MV rectifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I built the power supply for my 6L6 amp using an 83 rectifier tube.
Good thing too. It's been working just fine for quite a while now. but just today, I noticed something different.
As you all should know, MV rectifiers glow blue when current is flowing, and more blue equals more current. When HV is first applied there will be a very brief flash of blue, then almost none at idle. Well, today, I was in the shop, working on the T-Buzzard, listening to some music, when the power went funky. Lights flickered, then went off, then back on. The amp was on the other side of the building. It seemed to be playing ok. Then, at the last flicker, it went off, and didn't come back on.
I shut it down, and pulled the plug. I checked it's fuse, (3a Slo Blo) and it was blown.
I replaced the fuse with a 3a regular fuse (all I had in the 3a range) and re-fired it.
I warmed the filaments for a few minutes, then hit the HV switch. Rectifier flashed, and settled to it's normal barely noticeable glow as usual. Turned on the music, and went back to work.
Well, after a bit, the mains went wacky again. And fuse popped almost immediately. Well, dang.
I have a 4a slo-blo so I'll give that a go. (what's an amp give or take?) when I hit the HV, rectifier flashed bright blue for the whole tube, and stayed that way for a bit. IOW something's drawing a lot of current and not letting up. I suspect a bad cap, but not sure. I'll check into it tomorrow.
But my point, is that I doubt SS rectifier would show any indication before the majik smoke escapes.
 
You don't have a "point", just a doubt. What do you mean by 'indication' - a fuse blowing, or your room lights flickering?
I have a "point", that being the rectifier has indicated something has gone wrong with the internals of my amp. (abnormally high current flow) Whether it be due to the mains going wacky and overloading something, or something internal just decided it was time to act a fool. either way, the indicator of abnormal current draw is the rectifier tube.
 
Ok, so the indication process was initially the lights flickering, and after a few flickers, the amp went off. Then you did some sleuthing which involved stressing the power transformer and mains fuse again, and then went up a stress notch by increasing the fuse value.

Note that the 83 rectifier, like an ss diode, has a steady state peak plate current limit rating, but unlike even a basic 1N4007, the 83 rating is only about 0.7Apk, whereas the 1N4007 is a thermal junction temperature issue as it can allow up to transient 35Apk. Your stress testing may have in some way damaged the 83, but I'm not across MV degradation well enough to have a good view on that.

I'm also not sure if the MV itself may not have been causing the 'fault', due to PIV related short circuiting the power transformer secondary - inserting ss diodes in series with each MV anode (as is typical practise nowadays for common vacuum tube diodes) would help faultfind that scenario. Adding a fuse on the power transformer secondary would reduce stress on the power transformer if ever a fault occurred (such as in an output stage).
 

PRR

Member
Joined 2003
Paid Member
...the rectifier has indicated something has gone wrong with the internals of my amp.....

I wonder if the "something wrong" IS the mercury tube.

Mercury vapor can pass HIGH current. And unlike a 1N4007, they sometimes work again after the stress has passed.

(If you have worked in an old-time radio transmitter, or subway power station, you know the cut-outs protecting mercury rectifiers are often rigged to cut back in after a short delay, because sometimes that is enough to break a mercury "fit" and operation may be continued.)
 
MV rectifiers susceptible to arc back, especially with power line surges. Makes no difference whether it's an 83, a 816, 866A, 575A, 872A, 266B, 857B... Transients on the AC probably go right through the power transformer with peak voltage in excess of the PIV..That is what the fuse on your equipment is for. Tried a surge protector??
 
What size should the fuse be? Running one 83, two 12sn7, and two 6l6. About 350v on the anodes of 6l6 @~76ma. Looks like closer to 10amp fuse is needed. but a 3amp slo-blo has been doing well for a while now, so going to a 4a should not cause any troubles.

Start up time for MV tubes is a variable. The mercury needs to be vaporized completely prior to HV being applied. If not you will experience arcing, especially at a kv+. There is no set time for this, It's temperature dependent. As the filament warms the mercury, it will condense on the cooler parts, especially the glass. One must give it time to completely evaporate, and the tube to become clear, prior to turning on the HV. So you don't want to use the 5v leg of the HV transformer for the filament of a MV rectifier like you would a 5u4g.

That said. The MV rectifier can be used as a plate current meter, by paying attention to the amount of blue glow. More blue = more current flow, less blue = lower current flow.
 
Knowing your MV diodes are properly warmed up. Easy to see:
All pics are 872 MV diodes, full wave setup, 2kv power supply feeding an 813. (but any mv rectifier would show similar)
First pic is cold, prior to lighting the filament. Second pic is beginning to warm, Hg is condensing on the glass. Third pic HG is totally evaporated, and ready for HV to the anode.
 

Attachments

  • DSCN1123 (Copy).jpg
    DSCN1123 (Copy).jpg
    143.4 KB · Views: 261
  • DSCN1126 (Copy).jpg
    DSCN1126 (Copy).jpg
    130 KB · Views: 257
  • DSCN1128 (Copy).jpg
    DSCN1128 (Copy).jpg
    148.4 KB · Views: 257
here's what happens when HV is first applied, which immediately settles to a 20ma draw, followed by a 280ma draw when the xmit button is pushed. You can get an idea of current draw by paying attention to the blue glow.
Note the difference between the two diodes. That is also the difference between the Hg content of each. Guess which has more Hg.
YouTube
 
Schemo please?

I built the power supply for my 6L6 amp using an 83 rectifier tube. Good thing too. It's been working just fine for quite a while now. but just today, I noticed something different.

As you all should know, MV rectifiers glow blue when current is flowing, and more blue equals more current. When HV is first applied there will be a very brief flash of blue, then almost none at idle.

This, right here, suggests a problem. It looks like you're operating the 83 into a C-input filter. I've seen the same phenomenon with SS diode power supplies and C-input filters. Hg diodes don't have the Isurge capabilities of vacuum diodes, and are not to be used with C-input filters, but rather L-input to slow down that Isurge into discharged filter capacitors. Hg diodes are typically found in applications requiring high voltage and high steady state currents where good voltage regulation (the forward voltage of a Hg diode doesn't change very much with variations in current -- much like a SS diode) is required. This would be the case for transmitters or audio rigs that can fill a stadium with sound or plate modulate a high power AM rig. The really big ones such as a commercial AM transmitter will run off 3-phase AC (six Hg diodes).

Well, today, I was in the shop, working on the T-Buzzard, listening to some music, when the power went funky. Lights flickered, then went off, then back on. The amp was on the other side of the building. It seemed to be playing ok. Then, at the last flicker, it went off, and didn't come back on. I shut it down, and pulled the plug. I checked it's fuse, (3a Slo Blo) and it was blown.

I replaced the fuse with a 3a regular fuse (all I had in the 3a range) and re-fired it.
I warmed the filaments for a few minutes, then hit the HV switch. Rectifier flashed, and settled to it's normal barely noticeable glow as usual. Turned on the music, and went back to work. Well, after a bit, the mains went wacky again. And fuse popped almost immediately. Well, dang.

I have a 4a slo-blo so I'll give that a go. (what's an amp give or take?) when I hit the HV, rectifier flashed bright blue for the whole tube, and stayed that way for a bit. IOW something's drawing a lot of current and not letting up. I suspect a bad cap, but not sure. I'll check into it tomorrow. But my point, is that I doubt SS rectifier would show any indication before the majik smoke escapes.

This time of year, with the heat and lots of air conditioners running, mains drop-outs become rather common events. It's bad for electronics as it's the same as power cycling. Again, Hg diodes don't like hot starts like that. Who knows what happened here? Hot start and/or a high voltage transient that over volted and poofed a filter capacitor? Arc over inside the 83 that poofed something? Straighten that out, and if summer drop-outs are common in your area, consider investing in a surge protector/power failure backup.
 

PRR

Member
Joined 2003
Paid Member
....Hg diodes don't have the Isurge capabilities of vacuum diodes, and are not to be used with C-input filters....

Not disagreeing with a dislike of merc with cap-input.

But the '83 datasheet specifically calls out a condenser-input condition, with small series resistance and reasonably large (huge for the day) C value.
 

Attachments

  • '83-cap-in.gif
    '83-cap-in.gif
    9.2 KB · Views: 227
Schemo please?



This, right here, suggests a problem. It looks like you're operating the 83 into a C-input filter. I've seen the same phenomenon with SS diode power supplies and C-input filters. Hg diodes don't have the Isurge capabilities of vacuum diodes, and are not to be used with C-input filters, but rather L-input to slow down that Isurge into discharged filter capacitors. Hg diodes are typically found in applications requiring high voltage and high steady state currents where good voltage regulation (the forward voltage of a Hg diode doesn't change very much with variations in current -- much like a SS diode) is required. This would be the case for transmitters or audio rigs that can fill a stadium with sound or plate modulate a high power AM rig. The really big ones such as a commercial AM transmitter will run off 3-phase AC (six Hg diodes).
Running an LC input filter/voltage divider. Yes the immediate flash is the caps charging. (what's supposed to happen) What happened during the power anomaly was the steady higher than normal current draw, as indicated by the glow of the rectifier. Probably around 100ma+ instead of 76ma. It finally settled down after a while. No majik smoke escaped, and all seems to be well now.





This time of year, with the heat and lots of air conditioners running, mains drop-outs become rather common events. It's bad for electronics as it's the same as power cycling. Again, Hg diodes don't like hot starts like that. Who knows what happened here? Hot start and/or a high voltage transient that over volted and poofed a filter capacitor? Arc over inside the 83 that poofed something? Straighten that out, and if summer drop-outs are common in your area, consider investing in a surge protector/power failure backup.
I've noticed from time to time, that the mains will fluctuate from 120v to 130v RMS. I called a guy that I know who is the engineer for this area at the power company. He told me that it's not unusual to find anywhere from 117 to 127 or more, depending on demand, and location.
Dropouts are not all that common, but wind, and trees seem to cause fluctuations in the power, from time to time.
 
Here's the PS with the 83:
I'm no artist, and forgot the resistor values, but I can get them in a day or two.
Sorry I didn't include it earlier.

Oh, the secondary on the hv xformer is 450v. With also 6.3, and 5v. 5v is unused, and 6.3 feeds the heater on the 6L6.
 

Attachments

  • power supply.jpg
    power supply.jpg
    637 KB · Views: 209
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.