[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [TCML] OLTC midi problem

[embedded thoughts:]

> A small problem that I would like to present to more experienced
> builders...

> I have been building a small OLTC that is still in testing phase.
> The coil produces about 2.5 feet long arcs and runs great with the
> basic PWM controller I have made to give the 0-500bps.

> IGBT is a 1200v 600A, and is driven by a simple 30A driver IC and
> circuit, with isolated PSU, and opto coupler isolates between the
> driver and the PWM control box.

> PWM control box is metal, negative to case, powered by battery and
> connected via a coax cable with the shield connected to mains supply
> earth at the driver end.
> So with this set up, the coil runs great...
> I then bought a basic midi interface from EVR, and assembled, tested
> etc the board and set pulse output to suit the needs of my setup.
> (also running off battery...)

> Connected in place of my manual controller and runs great
> up to about 75% power input to the coil.

> Even a fraction above the 75% mark and suddenly, nothing but the
> occasional splutter...

> Moved controller further away... still faults at the same 75% mark.

> Scope across output of midi controller is clean and runs continuously
> even when spluttering, so I ruled out the controller.

> So I put scope across the Gate of the IGBT... and what do you know,
> the problem went away!

> Next I removed the scope and connected the Emitter side of gate to
> ground via a 1000pf capacitor (i can't ground it directly due to
> circulating currents) and it also solved the problem, running
> up to full power no problem...

> So my question is, what was the real cause of the problem?
> Any ideas?
    It sounds like a classic case of RFI getting into
    the gate.  The RFI is bypassed by the stray capacitance
    of the scope probe, or by the explicit capacitor.
    (I have vague memories of typical scope probe
     capacitance being 15 pF?  check probe specs.)

    Why does this show up at '75.111%'?  Speculation:
    somewhere a stray spark starts?  General rise
    in RFI as output rises exceeds some threshhold
    in the controller?

    I recall, tho it may not be applicable that having
    'some' resistance in series with gate to reduce
    sensitivity to stray RFI.

    Suggest seeing what other advice appears.

Tesla mailing list