Auricaps - 400V vs 450V (SimpleSE)

[Zap]

I recently placed an order for some Auricaps (0.22uF per the spec in my SimpleSE) and they were listed at 450V, but the parts that just arrived only say 400V. Has this happened to anyone else? The Audience website only list 400V and 600V values so I presume this was a change?

On George's SimpleSE page I see 450V's being installed, and without much thought that is what I ordered. Do I need to substitute and get 600V instead? Do the coupling caps need to be able to stand up to the B+ voltage, or can they be rated lower?

Thanks!

 

[Salas]

They just need to withstand the (commonly plate) DC bias voltage of the preceding stage. Check with your voltmeter if it is not noted on the schematic. Select max DC if its not auto range and measure between cap's signal input leg and ground.

 

[Tubelab_com]

When you first turn the amp on the tubes are cold and not conducting. If the rectifier tube comes up before the 12AT7 is hot the 12AT7's plate voltage will be equal to the full B+ voltage. The grid of the output tube will be at zero volts. At this instant the coupling cap will see the full B+ voltage across it. As soon as the 12AT7 warms up the plate voltage will drop into the 150 to 200 volt range. So good engineering practice dictates 450 volt caps or higher.

On the other hand, I have used the 400 volt caps in some amps without issue. I have been getting these from AES for several years. They have always been rated at 450 volts until early in 2007. I didn't even notice the changed rating until I had used them in some amps. I suspect that the caps have not changed, but some may have failed so the ratings were lowered.

Keep this in mind. Capacitors usually short out when they fail. A shorted coupling cap will make life very unhappy for the output tube, causing it to glow bright red and fry in a short time. This should blow the mains fuse, but if it doesn't and the amp is left on other components may be damaged including the power transformer. I have never seen this happen, but it is possible.

 

[dsavitsk]

When you first turn the amp on the tubes are cold and not conducting. If the rectifier tube comes up before the 12AT7 is hot the 12AT7's plate voltage will be equal to the full B+ voltage.

Doesn't that endanger the ixys chip? These things can only take 450V, and if the tubes are not conducting, without a regulated supply, a 450V B+ supply is going to be above 450V which should toast the chip long before the cap. At any rate, I'd think the 5ar4 should prevent either of these things from happening.

 

[Tubelab_com]

The 5AR4 should be the last to warm up, but my RCA branded Sylvania 5AR4 does seem to go live about 5 seconds before the EH KT88's start drawing the big current. For these 5 seconds the B+ is about 465 volts. The cheap Chinese ones are a bit slower.

The B+ is 465 volts for a short time, but the IXYS chip sees nothing across it until the 12AT7 comes up, then it sees 450 (or so) minus the plate voltage (150 to 200) or about 300 volts. There is some current through the chip as the coupling cap is charged, but the voltage across the chip doesn't go over 300 volts.

I have had the IXYS chips up to 500 volts and haven't blown one yet due to voltage. They will fry rather quickly if they are provoked to oscillate. I have been experimenting with higher voltages for another project 550 volts B+ using a CCS loaded 6EM7. I just got some of the new 900 volt IXYS chips to test, but haven't tried them yet.

 

[Zap]

Well, it looks like I am going to exchange the caps for the 600V versions. Better safe than sorry! Had I thought about it more when I placed the order, I would have just purchased the 600V caps from the beginning since the voltage runs so close to the max voltage limit. If Audience isn't certain enough that they will run at 450V, then I'm not going to push it. I'm not sure how the Sovtek 5AR4 compares to the other rectifiers mentioned, but that's what I'll be using for now. Once I know everything is working properly, I'll probably try some higher end tubes.

I still need to get some of the items external to the board such as decent hookup wire, switches, fuse plug, etc.. so I guess it's still going to be a while before I get to hear it. Hopefully I'll get access to the shop soon so I can fabricate the top mounting plate for everything.

 

[Salas]

The over voltage instances across the caps are going to be very short timed. I would not worry that much. But ''better safe than sorry'' is a good argument, especially if they can be traded in for the 600V ones. A side benefit maybe that the thicker 600V film sounds better too.

 

[rdf]

One thing I always do with an IXYS is drop a large portion of the 'unused' voltage potential across a series resistor. If the AC plate swing required is, for example, 50 volts p-p and the IXYS alone would have 400 volts between plate and B+ I would drop at least 200 across a series resistor. It's good protection, the chip runs cooler and, shifting more towards the voodoo side, I think it sounds better. Why is another question. Two guesses, the chip is reputed to be a FET with an error amp front end. Perhaps it prefers seeing a more resistive load, without the dropper the IXYS sees an active device in series with the power supply cap. Tin foil hat theory 2, a dropping resistor to some degree buffers the plate from the IXYS's input capacitance.

 

[Tubelab_com]

One thing I always do with an IXYS is drop a large portion of the 'unused' voltage potential across a series resistor.

The SimpleSE has a resistor in series with the feed to the IXYS for exactly these reasons.

Tin foil hat theory number three. The IXYS chips capacitance varies with the applied voltage. The resistor isolates some of this changing capacitance from the plate.

I have been "advised" that I "need" a decoupling cap from the junction of the resistor and the IXYS chip to ground. Addidn this cap would remove the sonic benefits of the resistor, but I have tried it anyway to see if there was any difference. I don't hear any, but I don't claim to have "golden ears". Too many rock concerts in my younger years.

The over voltage instances across the caps are going to be very short timed

Yes, but even a temporary short will dump the 10 mA from the CCS into the grid of the output tube. If this happens during warm up the cathode may not like it. After the tube is conducting a full scale fireworks display inside the tube is possible that may continue after the short has cleared.

I know for a fact that a very momentary OOPS with a screwdriver shorting G1 to G2 caused a really cool fireworks display. Since all parts used in that experiment were essentially junk (old Chinese KT88 tube) I just watched. The arcing and sparking went on inside the tube for 10 or 20 seconds and then it just died.

 

[zigzagflux]

The IXYS chips capacitance varies with the applied voltage. The resistor isolates some of this changing capacitance from the plate.

Not sure how similar the IXYS is to the DN2540, but it is clear from the DN2540 datasheet that at least 20V is desired across the device before Coss capacitance is at its lowest. After that, it (by the datasheet, at least) is fairly flat.

My experience (which is very little) with the IXYS part in a LTP is it wants lots o volts across it for best performance. I had disappointing results with the series resistor. Similar experience with bipolar current sources. I don't know why, but Tin Foil Hat Theory number 4 attributes the improved performance to lower capacitance as a result of higher device voltage.