I have been asked to replicate a design of a LED indication board for the company I work for. I mostly copied the board from the company that has been manufacturing it for us, except for using an extra resistor that for the life of me I can’t figure out how they did not use. That purchased board is a ‘critical component’ so we will have to work with our certifier to change it. I am unsure if the purchased version was separately certified or if our certifier just accepted it because of that company’s reputation.
The main part of the board is 6 LEDs that are powered from 120VAC contacts. It goes as below
I have met the clearance requirements in the layout with lots of room except for the LEDs themselves in which I question how effective a conformal coat will be when they are flush to the board. Without the conformal coating the LED pins are too close together (0.74mm). The board will have conformal coating but I’m not sure if that will be sufficient for those LEDs.
The other part of the board is 24VAC and I’m less worried about it because I’m sure the certifiers will be less concerned about it. It is in this part where the previous company used one less resistor and it was for a very crude voltage regulator to power a 555 timer. They somehow did not use a current limiting resistor (R8 below) before the Zener diode as pictured below.
I have reached out to a ‘sister company’ to the company I work for about getting them to manufacture and review the board. They suggested converting it to SMD components which I think would be a mistake, especially for the 5W resistors included in the design as current limiters to the LEDs.
Primarily I am trying to gauge
How big of a headache it will be to get this board approved for use in my employer’s product?
Secondly whether my hesitation about switching to SMD components is justified?
The power supply hanging out after the zener will draw some current, but that will only be a limited amount. The zener drops the voltage at Port P to a suitable voltage for the regulator and will likely dissipate a fair bit of power, so I’m surprised it or the power supply are not also designed as critical.
The clearance of the legs on the LED sounds like it could turn out to be an issue, though.
I’ve done LEDs from mains power and gotten it safety-approved for 240VAC.
It’s much more efficient to use a capacitor to limit current instead of a resistor.
You need to protect the LED against reverse voltage with a diode.
If you really want to use a resistor, there’s nothing wrong with SMD - the PCB and copper will act as a heat spreader.
The 330Ω 1/2W resistor must be surge-rated. You might be able to eliminate the 100Ω resistor, I forget why it was deemed necessary - maybe for compliance under single-fault conditions.
The test lab will be very concerned indeed with what happens if the transformer is overloaded, e.g. if the secondary gets shorted somehow. You will want a fuse (single-use or resettable PTC) in either the primary or secondary, I believe you get more coverage by putting it on the primary side.
They are directly connected to the mains connection with only the single diode and resistor and then each are powered separately by a circuit that monitors a part of the system (HVAC). The LED can then be used for troubleshooting problems.
The zener diode is only used in one place where a 555 timer is being used to blink one of the LEDs on and off if it is powered. Below is the Falstad circuit I played with while making it.
It has been discussed whether we should just drop this part out as on the current design it is the first part that fails. It doesn’t actually have a power supply either. Its not really something I’d have considered doing until I started trying to replicate this design based on the current board.
Again I want to mention that somehow the current board does not use the 470ohm resistor I have included. I don’t have access to their design files and such (just mechanical drawings for mounting and the board itself). Its not that important to me to know how they got away without it but I’m curious.
I think you need to provide an actual as-built schematic for this to get answered more fully and correctly. There are lots of potential explanations but they would all be guesses without knowing what you have.
Thank you for this. I will keep in mind both of your comments. I have read that using a capacitor is more efficient. Honestly I went with solely the resistors because I was trying to stick as close to the current board as possible.
For the certification process itself do you have any advice for it? Are there any useful guides online or something?
I can already see from the photo you shared that my documentation likely needs to be improved even more than I have already. I am thinking of asking my employer to schedule an informal meeting with the certification body we work with (I heard that’s a thing) to get a better idea of what they want before getting it tested.
Now it does connect to a much larger circuit, but I won’t be able to share that.
Also not shown here is all the power calculations and such for each component. I know that I will have to share all that with the certification body, but I am confident the parts I chose meet the requirements. Right now I am primarily trying to determine if the LED clearance will be a problem and (after reading these comments) if fuses should be included (even though there are none on the current board).
I was trying to share the parts that were necessary for the questions about clearance of the LEDs and general questions on how to get a board certified to be used in an existing product without confusing the issue too much.
I more wanted to ask about advice/general experience with getting PCBs that interact with power mains certified and not ask about a particular design or for a design review.
Primarily I didn’t want to ask for a design review because I do think there are many things (primarily efficiency wise) that can be found wrong with this design and didn’t want to get too bogged down in that.
Find out what standards are applicable, buy them, read them. Then read them again
The Critical Component List or Constructional Data File (different NRTLs have different terminology) has a pretty standardized format: columns for Description, Manufacturer, Mfr Part Number, Ratings/Specs, standards certified to (the component) and NRTL listing(s) for the component, with file number(s).
Some NRTLs will do a preliminary design review for you, for a fee. They generally will not (cannot, legally) offer advice per se - they’ll only tell you where your design is not in compliance with the standard(s).
UL, being The King, is often a pain to work with. Intertek tends to nickel-and-dime you, but I’ve done projects with them. TUV Rheinland is my go-to when possible.
So let me see if I get this right – the first schematic, you have 120VAC hot coming in separately on pins 1 to 6, and a shared neutral return on pin 7?
The second schematic, you have 24VAC with return common on pin 1, and sources on pin 2 and 3?
But you also said there is no 470 ohm resistor, but you have drawn a 470 ohm resistor – so is that the actual as-built circuit that exists currently?
The circuit I built has the 470 ohm resistor. I do have a tested prototype.
The circuit I am replicating (that I do not have any engineering files for) appears to not have that resistor. I diverged from that circuit (or what I can see from it) because I could not for the life of me figure out how they did not include a resistor there. I have read online that they may be using a current limiting diode or something. I have not checked that option, because I prefer using the resistor and did not want to cut the diode off the board (we still need them).
All we were ever given by the 3rd party manufacturer for that board (that they make for us) is mechanical mounting and dimensions. No schematic or layout documentation. This is the primary reason I was asked to work on it.
Sorry I guess I was not completely clear that we do not have any documentation for the one circuit in production. It was designed for us years ago and no-one ever thought to ask for a schematic and documentation. When I came here I did ask for that and the 3rd party company only supplied mechanical dimensions but not any schematic.
The also did not acknowledge that it was not a schematic or out right refuse to supply it. Kind of just ignored us.
The 24VAC comes from a transformer that is used to power other things as well as this board. I think it is current limited, but the outside company would not necessarily know that.
I’m trying to find a defective board so I can cut that one diode out but it looks like all the ones we had have been thrown out.
Class 2 are supposed to be current-limited but I have had them melt on me (Triad, no less). It didn’t burst into flame and maybe that’s enough for the classification, but it stank and took a chunk of circuitry with it.
