[Amps] Amps Digest, Vol 93, Issue 11

[Marius Hauki]

Good idea to use diodes instead of drop resistor Bill. Thank you for
helping out! I would think that the diodes should be able to withstand
very large surges, as you mention in case of a glitch. If the back to
back diode combination should become open circuit (in case of a very
large glitch) and someone (not me) then removes Bminus before Bplus,
there is no return current via the chassis path to Bminus. This would
place the RF deck at Bpluss. This also would apply to a defective grid
drop resistor of course. However I am not sure if a resistor is more
likely to survive a glitch than a 6A10. I would tend to trust the
resistor, but thats just me.

>From all advice here in the last mails and a small FMEA analysis I
have carried out, here are my list of safety measures  (please review,
 comment and expand on it):

-Use a connector that cannot disconnect Bminus before Bplus

-Isolate B- from chassis ground if a SHV connector is used

-Use HV rated cable (where do I get hold of proper specced HV cable
that will match the SHV connectors. I dont trust RG8 coax, or should
I?)

-Make sure there is an extra PSU to RF deck chassis ground strip

-Use either a large wattage low resistance resistor or a set of back
to back 6A10s from Bminus to chassis ground in the HV psu chassis

-Never run meters on the "hot" side of the tube due to risk of arcing
from the meter to the operator

-Run a separate spun metal "sock" outside the HV cable and ground that
metal shield in the HV PSU chassis so that in case the cable
insulation breaks down and Bplus arcs over, the path of lowest
restance is via the sock back to the HV psu and not via any other path
(for example the operator).

-Design in a safety interlock so that the PSU cannot be engaged w/o
the controller in the RF deck commands it to engage. Run an extra
chassis ground in this interlock cable for safety.

-Design in a RF deck safety interlock so that in case the RF deck
cover is opened, the PSU is cut immediately (and possibly the HV is
cut via a vacuum relay to avoid waiting for the cap bank to
discharge). Any vacuum relay must not disconnect the bleeding
resistors over the cap bank.

-Make sure no holes in the PSU chassis or RF deck may expose any
humans to HV in case something is inserted accidentically in the holes

-Use MOVs over the primary of the HV trafo to prevent spikes arriving
from the mains causing overvoltage and exceeding the cap banks max
voltage, damaging the tube or causing arcs.

-Use MOVs from L and N side of the primary to chassis ground to
prevent any dangerous voltages to be present on the L and N lines in
case both the diode recifier bank fails (short) and the HV trafo
insulation between secondary and primary fails at the same time. If
this happens B+ may be present on the L or N lines and the insulation
in the mains wiring in your house may have insulation enough to tackle
that for some time. This may cause dangerous overvoltages to be
present on the mains system damaging sensitive equipment (e.g. your
spouses TV) eventually causing an arc to the ground connector
somewhere on the mains system in your house when the insulation breaks
down. Note: In norway the Neutral (N) wire is not connected to earth
on the supply side of the main breaker at the house (as in the USA)
since we run an isolated terra system. Both L and N are routed to the
main distribution transformer here and the center tap of the main
distribution transformer is connected to earth via a disneuter (spark
gap). Earth current return happens via capacitive coupling.

73
Marius
LB3HC





>
> REPLY:
>
> You are correct to leave the B- lead isolated from chassis, but you
> don't need the high wattage resistor you mentioned. Instead, you
> should connect a pair of diodes back-to-back from the B- lead to
> chassis ground inside the PS chassis and another pair across the grid
> meter inside the amp chassis as described below.
>
> Use high surge current rated diodes such as the 6A10 or equivalent.
> With the diodes present, you do not need the high wattage resistor you
> mentioned above. For GS-35b I would use a 500 mA meter. You should
> protect the meter with a low value resistor in series with the meter.
> The value of the resistor should be chosen so the diodes begin to
> conduct when the drop across the meter + resistor equals about .7
> volts.  This combination of series resistor and the back-to-back
> diodes mentioned above will protect the grid meter against even
> massive arcs. The negative side of the grid current meter is connected
> to ground and the positive side is connected to the positive side of
> the cathode current meter, below.
>
> The cathode current meter is placed in the B- lead and should also
> have a series resistor on it's negative side and protective diodes
> back-to-back across the combination of meter and series resistor. The
> value of the series resistor is chosen by the same means as above for
> the grid current meter.
>
> If any of this is not clear, please ask questions. Good luck and keep
> us posted on your project.
>
> Bill, W6WRT
>