Hi. Went to the store to buy some light bulbs and decided to grab a LED night lamp just to see how the circuitry has progressed.

Dismantled, found the AC in series with tiny resistors and a diode? stamped A7 to the single LED. Can be a 1N4007.
By some reason there is two resistors, allowing to simplify further into only one. It is not the circuitry seen in older, elaborated
night lamps. What do you think ?

 

By some reason there is two resistors, allowing to simplify further into only one.

I think they used 2 resistors in series because the working voltage of one isn't high enough.

 

Judging from the two electrical prongs (the plug, which appears to be American type, not European) suggests a 120VAC source. Assuming the 101 resistor is 100Ω and the 243 is 24KΩ, that's 24,100Ω total. For a moment with such a high voltage I'd not even bother with the forward voltages of the LED or the diode, that's approximately 5mA through the LED. Just for sniffs and giggles, accounting for the Vf of both diodes, the 0.7Vf and the probable 2.3Vf for a total of 3Vf, (120-3)÷ 24K1 = 0.00485477A (whereas disregarding the Vf's is 0.00497925A) (so little difference it's not worth including in the calculations).

Off hand I don't know what those resistors are rated for voltage wise, but can a 100Ω resistor make enough of a difference to account for the voltage? They're small, but I don't see any concern for size of the components lending themselves to a spark gap for 120VAC. Maybe, I don't know. But my first inkling is that it probably isn't the reason. But not being able to account for any other reasoning, perhaps I'm wrong.

In the past whenever I was wrong a single hair follicle would die and the hair fall out. More skin than hair paints a pretty sad story.

 

Of course my chin seems to be on a "Right" streak. Lots of hairs there. Maybe the hair hasn't been falling out. Maybe it's just sliding down my face. Probably should update my avatar.

 

Might one of them be a fusible resistor?
I‘m surprised that there isn‘t noticeable flicker on a half-wave rectified supply.

 

Are you sure one of those is not a capacitor?

 

I wondered about a cap (the usual voltage dropper), but wouldn't that need the A7/LED combo to be bidirectional so that the cap can both charge and discharge? Perhaps the LED is in fact two LEDs back-to-back in anti-parallel (though it doesn't look like it), or else the A7 diode is anti-parallel to the LED rather than in series with it.

 

Yep that is what I was thinking. Capacitor, resistor in series, diode anti parallel to LED.

Can we see a photo thst shows the components?

 

Confirmed resistors, read the values on meter OK. No capacitors, no bridge rectifier.

AC-------------switch -------------/\/\/\/\-----------------/\/\/\/\-----------------|>|------------------------LED--------------------AC

 

It’s so they can use the same pcb for a 230V version which would require two series resistors to meet the voltage spec.

 

I‘m surprised that there isn‘t noticeable flicker on a half-wave rectified supply.

Most of my video cameras film at a 30 fps rate. Half of a 60Hz sine wave would be equivalent to a 30 fps video. You can start to notice flicker at 27 fps but it's hard to detect visually. Of course, moving your eyes across the field of vision will definitely display dashes in space where you see the LED when it's on and when it's off. Something like this ---....---....---....--- (ignore the dots - this thread won't allow more than a single space between characters and I don't know how to produce a blank character where the thread will see it as a legitimate character)

Here's what I get from the photograph:

 

Most of my video cameras film at a 30 fps rate. Half of a 60Hz sine wave would be equivalent to a 30 fps video. You can start to notice flicker at 27 fps but it's hard to detect visually. Of course, moving your eyes across the field of vision will definitely display dashes in space where you see the LED when it's on and when it's off. Something like this ---....---....---....--- (ignore the dots - this thread won't allow more than a single space between characters and I don't know how to produce a blank character where the thread will see it as a legitimate character)

Here's what I get from the photograph:
View attachment 273136

The fact is that at 60 Hz we have 60 positive half cycles every second not just 30 half cycles Hence no 30Hz flicker.

 

AFAIK non-isolated AC circuits like that tend to use two series resistors for safety, just in case one fails short, (hopefully) the other would still be intact. A cap drop supply uses a capacitor in series as well as a resistor but the cap needs to be high voltage type which would make it big also generally if good design practice is exercised, it should be an X2 cap.
So that design looks like a cheap and minimalistic design with no thought for spikes and surges which generally is the cause of AC LED failures.

 

60 Hz ••• have 60 positive half cycles every second

Actually - yeah. Thanks for pointing that out.

 

Confirmed resistors, read the values on meter OK. No capacitors, no bridge rectifier.

AC-------------switch -------------/\/\/\/\-----------------/\/\/\/\-----------------|>|------------------------LED--------------------AC

View attachment 273043

You say you measured the resistors - are they really 24k and 100 ohm, totalling 24.1k?

I can understand the value of two separate resistors for safety, but only if their values are close enough together to significantly reduce the voltage across each of them.

As it is, the voltage drops at 120V would work out to only 0.498V across the 100 ohm resistor and 119.502V across the 24k resistor.

i have to say that I don't feel much safer with 1/2 volt less across the big resistor! If those are the real values, I see no reason for two separate resistors instead of one.

... Unless maybe this PCB is a shared layout that is also used to make 230V models, and they put a much larger resistor in place of the 100R for the higher voltage versions. That would make sense to me. In that case, the 100R here is doing essentially nothing, and could just add easily be 0-ohm resistor (aka jumper resistor?)

 

AFAIK non-isolated AC circuits like that tend to use two series resistors for safety, just in case one fails short, (hopefully) the other would still be intact.

Was going to say what ebeowulf said.

You say you measured the resistors - are they really 24k and 100 ohm, totalling 24.1k?

I can understand the value of two separate resistors for safety, but only if their values are close enough together to significantly reduce the voltage across each of them.

As it is, the voltage drops at 120V would work out to only 0.498V across the 100 ohm resistor and 119.502V across the 24k resistor.

i have to say that I don't feel much safer with 1/2 volt less across the big resistor! If those are the real values, I see no reason for two separate resistors instead of one.

Have to agree. 100Ω failing short - unlikely as it is - won't change much in the overall operation. I'm more in favor of:

It’s so they can use the same pcb for a 230V version which would require two series resistors to meet the voltage spec.

While in more favor of this explanation I still don't see the reason for 100Ω. It contributes virtually nothing to the circuit. They could have used a 0Ω jumper instead. Depending on the production run, at 120VAC R1 could be 0Ω and R2 could be (as is) 24KΩ. In the case of 240VAC, both R1 & R2 could be 24KΩ each. But perhaps they have an overstock of otherwise useless to them 100Ω resistors. Since it adds so little to the function but they need something in that location - why not use what you have?! It's probably just as expensive as a 0Ω jumper, but if you have to stock up on 0Ω and you have tons of 100Ω; why carry the inventory?

Neither resistor looks like a fusible resistor. But looks don't always tell the whole story. We have no idea why there's that 100Ω resistor in circuit. At least I don't. And the difference in current is 5mA versus 4.979mA. A difference of 21µA.

 

Doh! I didn't realize that @Ian0 had already mentioned the dual-use 230 application where they would use different resistor values, back in post 10.

Hehe, I thought I was being clever, but he beat me to it!

 

It's about inventory. 2 x 24k is correct for 230V. So for 115V, you need 2x 12k but you might not have 12k in stock, but you do have 24k because it is used in the 230V version. So fill the gap with the lowest value resistor you stock. If you normally do double sided boards, you won't have zero ohm links.

 

Probably the 100 ohm resistors were cheaper than the zero ohm ones. And if the 100's were of the +-20% type they may have been really cheap.

 

Probably the 100 ohm resistors were cheaper than the zero ohm ones. And if the 100's were of the +-20% type they may have been really cheap.

You're right - they charge a premium for low value resistors. 100Ω would cost less than 10Ω