of your batteries. (But that takes carefull design! ).

Paul van der Hoeven.
Bunch of old projects with AVR's:
http://www.hoevendesign.com

Last Edited: Fri. Aug 18, 2017 - 05:10 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I absolutely am not ignoring anything being said. I have not commented on the load dump situation because I am reading all of the provided links and any other material I can find and gathering knowledge on my own before I ask any annoying questions.

Due to advice in other threads I have in fact ordered an Arduino in order to play around and learn and get practice programming.

I work in mechanical engineering and it has literally been years since I have taken an electronics course. So, your evaluation of my level is appropriate.

The project actually started as a means to fill a need for myself and some friends and a way to learn new things. I am fully prepared to fail and redesign and fail and redesign and repeat until it works.

The LED in the schematic is just a representation. The actual load is much larger being an LED cargo light that could be up to 10A of draw. Draw while the load is on is of minor concern. The load while the system is idle is a concern. I will go back to the research phase before I ask more questions.

And I do realize the limitations here and very much appreciate all input and advice!

EDIT - however, powering the control unit (the MCU and MOSFET) from a battery on the device is a possibility.

Last Edited: Fri. Aug 18, 2017 - 05:45 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

LsD wrote:
The load while the system is idle is a concern.

From the description, the app is simple.  In fact, when "idle", it could well be that the system remain in sleep mode pending a trigger signal.

 

But let us consider a probably more usual case.  The app has work to do, at least periodically.  Gather input status and enter into debouncing engine.  Do some coarse or fine timekeeping,  Whatever.

 

Let's say that this work can be carried out in one millisecond every 10 milliseconds, with the AVR running at 1MHz.  The rest of the time the app just spins, with nothing to do.  Pick a Vcc level--let's say 3V.  The datasheet says that the AVR's draw is then about 500uA.

 

Now, let's say we run at 8MHz.  The 1ms of work can now be done in 1/8 the time, or 125us.  Active current draw at 8MHz is about 3mA or a little less.  Active time is 1/80 of total time.  So active draw each second is 1/80 of 3mA, or 37.5uA.

 

The rest of the time, 79/80 of the time, the system can be in a sleep mode.  The "best" if the app can tolerate the restrictions is power-down sleep, with the watchdog timer giving a periodic wakeup about every 10ms.  The datasheet says that power-down draw with watchdog enabled is 4uA.

 

Thus, with this simple example, we see that draw over a time period averages right around 40uA, about 1/10 of the draw doing the same work at 1MHz without sleeping.

 

===========

[edit] Inquiring minds will ask "yahbut, why not run at 1MHz and also sleep?"  A very good question; what do the numbers say?  Awake for 10% of the time at 500uA; 50uA there.  90% of the time sleeping; about 4uA draw.  [no need to even subtract the 10% if you don't want to; it doesn't matter much.  But if you care to, call that 3.5uA]

 

50 + 3.5 => 53.5uA average draw; 1MHz clock; 3V Vcc; 90% of time in power-down sleep with watchdog enabled.

37.5 + 4 => 41.5uA average draw; 8MHz clock; 3V Vcc; 98.7% of time in power-down sleep with watchdog enabled.

 

LsD wrote:
The load while the system is idle is a concern.

So, which is better?

 

 

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

Last Edited: Fri. Aug 18, 2017 - 06:27 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You've given me a lot of helpful information and a lot to thank about. I am going to do a lot more research and reading on my own and investigate the idea of powering the controller with an on board battery. I'm sure I'll be back with questions after a bit, but I'm going to gather as much data as I can on my own.

Thank you!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

LsD wrote:
I am going to do a lot more research and reading on my own and investigate the idea of powering the controller with an on board battery.

I don't see why you would want to do that, but suit yourself.

 

Some years ago, user Chancy99 led a series of threads about an automotive-environment datalogger.  Some of the threads thoroughly discussed many of the issues brought up in this thread.  Start here and scroll through to find them...

http://www.avrfreaks.net/search/...

...or here, "In-Vehicle Datalogger..."

http://www.avrfreaks.net/search/...

 

 

 

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

Last Edited: Fri. Aug 18, 2017 - 07:05 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thank you again, any and all references and research material are great and highly welcome!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hello
With  permission, I had a few questions.
Did you use a linear regulator?
Is this the relay on the board?
Do not you want to install a resistor before the protective diodes?  Of course, a  “PPTC”   with a resistor is better.   PPTC  “  https://en.wikipedia.org/wiki/Re...

 

Last Edited: Fri. Aug 18, 2017 - 07:50 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes, that is a linear regulator in the drawings. I was originally going to go with a switching regulator because I was concerned about heat with a linear, but was told that at the low current draw in this application that was not a concern and that switching regulators would caue much more noise and possible interference with the MCU.

 

Yes, the relay is a surface mount device. Qualified for automotive applications, that is what I didn't do much of anything to protect it from the power spikes and transients in the 12V supply.

 

I was lead to believe that a PTC fuse was a "no go" in an automotive environment due to heat sensitivity. Is that false?

 

I will still be monitoring this thread. In the mean time I am reading the provided links and researching the suggested topics. 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well, I did a few quick calculations, and a 40 Ah 12V lead battery has a self-discharge current of 2-3 mA. So the impact of the MCU is not very large, I'm sure it will not be responsible for draining the battery, unless something goes terribly wrong.

 

Edit: The quiescent current of the LDO you chose is quite high (typ. 4mA), maybe you should choose a different one.

Last Edited: Fri. Aug 18, 2017 - 08:01 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In an automotive electrical system a 25mA draw is considered about normal and anything that exceeds 100mA indicates an electrical issue that needs to be addressed because it can drain the battery if the car sits for a few days. I have always followed the rule anything over 85mA needs to be investigated as abnormal. These are just the general guidelines we followed when I worked in the automotive field. That being said, there are already multiple computers and memory devices drawing on the batteries of the vehicles in question and I would like to keep any current drain I add to that to a minimum. Am I worried about a non-issue? Maybe...

 

I am just a little gun shy about anything that will be drawing power all the time when the vehicle ignition is not on and the vehicle is parked.

 

EDIT: thank you, that is something else I can look into further.

Last Edited: Fri. Aug 18, 2017 - 08:10 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Perhaps the NCP1117LPST18T3G regulator would be a better choice?

Quiescent Current: 550 uA

http://www.mouser.com/ds/2/308/N...

I am still reading about how to deal with the surges and transients caused by load dumps...

Then again, everything may change after I am done reading if I end up changing the VCC. I won't know until I have finished all of the suggested reading.

I just didn't want to not respond.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You will be successful with the information you have.
Have a resistor before the "protective diodes". To keep them at high voltage transient time.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yeah, that LDO is a good improvement. I just wanted you to be aware that regulators use power too, don't focus just on the MCU. There is not much more I can help, my knowledge of automotive electronics is approximately zerocheeky

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

El Tangas wrote:

 

Yeah, that LDO is a good improvement. I just wanted you to be aware that regulators use power too, don't focus just on the MCU. There is not much more I can help, my knowledge of automotive electronics is approximately zerocheeky

 

It appears that I'm not far off from that. I have quite a bit of knowledge and experience in automotive electrical systems, but as everyone here is well aware, electrical and electronics are two vastly different things!

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

LsD wrote:
Yes, that is a linear regulator in the drawings.

???  You are going to buck from 12V to 1.8V, and you are concerned with the "idle" current draw?!?  How efficient >>is<< this magic regulator?

 

The reference design you posted above looked like a switcher to me...

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

Unless I'm over simplifying it, it seems like placing a Surface Mount Multilayer Varistor (MLV) specifically designed for automotive transients such as load dumps (like the Littlefuse AUML series or similar) in the circuit before the other TVS diodes would solve that part of the issue.

 

If I'm reading the data sheets right it doesn't quite look like they function exactly the same as a regular TVS and it may still be a good idea to keep one TVS there to prevent over voltage to the regulator.

Or at least, their maximum clamping voltage is 40V, so they may allow voltage spikes up to 40V to get through to the regulator.

I would assume that most normal TVS/ESD devices would be able to handle that after the MLV has done what it needs to do? 

 

I knew automotive power systems were unstable and "spikey", but I never realized the amount of power actually involved. It's rather eye opening!

 

Am I on the right track? 

Last Edited: Fri. Aug 18, 2017 - 09:22 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

theusch wrote:

 

LsD wrote:
Yes, that is a linear regulator in the drawings.

???  You are going to buck from 12V to 1.8V, and you are concerned with the "idle" current draw?!?  How efficient >>is<< this magic regulator?

 

The reference design you posted above looked like a switcher to me...

 

Like I said above:

"I was originally going to go with a switching regulator because I was concerned about heat with a linear, but was told that at the low current draw in this application that was not a concern and that switching regulators would cause much more noise and possible interference with the MCU"

 

If there is something inherently wrong and I was lead astray, please enlighten me. I'm not looking for spoon fed answers. I've been happily reading about load dump and jump start transients in automotive systems for hours now. 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Vehicle wiring for add-ons can be a real pain in the ...

 

(Photo is the front passenger's floor.)

 

JC

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

switching regulators would cause much more noise and possible interference with the MCU"

Compared with a linear regulator which is not switching, the higher noise is likely true.  Of course the switcher's absolute noise level can be quite small, & some LC filter & proper gnd layout (and layout in general)can be quite helpful.  At 100 ma current levels, it is rather unlikely a switcher would upset anyone, unless you are right next to some precision analog processing.  Then things like filtering, shielding, gnd plane/layout might help sop up some of the mess. It's entirely different with a 700 watt switching circuit located 10mm away, or even further away, but in the same gnd path.

 

 

When in the dark remember-the future looks brighter than ever.

Last Edited: Sat. Aug 19, 2017 - 04:55 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've been doing a LOT of reading and there are plenty of solutions out there for load dump and transient suppression. I just have to decide which ones I want to use and how to implement them. Once I do some more reading and get what I think is a good plan I would like to post a modified schematic up here and get some input if that's acceptable?

 

The distinct possibility of the vehicle battery voltage supply dropping as low as 2.8V for up to 15ms on a cold start are pushing me back toward staying at the 1.8V operating voltage for the MCU so it doesn't self reset during these events and to avoid having to use a combination of a boost converter and a regulator. Am I going way off the edge?

 

I'm not hoping to design a "perfect project" on paper without building and testing anything, just wanting to avoid any obvious simple beginner's mistakes on my first go-round.

 

Thank you everyone for your input!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Why can't it just reset, then restart? 

It will probably happen anyway at times, plan ahead

When in the dark remember-the future looks brighter than ever.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I suppose it wouldn't be the end of the world, but I can see instances where I would have the external load (cargo/work light) that is controlled by the MCU on and would desire that it not shut off during vehicle cranking.

 

EDIT - well that was just a stupid thought... if the voltage dips to 2.8V it's not going to stay on anyway... 

Nor would I WANT a 10A load pulling power away from my starter during a cold start anyway!

(facepalm)

Last Edited: Mon. Aug 28, 2017 - 05:02 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The relay isn't bad, but if driving only LEDS a mosfet is much more reliable (non-mechanical)--Especially a tempfet, or omnifet that is thermally & current protected.

 

For regular bulbs there is a very high inrush that makes relays a somewhat hardier solution (though some fets are made to limit inrush too)....gotta watch that bulb inrush currents don't weld any relay contacts!

A 10Amp bulb can draw 100amps at cold turn-on (super high peak rapidly declining over a handful of ms).

 

this works best with at least a 5V control signal  (perhaps even 3.3v

http://www.bdtic.com/datasheet/ST/VNP35N07.pdf

When in the dark remember-the future looks brighter than ever.

Last Edited: Mon. Aug 28, 2017 - 05:51 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thank you.

That is a very interesting possible alternative.

I will have to research more to ensure I understand all of the specifications and applications.

I went with the relay because it is something I know I understand and know it is robust enough for whatever external light source I choose, but that would certainly shrink my physical footprint quite a bit! 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Could someone tell me where I'm wrong in this analysis?

 

I see the +12V coming out of the battery then through the diode and fuse to the input of the relay's energizer coil.   Power flows through the coil to the MOSFET.   When the MOSFET is ON (a 1.8v logic signal from the CPU), then current flows through it with little or no resistance.  So there is effectively no resistance from the battery to the ground when the MOSFET is switched ON.  The designer is depending on the internal resistance of the relay to prevent large amounts of current to flow through the relay.   It seems like that the only thing that the relay does is turn on the external LED.  The CPU is always powered at 1.8V through the regulator.   Is the MOSFET supposed to act as an off/on switch for the CPU circuit?

 

Plus the negative leads on the 10uF capacitors on the 1.8v regulator should be connected to ground instead of each other and floating.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

So there is effectively no resistance from the battery to the ground when the MOSFET is switched ON. 

it goes through the relay coil    batt-->diode-->fuse-->COIL-->mosfet-->gnd

 

good call on the caps.

 

You could scrap the relay & just put the LEDs in its place (using a high power fet)...also better if you want to "strobe"  the leds or dance to the music.

 

The real question is does current flow from + to - or - to +    !!!!!!!!

 

 

 

 

 

When in the dark remember-the future looks brighter than ever.

Last Edited: Mon. Aug 28, 2017 - 08:08 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:
The real question is does current flow from + to - or - to + !!!!!!!!

 

Does current even "flow"? We are talking about charge carriers that are quantum entities. I prefer not to even think about it, makes my life easier.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Current (Amps) 'flows' positive to negative but is usually electrons (negative charge) moving from negative to positive.  Holes move the same way as Amps but are created by electrons moving the other way.  I am sure there will be other mechanisms for current flow at the quantum level.

 

David

Pages