AC mains on a PCB ?

Question

A total newbie to eagle design and PCB fab. So plase bear with on my silly questions, trying to learn.

I have pcb that is schematically completed with the layout. Before i start the fabrication process i need some expert advise if the components placed and the wires routed are ok for the ac mains and the others. The load here will be a transformer. The ac mains are 240VAC and all works well as designed in the schematic on a bread broad except for the load for which MOC3023 is yet to arrive from where i've ordered.

Thanks in advance.

rachaelp · Accepted Answer

Hi George, See my earlier post with some initial findings. I've been playing with your layout including re-pinning your microcontroller as it looked like that should be fine and you can just change the pin setup in the software as the functionality all seems fairly simple. Here is an image of how the layout works after I have altered everything. I could just upload the design files back but then I think maybe you should be trying this some more yourself as a learning exercise so you can work out what you need to do to achieve this for yourself So this isn't perfect as I was just having a quick play around with it, if I were doing this for a real project of my own then there is probably quite a lot more I would change still, but within the confines of your existing circuit with some minor tweaks this is my first stab at a layout. You can see I have got the +5V and GND connected with polygons and all the other nets are routed much more cleanly as I tidied everything up in the placement first. There is much better isolation and clearances to the mains conductors are improved. I see your connectors are a little wider pitch than I had initially thought but personally I would probably swap those out for ones with a wider pitch still and I would add surge protection and EMI filtering. Best Regards, Rachael UPDATE: Moved triac slightly to reduce high current trace lengths.

autodeskguest · Answer

On 11/09/16 12:02, George Thomas wrote: Why two thrustirs to control the load and am trouble witching on yhe Triac. Triacs can suffer commutation problems with certain types of load - highly inductive (like a transformer) is one of the problem cases. The problem exhibits as a failure to turn off at the end of a half-cycle. I think this is more of an issue when PWMing the mains (light dimmer style), which is why I asked that question earlier. For that type of application a pair of back-to-back thyristors is more robust. It may not matter for your application if you occasionally get an extra half-cycle (10ms) as long as that is all you suffer. If you find that you get worse than that, the two-thyristor design is what you need. http://link.springer.com/chapter/10.1007%2F978-1-4612-9877-9_12 is worth reading on this.

michaelkellett · Answer

For mains input spike suppression I think you are much better off with this kind of device: http://uk.farnell.com/epcos/b72214s0231k101/varistor-60-0j-230vac/dp/1004389 Farnell 1004389 This one is rated at peak current of 4500A. Make sure you pick one rated and intended for the mains voltage it will experience. This kind of part has a much larger volume of material to absorb spike energy than a 'transorb' zener diode like device at a similar price point. Generally speaking you should not expect the transient suppressor to blow fuses. It is not normal to protect mains powered equipment from continuous application of excessive voltage - but it should be designed to be happy with the maximum allowed in the locale in which it is used. (I would go for nominal +20% if possible - in the UK the mains is supposed to be 230V AC within -6%, +10% tolerance). MK

D_Hersey · Answer

I'm in a hurry, late for Shoe Fest. You might want to throw a (sb) fuse in. You seem to be using a chassis ground symbol for signal ground.

autodeskguest · Answer

Ooops - looks like I accidentally hit the wrong reply button before... On 03/09/16 14:57, George Thomas wrote: A total newbie to eagle design and PCB fab. So plase bear with on my silly questions, trying to learn. I have pcb that is schematically completed with the layout. Before i start the fabrication process i need some expert advise if the components placed and the wires routed are ok for the ac mains and the others. The load here will be a transformer. The ac mains are 240VAC and all works well as designed in the schematic on a bread broad except for the load for which MOC3023 is yet to arrive from where i've ordered. You have a lot of traces running unnecessarily close to each other and to pads. That's fine to low voltage stuff but undesirable when high voltages are present. Your placement of parts results in quite a few traces being quite long. That's undesirable in all domains. I particularly note that the trace from C7 to R20/21 goes all around the houses, when a simple rotation of C7 could reduce it to nothing. I don't know how much current that BTA08 is expected to drive, but the text describing the board as a "spot welder microcontroller" hints that you might be looking at several amps. If so, your traces look rather thin to me, especially given how long they are.

autodeskguest · Answer

On 04/09/16 00:19, George Thomas wrote: Thanks a lot to take the effort to correct me. I've removed C7 and R21 which are actually used to protect the optocoupler but i don't see they being necessary here and i have rearranged the the remaining too. Definitely better... but I still think you want more "meat" on the important tracks. So, thinking about your design, it seems to me that the crucial part - the bit you need to worry about - is the path through the triac and between the connectors. That's where ALL the current is flowing - the rest is negligible. So, if it were my design, I'd start by placing those three components close together for optimal (i.e. very short) tracks. Then I'd make those tracks as wide as I could reasonably achieve. The rest can then fit around that. Also, the track from IC5 to R20 may be a low current track but it's at mains voltage. I'd rotate R20 to keep that track away from the low voltage ground plane. Would I be right in thinking you're using the triac as a pure switch? That is, you're either triggering it as early in the cycle as possible or not at all? I only mention it because I have some experience of driving a transformer in "light dimmer" mode (PWM of the mains cycle) and it's... tricky.

rachaelp · Answer

Hi George,

Your latest layout is definitely improved over earlier versions although I still think there are quite a few things that can still be done to improve things. The clearances are better but not perfect, I'd also try to ensure there are no planes encroaching underneath, IC4/5, B1. I see you still haven't changed the connectors although you had selected some 5.08mm ones in an earlier post if I recall. Why have you chosen not to go with these?

You should also think about the "what if" cases for faults which could allow dangerous mains voltages to appear on the low voltage circuitry (e.g. a fault in TR1). Do you need to provide a star point connection between 0V and mains earth? Should you use a mains isolation transformer outside of this unit? Etc....

Best Regards,

Rachael

autodeskguest · Answer

On 04/09/16 19:23, George Thomas wrote: Now about placing the components, will keep the AC components as close as possible with the widest copper. The space between the pins is short so that would limit the copper size there would that be a problem ? I mean the route to the TRIAC would be. I've rearranged the AC components closer. how are the track size for the other components looking ? The shorter those tracks are, the less it matters if they're not as wide as you'd like. Basically, the copper track acts as a (low value) resistor, where R = k.l/w (give or take). Short and fat is the lowest resistance and hence gets least hot. It looks like you've changed to a different triac in a wider (TO3-P) package, which will help with wide tracks! If it were my design, I would move the small transformer to the right a fraction and put the triac between it and the connectors. That would, I think, allow the shortest track from the "AC" connector to the middle pin of the triac. Well, other than placing the triac between the connectors and swapping the connectors so that AC is towards the top.

rachaelp · Answer

George Thomas wrote: Hi Rachel, I've started off with the routing and now how do we separate the +5v polygon from the GND. Hi George, You need to adjust the polygon ranking. Change the ranking of the GND polygon to be a lower rank (1 is highest, 6 is lowest) and then they will be separate. Best Regards, Rachael

autodeskguest · Answer

On 05/09/16 23:57, George Thomas wrote: Well, after a while, i've managed to move IC5 away from R20, i've also changed the angle for the R23 and IC5 making them more routable during Auto tracking . No, I think you misunderstood. There's no problem with IC5 being close to R20, there's a problem with the tracks connected to IC5 being close together. You absolutely don't want them passing between the pins! Rotate IC5 by 180 degrees and move it left by 0.2" or so. Then rotate R20 by 180 degrees and move it due south of R22. IC5 is an optoisolator - its whole purpose is to isolate - so make sure you maintain an isolation gap across it.

michaelkellett · Answer

It might help you with getting the separation between mains and low voltage side stuff if you make sure that there is a gap of at least 4.5 mm between mains and low voltage tracks (this is about the separation on the opto isolator package). Here's a useful guide to this kind of thing: cache.freescale.com/files/analog/doc/app_note/AN3962.pdf MK