Platform forum

Circuit Design for a simple LED

allee.khaan , 09-05-2023, 08:00 AM
I hope everyone doing well.
I need some suggestions regarding isolating my circuit.
I attached the schematic of the circuit. The function of circuit is as follows.
1: 1s lipo battery charging circuit
2: LED ON OFF function

This circuit is replicated multiple times on different boards and all boards connect together in parallel to each other through J2 connector.
There are 4 signals on J2 connector
1: VCC
2: GND

At the moment, there is zero volt at VCC pin but somehow leakage voltage appears from battery to VCC pin that are 20mV. I already selected a reverse protection IC instead of TPS2051. I hope it will resolve in the next version.
The main problem is when all boards are connected with each other. I want to isolate battery power from other boards.
Let's say, i connect 4 these boards in parallel. Each board will have it's own battery .
The Dock_Sig pin shows 1.4V with respect to GND when connected with other boards. When it is not connected with other boards then it shows less voltage than that i.e. 1.13V
Here comes main problem.If somehow any board battery dry out then it start influencing other boards LED and glow them at very low brightness even if ON condition is not present. The problem appears across R11 that's takes voltage from DOCK-Sig pin and pass to the 3.3V node. As it has itself no voltage at 3.3V so it act as a power sucker.
I need help to isolate DOCK_Sig pin from other board.

1: Dock_Sig board connect with GND when it attach with other boards and Switch Node connect with a switch that either stay in floating state or connect with GND.

Plz, let me know if you have any questions about it.

allee.khaan , 09-06-2023, 03:19 AM
Hi, Can anyone help me plz?
qdrives , 09-07-2023, 03:08 PM
How can pin 1 of J2 be Vcc? Vcc is a net between U3 and U4 and pin 1 of J2 does not have a name, so is not connected to Vcc.

Leakage in TPS2051 is in the dataaheet:

Your Dock_Sig is directly connected to the anode of the LED in the solid state relay (MOSFET relay).

What is the function of the DockSig?
How do you want to drive it so that the LED is on?
Should it be on by default?

allee.khaan , 09-07-2023, 04:10 PM
Yes, you are right.
I should write VDD pin instead of VCC.
The J2 pin 1 connect with external 5V that after passing through U1 appears at VDD node of the TP4054.
I attached a block diagram, how multiple boards are connected with each other.
There is only single Board C in whole design that provide 5V and GND and SW signal.
The dock signal function is to turn OFF SSR when it connect with GND otherwise keep it ON.
Here is the function of whole design.

Board A or Board D when not connected with it's corresponding Board B or Board E

The SSR always ON and 555 timer control the LED using a vibration switch. This part working fine

Board A or Board D when connected with it's corresponding Board B or Board E through J2 connector

Now Dock_signal will turn OFF SSR and eliminate the function of 555 timer to control LED. Now the Switch signal will control the LED directly by connecting it with GND or floating(ON or OFF).

Let me know if you have any more questions.
qdrives , 09-08-2023, 02:17 PM
I would replace your Solid State Relay and replace it with a P-Channel Mosfet.

Rleds if where your LEDS are.
R5 represents a switch to Gnd.

allee.khaan , 09-08-2023, 02:49 PM
Hi @qdrives
Thanks for your reply.
I also posted this question to another platform a month ago. I got some suggestions that indicate the actual problem but couldn't find the solution yet.
It is Switch NODE.
The switch node is connected in parallel with other boards as well.
When the power is good in all other boards then Switch node stays at 1.4V but as soon, one board power fails, it create a 0V at it's Switch node that causes the led glow at very low brightness i.e. 2.3V Vf. The typical Vf of LED is 2.75V.
When switch node connected with GND then it work fine. The switch node create problem only in two conditions.
1: When other parallel board power goes to 0
2: When it is in floating mode
They suggested me to connect a pullup resistor to avoid floating.

Now if i add a pull up resistor then it can solve this issue but it can add up another challenge.
The pull up resistor will create a voltage at switch node i.e. 3.3 - Vr pull up.
The 555 timer act as a sink in this operation. When it create logic 0V(SSR is ON) then to LED glow, it must have voltages greater than 3.3 = Vf led + Vr pull up.
I need a way to isolate switch node from other boards in floating mode and when i have to apply GND signal then all boards can connect with GND through a single switch.
This is where i am stuck now after long discussion with Texas instrument team.
allee.khaan , 09-08-2023, 02:51 PM
​This is when other boards get to 0V and create a path for leakage
allee.khaan , 09-08-2023, 02:52 PM
​Voltages at different node in normal operation
allee.khaan , 09-08-2023, 02:54 PM
When all boards has good battery power and no issue.

When one board battery fail then voltages across switch node.
qdrives , 09-09-2023, 10:15 AM
So the first thing is to describe what you want to do.
A bit from your problems and connections I assume you want to control the brightness either locally (dock sig = high) or remotely (sock sig = low).
The way you have done will cause EMI issues as the LED current goes from switch, through board B, to board C to Gnd. Then power from board C through board B, U1, U3 and U4 to reach the LEDs again. That is a big loop.

Change the architecture so that:
1) The dock sig powers on/off the 555 timer. You could use the same circuit as I made but this time Rleds is the 555 timer.
2) A kind of OR functionality that turns on the LEDs - so either Switch or the 555 timer can turn the LEDs on.
3) Make sure that the levels are that when the 555 is not powered, it does not turn on the LEDs.

For 2 and 3 you could simply have 2 FETs or transistors in parallel.
allee.khaan , 09-09-2023, 10:43 AM
Your 1st assumption is correct.
I have to run the board locally as well as remotely.
In remote condition, i am doing two functions.
1: Charging multiple boards(Board A, Board D and so on) through a single dock station. When Board A(Board D and so on) connected with Board C through B. It must transfer it's control to Board C to turn ON/OFF LED.
The dock signal simply define when Board A is connected with Board C through Board B and eliminate the usage of 555 timer through SSR.
Regarding, your suggestion about loop, unfortunately, i have to go with it because dock station main switch is connected with boards through a cable. I can't move that switch location to anywhere else.
Now about your suggestions.
1: I couldn't see your circuit. Did you upload it somewhere?
2: Yes this is what i need. But there must be some signal that define when to shift control over 555 timer and when to switch. This is what i did through Dock signal. It shut off SSR when dock signal is low otherwise keep it High. The dock signal is only low when it connected with Board B.
3: you mean logic level of FETs?

I previously used FET but due to it's leakage current , it switching created issues so i moved to SSR for dock signal.

The dock signal only loosing power from battery i..e (3.3 V)/2K = 1.65mA when connect with GND if i am not wrong.
I can live with it but if it can also avoid, that would be great.
The main problem is Switch NODE.
I want to isolate this node from Board B but in the meantime, also want to control this node through switch from Board C.

qdrives , 09-09-2023, 01:10 PM
I am not suggesting to move the switch. I am suggesting to have the LED current remain local. The control signal to drive the LEDs can be remote.

1) Don't you see a schematic here: https://designhelp.fedevel.com/forum...2357#post22357
2) you can drive the LED like this:

The big difference in this circuit is that it is switching the LEDs on when supplying voltage/current and not connecting to Gnd. But it is a bit less error prone in low battery conditions.

The dock signal only loosing power from battery i..e (3.3 V)/2K = 1.65mA when connect with GND if i am not wrong.
You are wrong. True Dock signal draws 1.65mA when connected to Gnd. However, If not connected, current goes thru the LED of the SSR. Lets assume 1.5V, then you draw (3.3-1.5)/2k=0.9mA.
allee.khaan , 09-09-2023, 02:02 PM
Now i got your point and it make sense to me.
I rephrase it here to make sure we are on the same page.
1: Control 555 timer power using this circuit. Rleds will be supply pin for 555 timer. The Rswitch will be dock signal that will control the power to 555 timer. If floating or open then 555 timer is supposed to get power because i have only one option to connect with GND of dock signal to turn OFF power
2: If dock signal turn OFF 555 timer then Switch will control LED using this circuit because 555 timer is not working anymore. Connecting with GND or Floating, Switch node suppose to turn ON/OFF LED. I need switch node connect with GND to turn ON LED. In the floating stage, it is supposed to be OFF because this condition will also available when 555 timer power ON.

qdrives , 09-10-2023, 09:16 AM
1) Correct.
2a) Both switch and timer can switch on LEDs. But I do assume that switch and dock are either both connected or neither of them.
2b) If you also need the LEDs to turn on when switch is connected to Gnd, you need to change it to PNP transistors, swap LED and transistors, change pull-down to pull-up and diode to pull-up like in the first schematic.
allee.khaan , 09-10-2023, 09:23 AM
Yes, the switch and dock will either connected both or neither of them
I will update the circuit and will send you for final review.
Thanks for all your help.
allee.khaan , 09-11-2023, 02:32 PM
A quick simulation of your circuit shows that grounding or floating the Switch node does not turn on/off the LED.

Furthermore, the 555 node is being pulled to a voltage of ~500mV in the simulation. I would be cautious with any circuit configuration that can drive the 555 nodes to any voltage other than ground while the device is powered off, as this can cause damage to the device. Remember, just because the 555 timer is powered off does not mean that it is no longer in the circuit!​
qdrives , 09-11-2023, 02:51 PM
As I stated, the switch signal in the simulation I created needs to high in order to switch on, NOT to Gnd.
If you want it to switch on when connected to Gnd, you have to use PNP or 2x NPN to invert (not darlington).

I do not think that the 8nA going to a switched on 555 timer would be a problem. In your simulation nothing is connected there. So if you still want to reduce the voltage, you could add a 100k pull-down on the 555 port too.
allee.khaan , 10-28-2023, 12:20 PM
Hi @qdrives
I modified the circuit. Sorry for being late. i was busy in different things.
Can you provide review about it plz?
I know it is different than what you suggested because i was facing some challenges to satisfy all functions. This one seems to work but not sure if there is some technical issues like forward voltages or cutoff and saturation regions of mosfet using optocoupler.
qdrives , 10-29-2023, 08:57 AM
My review is very short. It is a blank sheet.
Upload it again.
allee.khaan , 10-29-2023, 11:24 AM
Sorry about that. Somehow print message data.
I uploaded again. Plz, check it now
allee.khaan , 10-29-2023, 11:29 AM
i have main concern about optocoupler pullup resistance at input side or maybe you can find some other issues as well.
allee.khaan , 10-30-2023, 07:46 AM
One quick correction, I connected the Switch port directly at 3.3V at optocoupler input. It is supposed to be at Anode pin.
qdrives , 10-30-2023, 02:33 PM
Let me ask as a teacher: why/when do you use/need an optocoupler?
allee.khaan , 10-30-2023, 02:37 PM
To isolate supply. I know what you are talking about.
My concern isn't supply.
My only issue is to not share rest of the circuitry control with similar other boards that will connect with switch node,
and dock station node
qdrives , 10-30-2023, 02:39 PM
Will these boards be produced in high quantities (> 500 pcs)?
allee.khaan , 10-30-2023, 02:41 PM
yes. maybe.
Is there a problem if we go for higher quantity?
qdrives , 10-30-2023, 03:19 PM
Optocouplers are expensive.
Most of what you want to do can be done with a simple $0.03 costing transistor.
U7 =~$0.30 U6=~$0.90

Are the dockswitch or switch normally active when it is battery powered?
Perhaps it is also better to give each LED it's own resistor so that variations in forward voltage would not cause one to be bright and the other soft.

And if you are going to use Gnd symbols, have them pointing down.
allee.khaan , 10-30-2023, 03:34 PM
Dock signal can only connect with GND and same for switch port
allee.khaan , 10-30-2023, 03:38 PM
As u said transistor a cheap way..can you tell me how I can replace opto and mosfet for switch node only..
consider switch port can be only connect with GND through a switch.
allee.khaan , 10-31-2023, 06:52 AM
Hi @qdrives
I updated the schematic further.
Removed Mosfet and now using a single SSR to do the same function
qdrives , 10-31-2023, 04:11 PM
Recreate this simulation and see what you think...
You figure out what you:
1) do not need
2) need more off
3) which is which
Use our interactive Discord forum to reply or ask new questions.
Discord invite
Discord forum link (after invitation)

Didn't find what you were looking for?