Thermistors & diodes to protect rectifier Tube.

[HaroldBrooks]

Here's the schematic to the amp in question : https://el34world.com/charts/Schemat...supro_1600.pdf

Now for my actual questions :

1.) Does the Field coil in any way act as a choke to limit the inrush current for the first 40uf filter capacitor ? Doesn't look like it to me, but I am still too green at all this to determine.

2.) The 40uf Filter capacitor in this amp exceeds the 20uf spec sheet limit for a 5Y3, and the wall voltage is higher than years ago, so can I use thermistors in series with the 5Y3 plates to protect (as you would with ordinary resistors) from too much inrush current ? (this might not be the correct Thermistor) - https://www.mouser.com/ProductDetail...YaAtsiEALw_wcB

3.) I added two 1N4007 diodes in series with each of the plates on the 5Y3, and I am told that this increases the 'Peak Inverse Plate Voltage' the 5Y3 can handle. Will this extend the life of the 5Y3 rectifier tube overall ?


Thanks for any insight and help !

[R.G.]

1.) Does the Field coil in any way act as a choke to limit the inrush current for the first 40uf filter capacitor ? Doesn't look like it to me, but I am still too green at all this to determine.

No. The path from the rectifier to the filter cap does not include the field coil.

2.) The 40uf Filter capacitor in this amp exceeds the 20uf spec sheet limit for a 5Y3, and the wall voltage is higher than years ago, so can I use thermistors in series with the 5Y3 plates to protect (as you would with ordinary resistors) from too much inrush current ?

(1) Maybe, if inrush current was the actual problem; (2) no, as the real problem is the constant high current peaks from the rectifier to the filter cap under normal operation, not turn-on inrush.

The limitation in the tube data books on filter cap size is a translation of the fact that tube rectifiers have a maximum safe instantaneous peak current. Making the go above this damages the cathodes eventually, although minor excursions above the maximum may not show instant damage. It's a long term process of wearing them out faster. Inrush at power on does this, but there are few AC cycles of inrush. Too-high capacitance makes the peaks of current too big every single AC power line cycle, and this is in addition to the inrush high currents. So limiting power-on inrush may help a little, but doesn't solve the bigger problem.

There are at least five "solutions" to this issue.
(1) Use the bigger cap and just let the user pay for more-frequent tube changes. If it takes a long-ish time, the company will probably not get bad press.
(2) Use a "bigger" rectifier tube. Costs more, runs up the per-unit cost of production, so not a good choice for the bean-counters. Might be a good choice for you.
(3) Use a smaller cap. Increases power supply ripple hum, though.
(4) Use a new, smaller "first filter cap" after the rectifier, separated from the existing 40uf cap by resistances. This smaller cap lowers the peak currents in the rectifiers, and the separating resistors help filter out the higher ripple. Good filtering, lower stress on the rectifiers; the payback is a smaller B+ voltage and lower power from the amp, but may not be too bad.
(5) Use a current clamp and force the current peaks to not be too high. There is an article at geofex.com on how to do a simple MOSFET current limiter that can be put in series with the rectifiers so they ...never... go over the peak current you set. Cures inrush issues and capacitor size issues, but takes some circuits skills to get right.
(6) Replace the rectifier tube with silicon diodes plus a series power resistor to "fake" the existing rectifier tube. This actually works pretty well. It's how many "solid state tube rectifier replacements" are made. It is between very hard to impossible to tell by listening to the amp that such a replacement has been done. SS rectifiers are essentially immune to the too-big-first-cap issue.

3.) I added two 1N4007 diodes in series with each of the plates on the 5Y3, and I am told that this increases the 'Peak Inverse Plate Voltage' the 5Y3 can handle. Will this extend the life of the 5Y3 rectifier tube overall ?

No. However, it may make it appear to be immortal. The overcurrent pulses will still age the rectifier tube, regardless of the SS diodes. If the rectifier tube ages into low emission, the rectifiers do nothing. However, if the rectifier shorts, the SS diodes will simply keep rectifying in spite of the shorted rectifier, and all will be well. You might not even notice the rectifier tube being dead.

[eschertron]

There are at least five "solutions" to this issue.
(1) Use the bigger cap and just let the user pay for more-frequent tube changes. If it takes a long-ish time, the company will probably not get bad press.

I know RG knows this, but HB may not:
In the golden age, electrolytic caps were often spec'd with a uF value +80%/-30%. The engineers who spec'd the rectifier tubes, then, could not be sure that any nominal 20uF cap didn't actually have a capacity of 36uF. Reason suggests that the 20uF limit is conservative based on modern tolerances of +/-20%.
I wanted to mention this to make option (1) a bit more palatable.

[HaroldBrooks]

I am a tiny bit suspicious about RCA datasheet maximums in particular, as I own more than a few RCA 6973 tubes that were run at awfully high plate currents, well in excess of the 12 watt maximum, and they are still alive and work as intended. Could be those were exceptions though, and I am playing with fire if I adopt a cavalier attitude as a general principal.

Same with running higher plate voltages above maximum on other tubes I've seen.

However, I would still like to keep things within the spec limits if it's an easy thing to do and doesn't change the original amp character. It's very recent that I started paying attention to the maximum capacitance of rectifier tubes, and I've seen more than one run with higher values, starting with mine.

I like the concept of a THERMISTOR (limiting current and not blowing like a fuse), does anyone know where, and when you would use that as a protective element in an amplifier, above and beyond using a fuse ? I seen it used on AC lines, but how about HV DC ?

Thanks for the help !

[pdf64]

As RG notes, an NTC thermistor is only of benefit for the initial, start up current surge. After a few seconds it heats up and its resistance drops to a negligible level, thereby exposing the rectifier plates to excessive current every half cycle, in order to keep the overvalue reservoir cap charged.

[HaroldBrooks]

Ok, now I think I see it. I was comparing it to a Varistor in clamping, but a Thermistor is different not only in that it controls current instead of voltage, but there is no clamping.

I don't think it has too much value for consistent DC current spikes. Just making a bit of trouble for myself, as I am on this new kick to ensure things never fail, or at the very least never fail catastrophically, but I think I have most of it covered, and the day my 5Y3 signs off, so be it. I have spares and the diodes will stop any carnage if the rectifier were to short.

Thanks !

[mozz]

"added two 1N4007 diodes in series with each of the plates "

This is a common thing to do for Dynaco Dynakit ST-70 amplifiers. They are hard on the 5ar4 and supposedly if you lose power and do a fairly fast restart, it takes out the rectifier tube. Just fixed one with a totally dead 5ar4, and placed 2 1n4007 diodes in series with the plate wires. This was argued over many times on a few forums and the general consensus was it does help prolong the life of the tube when power is interrupted.

[HaroldBrooks]

If it's a pre-rectification of the AC, that only allows a positive voltage through to the plates, how could in not benefit the tube rectifier in some manner ?

Maybe it doesn't. I would like to understand the details, and I probably lack the deeper experience of nearly everyone here.

I know that if the Rectifier tube were to short and allow direct A/C to make it to the filter caps it would wreak havoc for sure.

One thing I've not yet gotten used to is how a super tiny piece of silicon in the diode can stand 1A of current and huge amounts of voltage ! ... But there it is !

Thanks for your help and insight !

[nosaj]

If it's a pre-rectification of the AC, that only allows a positive voltage through to the plates, how could in not benefit the tube rectifier in some manner ?

Maybe it doesn't. I would like to understand the details, and I probably lack the deeper experience of nearly everyone here.

I know that if the Rectifier tube were to short and allow direct A/C to make it to the filter caps it would wreak havoc for sure.

One thing I've not yet gotten used to is how a super tiny piece of silicon in the diode can stand 1A of current and huge amounts of voltage ! ... But there it is !

Thanks for your help and insight !

If the Diodes were installed on the tube socket if the tube rectifier shorted the diodes would take over the rectifier duties.
Go to the section on rectifier switch has the backup diodes picture. Another great site https://robrobinette.com/5F6A_Modifications.htm
nosaj

[trobbins]

If it's a pre-rectification of the AC, that only allows a positive voltage through to the plates, how could in not benefit the tube rectifier in some manner ?
....
One thing I've not yet gotten used to is how a super tiny piece of silicon in the diode can stand 1A of current and huge amounts of voltage !

The silicon diode, added in series with the valve diode anode, is a rectifier diode that could be safely used in place of the valve diode - it is typically a 1kV 1A rated diode that is used in amps with just silicon diode rectification.

The silicon diode in this application is being used only for its backup capability - its role is benign until the valve diode starts to show signs of leakage during peak inverse voltage stress - the valve rectifier may have become a bit gassy from recent outgassing due to high temperature operation, or end of life getter, or ... If the silicon diode was not in place, then a slightly gassy diode could show some internal arcing, and in bad conditions may cause collateral stress/damage to the power transformer and filter capacitor.

[Chuck H]

I've used R.G.'s suggestion 4 in post #2 above in a few amps with no noteworthy consequences. No voltages are shown in the schem so I made a guess, but if you start with a 20uf cap in front of the 5y3 and follow with a 150R resistor to the 40uf cap (still using the 40uf cap as the plate node) you will lose about 13 volts (or so). That's not a deal breaker for me and there's no need to wonder about thermistors, varistors or diodes. None of which offer a solution since it's not voltage that is of concern here, it's current. But on that note...

eschertron is spot on with his observation that caps from back in the day had a wide +/- spec. Typically 80% on the plus spec. And in my experience older caps were most often higher in UF than their rating. Obviously tube manufacturers would be aware of this and, at the very least, if plugging a spec'd cap value into a design was blowing up tubes I think the industry as a whole would have noticed

Modern caps are normally much more accurate to spec and in light of the above I've used 33uf in front of 5y3's on a couple of occasions without reservation. One amp that I'm still able to track that I did this with is probably overdue for a recap and is using the same 5y3 I sent it out with.