EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: GerGa on August 12, 2016, 02:15:33 am
-
Hey so I had a thread in the newbie section on this, had a bit of valuable discussion. If you are interested in the whole story, go here: https://www.eevblog.com/forum/beginners/vehicle-transient-protections/ (https://www.eevblog.com/forum/beginners/vehicle-transient-protections/)
The issue was a Suzuki DR650 motorcycle that blew up two garmin gps' in a row. They are powered by a Garmin 12 to 5v power supply. Maybe due to load dump from the starter motor, maybe due to an old battery that previously died and had been charger, or other causes.
One step I looked into was a protection circuit that I read many discussions about on this forum using TVS diodes. The problem is I just do not have the circuit design skills
After posting here in the newbies section, I went to another message board and someone generously offered to create the circuit for me.
The scheme is here: http://imgur.com/rIQSZPg (http://imgur.com/rIQSZPg)
Digikey cart here: http://www.digikey.com/classic/Ordering/AddPart.aspx?WT.z_cid=Shared_Cart (http://www.digikey.com/classic/Ordering/AddPart.aspx?WT.z_cid=Shared_Cart)
I am hoping some of you can take a second look at the scheme and let me know what you think.
I unfortunately do not have an oscilloscope, so can not really know exactly what kind of condition is happening.
Thanks in advance
Bonus relay problem: As a secondary protection, I also plan to incorporate a relay to drop the gps supply, triggered by the starter button, so that the gps is out of circuit while on startup(+a time delay relay).
So relay R1 drops gps supply via momentary NO start button. Button released, R1 supplies R2 time delay relay coil(using cap & transister delay+relay circuit). R2 closes in gps supply after xx time.
I was advise that if relay R1 is 12 or 14V, I may get relay chatter because of the significant voltage due to the starter motor. Possibly a relay with a lower coil voltage rating would suffice, as it is only energized during startup. Any advise on that?
-
What is really important to you is what are you trying to protect and control (types/kinds
of transient voltage spikes and occurances).Since you do not have the testgear or knowhow
even then you may fry your own multimeter (current wise) so it's better if you understand the theory and properties/parameters of the devices you want to use.You are essentially trusting the device datasheet to make your decision unless someone can simulate the circuit for you. Good luck even then it can only go so far without actual testing.
Page 10-50:
http://www.littelfuse.com/~/media/electronics_technical/application_notes/varistors/littelfuse_suppression_of_transients_in_an_automotive_environment_application_note.pdf (http://www.littelfuse.com/~/media/electronics_technical/application_notes/varistors/littelfuse_suppression_of_transients_in_an_automotive_environment_application_note.pdf)
------------------------------------------------------------
Important parameters: http://www.vishay.com/docs/88490/tvs.pdf (http://www.vishay.com/docs/88490/tvs.pdf)
surge and transients: http://www.vishay.com/docs/49749/49749.pdf (http://www.vishay.com/docs/49749/49749.pdf)
http://www.littelfuse.com/~/media/electronics/application_notes/tvs_diodes/littelfuse_tvs_diode_automotive_circuit_protection_using_automotive_tvs_diodes_application_note.pdf.pdf (http://www.littelfuse.com/~/media/electronics/application_notes/tvs_diodes/littelfuse_tvs_diode_automotive_circuit_protection_using_automotive_tvs_diodes_application_note.pdf.pdf)
An example protect circuit by TI: http://www.ti.com/lit/an/snva717/snva717.pdf (http://www.ti.com/lit/an/snva717/snva717.pdf)
How to select device:You mentioned 1500W see pg.2-
http://www.microsemi.com/document-portal/doc_view/14650-how-to-select-a-transient-voltage-suppressor (http://www.microsemi.com/document-portal/doc_view/14650-how-to-select-a-transient-voltage-suppressor)
Another how to select guide:
http://www.completepowerelectronics.com/tvs-diode-selection-tutorial/ (http://www.completepowerelectronics.com/tvs-diode-selection-tutorial/)
Introduction to transient devices by Onsemi: http://www.onsemi.com/pub_link/Collateral/AND8229-D.PDF (http://www.onsemi.com/pub_link/Collateral/AND8229-D.PDF)
-
L1 and C1 form a resonant circuit that actually amplifies certain frequency components of your transient. Better use a ferrite instead of an inductor. And make sure the capacitor is low ESR and high bandwidth, like ceramic caps. Use as physically small as possible capacitor, maybe parallel some of same type and value.
-
I would like to add what the person initially said about their design. No time at the moment to look at your posts though.
Okay I threw some stuff together. Diode D2 is only needed if the relay is used (it prevents the inductor from trying to develop a high voltage across the relay contacts if the relay disconnects it while current is flowing). While there is a fuse, it's main purpose is to protect against short circuit type events, including the TVS absorbing too much power and shorting out (the fuse would maybe protect the TVS diode, but would definitely protect the wiring from a blown TVS diode).
The fuse won't protect the inductor in the 1.5-2.5A current range; in this amperage range it is possible for the inductor to overheat. Don't use a splitter and connect 2 power adapters at the same time - this could be too much current for it.
It was kinda a delicate balancing act to have the inductor have enough resistance to limit current effectively but not overheat during normal use, but I think this would work well for your application.
-
I'd replace D2 with D1 (the TVS).
-
L1 and C1 form a resonant circuit that actually amplifies certain frequency components of your transient. Better use a ferrite instead of an inductor. And make sure the capacitor is low ESR and high bandwidth, like ceramic caps. Use as physically small as possible capacitor, maybe parallel some of same type and value.
Can you elaborate? What value ferrite and cap would you put in their place? Why a physically small capacitor?
-
GerGa; I would like to suggest a different approach for you to consider. I can't draw a schematic here so will try and explain this.
Consider using a P type Power MOSFET, such as the On-Semi = MTB50P03HDL. You will need a total of 4 surface mount components.
1) Power MOSFET - Such as - MTB50P03HDL
2) 2x 4.7K 0805 resistor (or you could use a 0603, or even 0402 if you want)
3) 1x 0.01uF Capacitor 0805 (or you could use a 0603, or even 0402 if you want).
You could solder the other 3 components onto the MOSFET (connected as explained below), if you choose the 0402 size components for the 2 resistors and the 1 cap (if you are good at soldering). You will, however, have to make some kind of covering to keep water and such off of ALL these components. I don't know what the current is of your load, so I just choose a MOSFET that would certainly power the load of most any GPS. However, you are powering the load of the brick, or the 12V to 5V device which connects to the Garmin GPS. The current this draws will be less than that of the Garmin.
And speaking of the brick (12V to 5V device connecting to the Garmin), I would suggest if you still have the 2 assumed bad GPS devices to try them with a known good brick as most likely (guessing here) the spikes and low voltage on the brick when the bike is started may have taken out the brick and the GPS may be fine, or it may have taken out both or either. These cheap bricks are poor DC-DC regulators, so a better one might help some as well.
Now, to the connections. First, you will need to run a wire from the red wire on the starter solenoid (this should be a smaller wire not the big one on the starter), and the one that gets 12V when starting and then GND when started or not running, to the location you will be putting your circuit. The MOSFET has 3 pins, Gate (pin 1), Source (pin 3) and the Drain (pin 4, and 2 but you won't need to do anything with pin 2). You want to connect one resistor on the MOSFET between pin 1 (Gate) and pin 3 (Source). Next, connect the cap between the MOSFET's pin 1 (Gate) and pin 4 (Drain). Next connect the MOSFET pin 3 (Source) to your bikes battery (12V). Connect the MOSFET's pin 4 (Drain) to your Garmin 12V to 5V brick. Finally, connect one end of the final resistor to the MOSFET pin 1 (Gate) and the other end to this wire you ran from the starter solenoid.
This circuit is a Load Switch and will disconnect the 12V to your brick (12V to 5V to Garmin GPS) when the starter is being used, and reconnect it when the bike has started. Also, the GPS should work with the bike not running. This circuit, if connected properly, should work fine IF the cause of your problem is related to spikes or dips in your voltage when starting the bike. Please note that you are putting the MOSFET in SERIES with your battery positive terminal (12V) and the Garmin 12V input on the brick. You would NOT be connecting your bikes battery directly to the Garmin brick in this circuit. Try drawing this out on paper (find the data sheet for the MOSFET and use the symbol used in it), and this might become clearer if I have confused you. Hope I haven't.
Good luck.
Omegaman
-
L1 and C1 form a resonant circuit that actually amplifies certain frequency components of your transient. Better use a ferrite instead of an inductor. And make sure the capacitor is low ESR and high bandwidth, like ceramic caps. Use as physically small as possible capacitor, maybe parallel some of same type and value.
Van you elaborate? What value ferrite and cap would you put in their place? Why a physically small capacitor?
In genera, you have to think about a transient as a wideband signal, with lots of high frequency content. You want to filter out as much of that as possible, so you can consider a transient protection as a low pass filter. DC should pass, everything else not. Ive put your circuit into LTSpice and here's what comes out:
(https://www.eevblog.com/forum/projects/critique-this-transient-protection-schem-and-relay-question/?action=dlattach;attach=247979)
It is working nicer than I expected, and that is because the inductor has a high ohmic part, 844 Milliohms. This is sufficiently damping the LC resonant circuit that I mentioned. If I take that out, then you see what would be happening:
(https://www.eevblog.com/forum/projects/critique-this-transient-protection-schem-and-relay-question/?action=dlattach;attach=247981)
You can see, that the circuit would amplify frequencies around 100 Hz.
This is because an inductor does not dissipate the energy of your transient, but it bounces it back. A ferrite is a dissipative device, and 'eats up' this energy, converting it into heat. The choice should be: as much impedance as possible over the interesting frequency range, and enough current handling capability.
Chosing a physically small capacitor is important for higher frequencies, because every inch of conductor is also an inductor. If you have too much parasitic inductance in a capacitor, then it loses its function at high frequencies.
-
GerGa; I would like to suggest a different approach for you to consider. I can't draw a schematic here so will try and explain this.
Consider using a P type Power MOSFET, such as the On-Semi = MTB50P03HDL. You will need a total of 4 surface mount components.
1) Power MOSFET - Such as - MTB50P03HDL
2) 2x 4.7K 0805 resistor (or you could use a 0603, or even 0402 if you want)
3) 1x 0.01uF Capacitor 0805 (or you could use a 0603, or even 0402 if you want).
You could solder the other 3 components onto the MOSFET (connected as explained below), if you choose the 0402 size components for the 2 resistors and the 1 cap (if you are good at soldering). You will, however, have to make some kind of covering to keep water and such off of ALL these components. I don't know what the current is of your load, so I just choose a MOSFET that would certainly power the load of most any GPS. However, you are powering the load of the brick, or the 12V to 5V device which connects to the Garmin GPS. The current this draws will be less than that of the Garmin.
And speaking of the brick (12V to 5V device connecting to the Garmin), I would suggest if you still have the 2 assumed bad GPS devices to try them with a known good brick as most likely (guessing here) the spikes and low voltage on the brick when the bike is started may have taken out the brick and the GPS may be fine, or it may have taken out both or either. These cheap bricks are poor DC-DC regulators, so a better one might help some as well.
Now, to the connections. First, you will need to run a wire from the red wire on the starter solenoid (this should be a smaller wire not the big one on the starter), and the one that gets 12V when starting and then GND when started or not running, to the location you will be putting your circuit. The MOSFET has 3 pins, Gate (pin 1), Source (pin 3) and the Drain (pin 4, and 2 but you won't need to do anything with pin 2). You want to connect one resistor on the MOSFET between pin 1 (Gate) and pin 3 (Source). Next, connect the cap between the MOSFET's pin 1 (Gate) and pin 4 (Drain). Next connect the MOSFET pin 3 (Source) to your bikes battery (12V). Connect the MOSFET's pin 4 (Drain) to your Garmin 12V to 5V brick. Finally, connect one end of the final resistor to the MOSFET pin 1 (Gate) and the other end to this wire you ran from the starter solenoid.
This circuit is a Load Switch and will disconnect the 12V to your brick (12V to 5V to Garmin GPS) when the starter is being used, and reconnect it when the bike has started. Also, the GPS should work with the bike not running. This circuit, if connected properly, should work fine IF the cause of your problem is related to spikes or dips in your voltage when starting the bike. Please note that you are putting the MOSFET in SERIES with your battery positive terminal (12V) and the Garmin 12V input on the brick. You would NOT be connecting your bikes battery directly to the Garmin brick in this circuit. Try drawing this out on paper (find the data sheet for the MOSFET and use the symbol used in it), and this might become clearer if I have confused you. Hope I haven't.
Good luck.
Omegaman
If you are interested, here are some views of the power supply. I had to scrape off the waterproofing silicone off to see it. http://imgur.com/a/B60qi (http://imgur.com/a/B60qi)
What would you consider a better power supply? I have been using this power supply for charging my phone in my car(better than plugging into the cigarette lighter all the time) with no issues(although it is chinese): https://www.amazon.com/DROK-Converter-Step-down-Transformer-Waterproof/dp/B00CE75K0W/ref=sr_1_18?ie=UTF8&qid=1471311553&sr=8-18&keywords=5v+converter (https://www.amazon.com/DROK-Converter-Step-down-Transformer-Waterproof/dp/B00CE75K0W/ref=sr_1_18?ie=UTF8&qid=1471311553&sr=8-18&keywords=5v+converter)
However I don't really want to use that one because it is a bit heavy to add to a motorcycle considering that I also want to add a second supply for charging my phone on the bike. Reason for putting a second supply is because the TVS circuit is only rated for about 2A@12V and the GPS converter is fused at 2A, so I can't add any more load.
By the way, not sure if you knew from my previous thread, but I was calling my gps and converters "bricked" meaning dead, like paper weights. As for the broken GPS's, Garmin makes you go through some combinations of button presses and charging them to try to revive the devices, but mine were completely dead. They also make you send it back to them.
Thanks for the suggestion on the load switch. I will likely buy the parts for the load switch to have a sort of back up circuit for dropping the GPS load to relays. I am running out of time on my GPS warranty so will probably just assemble relays and TVS first for testing though.
Edit: Regarding my original relay question, I realized I should not have any issues with the voltage drop because relays have a coil turn on voltage much lower than the nominal voltage, and once "on" they will not turn off until the voltage dips quite low. A lot of the 14v nominal automotive relays turn on at 7V and off at around 2V.
-
tatus1969,
I showed him your previous post and his explanation mirrors what you had to say about choosing an inductor with a high resistance. And thanks for running that in LTspice. Here is what he had to say if you're interested:
Well, that inductor, even if saturated, I think can only pass high frequency via parasitic capacitance, and this would be counteracted somewhat by the aluminum electrolytic, and also capacitance inside the TVS which is probably a lot faster to respond than the capacitor and I assume a good bit larger than the parasitic capacitance in the inductor.
I included the reverse direction diode also to prevent resonance, since if the current through the inductor can't go negative, I'm not sure how it's really said to be resonating, but maybe I'm misunderstanding what they're meaning.
But using a lossy ferrite would be an improvement - what happens when a large voltage pulse shows up, is the inductor starts storing magnetic energy as the current through it increases. This magnetic energy will be released as the current decreases later, and manifests as the current decreasing more slowly than the rate of the disappearing pulse.
A noise suppression ferrite has a property where a lot of the magnetic energy is lost in the core material; meaning the core material will warm as the molecules twist inside it in response to the magnetic field increasing and decreasing, bleeding some of the energy away. So if the inductor sucks up 1 unit of energy, it might only release .5 units electrically afterward (the rest into heat).
However, I'm depending a lot on the inductor's resistance in this design to keep things simple and fewer part count. It's really the resistance that's providing most of the protection here. If you just buy a bare ferrite and wrap a wire through it, it would be extra work to control for the resulting resistance. But, "some" inductance helps keep the rate of change of voltage and current from surpassing what the capacitor and TVS might instantaneously handle.
I guess to sum up, I think the high frequency aspects are handled okay enough, but picking that particular inductor provides resistance that improves lower frequency protection, and if you go with a ferrite instead, you'd have to be a lot more careful to control the resistance (maybe as simple as adding another resistor, but not all meters are good at measuring < 1 ohm to see where you're at with a ferrite+coiled wire setup)
That's my 2 cents on it; this isn't really an area of expertise for me, but I feel pretty confident that it would work with those specific parts. If this were a professional product, I probably wouldn't recommend this design because it's not clearly separating different roles into different components, and if a part becomes unavailable it would be more challenging to find a replacement.
-
tatus1969,
I showed him your previous post and his explanation mirrors what you had to say about choosing an inductor with a high resistance. And thanks for running that in LTspice. Here is what he had to say if you're interested:
Well, that inductor, even if saturated, I think can only pass high frequency via parasitic capacitance, and this would be counteracted somewhat by the aluminum electrolytic, and also capacitance inside the TVS which is probably a lot faster to respond than the capacitor and I assume a good bit larger than the parasitic capacitance in the inductor.
I included the reverse direction diode also to prevent resonance, since if the current through the inductor can't go negative, I'm not sure how it's really said to be resonating, but maybe I'm misunderstanding what they're meaning.
But using a lossy ferrite would be an improvement - what happens when a large voltage pulse shows up, is the inductor starts storing magnetic energy as the current through it increases. This magnetic energy will be released as the current decreases later, and manifests as the current decreasing more slowly than the rate of the disappearing pulse.
A noise suppression ferrite has a property where a lot of the magnetic energy is lost in the core material; meaning the core material will warm as the molecules twist inside it in response to the magnetic field increasing and decreasing, bleeding some of the energy away. So if the inductor sucks up 1 unit of energy, it might only release .5 units electrically afterward (the rest into heat).
However, I'm depending a lot on the inductor's resistance in this design to keep things simple and fewer part count. It's really the resistance that's providing most of the protection here. If you just buy a bare ferrite and wrap a wire through it, it would be extra work to control for the resulting resistance. But, "some" inductance helps keep the rate of change of voltage and current from surpassing what the capacitor and TVS might instantaneously handle.
I guess to sum up, I think the high frequency aspects are handled okay enough, but picking that particular inductor provides resistance that improves lower frequency protection, and if you go with a ferrite instead, you'd have to be a lot more careful to control the resistance (maybe as simple as adding another resistor, but not all meters are good at measuring < 1 ohm to see where you're at with a ferrite+coiled wire setup)
That's my 2 cents on it; this isn't really an area of expertise for me, but I feel pretty confident that it would work with those specific parts. If this were a professional product, I probably wouldn't recommend this design because it's not clearly separating different roles into different components, and if a part becomes unavailable it would be more challenging to find a replacement.
I basically agree to that, the circuit should work fine as it is proposed. Also the inductor's parasitic capacitance will only kick in at very high frequencies (typical value is maybe 20pF for that size). The TVS has enough capacitance to short that energy (typically in the nanoFarad range).
There are a few general things where I cannot agree:
- a saturated inductor (in terms of magnetic saturation) loses its filtering property, because it loses its inductance.
- yes, impedance of an inductor is proportional with frequency, but only if there is no capacitor around. In this circuit, L1 and C1 form a resonant circuit (https://en.wikipedia.org/wiki/LC_circuit (https://en.wikipedia.org/wiki/LC_circuit)), creating amplification. This is correctly counteracted here by choosing an inductor with a high resistance, which is damping the resonance. I have seen many cases where this was not the case, for example a power inductor and a capacitor at the output of a regulator, hoping that it will reduce ripple. It does not, it amplifies a certain frequency.