Picking correct varistor and caps for in-wall power supply based on Hi-Link HLK-PM 01/03?

Thread Starter

Wookbert

Joined Sep 24, 2020
10
I'm an electronics NOOB, working on my very first own PCB. The attached layout and schematic shows a triple-fused in-wall power supply (supposed to fit into an enclosure for a 60 mm diameter standard EU backbox). Next to the 100-240V PSU, which is either a Hi-Link HLK-PM01 (5V, 0.6 A) or a Hi-Link HLK-PM03 (3.3V, 1A) sits in this case a Wemos D1 mini ESP8266 module.

My questions are as follows (sorry for asking that many):

F1 - Subminiature fuse
Because of the very limited space I chose a subminiature instead of a classic glass fuse.
I'm considering two options: 250V, 200 mA, time-lag but different breaking capacity
  • a) 35A (Eska 883.010)
  • b) 130A (Eska 883.110)
See attached data sheet for details. Which one to pick?

MOV - Varistor
I have too little understanding on how to pick the correct varistor in terms of ratings and specifically size (S07, 10, 14 or 20??). Can you tell me which one to pick from the Epcos range of disc varistors?

C1/C2 - Caps
A found these output capacitors in a comparable schematic, where C1 was 0.1µF and C2 10µF. I intend to pick the following two
  1. C2: Würth Electrolytic capacitor 10 µF 16 V 20 % (Ø x H) 5 mm x 11 mm
I'm a bit concerning here regarding the voltage. According to what I've learned caps with higher voltage rating last longer, but if the rating is significantly higher than what’s actually used, it can cause ESR problems. I can't find C1 (0.1µF) however in less than 50V.

So is the combination of these two caps in terms of both µF and voltage OK?

Finally: Ground-area in PCB or not?
When looking at my 2-layer PCB design, do you see any need for a ground/copper area? If so, where? My worries are that, with the AC being so close, a copper area could cause leakage current, or that ground areas might block the Wifi signal from the ESP8266.

Thanks in advance for your patience and not hitting me right away :)

1600993812467.png1600993765343.png
 

Attachments

Hymie

Joined Mar 30, 2018
1,277
The fuse should be able to withstand the fault current that flows in the event of the MOV failing short circuit – this will certainly be more than 35A and could exceed 130A. For this reason HRC fuses capable of handling up to a 1500A fault current are common in primary circuits.

The AV/IT product safety standard (IEC/EN 62368-1) requires MOVs (across the mains) to have a continuous rating of at least 1.25 times the equipment rated voltage. Therefore your MOV should be rated at least 300V. The same standard requires the MOV to withstand a 8/20µs current of at least 3kA

In terms of the capacitor reliability, the use of 105⁰C rated electrolytics is recommended – if worried about the 0.1µF capacitor, use a non-electrolytic type.
 

Thread Starter

Wookbert

Joined Sep 24, 2020
10
F1 - Subminiature fuse
The fuse should be able to withstand the fault current that flows in the event of the MOV failing short circuit – this will certainly be more than 35A and could exceed 130A. For this reason HRC fuses capable of handling up to a 1500A fault current are common in primary circuits.
I've checked more of these sub-miniature fuses in TE5 and TE7 form factor. The typical breaking capacities are 35, 50, 100, 130 and rarely 150 A. It is virtually impossible to find these with 1500 A (I found one SMD fuse with 1500 A breaking capacity ... but that's about it). A 5 x 20 mm ceramic tube fuse would also come with your recommended 1500 A breaking capacity, but that one is way too bulky to fit on my board.

I'd simply go with the „strongest“ 130A sub-min fuse in this case. If I understand your explanation correctly, it simply means, that if the MOV (varistor) is „triggered“, the F1 fuse with 130A breaking cap. might get killed, too, correct me if I'm wrong. If that’s the case: Well, who cares? (As long as the house doesn't catch fire and burns down...)

MOV - Varistor
The AV/IT product safety standard (IEC/EN 62368-1) requires MOVs (across the mains) to have a continuous rating of at least 1.25 times the equipment rated voltage. Therefore your MOV should be rated at least 300V. The same standard requires the MOV to withstand a 8/20µs current of at least 3kA
Which brings me to the Epcos SIOV-S14K300 with 300 Vrms and 4.5kA imax (8/20 μs), as the smaller S10K300 has only 2.5kA imax (8/20 μs). (Part No. B72214S0301K101). (Perhaps you can confirm that you've meant 300 Volt rms, not 300 Varistor Voltage (=Vv)).

C1/C2 - Caps
In terms of the capacitor reliability, the use of 105⁰C rated electrolytics is recommended – if worried about the 0.1µF capacitor, use a non-electrolytic type.
What do you mean by „if worried“? That it breaks? Or that the 50V rating of the 0.1µF might causes the mentioned ESR issues? If the latter is the point: I can find 16 and 25V versions, but those are SMD. When looking for through hole the lowest I can find is a 35V 0.1µF tantalum capacitor.

So? 35V 0.1µF tantalum capacitor or 50V 0.1µF electrolytics capacitor?

Thank you again!
 
Last edited:

Hymie

Joined Mar 30, 2018
1,277
If the fuse fault current exceeds the fault current rating, it could fail catastrophically, exploding – spreading conductive debris within the enclosure. At 130A, this would require a short circuit impedance of less than 2 ohms – so is likely to be adequate.

The varistor rating is 300V is rms (1.25 x 240V~).

The purpose of the 0.1µF capacitor is to minimise any high frequency noise through its low impedance response (not to achieve a 10.1µF capacitance). Although the PSU data sheet specifies a 0.1µF value, I would expect a better performance with a 22nF ceramic capacitor.
 

Thread Starter

Wookbert

Joined Sep 24, 2020
10
If the fuse fault current exceeds the fault current rating, it could fail catastrophically, exploding – spreading conductive debris within the enclosure. At 130A, this would require a short circuit impedance of less than 2 ohms – so is likely to be adequate.
The fuse sits in a corner, blocked by the housing of the PSU, the housing of the 2-pin terminal block and the 3d-printed enclosure for my device. So if it explodes, I can kind of make sure that conductive debris doesn't get anywhere.

At 130A, this would require a short circuit impedance of less than 2 ohms – so is likely to be adequate.
I don't get this one, sorry. The 130A version would be adequate for what?
On „less than 2 Ohms“: Max. rating 250V / 130A (= 1,92 Ohm)? Or the actual 230V / 130A (=1,77 Ohm)?

What confuses me with your recommendation for a fuse with 1.500A breaking capacity, is that one – correct me if I'm wrong here – hardly can find these in such small, low-powered devices. Glass and ceramic tubes and these subminiature fuses are common in such small devices, yet these fuse type typically don't have more than 130 or 150A breaking capacity.
 

Hymie

Joined Mar 30, 2018
1,277
My statement that the 130A breaking capacity fuse is likely to be adequate is based on the fault resistance not being less than 2 ohm, which at 230V would result in a fault current of 115A. You have calculated the resistances that result in the 130A limit fault current for differing input voltages.

Bear in mind that the line impedance at a mains socket outlet is typically around 0.5 ohm with a possible resultant fault current in excess of 500A, which is why HRC fuses are normally used for fault current protection in primary circuits.
 

Thread Starter

Wookbert

Joined Sep 24, 2020
10
@Hymie What I don't understand regarding this whole over-current/breaking capacity thing: Doesn't the 230V/16A main fuse in the fuse box, which covers the wall socket kick in at one point?

As for the filtering caps: I'm combining now a 22nF ceramic cap with a 10µF electrolytic cap as per your suggestion. Does the order of these two matter? In my layout the 22nF comes first, followed by the 10µF:

1601599126905.png
 

Hymie

Joined Mar 30, 2018
1,277
The consumer unit/fuse box will have protection for a high fault current – this will be to protect of the wiring installation. In the UK each ring circuit will have a fuse rating of 30A/32A which will pass considerably more than its rating before breaking/operating. So a reduced fuse rating is required within individual products.

The same high fault current will flow through both the consumer unit fuse/breaker and the product fuse, which is why both need to be capable of withstanding the peak fault current. With the lower rated fuse within the product, this should operate before the consumer unit protection.

Theoretically it should not matter which capacitor is closest to the PSU output as they are electrically connected at the same point. But since there will be some minimal resistance between the components, the 22nF should be closest to minimise any high frequency current travelling over a long pcb track.
 

Thread Starter

Wookbert

Joined Sep 24, 2020
10
Thank you very much for your patience and support. This is what the final design looks like, currently in production at JLCPCB.

1601667241645.png

The F2 board cutout is for the thermal fuse which sits on the top side of the PCB, underneath the PSU in will make thermal contact through a drop of silicone (the cutout allows it to sink into the PCB, so the enclosure of the cast-in PSU should theoretically still sit flush on the board. Can't wait for the boards to arrive, especially when recalling my attempts to etch PCBs myself in the mid-80s.

So, thanks again.
 

Hymie

Joined Mar 30, 2018
1,277
A further observation on your project is that the distance between the primary and secondary pcb tracks should be at least 5mm to provide reinforced isolation – ensuring that there is no electrical hazard in accessing/touching the low voltage circuit.

With no dimensions on the pcb, I do not know if the above important requirement is met.
 

Thread Starter

Wookbert

Joined Sep 24, 2020
10
The PCB outline is 52 mm in diameter. I really thought I paid attention to the distances and accordingly added board cut-outs, but only with creepage in mind. There’s such a cut-out between the F2 thermal fuse and the Wemos D1:
1601727217193.png

I just realized after your comment however that the cut-out only covers the 230V solder pad and I idiotically forgot the traces. Revised version looks like this (attempting a last minute change at JLCPCB):

1601727688671.png
 

Thread Starter

Wookbert

Joined Sep 24, 2020
10
The whole board will sit inside a 3d-printed enclosure (not made from PLA of course). To prevent accidental touch e.g. during repair, I've already added a silk-screened warning at the bottom side:
1601730153383.png
 

Hymie

Joined Mar 30, 2018
1,277
Provided the pcb slot is at least 1mm wide, then the requirement is for a minimum clearance of 3mm separation (primary to secondary).

Prior to the fuse, the separation between Live & Neutral should meet functional insulation (minimum creepage 2.5mm, minimum clearance 1.5mm); post the fuse there is no requirement to meet functional insulation since with the insulation short circuit the equipment will fail safe with the operation of the fuse.
 

Thread Starter

Wookbert

Joined Sep 24, 2020
10
The closest the primary gets to secondary is the shown spot between F2 and the Wemos D1, which is 2.3 mm, separated by the 1.4 mm wide slot (width of all slots). The distance of the Live & Neutral AC pins on P1 and the Hi-Link PSU are beyond my control, so I've added another 1.4 mm slot.

I know it's all very close, but I had to squeeze the components onto that 52 mm PCB, so it fits the printed enclosure, which itself has to fit into the wall box.
 

Hymie

Joined Mar 30, 2018
1,277
If separation between Live & Neutral at P1 does not meet the minimum requirements – then it is not being used in correct application; you should select an alternate meeting separation requirements.

Remember if adding a pcb slot at P1 to achieve the required separation, the mounting of the connector housing on the pcb surface may defeat this measure.
 
Top