[TenTec] (no subject)

[Dr. Gerald N. Johnson]

On Sun, 2007-09-23 at 10:32 -1000, Ken Brown wrote:
> Dr. Gerald N. Johnson wrote:
> >  TenTec PAs are relatively unique in being
> > able to handle a great deal of SWR without frying instantly and so a
> > simple DC current protecting is adequate whether fuse, fast circuit
> > breaker, or power overcurrent shut down. In any of these the protection
> > works in tens or 100s of milliseconds while without TenTec's circuit and
> > thermal care the lifetime of the over voltaged RF power transistor may
> > be half a cycle at 30 MHz, something like 17 nanoseconds.
> >   
> I'm trying to understand this. What part of Ten-Tec's circuit keeps the 
> PA transistors from frying in a half cycle at 30 MHz? Looking at my Omni 
> VI PA schematic C25 a 33ufd capacitor on the PA board would hold enough 
> charge to supply the PA board for well over 17 nanoseconds, by my 
> calculations.

Only an RF over voltage large enough to put the transistor into
avalanche mode will kill the transistor that fast.
> 
> PA system load on the 12 VDC supply is roughly equal to 0.5 ohm if we 
> use the rough figures of 12 volts and 24 amperes. Time constant of 33 
> ufd and 0.5 ohms is about 16 microseconds. Or if in a high SWR condition 
> if the effective DC load resistance of the PA goes down to 0.05 ohms 
> there is still a full microsecond of charge available in that capacitor. 
> So even if the power supply current limiting, whatever method is used, 
> could shut down the supply in a nanosecond, there is enough charge in 
> C25 to keep the PA transistors running for over 17 nanoseconds. It looks 
> to me like other electrolytics in the bias circuitry would also keep 
> things going for longer than 17 nanoseconds, even if some fast shutdown 
> of the bias is attempted.

We really want biases to stay fixed for longer than syllabic rates,
probably with a few second time constants. Else we get distortion of the
SSB or CW signal.
> 
> Is it just beefy PA transistors that keeps them from frying in a half 
> cycle at 30 MHz?

I think that's the case. Can't locate my Corsair II manual today to
check on the part numbers. But at some times, transistor makers have
rated transistors to survive medium and large load SWRs. Clearly when
the load perceived a the transistor collector goes down in impedance the
transistor current can rise and so can the heat so having more than the
minimal heat sink size is important. And having a substantial transistor
case to transfer heat from the die to the heatsink is important. Those
are things possible to select at the design stage. And then another way
to get tolerance for high SWR is to use transistors rated for
significantly more power than the radio rating. That doesn't mean we can
pour the drive to the transistor and get the radio to do more output
because the output matching has a fixed ratio and doing more power
requires a lower load R at the transistor collectors. But it can mean
better transistor longevity with mismatched loads.

I don't think the Corsair II has the Paragon's drive reduction when the
PA current rises. At least, I've taken out a few of my protective fuses
while tuning the antenna tuner under power.

Many rice boxes use SWR, specifically reflected power, from an internal
SWR bridge (or approximation of one) to reduce drive, generally hooked
to the ALC circuit. This can work well. I know from experience that the
Kenwood TS-120 and 130 tolerate laying a brick on the key with full
drive and then tuning the antenna tuner for maximum RF output. They
never draw excessive current because of their reflected power back off
on drive. But Corsair II vintage TenTecs don't work from the RF for
protection. They depend on power supply current limiting and a sturdy PA
design.
> 
> DE N6KB
-- 
73, Jerry, K0CQ,
All content copyright Dr. Gerald N. Johnson, electrical engineer