View Single Post
Old 19th May 2019, 10:25 pm   #1
GrimJosef
Dekatron
 
Join Date: Sep 2007
Location: Oxfordshire, UK.
Posts: 4,310
Default Simple reformer for HT capacitors in valve-rectified equipment

The forum has had quite a few descriptions of reformers for electrolytic capacitors posted in the past. The circuits generally consist of two parts - a DC power supply of appropriate voltage, and something to limit the current so the capacitor's insulating layer has time to reform before the voltage gets high enough to punch through and permanently destroy it.

I've put together a simple circuit which plugs into the rectifier socket of a piece of equipment and allows the existing mains transformer to be used as part of the DC power supply. The current limiter consists of just three series resistors, two of which can be shorted out by a rotary switch to maintain the reforming current at a reasonable value as the capacitor charges and the voltage difference between the supply and the capacitor falls.

As well as making use of the mains trasformer, connecting via the rectifier socket also allows the HT capacitors to be reformed in situ, so obviating the need to disconnect and reconnect them. Another advantage is that this reformer also carries out a simultaneous high-voltage insulation test of any other components connected to the HT rail.

The circuit diagram is shown in the first picture. The black wiring is the reformer itself, the red shows what it's connected to via the rectifier socket. It uses components I happened to have to hand, so values haven't been tightly optimised. It was designed to deliver a low reforming current - less than a milliamp - since in my experience such very gentle reforming hardly ever overcooks the capacitors. The DC supply consists of two 1N4007 rectifiers, a 1uF reservoir capacitor and a series resistor/Zener diode chain feeding a 33uF smoothing capacitor. This chain forms a potential divider with the smoother's 390k bleed resistor, the division ratio setting the maximum voltage that the downstream capacitors will be reformed to. This matters because the combination of an off-load transformer and solid-state rectifiers with low forward voltage drop would otherwise raise the reforming voltage far beyond what the HT capacitors are used to.

The Zener diodes are small, which is helpful given that I wanted to house the whole circuit in a package compact enough to fit where a valve rectifier normally goes. But some resistance is needed both to reduce the ripple current into the smoother and also to allow the division ratio to be easily adjusted. This particular circuit was set up for reforming the capacitors in a (pretty scruffy) Quad II amplifier, where the HT rail is usually in the region of 360-385V, depending on the mains voltage, the condition of the rectifier and whether we are measuring before or after the amp's smoothing choke. The resistance needed to deliver this voltage turned out to be 27k.

The second picture shows the reformer in use. It's housed in a stack of 40mm solvent-weld PVC pipe fittings with a scavenged octal valve base glued into the bottom and a screw-fit inspection cap on top. The latter serves as a mount for the rotary switch and also allows the circuitry to be inserted and removed if/when necessary. There are three 4mm sockets for metering the capacitor voltage and the reforming current (via the voltage across the 36k final charging resistor). In the picture this current is 0.23mA (8.24V/36k) with the Quad II's 16uF reservoir and smoothing capacitors charged to 253V. You can see the reformer's ground connection made via a trailing lead into one of the speaker sockets.

I chose valve era resistors (mostly Piher 1W ones) since these are rated for typical HT voltages. The exceptions were the two 18k current limiting/sensor resistors which are 5W wirewounds. In the event of an HT rail short to ground at maximum voltage, when the rotary switch would probably have taken all the 220k resistors out of circuit, I wanted the 18ks to last long enough for me to notice the fault and switch the power off. 375^2/36k is about 4W which shouldn't fry the wirewounds too quickly, even given the limited ventilation provided by a few air holes round the back of the PVC pipe.

Cheers,

GJ
Attached Thumbnails
Click image for larger version

Name:	Octal cap reformer small.jpg
Views:	546
Size:	42.4 KB
ID:	183433   Click image for larger version

Name:	Octal cap reformer Quad II small.jpg
Views:	267
Size:	104.9 KB
ID:	183434  
__________________
http://www.ampregen.com
GrimJosef is offline