| Electronics > Projects, Designs, and Technical Stuff |
| 'HV' switches voltage ratings: AC vs. DC |
| (1/3) > >> |
| TinkeringSteve:
For things like rotary 1->N switches, flip switches, push switches, or relays - you name it: Their voltage ratings usually look like this: AC: 250V DC: 30V While looking for switches to do the mapping between a vacuum tube measurement circuit (400V) and the tube socket(s) for tubes of arbitrary pinout, I noticed that there are no switches to be found which are rated for anywhere near such a high *DC* voltage, which would be needed here. Now I'm not wondering anymore why some people use such a dangerous setup with a lot of banana jacks & patching cables for that... Why is that? Because (I guess) AC makes arcing harder due to the direction changes? *But* there are a lot of old technicians' tube testers to be seen in YT videos where roughly ordinary looking (if somewhat clunky) rotary switches are fit. Were those actually special switches that were only made back in the age of tubes? I am wondering because I am also wondering if the voltage a switch could be rated for would change if you lift the requirement of it switching *under load*, with stuff actually moving while the power is on. E.g., say there is one main switch that turns the 400V on or off, and those pinout-mapping switches would never be moved while it's on. |
| Neilm:
Your guess is correct IMO. The contact opening will draw an arc. An AC system that arc will only have certain characteristics and will extinguish in a resonably known time. A DC system the arc could be maintained and not extinguish leading to a fire risk. Generally, the above means the contacts on the switch remain very close. Older testers have switches where the contacts move far enough apart for this not to be a hazard. Also, the required safety standards were less then. |
| Ian.M:
An AC arc goes through zero current 100 (or 120) times a second. That makes it *MUCH* easier to quench, hence the higher AC voltage rating. However for a DC rating, the switch manufacturers must allow for the possibility of an inductive load, and as you can arc weld with as little as 24V if you've got a big enough series ballast inductor, its very hard to find compact switches rated for >30V DC. Back in the day, the rotary switch wafers were typically around double the diameter of modern ones, so there could be far greater contact separation, and also they typically had very stiff ball detents so could guarantee rapid contact opening. Probably the best option would be to use appropriately rated relays, and if you aren't going full retro, using a MCU with a LCD display for valve pinout selection. If you are going full retro, consider 'personality modules' for each valve type, built on a D connector, which if you use the first and last pins to interrupt power to a relay switching on everything else, should be very resistant to abuse including hot-plugging. |
| TinkeringSteve:
Thanks for the answers. I have no need for full retro, although my father, who'd like a tube tester also, is kinda allergic to all that computerized newfangled jazz, lol. Myself I'm aware of e.g. the uTracer project by that Dutch professor, and its clever PSU scheme of "load a fat cap from a minimalist switching PSU and then discharge it during the measurement for the anode current, compensate for voltage drop in calculation". So I myself would want to trace curves, not just get "replace / not replace" kinda readings. His example materialization of the project uses... a banana plug patch board, though :D I would like to avoid that. I was already planning to make tube selection at least supported by MCU + display. I imagined that if relays are too expensive, I'd use rotary switches with double contacts, where the 2nd contact route has a low voltage for the MCU to track switch positions and not release the HV when, for a selected tube type, any switch position is off. The correct switch positions would be shown on the display. Alas so far I've not seen relays with high DC rating either. At some point I wasn't looking further because of the price, with regards to the number of relays needed. (I don't know what can be done / what's out there, but I made pencil scribblings where I'd have C->A/B selecting relays with 1 set of contacts, 2 sets, and 4 sets, to make a 1:8 mux switch out of a minimum of 3 relays, or 4 pcs if at max 2 sets of A/B contacts are available. For 9 such 1:N switches, that's 27 or 36 relays, quite a bunch.) Maybe I should go on a longer term hunt for broken old gear to salvage parts from. That "mapping function as pluggable module per tube type" idea sounds neat. For me, I only need a very small number of tube types, so far, so that might be viable. My father, IDK. I have no idea how many different pinouts there are. I shall ask. (I think he's also into almost-unobtainium early days stuff...) Wild side note: motor-driven rotary switches would be funky, as a bunch of 1:N switches need far fewer units than relays. Or even rotary switches which *are* e.g. a reduced-to-that-task stepper motor (8 coils 1 magnet?) - but I guess without proper knowledge about contact materials etc., one better does not try to 3D-print some base for a switch and bolts contacts to it, to then operate it at 400+V... Probably not even legal to operate, or at least insurance wise rather unwise, eh? |
| TinkeringSteve:
Oh! One thing that got neglected and I'd almost forgotten about is this aspect: So if the much lower DC rating of switches, relays is due to it being more difficult to extinguish arcing: How about not creating arcs in the first place? I.e. making sure (as best possible) that the switches will not be actuated when power is ON? Remember, this is to form a mapping function from measurement circuit pinout to arbitrary tube pinout. Not for power switching per se. Could I then use a higher voltage than the DC rating, which is meant with arcing in mind? If relays are use, I guess I could also make a trigger-happy protection circuit which looks on the coil voltages of the relays - and, if during power is ON, any relay state changes, the main pwer is immediately shut OFF. Due to looking at the coil voltages, maybe this is even faster than anny mechanical part actually moving? So even then an arc would not be created, or almost. With rotary manual switches, I guess one would need them with doubled contacts, to use one contact path with a low voltage, as a position indicator, and shut power off when a change is detected that way - although slower than in the other scenario. Or maybe get a bit adventurous and fasten a plastic disc to the axis of the switch, and connect something there that will detect any movement - there are several way that come to mind. (optical mouse sensor IC? probably not adding much to the price of a fat rotary switch) |
| Navigation |
| Message Index |
| Next page |