Coupling Cap Voltages?

jaymanaa

RIP 1961-2018
Got a dumb (probably) issue I'm trying to get my head around. On a typical, say, EL84 amp, the coupling caps are always a minimum 400 wvdc, and often times 600v. On epay, the 400 or 600 v caps (NOS PIO) are way more money than the 200 or even 250 wvdc jobs. When I measure the voltage that the cap sees, it's usually always under 200 (maybe close). The 400 and 600 volt jobs are much larger too, and sometimes a tight fit, so why can't we use 200v rated caps If it was a manufacturer that was known for cheap stuff, I wouldn't consider it, but why wouldn't TOTL caps do fine, and save us money and room. This is considering the use of a modern inrush limiter like a CL90 or the like. First thing is, how would it ever see a higher voltage, and if it could, wouldn't it be a temporary thing? Thanks in advance fellas.
 
Before the tubes start to conduct there could be more than 200v across them. Especially true with SS rectifiers or direct heat ones.
 
I agree with gadget73, I wouldn't recommend using anything under a 400v cap for coupling. Most of the caps I keep on the shelf for service work are either 400 or 600 volts. I do the same thing with electrolytic caps. I do a bunch of repairs to tube equipment, so I don't see any reason to stock lower voltage caps just to save a bit of money.
 
I design for fault conditions including operator error such as some tubes not in place with power applied. I’ll also prioritize safety and reliability above sonics, style and fashion if a compromise must be made. This usually results in caps rated for max unloaded PS voltage, typically 600/630V.
 
Signal voltage peaks can go up to the B+ voltage, gotta consider AC too here. Don't use a coupling cap rated less than B+

Also if you operate the amp with a tube pulled, say to trounleshoot or disable a feature, again full B+

Have a look at the voltages used by RCA, Philco, etc back in the old days, they wouldn't have spent 5 cents on a cap if 3 would do.
 
Last edited:
Jayman - The main coupling caps to the power tubes are especially vulnerable because there is negative bias voltage on one side, which adds to the overall voltage across the cap. On the other hand, Orange Drops (and maybe others) are tested and rated for 150% @ 85deg.C. for 1000 hrs. So a 400v cap can run at 600v for quite a while without breaking down.
In manufacturing, when most of your parts have to be rated at say 400v, it doesn't pay to buy some 400v and some 200v. So, it's cheaper to buy a large qty. of the 400v versions and spread them around.
In your case, if you limit how fast the amp comes up to full voltage with a CL-90 (read as a SLOW start), and a 400v cap never reaches 200v, I would use a 200v cap.
 
It just seems that if you measure the voltage that coupling caps actually see in most commercial 50s and 60s vintage amps, it's always under 200. Knowing that most parts are tested at 150% just got me to wondering. It's not just pennies I'm talking about either when you get into the old NOS really sought after stuff. Surely the 400+ parts will eventually dry up, so if a 200v part can safely be used, it seems to make sense. I've played around with my TO6 before trying to blow up caps, and it takes a lot to blow up jobs like Hyrels etc...... Maybe closer voltage caps would best be used in a techs personal gear, or better yet, in shop use gear? Don't get me wrong, I tend to overbuild a ton, but I'm just thinking about harder to find stuff here. I appreciate the responses, they've given me lots to think about.:thumbsup:
 
I would much rather work on tube equipment than solid state. I really dislike printed circuit boards, IC"s and the dozens of small electrolytic caps that have to be replaced when doing a restoration on a piece of solid state gear. I keep parts on the shelf to service solid state equipment if the need arises, but my stock of parts for tube equipment is a bunch bigger.
 
which is a big part of why I don't go for that old stuff.

Good point Gadget, seems I go back and forth on this issue myself. You have the group that swears by an old ugly output transformer, while others wouldn't use it even if it was potted and painted. I suppose I fall somewhere in the middle, except for tubes, where I almost always prefer vintage over new.
I've been thinking about what Max said, and while that would certainly be a safe bet, I don't fully understand. Saying we are talking about "most" older designs, wouldn't the highest voltage you could expect be whatever your plate voltage to final driver or splitter is. The coupling cap goes from that plate to the grid of the output tube, with a resistor sharing the output of the CC with the grid of the output tube. So, knowing that resistors fail open most of the time, we will figure ours failed closed and took that side of the cap to ground. This, in my mind would be the highest potential for voltage, or am I missing something? So, say you're using the 12AU7 for your splitter. The max plate voltage is 300, but we have at least 2 resistors, usually 3 or maybe more between that and our B+. Wouldn't whatever voltage that plate is running at be the highest voltage we could expect to see? Again that's with a slow turn on device, especially if SS rectified. So even if they were running the 12AU7 at the maximum plate voltage allowed by law, and we had that one in a few hundred resistors that failed shorted, our 200v cap, that is really tested to 1.5 times rated value would surely survive a long time. This brings up another voltage issue that's always bugged me. Why did all the EEs disregard the max plate voltage ratings on the EL84 especially, or so it seems. How many vintage amps run these tubes at, or under 300 volts? I can't think of any right off, but that doesn't seem to bother anyone, or else we would have all kinds of threads about how to get it down then re-calculate the cathode R. I agree totally with playing it safe, and I sure wouldn't skimp without a dang good reason, just playing devil's advocate for the sake of discussion. Let me sneak this in too, as this looks to be a good learning thread for me. What do they really consider when coming up with a max plate voltage. I would guess it would be things like proximity of the plate to other components to prevent an arcing condition. I remember reading something about Mullard being very excited about the super high voltage rating they achieved while developing the EL34, something like 8 or 900 volts IIRC. Seems like it was the way things were laid out and how it allowed the internal wiring to be physically distant so that arcing wasn't an issue there?
 
Last edited:
I would much rather work on tube equipment than solid state. I really dislike printed circuit boards, IC"s and the dozens of small electrolytic caps that have to be replaced when doing a restoration on a piece of solid state gear. I keep parts on the shelf to service solid state equipment if the need arises, but my stock of parts for tube equipment is a bunch bigger.

Amen, and it seems like every day I need more light and more magnification to see anything. Plus, circuit boards don't lend themselves to mods like PTP does. I'm doing an SCA35 right now, and no way would I leave it using 7199s if the sockets weren't on circuit boards. I did mod my personal ST35 to use 10v output tubes even though they were on boards, and it was a huge PITA.
 
One thing you may wish to consider is the popular myth that coupling capacitors with higher voltage ratings ''sound better''. Maybe that's why they're getting harder to find....

Personally, I rank this one right up there with audiophile power cables and crystals:rolleyes:
 
The higher the voltage rating, the thicker the film between the electrodes and the lower the voltage gradient. The MKV 660VAC caps I really like for couplers sound quite good, and that is just a metalized polypropylene. I wish I could find the same in film/foil with a low inductance construction. The Russian MBGO and KBG oil caps work quite well too...but they're not small.

This is all a fair justification for me to stay away from commercial stuff. If I need more room for the right cap, it gets built in. I also don't use less than 1 uF for coupling caps.
cheers,
Douglas
 
DSC_0348-1024x578.JPG Ok the attached is a diagram of a basic resistance capacitance stage.

If the AC input signal is very strong, it will bias the tube "off", this will bring the voltage at "B" to the same as the voltages at the top end of the plate load resistor.

If we pull the tube, the voltages at "B" will be the same as the voltage at "A", which in many amps is the full B+ voltage.

I know tube ratings were regularly exceeded in the old days, but capacitor ratings much less so. I've seen some companies cheap out to the extent of using a different voltage rating for the top and bottom coupling caps on a split load phase inverter.DSC_0348-1024x578.JPG
 
One thing you may wish to consider is the popular myth that coupling capacitors with higher voltage ratings ''sound better''. Maybe that's why they're getting harder to find....

Personally, I rank this one right up there with audiophile power cables and crystals:rolleyes:


Dang man, I have that same shirt, and had that haircut, and that's a fairly close match to Brothers Fender Jazz that he would let me use. Thanks for the drawing Duglas, you always give me extra stuff to consider. Everyone did so far in this thread, I need to copy and file it I reckon.
 
Back
Top Bottom