Dmills Posted May 18, 2006 Share Posted May 18, 2006 Ok. just to clarify my position here. I'm not saying that I agree with ETC's suggestions, I'm mealy putting it up for discussion... One of the points they made was that the size of a 2.5Kw sine wave dimmer is about 4 times that of a 1.5Kw one and significantly more expensive.I think what they are saying is that to make sine wave dimming a feasible alternative to thyristor dimming they need to consider 1.5K for the majority of circuits with 2.5K and above only where necessary. Or wait a year or so for better FETs! It seems likely that the heat load generated by a sine wave dimmer is a linear function of the power being supplied to the load (unlike a thyristor rack where is is close to a linear function of the **current**). Thus going to a lower per channel current rating may not facilitate more channels as the limit may be the ability to dissipate Rds(on) * Current^2 per channel in the available space. Now there is a question buried in there, if you parallel two identical FETs you halve the current each handles and I don't think that even fast HV fets are a major part of the cost of a sinewave rack. Sure they are more expensive then a thyristor pair or IGBT, but given the ready availability of 500W power factor corrected PC power supplies for <25 quid retail, even a bank of half a dozen fets of the type used in the PFC network (plus drivers) cannot be a major cost. Actually, you could probably finesse this further and use a PFC controller chip to implement most of the dimmer control logic! I think I will wait for the better high speed switching devices before going for sine wave, it mostly is not a compelling win for me with this tradeoff. What is interesting in all this is that apparently 1.2Kw @120V is possible, but 2.4Kw at 240V is a problem (10A in both cases). Now it might be that finding fast, low RDS(on) FETs with the appropriate Vds rating is a real problem, indeed a quick check with international rectifier reveals lots of suitable options in the Vds ~ 500V region (OK for 1110V line) with far less up at the 800 ~ 1Kv rating required for 230V service (and those have rather high Cgd values making gate drive a problem at more then a few Khz (Due to the Miller effect)). It looks to me like waiting for better switches is the answer! I agree that pairing channels is a necessity at the moment for many situations but with 4x1.5K dimmers in the space of 1x2.5k would you still need to pair up.Discuss... :) Without some way to easily interchange between the 4 * 1.5 and the 2 * 2.5 modules this wouldn't fly for me and if you can interchange then the modules need to mount a power patch system locally.... Just my thoughts. Regards, Dan. Link to comment Share on other sites More sharing options...
OB2 Posted May 19, 2006 Share Posted May 19, 2006 one point also is the fact that unless this would eb a new install, either refurb or new theatre most companies dont have the money to just replace thier dimmers, I know we definatly dont!!! Link to comment Share on other sites More sharing options...
Tomo Posted May 21, 2006 Share Posted May 21, 2006 Or wait a year or so for better FETs!MOSFETs aren't really the issue here - MOSFET technology has been pretty stable for several years, and isn't likely to change much in the near future.I've built a couple of fairly high power (36V, 20A) SELV DC dimmers as proof-of-concept prototypes, and it would be a simple problem to upgrade them to around 100A current handling - essentially a bigger heatsink is the only difference, as gate drivers are cheap and pretty reliable. However, as I mentioned elsewhere, MOSFETs generally aren't used for this purpose. A MOSFET is primarily a DC control device - there is a possible design whereby the incoming 3-phase is rectified to twin DC rails using existing SCR/thyristor technology, which are then chopped via MOSFET H-bridges into a variable sinewave (big inductors are needed, but all good dimmers have them anyway)MOSFETs that are good for 20kHz switching have been around for quite a while. Sure they are more expensive then a thyristor pair or IGBT, but given the ready availability of 500W power factor corrected PC power supplies for <25 quid retail, even a bank of half a dozen fets of the type used in the PFC network (plus drivers) cannot be a major cost.Neither silicon is particularly expensive.However, IGBTs are much newer technology than MOSFETs and are consequently less reliable and available - and IGBTs are the ones that are generally used in these dimmers as it's less silicon than doing a rectification followed by regeneration. Actually, you could probably finesse this further and use a PFC controller chip to implement most of the dimmer control logic!The control logic is the big issue at present. ETC have some good sine-wave dimmers, but nobody else does (yet), primarily because of the difficulty of the control software. The silicon is here already, but the control is very, very difficult. Link to comment Share on other sites More sharing options...
Brian Posted May 21, 2006 Share Posted May 21, 2006 The silicon is here already, but the control is very, very difficult.And if it's not difficult it's covered by someone patent. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.