Author Topic: Power line filtering in a vintage car  (Read 4630 times)

0 Members and 1 Guest are viewing this topic.

Offline robzyTopic starter

  • Regular Contributor
  • *
  • Posts: 155
  • Country: au
Power line filtering in a vintage car
« on: January 05, 2020, 10:00:09 am »
So I have a problem: I connected an ESP32 to the 12v voltage rail of a vintage car (~70 years old) and let the magic smoke out. The module's AMS1117 voltage regular is rated up to an input voltage of 20v, but the module stopped working when the car was running, and the AMS1117 literally snap crackled and pop.

I want to solve the problem.

I think that the problem is the huge transient voltages caused by oldschool ignition systems. Even my multimeter can't handle voltage measurement of the car's power line, it shows around 12V, but then goes all glitchy.

What is the best way to create a cleaner 12V rail from an old car for an ESP32?

I have done some searching and I haven't found a lot of information. These are the thoughts that have occurred to me so far:
  • Coils and capacitors: Currently my preferred option, but I can't find any information on the design of such filtering circuits
  • MOVs: They don't seem ideal, because it looks like they are generally used for irregular voltage spikes and the car would just hammer them
  • Industrialised voltage regulator ICs: I haven't found anything that seems rugged enough for this use case, I've found some IC's that claim to be able to protect themselves against higher voltage input levels, but nothing that seems ideal for this particular case
 

Offline Psi

  • Super Contributor
  • ***
  • Posts: 10385
  • Country: nz
Re: Power line filtering in a vintage car
« Reply #1 on: January 05, 2020, 10:13:10 am »
LM2937IMP regs are automotive rated and have built in 60-V transient protection. (buy them from legit source, there are fakes out there which are just normal regs)


If you want a bit more protection, add a 0.1 ohm 5W series resistor and chunky ~3000W  ~40V TVS diode to GND on the input side of that reg.
« Last Edit: January 05, 2020, 10:17:25 am by Psi »
Greek letter 'Psi' (not Pounds per Square Inch)
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 29810
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: Power line filtering in a vintage car
« Reply #2 on: January 05, 2020, 10:15:40 am »
Not only are ignitions the noise source, generator brushes and electromechanical regulators are noisy too.
Generally the solution was to suppress at the source and just the addition of caps was sufficient.
IIRC 0.22 - 0.47uF 200V caps were used.
Avid Rabid Hobbyist.
Some stuff seen @ Siglent HQ cannot be shared.
 

Offline DBecker

  • Frequent Contributor
  • **
  • Posts: 326
  • Country: us
Re: Power line filtering in a vintage car
« Reply #3 on: January 06, 2020, 12:19:11 am »
Many '1117' regulators are rated for *far* less than 20V.  Some are rated at 6V or 8V -- fine for dropping a regulated 5V supply down to 3.3V, but few other use cases.  The real AMS1117 is rated for a maximum of 15V, but I've blown them out (more likely, a counterfeit) with a 12V wall wart supply.

Some are also thermally vulnerable, and dropping 14V down to 3.3V might result in the thermal limit circuit not responding in time (assuming that a generic regulator even has a thermal shut-off).

Automotive power supplies should be rated to survive at 28V continuous, with a brief load dump to 35V or more.  There are plenty of medium-priced automotive chips that can reliably do this.  What are your voltage and current requirements?  Do you need the module to work during cranking transients?


 

Offline duak

  • Super Contributor
  • ***
  • Posts: 1048
  • Country: ca
Re: Power line filtering in a vintage car
« Reply #4 on: January 06, 2020, 12:44:53 am »
There are usually two capacitors (AKA condensors) on a Kettering or points style ignition system.  One is across the points to reduce arcing.  The second is connected to the other low voltage terminal on the coil to reduce RFI.  This vehicle probably has a generator with an electromechanical regulator.  These systems usually have caps on the generator and/or the regulator to reduce RFI.  These caps usually have a value of 0.18 to 0.22 uF.  I think I've seen these caps on horns and electrical fuel pumps too.

For a filter into the electronics I would be looking for something like a 1 to 5 mH inductor in series and something like 10 to 22 uF in parallel with 1 uF to circuit common.  These were available at car stereo shops at one time.

 

Offline robzyTopic starter

  • Regular Contributor
  • *
  • Posts: 155
  • Country: au
Re: Power line filtering in a vintage car
« Reply #5 on: January 06, 2020, 02:12:54 am »
Thank you everyone! Each of those posts was super helpful.

I need 50mA at 3.3V, so my plan at the moment is to pick up an 'industrialised' voltage regulator, having its input protected by a resistor&TVS, and all this being fed an LC low-pass filter. I'll need to do some efficiency calcs to work out if I should do a 3.3V regulator or a 12V regulator (followed by a cheap 3.3V switcher).

duak, when you say a 1uF to circuit common what do you mean? I.e. if I put a 15 uF from 14V (unregulated) to ground, where exactly would I put this 1uF?

DBecker, I do not need the module to work during cranking transients, but I do need the module to not blow-up :P I assume that the above-mentioned plans should be sufficient for that, but do tell me if I'm missing something. Incidentally, thanks for sharing the guidelines re 28V/35V. Super helpful.


A few other FYIs:
- The 1117 regulator came with the cheap-as-chips aliexpress ESP32, so it is likely very suspect here. Lesson learned.
- I want to avoid messing with the car's original electronics (e.g. by adding capacitors). The distributor contains a condenser between the coil and ground, but afaik there's nothing on the other side of the coil, and I'm not sure that the Lucas RF95 regulator has any capacitors in it.
- You're correct in noting that it isn't just the ignition system, but also generator, regulators, and fuel pump. I confirmed that my multimeter still glitches when the generator is disconnected, but that doesn't mean it wasn't a problem in its own right.
 

Offline basinstreetdesign

  • Frequent Contributor
  • **
  • Posts: 466
  • Country: ca
Re: Power line filtering in a vintage car
« Reply #6 on: January 06, 2020, 05:00:35 am »
OK, about designing electronics to work within a vehicles electrical environment. It's not pretty. In fact automotive electrics are acknowledged by the experts to be one of the most vicious and nasty electrical environments going. You may hear everybody talk about a "12V system" but that value is a nominal value at best. Because the system also has to contend with engine cranking, accessory noise, inductive kickbacks from things like electric window motors when they stop and something called "load dumps", anything plugged into the system may be subjected to negative voltage spikes of over -100V, polarity reversals, positive voltage spike of several 100 V in amplitude as well as 24V jump starts, and accessory noise and RF from several Hz to the MHz range.

It's kind of like Alice walking into Wonderland if you are not prepared for it. It's not a nice well-behaved place. You may find that your gizmo dies the first time you open the window. Here is an app note from Littlefuse that talks about it more.

https://m.littelfuse.com/~/media/electronics_technical/application_notes/varistors/littelfuse_suppression_of_transients_in_an_automotive_environment_application_note.pdf

You should try to research what protection and filtering circuitry is used between the 12 source and the first component of your circuit. It usually amounts to an inductor/capacitor pair on input followed by a zener diode, silicon diodes to limit input to -1V to +15V range and more decoupling capacitors to filter out more noise.
Here is a circuit that I used with a gizmo (car sign) that ran from cigarette outlet power on a Ford truck.  It worked without any problem at all.  I admit that the input inductor could have been somewhat bigger.  Ignore that buck regulator - it is irrelevant but worked fine, too

Good luck.
« Last Edit: January 06, 2020, 05:03:12 am by basinstreetdesign »
STAND BACK!  I'm going to try SCIENCE!
 
The following users thanked this post: I wanted a rude username

Online langwadt

  • Super Contributor
  • ***
  • Posts: 4857
  • Country: dk
Re: Power line filtering in a vintage car
« Reply #7 on: January 06, 2020, 05:38:58 am »
LM2937IMP regs are automotive rated and have built in 60-V transient protection. (buy them from legit source, there are fakes out there which are just normal regs)


If you want a bit more protection, add a 0.1 ohm 5W series resistor and chunky ~3000W  ~40V TVS diode to GND on the input side of that reg.

not need to use as little as 0.1 Ohm when you only need maybe 100mA at 3V
 

Offline Psi

  • Super Contributor
  • ***
  • Posts: 10385
  • Country: nz
Re: Power line filtering in a vintage car
« Reply #8 on: January 06, 2020, 06:00:05 am »
LM2937IMP regs are automotive rated and have built in 60-V transient protection. (buy them from legit source, there are fakes out there which are just normal regs)


If you want a bit more protection, add a 0.1 ohm 5W series resistor and chunky ~3000W  ~40V TVS diode to GND on the input side of that reg.

not need to use as little as 0.1 Ohm when you only need maybe 100mA at 3V

That was to protect the TVS, but if his overall power needs are that low then the thin power wire to Vbat is enough to add some resistance.

We have 36V 4000W TVS diodes on our automotive gear and, even then, they occasionally do pop.
However we need to drive 10A loads, so our cables to Vbat are quite thick.
« Last Edit: January 06, 2020, 06:04:56 am by Psi »
Greek letter 'Psi' (not Pounds per Square Inch)
 

Offline duak

  • Super Contributor
  • ***
  • Posts: 1048
  • Country: ca
Re: Power line filtering in a vintage car
« Reply #9 on: January 06, 2020, 06:14:57 am »
robzy, the two capacitors I mentioned above should be connected in parallel.  I used the term common because I don't know how your circuit common is connected to the vehicle as It could be a wire or a mounting screw to some sheet metal.  The filter's effectiveness is the greatest with the caps' connection to circuit common as short as possible.

Basin's design looks good too.  I'd increase the value of C31 to something like 100n to 1u and L3//L4 to what I mentioned above.  It sounds like the electrical system in this vehicle has some industrial strength transients.  It's also about 70 years old - there's no guarantee that there's one solid, low impedance ground between the battery, charging system, ignition system, all the other bits n' pieces.

That being said, if you have a 'scope, it'd be a good time to use it.  We've all proposed solutions to what we think are the problems.  It may be that there's something else that needs special attention.

 

Offline robzyTopic starter

  • Regular Contributor
  • *
  • Posts: 155
  • Country: au
Re: Power line filtering in a vintage car
« Reply #10 on: January 06, 2020, 06:57:09 am »
Good luck.
Thank you very much for your post! That is very helpful. The addition of a diode in series with the circuit makes a lot of sense when it comes to protection from negative transients.

robzy, the two capacitors I mentioned above should be connected in parallel.
Ah, that makes sense. I was totally misreading your post.

That being said, if you have a 'scope, it'd be a good time to use it.  We've all proposed solutions to what we think are the problems.  It may be that there's something else that needs special attention.
If I had a scope I wouldn't dare to put it anywhere near this car ;) The proposed solutions have been very helpful, I've just ordered some parts, and I'll put together a new circuit on the weekend.
« Last Edit: January 06, 2020, 06:59:54 am by robzy »
 

Online jfiresto

  • Frequent Contributor
  • **
  • Posts: 896
  • Country: de
Re: Power line filtering in a vintage car
« Reply #11 on: January 06, 2020, 10:09:19 am »
... If you want a bit more protection, add a 0.1 ohm 5W series resistor and chunky ~3000W  ~40V TVS diode to GND on the input side of that reg.

Is the tenth ohm resistor necessary? I ask because an ISO-16750-2 load dump pulse generator for verifying such things has an internal resistance of 0.5–4 ohms. The last unsuppressed generator I designed for was 80V @ 2 ohms.
-John
 

Online langwadt

  • Super Contributor
  • ***
  • Posts: 4857
  • Country: dk
Re: Power line filtering in a vintage car
« Reply #12 on: January 06, 2020, 05:42:21 pm »
LM2937IMP regs are automotive rated and have built in 60-V transient protection. (buy them from legit source, there are fakes out there which are just normal regs)


If you want a bit more protection, add a 0.1 ohm 5W series resistor and chunky ~3000W  ~40V TVS diode to GND on the input side of that reg.

not need to use as little as 0.1 Ohm when you only need maybe 100mA at 3V

That was to protect the TVS, but if his overall power needs are that low then the thin power wire to Vbat is enough to add some resistance.

We have 36V 4000W TVS diodes on our automotive gear and, even then, they occasionally do pop.
However we need to drive 10A loads, so our cables to Vbat are quite thick.

sure for high current, but if the load is only 100mA@3V you might as well  use something like 10-20 Ohm

 

Offline amyk

  • Super Contributor
  • ***
  • Posts: 8526
Re: Power line filtering in a vintage car
« Reply #13 on: January 07, 2020, 02:51:37 am »
- You're correct in noting that it isn't just the ignition system, but also generator, regulators, and fuel pump. I confirmed that my multimeter still glitches when the generator is disconnected, but that doesn't mean it wasn't a problem in its own right.
70-year-old car with an electric fuel pump? :o
 

Offline Gregg

  • Super Contributor
  • ***
  • Posts: 1156
  • Country: us
Re: Power line filtering in a vintage car
« Reply #14 on: January 07, 2020, 03:15:47 am »
- You're correct in noting that it isn't just the ignition system, but also generator, regulators, and fuel pump. I confirmed that my multimeter still glitches when the generator is disconnected, but that doesn't mean it wasn't a problem in its own right.
70-year-old car with an electric fuel pump? :o
In post #5 OP stated the vehicle has a Lucas voltage regulator.  That should be a clue as to why it has an electric fuel pump.  Lucas electrics in vehicles are legendary (not in a good way).  I know from experience, I have a 1957 MGA.
Maybe Lucas and Murphy were drinking buddies.
Edit: Additional information.  Perlux ignitors (may be listed as Pertronix  Ignitor) are available for most older vehicles and use hall effect sensors to replace the points.  Resistor ignition wires may be of some help.
« Last Edit: January 07, 2020, 03:24:53 am by Gregg »
 

Offline John B

  • Frequent Contributor
  • **
  • Posts: 865
  • Country: au
Re: Power line filtering in a vintage car
« Reply #15 on: January 07, 2020, 08:25:59 am »
I run a few microcontrollers in my car, and on each design I have put an RC filter on the input. R is generally ~100R. That makes it easier to implement basic clamping with zeners and caps, without having to worry about high instant power dissipation.
 

Offline robzyTopic starter

  • Regular Contributor
  • *
  • Posts: 155
  • Country: au
Re: Power line filtering in a vintage car
« Reply #16 on: January 07, 2020, 10:41:41 am »
I know from experience, I have a 1957 MGA.
Very nice! I've got a 1949 Morris Minor MM.

Edit: Additional information.  Perlux ignitors (may be listed as Pertronix  Ignitor) are available for most older vehicles and use hall effect sensors to replace the points.  Resistor ignition wires may be of some help.
I'm planning on fitting an Accuspark to replace the points.
 

Offline DBecker

  • Frequent Contributor
  • **
  • Posts: 326
  • Country: us
Re: Power line filtering in a vintage car
« Reply #17 on: January 07, 2020, 08:19:54 pm »
Zero surprise that the generic 1117 blew up.  It had no chance of surviving.  Did it fail shorted so that everything attached died as well?  Because that is their superpower.

For an antique vehicle with DC dynamo you'll need to design for worse conditions than modern guidelines expect.  If the battery cable is making intermittent contact at high RPM with no other electrical load, the voltage transients will be far above what happens with an alternator.

Luckily you only have to supply a modest amount of power, and it's probably a carefully operated antique, so protection is far easier than the general case.

The simplest solution is a fast-blow fuse, TVS and the circuit board out of a switching USB lighter adapter, feeding into the 1117.  The next step up is a robust automotive linear voltage regulator on a heatsink.  A high end design would be using a SM15T33C TVS and the reference design of the TLE9471V33.

You'll have an excellent chance of surviving typical mechanical regulator transients. 
 
The following users thanked this post: SeanB

Online floobydust

  • Super Contributor
  • ***
  • Posts: 7679
  • Country: ca
Re: Power line filtering in a vintage car
« Reply #18 on: January 07, 2020, 08:40:19 pm »
OP there is also a great danger if you use a car cigarette lighter phone charger, these are basic buck-converters that can't take much for automotive transients. When they fail short, your phone will be destroyed. So maybe put more in than just for the ESP8266?

In my own car I have a transient-protected module using series diode, industry standard 6,600W 27V TVS SM8A27, polyfuse etc. It clamps to 40V worst case so any Vreg IC should be good to that or a bit less.

Lucas RF95 is harsh on electronics, it could use capacitors across the contacts for longer life.
 

Offline DBecker

  • Frequent Contributor
  • **
  • Posts: 326
  • Country: us
Re: Power line filtering in a vintage car
« Reply #19 on: January 08, 2020, 06:31:58 am »
OP there is also a great danger if you use a car cigarette lighter phone charger, these are basic buck-converters that can't take much for automotive transients. When they fail short, your phone will be destroyed. So maybe put more in than just for the ESP8266?


Many of the cheap lighter-to-USB buck converters are quite robust.  They are designed to work with either 24V or 12V systems, with margin.  They are certain *way* better than the sketchy low end linear regulators.

While the 27V TVS you suggest is OK in real life, it won't pass the tests that target a 28V jump start.  Those tests are why a 33V TVS is the usual recommendation.

A 28V jump start used to be a common tow truck trick to allow a quick start when long jumper cables were used with a good-but-discharged battery.  It became far less popular once cars started using ECUs that could be damaged by high voltage. Today it would be quite uncommon to see a mobile service truck use anything but a high end jump start box.
 

Online floobydust

  • Super Contributor
  • ***
  • Posts: 7679
  • Country: ca
Re: Power line filtering in a vintage car
« Reply #20 on: January 08, 2020, 08:06:23 am »
Would you trust your $500 cellphone/tablet to a generic car USB charger?
For a few bucks, I'll add a TVS. I find most chargers are very cheap and using old IC's with little headroom on Vin.

The TVS I mention is prolific in the auto industry, it's in many if not all ECU's and BCM's, ABS and in my car. I will confirm the 27V part but it's odd SM8A27 specs compared to a 26V SM8S26A. The boost is a 24V spec and I think in the SAE standards, not ISO. Where do you see it? In reality many automakers have their own vehicle transient standards, not all following ISO or SAE. It's perfectly fine to pop the fuse here with load dump or 24V boost, which are extremely rare occurrences in the life of a car.

A problem with going to a 33V TVS is most voltage regulators (LDO or buck) can't take the 53V clamp (vs 40V), and the TVS dissipates more energy. It's a tradeoff. You want the lowest possible clamping voltage, including over automotive temperatures (a TVS tempco is large).

Many people want to run an Arduino, Raspberry Pi, ESP32/8266 in their car.  There is a need for a basic recipe.
But any TVS is going to clamp fairly high and that eliminates all of the usual jellybean regulators out there.

https://www.emtest.com/what_is/glossary/standards/ISO-7637.php
https://www.teseq.com/product-categories/automotive-electrical-disturbances.php
 
The following users thanked this post: I wanted a rude username

Online jfiresto

  • Frequent Contributor
  • **
  • Posts: 896
  • Country: de
Re: Power line filtering in a vintage car
« Reply #21 on: January 08, 2020, 09:05:45 am »
... Many of the cheap lighter-to-USB buck converters are quite robust.  They are designed to work with either 24V or 12V systems, with margin....

It may not be as much as one might hope. I asked a manufacturer about their 12/24V vehicle USB charger before I added it to a car that has no central overvoltage (load dump) suppressor like the OP's. Up  to 36V was the number they gave me.

The best I could do was to add a fused MOSFET switch, upstream, that cuts out at 22–24V, followed by a 5kW 5KP24A TVS diode if the switch doesn't open or is a little slow. I can dig up and post the circuit if someone needs an example of how (or how not?) to do it.
-John
 

Offline DBecker

  • Frequent Contributor
  • **
  • Posts: 326
  • Country: us
Re: Power line filtering in a vintage car
« Reply #22 on: January 08, 2020, 05:46:25 pm »
Would you trust your $500 cellphone/tablet to a generic car USB charger?
For a few bucks, I'll add a TVS. I find most chargers are very cheap and using old IC's with little headroom on Vin.

Your lead-off argument is an appeal to emotion, rather than a technical point.

Billions of people trust their phones to car USB chargers.  The production volume of those chargers is so high that most are using recent purpose-designed chips.  They are optimized to work in harsh conditions with the cheapest external components, and they do an excellent job of it.

The specific TVS I listed, the SM15T33/33C, is an appropriate one to use with auxiliary equipment.  ST numbers their TVS as the nominal breakdown voltage, not the operating voltage.  The Vishay chart listed above shows that their "28" part actually has a 33V typical breakdown voltage, with a minimum breakdown of 31V, so it's quite close.

ISO 16750-2 is a short, easily read standard.  It's worth a quick read-through.
« Last Edit: January 08, 2020, 05:50:06 pm by DBecker »
 

Online floobydust

  • Super Contributor
  • ***
  • Posts: 7679
  • Country: ca
Re: Power line filtering in a vintage car
« Reply #23 on: January 08, 2020, 08:18:00 pm »
Your technical point relies on faith unless you disassemble the car USB charger and see what chip is used and if there is a TVS. I've seen enough of them to know they aren't always good products. My philosophy is not to trust it unless I've taken it apart.

Billions of people are not driving cars from 1949 with mechanical regulators.  Modern alternators incorporate zener diodes for rectifiers, the system transient voltages are lower.

One employer has product using a 78M05 on 24V truck electrical systems, with a SMBJ33A TVS. How do you think it would do? Vin max. is 35V and the TVS clamps well above that. About 20% of the power supplies failed in the field. Many with the Vreg IC dead or 22uF 35V tantalum input cap shorted.

If noobs can understand to make sure the TVS highest clamping voltage is below their voltage reg's maximum input voltage, then they are mostly covered. Another series diode to cover reverse-polarity spikes and best to have a some line resistance in the form of a polyfuse or resistor, to limit TVS current.
 

Offline m3vuv

  • Super Contributor
  • ***
  • !
  • Posts: 1738
  • Country: gb
Re: Power line filtering in a vintage car
« Reply #24 on: January 09, 2020, 12:36:06 am »
my friend had a 1935 derby bentley,the fuel pump was electric on that made by su.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf