I'm looking at the schematic for the HP 3048A option K23 DC blocking
filter. I've attached a simplified version of this schematic for
convenience. Note that I've omitted the circuitry related to reverse
polarity (switch and diode protection). But, I don't need that to
address my question.
The manual states that the 51.1ohm R1 acts as a fuse protecting the DC
blocking filter (C3-16) from a large input DC voltage (K23 is rated for
+/- 30 V). C3-16 are rated for 50 VDC. How does this fuse action work
and how can we ensure it's reliable?
The description of R1 in the manual is:
RESISTOR 51.1 1% .05W TF TC=0+-100
I believe TF stands for thin film. So, the power rating of the resistor
is pretty low. Maybe that gives a hint. There's no DC current path
through R1. But, the startup ramp of a DC voltage is AC. Broadly, it
seems like we need to identify the scenario(s) in which there is a
voltage across R1 creating a short-duration power dissipation that
greatly exceeds R1's power rating. Because there's no DC current path,
we need to rely on a large, short-duration excess of power rather than a
steady excess of power to blow R1.
Current can either flow through C1 or L1 out of R1. C1 will present a
low impedance, but only at high frequencies (|Z_R1|=|Z_C1| at
663MHz). L1 will slow the voltage rise across C3-16. Presumably, the
goal is for R1 to blow open before the voltage across C3-16 exceeds the
voltage rating. If I make the initial ramp rate very fast I can greatly
exceed the 1.6V R1 rating, but only for a very short time. How can I be
sure to destroy the resistor? And, if I destroy it, how can I be sure it
will blow open? The NASA reliability design handbook (MIL-HDBK-338B)
gives the failure mode distribution for film resistors as (p. 7-198):
| failure mode | probability |
| <l> | <c> |
|------------------+-------------|
| open | 0.59 |
| parameter change | 0.36 |
| short | 0.05 |
Wouldn't it be easy to bypass the protection and exceed the capacitor's
voltage rating by supplying a ramp that's just slow enough that C1 looks
like a high-impedance?
There's a limiter across the output, but no proper fuse to blow if one
of these diodes starts to conduct. It seems like the 11848 or 3561 input
is pretty exposed in this case.
Thoughts? Maybe R1 isn't just a normal resistor?
Thanks
Matt
