Re: SCR based coil

To: tesla-at-pupman-dot-com
Subject: Re: SCR based coil
From: "Malcolm Watts" <MALCOLM-at-directorate.wnp.ac.nz> (by way of Terry Fritz <twf-at-verinet-dot-com>)
Date: Sun, 31 Jan 1999 20:59:55 -0700
Approved: twf-at-verinet-dot-com
I really have to jump in and say something here:

> Original Poster: "Jim Lux" <jimlux-at-jpl.nasa.gov> 
> 
> Other problems would be
> The forward gain (hFE) of power transistors is pretty low (A 2n3055 is only
> 15, although, that probably isn't a great choice in this application (Vceo
> of 60V)). Driving it hard enough to keep it switched on, particularly as
> the current comes up, would be quite a challenge.

That gain is only valid for certain values of collector/emitter 
current. Gain usually falls off as the current approaches the device 
maximum. Check the characteristic curves out. 

> Bipolar transistors (like the horizontal output ones, for instance), don't
> switch all that quickly. So they do dissipate quite a bit of heat in HF
> switching service. Say the switch time is on the order of 1 uSec (typical
> for a big power transistor). That is 10% of the cycle time for a 100 kHz
> carrier (and, you have to switch twice). So, for 20% of your cycle, you
> will be dissipating a fair amount of energy.

Not only that, but you have to derate the device as die temperature 
rises beyond 25C. 

> BTW: The forward voltage drop between collector and emitter is less than
> the base-emitter drop. A typical Vce on might be 0.1 V, which for a 15A Ic
> would only be 1.5W.

Not necessarily true. Vce(sat) climbs as collector current does. It 
can easily exceed a volt near the maximum device current.

> The inappropriateness of the BPT for this application is why motor inverter
> manufacturers and switching power supply manufacturers (both of who make
> lots of Class D amplifiers using switching devices) have almost entirely
> gone to power FET's or IGBT's (more the latter for high power.) As a result
> the price on these devices has dropped dramatically. Check out the web site
> for firms like International Rectifier for info on these devices
> 
> ----------
> > From: Tesla List <tesla-at-pupman-dot-com>
> > To: tesla-at-pupman-dot-com
> > Subject: Re: SCR based coil
> > Date: Friday, January 29, 1999 12:04 PM
> > 
> > Original Poster: "Steve Young" <youngs-at-konnections-dot-com> 
> > 
> > Reinhard,
> > 
> > I'm no expert on transistor power dissapation, but I am quite sure the
> > transistor will only dissapate approximately the collector-emitter
> voltage
> > DROP times the current.  Thus, if one uses a Class D circuit (essentially
> > either completely on or completely off), then the dissapation will only
> be
> > a couple of volts times the current during the "on" state.  More power
> will
> > be dissapated during the times the transistor is transitioning from one
> > state to another, but this time could be kept as small as possible (fast
> > rise/decay times on the rectangular pulse drive).
> > 
> > --Steve

At 15A collector current, that amounts to 30W e.g.

> > ----------
> > > From: Tesla List <tesla-at-pupman-dot-com>
> > > To: tesla-at-pupman-dot-com
> > > Subject: Re: SCR based coil
> > > Date: Wednesday, January 27, 1999 10:38 PM
> > > 
> > > Original Poster: RWB355-at-aol-dot-com 
> > > 
> > > Hello Nick, 
> > >  
> > > You wrote:
> > > "I wondered if anyone has thought about using big bipolars for the
> > >  drive. The transistors used for scan coil drive circuits in TV are
> rated
> > a
> > > 15A
> > >  at 450V and cost u0.89 each.  They are therefore perfect for use in
> big
> > banks
> > >  to drive a tesla."
> > > SNIP
> > > 
> > > The transistor ratings you mention above are OR values not AND values.
> In
> > > other words, the transistors are able to handle either a MAXIMUM of 15A
> > or a
> > > MAXIMUM of 450V, but not both at the same time. This depends on the
> power
> > > disapation rating of your transistor. For example if it has a 150W
> > rating,
> > > then you could switch 15A (max), but only at 10V, etc. 
> > > 
> > > For a transistor that could switch 450V AND 15A at the same time, it
> > would
> > > have to be able to handle 6750 watts. That would be a VERY BIG
> > transistor.
> > > 
> > > Plus transistors are slow devices if you compare them to a SCR, etc. A
> > > transistor acts like a variable resistor, whereas a SCT acts more like
> a
> > > switch. Using a big trigger pulse on an SCR will turn on the complete
> die
> > > (semiconductor surface). This fact plus the fact that an SCR acts more
> > like a
> > > switch lets these devices handle very high ratings, eventhough the
> SCRs,
> > etc
> > > are not very large (size-wise).
> > > 
> > > coiler greets from germany,
> > > Reinhard

Unfortunately, SCRs have di/dt limitations to prevent hot spotting of 
the die. I have some Westcode devices rated at 2kA+ and 2.5kV. Even 
with rather impressive looking di/dt ratings, these devices could 
only replace the gap in a coil running hundreds of amps in the 
primary at very low frequencies in the 10 - 20kHz region. 

Malcolm