Be kind to your 5AR4 and input capacitor

[AmadeusMozart]

Build my amp, works fine and then fuse in mains lead blows. Turns out the cheap Chinese 5AR4 arced.

Put in the spare TAD 5AR4 and all powers up normal except that the B+ is now instead of 370V about 270V and the TAD gets very hot....

Check the PSU design in Duncan Amp's PSU designer and no design faults detectable. Researched internet and discover lots of complaints about modern day 5AR4, seems that at this moment only the Sovtek is reliable. Order a couple of new ones.

So far it looks like all problems are resolved however how to prevent / minimise the chance of happening again?

A few modifications have been done:

- Add a UF4007 with a 1nF 3kV capacitor parallel in series with each diode
- Reduce the filter input capacitance (before the choke) to 40uF
- Being aware of today's "low ESR" add an inrush resistor in series with the input capacitance (minimizes the start up current)
- Being aware of the limitations on the currents an electrolytic capacitor can handle use 4 * 10uF capacitors and make sure they are not getting too hot (although under the chassis I have created a ventilation draft across them)

Especially the inrush limiter makes a big difference on the start up current when there is downstream (after the choke) large capacitance.

AM

[Ty_Bower]

Originally Posted by AmadeusMozart

A few modifications have been done:

All your notes appear relevant and are good measures in my book. The only thing I might add would be to ensure correct operation of the amp. In other words, never apply power to a hot amplifier. Always allow all the tubes (and IRCL devices) to cool completely before re-applying the power. Ensure this either through procedure and discipline, or use relays and/or timers.

[indianajo]

My Hammond H182 organ has 50 uf filter caps, but the "400 ohm" choke comes first right off the cathode. The 5AR4's look like they lasted about 30 years, from the old ones left in the boxes in the back. The organs were produced in 1968. I replaced the electrolytic with 50 uf 1000VDC film motor caps (genelec), with extremely low ESR and 14 ga wiring. So far, the new Sovtek 5AR4 is handling it fine.
So I don't burn out my precious HV transformer, I took a terminal header and between the transformer and the tube I put 1/2" of fine wire teased out of a 19 stand 20 gauge. You can't buy 200 ma 600 VAC rated fuses, nor 1/4x1-1/4 fuse holders in that voltage. Hopefully the single strand will melt before the transformer guts if the tube shorts out. Geoelectro that owns an organ service puts 1/4x1-1/4 fuses in the primary of the HV transformer of tube organs he services, sized just big enough to blow if the rectifier has a problem. This is much smaller than the fuse that would be required to run the tonewheel generator motor.

[AmadeusMozart]

Originally Posted by Ty_Bower

... never apply power to a hot amplifier.....

Thanks for mentioning that - I do this through procedure and it has become second nature so I do not even think conciously about it. But for others that may be new to valve technology this may be worth mentioning since it is likely that they are not used to doing it in that way.

AM

[aardvarkash10]

Originally Posted by Ty_Bower

Always allow all the tubes (and IRCL devices) to cool completely before re-applying the power. Ensure this either through procedure and discipline, or use relays and/or timers.

Why?

[nicoch58]

the 5ar4 with cap input must see two 75r to two 200r(Rt) from power trafo to protect ....

The Valve Wizard
http://www.drtube.com/datasheets/gz34-philips1970.pdf

[Bigun]

Originally Posted by indianajo

Hopefully the single strand will melt before the transformer guts if the tube shorts out.

I didn't realize tube rectifiers can short ? I thought there was a physical gap between the conductors. In my amp I fuse the B+ after the psu choke/cap filter (the fuse doesn't see the pulsed current flow but the dc average).

[AmadeusMozart]

Originally Posted by nicoch46

the 5ar4 with cap input must see two 75r to two 200r(Rt) from power trafo to protect ....

The Valve Wizard
http://www.drtube.com/datasheets/gz34-philips1970.pdf

Thanks for mentioning this, this is another one that I forgot the mention (second nature again).

Measuring the DC resistance gives the ballpark figure (often on the high side) - one has to calculate the "reflected impedance" (this is easily done using PSU designer and otherwise one can use the Radio Handbook formula. With the "quality" of todays tubes it is better to go a bit above the recommended minimum value.

AM

[AmadeusMozart]

Originally Posted by Bigun

I didn't realize tube rectifiers can short ? I thought there was a physical gap between the conductors. In my amp I fuse the B+ after the psu choke/cap filter (the fuse doesn't see the pulsed current flow but the dc average).

If the anode gets hot enough as exhibited by red plating (e.g. cannot readiate enough of the heat as the result of too thin material being used) then it can short on the reverse peak. Similarly if the distance between the anode and cathode is not even then the reverse voltage capacity may not be enough.

[AmadeusMozart]

Originally Posted by aardvarkash10

Why?

If the valves are still hot and the capacitors are largely discharged then when applying power there is often a very large surge attempting to charge the capacitors in one cycle. This can have damaging effect, not only to the rectifier but also to the filter input capacitor and other parts down the line like the mains transformer or output transformer primary winding.

( How long does it take a tube amp to start working after applying power? That is how long it takes for the tube's cathodes to heat up
and it is a safe bet that it will take quite a bit longer for the cathodes to cool down. )

AM