[Vintage-Audio] Tube amps with no load

[Robert Nickels]

I found a description on the web at
http://www.nmr.mgh.harvard.edu/~reese/electrolytics/ that was good enough to
warrant sharing.   I use the same procedure only I have found it unnecessary
to replace tube rectifiers with solid-state plug-ins as suggested toward the
end of the following commentary.  Simply monitor the B+ voltage as you bring
the variac up from zero and  you will see that the common "5R4-type"
rectifiers do rectift DC even with heater voltages well below the nominal 5
volt level.   There is no difference to the capacitor in how reforming is
done, most of the time I just use my Sprague TO-6 cap checker to reform caps
in-circuit and out.

Go forth and reform!

73, Bob W9RAN


Reforming
The thin layer of aluminum oxide formed to insulate the capacitor foil
constitutes formation. Capacitor manufacturers use proprietary mixes of
chemicals and DC electricity to create this insulating layer, which
deteriorates with time and idleness. Often the oxide layer is in such bad
shape in older equipment that it must be reformed or else the capacitor will
fail catastrophically. All methods of reforming use the slow reapplication
of DC electricity to restore the oxide layer to its original thickness and
uniformity. In my opinion, there's no one proven way to reform - many
different approaches are available, but all have one element in common -
slowness. The reforming must proceed faster than the buildup of heat due to
the low resistance of the faulty oxide layer - this will at least take
hours, and can take days.

Current-Limited Method (from Angela Instruments): Here's a link to Angela
Instruments's http://test.angela.com/catalog/how-to/about_caps.html
instructions for reforming old electrolytics out of their chassis using an
external power supply. This method uses a large series resistor and a
high-voltage power supply to reform capacitors that are NOS (new-old stock)
or capacitors removed from the equipment's chassis.

Voltage-Limited Method 1: The voltage-limited methods use a handy device
called the variable autotransformer (a.k.a. Variac, General Radio's brand
name). Using a high-voltage external power supply, each capacitor is slowly
brought up to working voltage by slowly raising the line voltage to the
power supply. This can also be done with a variable DC supply with a range
from about 50V to 500V, but variacs are cheaper and more common. A resistor
can be placed in series to monitor the current, but watching the voltage
also can reveal reforming progress; at each variac setting, the voltage will
slowly rise until reforming at that voltage is complete.

It's easy to make a supply for this purpose from junk-box parts; the circuit
is a pair of 500mA 24V transformers connected secondary to secondary,
followed by a voltage tripling circuit. Total cost was about $10 (really),
including the box from the local Radio Shack. Being a voltage tripler,
regulation is weak and the voltage drops a lot as current increases. I've
exploited this characteristic to give a rough estimate of the current drain,
as shown in the chart on top of the supply. (The values were measured using
a rheostat and my DMM - a supply using a different collection of parts would
have similar behavior, but would measure differently). Typically I would
connect my supply across the electrolytics to be reformed, along with my DMM
set to its highest voltage setting. I plug my supply into the variac (turned
off, set to zero), turn the variac on and slowly increase to the 30 volt
setting. If the voltage reading on the DMM does not rise, or rises to less
than 95 volts, there's likely a short. If the voltage rises, the voltage
indicates the current drawn by the supply. As the capacitor starts to
reform, the leakage current will decrease and the voltage will continue to
rise. Once the leakage has decreased to an acceptable level, I go stepwise
upward with the variac setting until the operating voltage for the capacitor
is reached.

In the equipment's chassis, often capacitors of different voltage ratings
are connected by voltage-dropping resistors, and the equipment uses the
current demands of the circuit to keep voltages in operating range. You
could disconnect each capacitor from the circuit and reform individually, or
perhaps follow method 2.

Voltage-Limited Method 2: Using a two stage method, we can use the load of
the circuit to keep the voltages in all the circuit's power supply
capacitors within operating range. This is the method that I usually use,
and can be carried out by using the equipment's own power supply. Look at
the circuit and note the lowest voltage rating of all the capacitors that
connect to the high voltage (B+) supply. Remove the tubes from the chassis
and, using a variac, reform the power supply capacitors to this lowest
voltage. Now put the tubes in the chassis and raise the
highest-voltage-operating capacitor to this minimum voltage. This typically
gives about 60% of the B+ and enough of the filament voltage to provide a
load. Raise the line voltage slowly (using your variac) to reform the
resistor-connected power supply capacitors each to its own working voltage
(or slightly above).

This method has some more risk compared to reforming out of the chassis -
you'll need to watch the total current draw and raise the voltage more
slowly, since you have less information about the condition of the
individual capacitors. Remember that it's quite likely that all of the
connected capacitors except one will reform, but that one bad section will
draw lots of current. You cannot assume that, if the acceptable leakage for
one electrolytic is 1 mA, then it's ok for 4 electrolytics connected
together to have leakage of about 4 mA - your group of 4 electrolytics must
have a combined leakage less than that allowed for a single electrolytic
otherwise you've allowed the possibility of 3 of good quality and 1 clunker.
 If the equipment has a vacuum tube rectifier, you must jumper it with some
silicon diodes for this method to work. It's really easy though - remove the
rectifier and use some clip leads and a couple of 1N4007s as shown in this
picture.

Some final cautions:
Excess Current: you must keep a close eye on either the rate at which the
voltage is rising, or you must measure the current directly while reforming.
Either unsolder the connection between the rectifier and the capacitor and
insert a current meter, or insert a resistor (while measuring the voltage
across the resistor and calculating the current), or use the voltage drop
across a resistor already correctly placed in the circuit to follow the
current.

Vacuum Tube Rectifiers: These get their filament voltage from the same power
transformer as the B+ supply. Thus, at the low initial voltages that you'd
like to start the reforming at, they don't conduct. Observing the proper
polarity, temporarily replace them with silicon diodes using an old tube
base (with diodes soldered in place) or with diodes connected by clip leads.

Over-Fusing: To protect the power transformer while reforming, replace the
usual 2 or 3 amp fuse with a very low value, such as 0.25 or 0.5 A. Your
variac will prevent the turn-on surge that normally would open this sized
fuse.

Over Voltage for the Capacitors: Be careful of the operating voltage when
the tubes are removed from the chassis; without a load, the voltage
delivered by the B+ transformer will be much higher than the normal
operating voltage, and may exceed the capacitor's voltage rating.