Flashing Rectifier Tube- Bad Thing?

[Sherman]

The rectifier tube (5U4-G) on one of my new KT88 monoblocks has started briefly flashing white light on power-up. I saw it first two days ago and have powered it up twice since then. Each time it flashes then all seems normal. Voltages within normal range (within a percent or so of the other monoblock) and the sound is great. No fuse blowing or any other symptom than this white flash.

I've ordered two new tubes and don't intend to power it up again until I replace the tube, but as a newbie to tube rectification I'm wondering if this is a common harmless phenomenon or the harbinger of a magic smoke leak?

BTW years ago I was witness to a rather spectacular tube failure involving a flash, small explosion of the envelope and flame, something I'd rather not witness again!

[fdegrove]

Hi,

If it's a current production 5U4 then it doesn't come as a surprise, there have been reports of flakey ones for awhile....

However if this is a problem that only developped recently and only happens at turn on_when the PS caps charge up_ then it could well be that the cap's value is to high fro what the tube can handle.
The amount of max. capacitance for a cap input type filter is sometimes stated in the manufacturer's datasheet and is always directly proportional to the current drawn.
It's not unthinkable you're exceeding one or worse, both limits.

Cheers,

[Sherman]

Originally posted by fdegrove
Hi,

If it's a current production 5U4 then it doesn't come as a surprise, ..

... if this is a problem that ... only happens at turn on... it could well be that the cap's value is to high fro what the tube can handle...
Cheers,

Frank,
Thanks for the info. They are current production so that may be it. The ones I've ordered are also new production but will be Sovtek and my current pair are Chinese.

The data sheet says that for a cap input PS the first cap can be 40uF which is the value I've installed. The 5U4 on the other monoblock isn't showing any signs of this (yet ) and of course it is set up the same way. However I'll investigate purchasing some lower value ASC caps just in case.

[Dr. Photon]

40 uF for a 5U4 tube seems like a perfectly safe value. I've seen production amps from 20 to 30 years ago that exceeded the specification by 10 or 20 uFs that still have the original rectifier tube in them and are going strong. It IS an issue when you grossly exceed the maximum, such as with 100 uF or so. Your tube is most likely flakey. I'm not sure if the flash is just a harmless "thing", or if it is the precurser to an explosion of some sort. I have heard some bad things about the reliability of Chinese tubes. I have never used Chinese tubes, so i can't comment from personal experience. I use vintage tubes salvaged from new equipment, along with some SOVTEKs, JJ Teslas, and some of those russian marked 6(pi)14(pi) (Russian Military EL84, same as SOVTEK EL84M). SOVTEKs are very reliable tubes.

[skyraider]

I used 100uF input cap ahter the 6x4..
I checked the datasheet and it specifies only 10uF...
I am using too big cap I guess, and I didnt worry too much about it until I read this post..

Is this(too big input cap) bad for the tubes in the long run?

[Colt45]

keep in mind that a lot of caps may be as much more than rated value as it is easier to give a few bonus uF than it is to get the cap within 10% of rated value.

[Sherman]

Originally posted by Dr. Photon
40 uF for a 5U4 tube seems like a perfectly safe value. I've seen production amps from 20 to 30 years ago that exceeded the specification by 10 or 20 uFs that still have the original rectifier tube...
I have heard some bad things about the reliability of Chinese tubes... I use vintage tubes salvaged from new equipment, along with some SOVTEKs, JJ Teslas, and some of those russian marked 6(pi)14(pi) (Russian Military EL84, same as SOVTEK EL84M). SOVTEKs are very reliable tubes.

I figured that 40uF would be OK. It is listed in the data sheet as the max and given that values vary +/- 20% in a lot of caps, especially when this tube was designed, I assumed there was some safety margin in that spec. I hope I don't have to replace them.

This is my first time (and probably last) using Chinese tubes. As it turns out the Sovteks are only $1 more than the Chinese tubes I originally bought! I should have gone that way at the beginning.

I already miss listening to the system so I hope the tubes show up early this week!

[richwalters]

Originally posted by Sherman


I figured that 40uF would be OK. It is listed in the data sheet as the max and given that values vary +/- 20% in a lot of caps, especially when this tube was designed, I assumed there was some safety margin in that spec. I hope I don't have to replace them.

The specs of yesteryear hardly bear relation to current production tubes of today.......in another thread it was mentioned that manufacturers can't spend acres of time producing the NOS quality of earlier years. So with this in mind, the only solution is derate recent made tubes and operate well within limits.....if you want'em to have a decent life. That goes for rectifiers......so many rel new ones are fault-found casulties on my MI service bench .....simply because they can't take the stick. The worst is the DHeat 5U4G......and believe it or not the latter GZ34. My only solution is to reduce the Cin cap value....by a 1/3......i.e 40uF to 33uF and to put resistors equal to the secondary resistance per sec winding in each anode rect leg..that works for cathode bias fine. The case of fixed bias is to simply parallel another ident rectifier with another.....so long the heater can supply the guts.

richj

[Sherman]

Originally posted by richwalters


The specs of yesteryear hardly bear relation to current production tubes of today... So with this in mind, the only solution is derate recent made tubes and operate well within limits...

richj

Rich,
Thanks for the feedback. As I stated before I ordered new tubes. Today for the first time- shame on me- I noticed the tubes are slightly different inside. The envelope and bases are exactly the same but inside one has dish-shaped plates and the one that is flashing has smaller ones with a bump on them.

I also found a place to get smaller caps of the same type so if I need to lower those values I will do that as well. WHen I measured voltage and current after the tube all was within limits but I don't have a capacitance meter.

The problem I have is that the PS seems to be so darn clean right now that I hate to change it. If I do though I think I'll simply add the smaller cap, move my choke connections to the new smaller cap and the old 40uF cap and then make an RC network for the last 100uF cap. Should be even cleaner that way and I think I have the real estate on the chassis to add a third cap. No worries about too low voltage as my target B+ was 400V and I'm getting 429. I can lose a little across the resistor.

[richwalters]

Hi there.... this thingy about longevity of power tubes in general (+ rects) ..the strangest thing is I don't have any longevity problems with the minor stuff.....recent Svet ECL82's, EF86's and most of the 6SN7 class....they seem to take what comes.. year on year....I presume material costs aren't so much of an issue when compared to the large tubes.

The surge voltage from hv transformers is ghasty espec when a flash over occurs is stressful......and may contain ringing.......when using a center tapped secondary i.e bi-phase, an anti surge fuse placed in the center tap to grnd can offer protection if a flash over occurs. But it is a warning shot to do something about it.
The 5U4 or equiv U52
heater 5V @ 2.25A
max resevoir cap of 16uF (MOV spec)
min series resistance in each anode leg of 180 ohms non inductive.....(hard to get these days)
max Iout 250mA
Vin max AC rms 500-0-500

It looks that the same tube made by clone manuf will accept differing values of res cap values...The cue is really to seek an indirect rect capable of taking the same AC volts/ current...the GZ34 would give a lower I out and could be overrun. The GZ34 is quoted at
Iout = 160mA; (would be overrun in your circuit)
max resevoir cap of 60uF
series resistor each anode leg 175 ohms
Vin max AC rms 550-0-550

richj