The problem with the half-wave assumption is that even with a three prong plug, you're not *guaranteed *to connect your ground to neutral. I am not sure about UK specifically, but here in Poland it's only a guideline to have phase in specific hole in the outlet, and the double outlets sold in the stores simply make it impossible to implement in practice.
I think it's safer to assume half wave rectification doesn't make the circuit any safer and treat it as hot at all times. The increased safety will show up only when someone ignores the warnings. czwartek, 16 kwietnia 2026 o 07:55:30 UTC+2 gregebert napisał(a): > It's fine, as long as you know what you're doing. I've made several "hot > chassis" clocks before, with no problems. That said, you MUST be very > careful about what is exposed, so that nobody can get shocked, and be sure > to put warning labels on it. Make sure you have properly-sized fuses (ie, > smallest possible). I place small-ish series resistors in the power supply > to limit the surge current at power-up, and size them to double as fuses if > for some reason the actual fuses dont blow (1 one neutral, 1 on hot). It's > best to run SPICE simulations on your power supply to balance the inrush > current with the wasted energy (heat) of the series resistor. > > Half-wave vs full-wave rectification makes a **HUGE** difference. > Half-wave is likely to have your circuit ground connected to the neutral > wire, which is normally within a few volts of earth ground, hence minimal > shock hazard *under normal conditions. * However, things go wrong and > when your life or someone else's is at-risk, you cant take chances. > > If you used full-wave rectification, then **EVERY** point in the design is > a serious shock hazard. Basically, the (-) lead from the bridge rectifier, > which is likely to be your circuit ground, will vary from 0 to minus 310 > volts (for 220V mains), or from 0 to minus 180 volts (for 120V mains). I > have a few gizmos doing this; again you have to be very careful. > > Obviously, connecting anything like a scope will be disastrous unless you > have an isolation transformer, because the ground of your scope probe will > short-out your hot supply. > > But to answer your question, sensitive devices, like raspberry pi > computers, are fine in hot-supply systems. Since I use the onboard WiFi, I > do all of the software and logging via network, so there is never an > external physical connection. As long as everything is at the same > potential, there is no hazard. This is why utility workers can do their job > on 500kV power lines that are energized. > > On Wednesday, April 15, 2026 at 10:19:33 PM UTC-7 Richard Scales wrote: > >> ... and just checking something else: >> >> I have this rectified mains output powering the tubes. That has +V and >> GND. I have a small piece of electronics powered from a 5V supply which in >> turn is also connected to the same mains supply. >> >> Whilst it seems obvious that the GND from the control electronics needs >> to be connected to the GND of the rectified mains - everything about that >> scares me a little - am I right to be scared or is that absolutely fine? >> >> - Richard >> >> >> On Wednesday, 15 April 2026 at 16:09:01 UTC+1 Richard Scales wrote: >> >>> That is helpful - thankyou - I shall get some MPSA44 and give it a go! >>> - Richard >>> >>> >>> On Wednesday, 15 April 2026 at 16:02:52 UTC+1 gregebert wrote: >>> >>>> Your MPSA44 transistor is fine for 3 reasons >>>> 1- You are below BVceo >>>> 2- You are within the safe operating area (SOA) with 4meg of resistance >>>> that limit the max current below 1.5mA at any voltage (in your case, it's >>>> 85uA) >>>> 3- Bipolar devices, unlike MOSFETs, can actually sustain voltages above >>>> BVceo as long as your circuit limits the current. At higher voltages, >>>> there >>>> is reverse-junction breakdown resulting in current, but it's not >>>> destructive as long as the current is limited. MOSFETs, however, will have >>>> permanent oxide destruction at any current. >>>> >>>> Those IN-28 boards are cool !! >>>> >>>> >>>> >>>> On Wednesday, April 15, 2026 at 7:48:37 AM UTC-7 Richard Scales wrote: >>>> >>>>> [image: 7SegIN28.jpeg] >>>>> >>>>> On Wednesday, 15 April 2026 at 15:30:51 UTC+1 Richard Scales wrote: >>>>> >>>>>> Hello, >>>>>> >>>>>> I have a number of these '7 segment' panels which are made up of >>>>>> discreet groups of IN-28's. >>>>>> [image: 7SegIN28.jpeg] >>>>>> >>>>>> I am in the UK and am driving them with bridge rectified UK Mains >>>>>> (238V before rectification). >>>>>> >>>>>> Each 'segment' has all it's grids connected to VCC via a 3M, a 1M >>>>>> resistor is also connected so when the end of that is pulled to ground, >>>>>> the >>>>>> grid voltage is reduced to about 85V (assuming 340V peak). >>>>>> >>>>>> If I ground a segment via the 1M resistor, the tubes turn off, if I >>>>>> let it float, the tubes turn on - all good so far. >>>>>> >>>>>> [image: IN28.jpg] >>>>>> >>>>>> My question is this, what safe and reliable mechanism might I employ >>>>>> to make that switch? I was thinking of using MPSA44 NPN transistors >>>>>> which >>>>>> have a Vceo max of 400V, is it as simple as that or is there more to it? >>>>>> >>>>>> Right now I am using a hand made bridge of x 4 UF4007, I would use a >>>>>> proper bridge if this project ever gets anywhere near the finish line! >>>>>> >>>>>> Many precautions are being taken during testing to ensure that the >>>>>> rectified does not go anywhere near me (or anyone else for that matter!). >>>>>> >>>>>> - Richard >>>>>> >>>>> -- You received this message because you are subscribed to the Google Groups "neonixie-l" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion, visit https://groups.google.com/d/msgid/neonixie-l/61eb1b0a-e9a9-45a8-9d70-e66a323e0c8cn%40googlegroups.com.
