Small circuit design

[danadak]

It could be your charger has current limit, if decently designed.

If you use R you rely on charger design for safety and to prevent LED burnout.
And if charger puts out more V than you have LED current changing as well,
eg. not regulated.

If you use LM317 definitely need a heatsink on it. There is more protection
overall. But unlike a R it can have ESD concerns as pretty much all integrated
components have.

Do you have a datasheet for charger ? Or is it just a el-cheapo made in outer
Mongolia by sheep herders (they are good people by the way) ? With a datasheet
that can be written on a match head.......

As an aside I bought a benchtop power supply many years ago, and started burning
out parts/designs. When turned off it generated 50+ volt transients which fried any
LV circuits I had attached to it. Who knew ....? Point of this is what does your charger
do when turned on and off with a load attached....? A scope, a DSO, with single shot
triggering for excess V can quickly reveal this.

Don't get discouraged, good design requires a lot of investigation, you are a detective,
and its a constant learning process until you drop dead. Design life would be boring
if uncertainty were not a part of design.....

When confronted by a mish mash of conflicting pros and cons I employ a decision
matrix, with a column for pro, one for con, one for a 1 to 5 rating, and sum up each column.
Often can be an eye opener.

Lastly you can always throw in the towel, just use a R and keep your fingers crossed
the house does not burn down

Regards, Dana.

 

[Jefff]

It could be your charger has current limit, if decently designed.

If you use R you rely on charger design for safety and to prevent LED burnout.
And if charger puts out more V than you have LED current changing as well,
eg. not regulated.

If you use LM317 definitely need a heatsink on it. There is more protection
overall. But unlike a R it can have ESD concerns as pretty much all integrated
components have.

Do you have a datasheet for charger ? Or is it just a el-cheapo made in outer
Mongolia by sheep herders (they are good people by the way) ? With a datasheet
that can be written on a match head.......

As an aside I bought a benchtop power supply many years ago, and started burning
out parts/designs. When turned off it generated 50+ volt transients which fried any
LV circuits I had attached to it. Who knew ....? Point of this is what does your charger
do when turned on and off with a load attached....? A scope, a DSO, with single shot
triggering for excess V can quickly reveal this.

Don't get discouraged, good design requires a lot of investigation, you are a detective,
and its a constant learning process until you drop dead. Design life would be boring
if uncertainty were not a part of design.....

When confronted by a mish mash of conflicting pros and cons I employ a decision
matrix, with a column for pro, one for con, one for a 1 to 5 rating, and sum up each column.
Often can be an eye opener.

Lastly you can always throw in the towel, just use a R and keep your fingers crossed
the house does not burn down

Regards, Dana.

Wonderful words of advice.

There may or many not be sheep herders involved, I suppose in some instances this could be true.

Even with the Reed switch, and schottky reverse polarity diode, is there any reasonable risk that the supply from the charger would damage the circuit?
I had planned that the slow blow fuse would safely absorb any minor fluctuations arriving from the charger. Is this naive? I guess to be honest I am relying partially on the charger to provide some level of protection.

Currently I am inclined to believe in this solution an R is a solid choice in place of an LM317 for the various reasons mentioned.

Cheers,

 

[danadak]

Cheap chargers do occasionally fail, with no output to unregulated out of spec high
V out. Couple that with slo blow delay certainly possible to fry LEDs.

Its a roll of the dice.

 

[Jefff]

Cheap chargers do occasionally fail, with no output to unregulated out of spec high
V out. Couple that with slo blow delay certainly possible to fry LEDs.

Its a roll of the dice.

Thank you. Following the same trajectory, do you see any suitable alternative to the slow blow fuse?

 

[danadak]

A classic fast acting circuit is a crowbar circuit to blow a fast acting fuse.

Crowbar Circuit | Design using Thyristor, Working

A tutorial on Crowbar Circuit, which is an overvoltage protection circuit designed using Thyristor / SCR or TRIAC. Circuit Design, Working.

www.electronicshub.org

https://www.onsemi.com/pdf/datasheet/mc3423-d.pdf

https://www.littelfuse.com/assetdocs/how-to-design-crowbar-protection-in-high-power-applications?assetguid=62ede520-1813-40e8-a9d6-5ac1a14bcb43

 

[Jefff]

The crowbar circuit is very interesting, simple and adds good protections for the application. Out of curiosity, don't most oem mobile/ tablet chargers come with built in protections to protect downstream devices?
If I were to design a function-first circuit to evaluate component feasability and product functionality ( removing the slow-blow fuse and schottky ), whilst using an oem charger, do you feel there would be significant risk to damaging the components?
After validation of design, by adding the crowbar circuit after the PD, would that have an affect on downstream components, V and C demands/ dimensioning?

 

[danadak]

The crowbar circuit is very interesting, simple and adds good protections for the application. Out of curiosity, don't most oem mobile/ tablet chargers come with built in protections to protect downstream devices?

Do not know, cant speak for "most".....

If I were to design a function-first circuit to evaluate component feasability and product functionality ( removing the slow-blow fuse and schottky ), whilst using an oem charger, do you feel there would be significant risk to damaging the components?

Just use a current limited supply/charger. If using a bench supply set it's V then short its output and set the current
to a limiting value. Its a great way of powering up circuits you wire up on a protoboard and making sure they dont
get fried due to a wiring error. Not perfect but often sufficient. If its charger that is focus current limit its output with
a R in case its not short circuit protected. You can always use a DSO to evaluate charger transients, using a mosfet
to apply shorts to circuits, DSO trigger output to turn on the MOSFET, scope set for one shot, so transient captured.
Use a sampling R to probe for current, if used hi side in charger output use DSO in differential mode (2 channels.
2 probes, math set to subtract one channel from the other and present as trace).

After validation of design, by adding the crowbar circuit after the PD, would that have an affect on downstream components, V and C demands/ dimensioning?

The crowbar, generally speaking, does not draw any significant power until activated, eg., the transient
when its activated. If downstream you have a 3 terminal type regulator shorting its input to ground can
damage it with the reverse current created by its load C. So diodes are added as shown in this "typical"
circuit. Ignore the "obsolete product" reference, this is just typical for any 3 terminal regulator.

---

Note in your solution adding polarity input protection diode still a good idea if
battery can be inserted into product with reversed polarity.

 

[Joliane]

It needs at least 10V and 12V is too hot or inefficient.

 

[Sunnysky]

Here's a method using a Festoon car lamp that is close to the resistance <1 Ohm cold.
Reverse protection diode added (opt)
Disable can be made active Low with a simple change.






Simulation < press continue to destination = Falstad Interactive sliders for Vbat and Lamp Power rating at 14.2 on the right.

A power FET 0.1 to 1 Ohm with a sink works better

 

[ahsrabrifat]

• Suggestions on thermal considerations (LM317 power dissipation, resistor wattage, etc.).

I think you should use a heat sink for LM317.

 

[Sunnysky]

The most efficient solution is to choose a current-limiting resistor that drops X<< 1 V. The R will dissipate 350 mW at 1V, and you should select a power rating at least twice to reduce the temp. rise from 120 to 60 °C above ambient.

This demands you choose a power supply that is N* Vled + X volts = V+

For example 15V to two 7V LEDs requires 1V/350 mA = 2.9 Ohms or ~ 3, 1W or four 12 ohm in parallel 1/4W

 

[Delta Prime]

Deleted

 

[Delta Prime]

Currently I am inclined to believe in this solution an R is a solid choice in place of an LM317 for the various reasons mentioned.

I think you should use a heat sink for LM317.