On Fri, 6 Jan 2012 23:17:07 +1300, RussellMc wrote: :: - Uncontrolled leakage inductance between windings? Possible, although the end product is expensive ( $15 - $25K) things like=20 power supplies and inverter/ballasts are made from relatively cheap=20 components. These ballasts cost about AU$70.00 each. These ballasts are over 10 years old and no longer available new. There is= =20 only one company that provides the transformers world wide and they get=20 special production runs from a company in Korea. No manufacturers data is available. The main problem is probably ambient heat, poor tubes (they are only=20 changed when they fail) and noisy power environment. These are provided to an industry that is extremely poor and=20 where every penny counts towards their profits. Think, Las Vegas. =20 ::When they die, do the capacitors die too? Occasionally, the 33nF across the BJT's is often part of the problem - the= =20 cycle tends to be - caps go high or explode taking BJT's with them -=20 however often the BJT's die even though all caps (100uF or 220uF=20 electrolytic, 33nF primary and 3n3F or 2n7F across secondary have been=20 replaced at an earlier time). Changing the transformers doesn't always=20 solve the problem though often they develop either a shorted turns at AC=20 (DC resistance 0R1 to 0R2 is correct) or I suspect the insulation becomes=20 unwell.=20 ::Converter sounds self resonant. Yes? Yes. ::Is a "commutation failure" / stops oscillating / drops an occasional ::cycle possible No, it's just a multivibrator circuit with the R replaced with a=20 transformer ::Do both transistors fail together Varies. ::If not does the non dead one walk-funny afterwards? No, but I replace pairs at a time (there are two identical circuits per=20 PCB). =20 ::Is an occasional bad startup mode possible? Possible I suppose, the actual power supply that provides the 25vdc often=20 die themselves especially after power surges. ::Slow start? No start? funny mode start? No. ::Do you know when/where in cycle they fail? No, I really think the cause is poor quality transformers and a hot=20 environment. ::If not, can this be determined? Difficult, apart from upset caps or shorted transformer, they don't die=20 when put in the workshop test machines, but the workshop is much cooler=20 than the sites. ::Is it consistent? Not sure what you mean here. Oh, I see, well tube state is unknown but=20 could be poor to good, can die the moment they are connected on site even=20 though they worked fine in the workshop - note field techs do not switch=20 power off when changing these or tubes, only power supply problems or=20 logic/CPU boards will get the machine switched off. ::What other potentially irrelevant variables are known? (Murphy loves=20 ::these) The sites are electrically noisy, with WiFi, modems, displays, jukeboxes,=20 and irate punters attacking the equipment. This circuit is a good reference approximation of the actual inverter (note= =20 this isn't the actual circuit, just one I found on the Interent ready=20 ASCII'd) (Thank you Mr. Goldwasser of kellerstudio.de) +Vcc o T1 o Q1 +----------------+=20 | | ):: + B |/ C ):: L1 ::( +------| MJE3055T ):: C1 24T ::( | |\ E D 15T ):: +----------||---------+-+ #22 ::( | | #26 )::( .0039 uF | |=20 + | -_- )::( 600 V +|-|+ | | )::( | - | +--|-------------------------+ ::( | | | | )::( | | | | Q2 _-_ )::( | | | | | )::( O 600T | | FL1 | | B |/ E D 15T )::( #32 | | | | ----| MJE3055T #26 )::( | | | | | |\ C )::( | | | | | | )::( | | | | | +----------------+ ::( | - | | | | ::( +|-|+ | | | o ::( | | | | -----------------------+ :: +---------------------+-+ | | F 10T ):: | | #32 ):: | | +---------+ :: O =3D Output | | | F 10T ):: D =3D Drive | | | #32 ):: F =3D Feedback | +-------------------------+=20 | | | R1 | R2 +----------/\/\/\--+--/\/\/\--+ 220 22 _|_ Colin -- cdb, colin@btech-online.co.uk on 9/01/2012 =20 Web presence: www.btech-online.co.uk =20 =20 Hosted by: www.justhost.com.au =20 =20 This email is to be considered private if addressed to a named individual= =20 or Personnel Department, and public if addressed to a blog, forum or news= =20 article. =20 =20 =20 :: :: Who makes the transformers? :: Is leakage inductance known or quality of deign known? :: Are transformers off the shelf commerical items in general use or :: made :: by or for you directly? :: Is there a formal spec for the transformer and does it specifically :: address leakage inductance minimisation? :: :: When they die, do the capacitors die too? :: Is the fluro resonance cap dead o/c? s/c? walks-funny? :: :: Converter sounds self resonant. Yes? :: Is a "commutation failure" / stops oscillating / drops an :: occasional :: cycle possible? :: :: Is base drive controlled wrt voltage or energy into base? :: If so or if not does or may this come near or exceed spec? :: :: Do both transistors fail together? :: If not does the non dead one walk-funny afterwards? :: :: Is an occasional bad startup mode possible? :: Slow start? No start? funny mode start? :: :: Do you know when/where in cycle they fail? :: If not, can this be determined? :: Is it consistent? :: At startup / After hours of operation/ When hot/when cold/ At :: moment :: power is applied. / When power is turned off / Known random (as :: opposed to unknown)./ High voltage / Low voltage / New tube / Old :: tube :: / ??? / Spikes on supply/Variations in supply / ... :: :: What other potentially relevant variables are known? :: What other potentially irrelevant variables are known? (Murphy :: loves these) --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .