Selenium Rectifier - Conversion kit?

[Piper Blvd]

Hi Everyone!

Unfortunately, I am very much a "Drop In Replacement" kind of person (which, probably tells a little bit about my skill level). With the exception of the Dynaco PAS (for which I have bought power supply replacement board kits), I have always avoided buying / working on equipment with selenium rectifiers.

Here is what I do know:

- What they did.
- Their internal resistance drifts, causing them to fail.
- They should always be replaced.
- You cannot just simply replace them with 2 diodes alone, a resistor (of a T.B.D resistance value, but with a 10 Watt power rating) needs to be added because the selenium rectifier had a greater voltage drop than modern day diodes do.

But........What about this?:

http://www.ebay.com/itm/SARKES-TARZ...287819?hash=item3d35f584cb:g:WzUAAOSwTuJYuyix

If you check the pictures in the listing out, the kit came with specs, and installation instructions for both wiring it as either a half or full wave.

I guess in theory, this kit would work, but I am thinking that more details would have to be known about the selenium rectifier its replacing? Also, Capacitors and resistors are one thing, but what about old / NOS diodes, can they be used safely?

*Note, anything that I would install this in would be for my own personal use. I don't sell repair or equipment (or claim to be an expert!)

Any suggestions / advice / thoughts on this would be appreciated.

Thanks!

 

[gadget73]

what are you looking to replace it on? The trouble is that you may not have a one size fits all solution. There may be some external resistance you need to add depending on the circuit. Not always though. Low voltage applications had few sections in the rectifier, so there wasn't so much voltage drop. Very low current demand stuff just doesn't drop much because of the low current. If its in a bias supply for example, it may have enough adjustment range that you don't have to do anything special with it.

Also, those things are large and ancient, a modern 10 cent diode will do all that these will without any fuss. Functionally they're probably fine but for all the hassle of finding a spot to install them you can fit a small terminal strip with the diodes and resistor in the spot the original rectifier sat.

 

[OMGCat!]

It's not that hard to replace a selenium rectifier. You really just need to make sure at the end of it all your B+ is in spec, which usually means you'll need to add a resistor of some sort in series with your new diode to achieve the proper result.
I'll add the diode and too much resistance at first to get a baseline measurement. When that's done I can calculate what I need and finish it up. There are many tutorials on YouTube that can help you out. The size resistor you use depends on what it's powering. If it's a TV or something you'll have to use a higher wattage than if it's a VTVM or something along those lines.

 

[Piper Blvd]

Hi Gadget73,

I have the schematic attached here below. I am looking at a Ampex 403 preamplifier. The schematic is actually for a Ampex model 402 (as from what I have read, Ampex did not publish info on the 403 model, so this probably will not be exact).

The rectifier on the 403 is mounted out on the back side of the amp, and unlike many of them that I have seen, this one (the rectifier) only has 2 fins, or sections.

I took a closer look at the Sarkes-Tarzian conversion kit installation instructions, and it does indicate that a resistor needs to be added ("R in set must be 5 Ohms minimum"). The room for this old kit wouldn't really be a problem, but yeah, new diodes and a terminal strip would be far less expensive. Back to square 1.

Looking at the schematic, does it look as if a resistor would be needed to prevent the B+ from increasing?

Sorry, a little out of my comfort zone on this one!

 

[Bob]

the voltage difference between older seleniums and modern day diodes of any type
(schottkys for example) is so small (compared to AC line voltage differences that
are multiplied 2-4X after rectification) that you can ignore.

there are two 4K7 resistors tapped off the rectified end of the single diode in the
Power Supply (PS), and these can be adjusted to compensate for the diode
voltage difference (small) and the new AC line voltage (larger).

if the schematic has those voltages identified (DC volts after those two R15/R22
[not clear - resolution too low] ) either right there or indirectly as it hits a tube/cap/resistor
part of the circuit.

then you can up or down the resistance to bring the voltages into design spec. there are
three ways to do this

1. look at the schematic and find all references to voltages that reference back to those
resistors and note the voltage indicated on the schematic.
2. calculate what they should be based on circuits, operating conditions, etc
3. before you start replacing things, put the unit on a Variac, tune the AC input
back down to the lowest AC specified in your documentation (110, 117, etc)
measure the voltage at the two points AND pull those resistors and measure
the resistance to ensure they are within 1,2,5 % of its original value. or temporarily
replace them with 4K7 at 1% (up the wattage for the test since older CC
resistors at a given wattage are almost always larger than their modern counter
parts - when actually replacing them - use flameproof or other types to ensure/continue
your level of safety) and re-measure with the newer in-spec 4K7.
4. review the docs, back the day, they usually mention the size (wattage), precision
(1,2,5,10,20%), and occasionally the type (carbon film, metal film, wirewound, carbon
composite).

last note , there is no guarantee using a kit solves all the problems. the above testing
is recommended for all PS changes to bring the voltage back in line - no matter how
you replace the diodes, including replacing the diode with a tube rectifier.

 

[gadget73]

you'd have to measure it, but it looks like there is a resistor in there that can be changed if the voltage ends up too wrong. Its hard to read, the schematic got downsized considerably. It does appear that there is a voltage printed after the resistor, so you can simply confirm if its still in spec. Also keep in mind that most of these things are +/- 15% or so typically.

Looks like two sets of rectifiers in this. One is for B+, the other is for tube heaters. The one for the heaters would be a higher current, but since its low voltage it won't have so many sections. That one may or may not need extra resistance but its easy enough to calculate out if it checks in too high.

 

[H/K crazy]

I have simply replaced them with 1N4007 diodes in the past with no problems. Never needed to add anything.....

 

[CoastinHome]

I also have a 403 that I will be working on soon. All good advice above. The concern with an increase in voltage is generally when replacing a tube rectifier with it's solid-state equivalent. All else being equal, by replacing a tube rectifier with solid-state you might see as much as a 15% increase in voltage.

The minimum 5 ohm recommended resistor is likely not intended to compensate for a reduced voltage drop with the replacement diodes - but rather to absorb (dissipate as heat) the difference in voltage between the rectified transformer output and the voltage at the discharged capacitors at power up. Without it, the transformer, diodes and capacitors are unnecessarily stressed from the turn on surge.

I'm hoping to start a forum room specifically for the Ampex consoles. During their approximately 1957 -to- 1962 production run there were many undocumented changes to these consoles. I've done a lot of research to understand the different variations on the different models. I'm doing my best to nail down the differences in the Ampex Signature model.

 

[gadget73]

selenium rectifiers also have a greater voltage drop vs silicon. Typical silicon diode for up to 1000v is 0.6 volts. Each slice of selenium is good for about 20 volts reverse and loses about 1 volt forward. To get high voltages they simply stack up the slices, which of course stacks up the forward voltage loss. Figure 200 volts inverse means about 10 volts forward loss. Heater supplies aren't going to have a bunch of sections so you probably won't notice it. High voltage you will. Bias supplies probably have enough adjustment range to make up for it.

That forward voltage loss is also why they have big heat sinks. The voltage is lost in the form of heat. Same reason that silicon diodes typically do not need a heat sink until you get to very high current levels.

 

[C_D]

I can attest firsthand that installing silicon diodes in place of the selenium is a drop-in and go proposition. These ancient tube circuits were rated for +/- 20% voltages and line cord was getting 110 or maybe 115 volts depending on the era. US line voltage spec changed to 117 years later so other variables will swamp the voltage drop concern. Use a diode rated double the highest listed DC voltage on schematic and results should be fine. Selenium rectifiers typically pose substantial hazard only when they burn up and smoke (hazardous fumes are emitted).