My 50W plexi Screen Supply dilemma

[goldstache]

I built up a plexi-esque circuit with some iron I had laying around.  The B+ on the plates is around 525V.  Pretty high for moderately biased EL34's.  I had some winged C tubes in there and they lasted about 25 gigs.  Then, the other night upon flipping the standby I blew an HT fuse (.5A).  I replaced the fuse and fired it up, no heaters on one of the EL34's.  I turned it off, plopped the tube in the tube tester and it failed. 


So, I'm thinking that my power tube took a crap and that took out the fuse. 
It would be ideal to get the screen voltage lower.  I am wondering if I can develop a screen supply from the intersection of the totem'd reservoir filters?  That way I can lower the screen voltage and recalculate OT primary. 


Any ideas, thoughts?

[goldstache]

Or should I just keep the topology as is and use KT77's?

[shooter]

here's an example from one of Dummyload's creations;

note how the PT CT is utilized, I believe that's important

[pdf64]

Perhaps use a few 10V zeners in a string to drop the g2s down to a suitable level.

[DummyLoad]

ditch the standby switch. it's nothing more than small box full or transients waiting to blow shit up. check the failed tube's socket for arc trail from pin 2 to pin 3. if it has arced, you definitely want to replace it.

bias tubes colder - they'll last longer and sound just as good... with 525V on the plates, you'd want to be in the 60% of 28W arena: about 32mA per EL34.

--pete

[goldstache]

Thanks, I think I'm gonna go with the totem reservoir midpoint (balancing resistors included) as a tap for the screen supply.  Probably wil try without a 2nd choke, and just use a PI type filter setup.

[92Volts]

That won't work, the reservoir caps can't pass DC current so the only thing maintaining that voltage at 1/2 B+ is the balancing resistors. This effect is the reason balancing resistors are necessary to maintain an even voltage across 2 capacitors.

If you try to draw power from there it will be like trying to draw power through the balancing resistor (which I'm assuming is huge, like 100k, and cannot pass the current you need). Voltage will drop and this could even overvolt and destroy one of the caps, as the middle drops towards ground and the top cap has more/most of the B+ across it.

The only way it works is if a transformer tap is holding that location at a steady DC voltage. In "traditional" tube power supplies the center tap is grounded so nothing is available to serve this purpose. The circuit shooter mentioned may or may not be what you're using

[goldstache]

I also would like to note that the zener trick wasn't working for me although I did use it to dim the HT before the reservoir.  Smoked them, probably because the current from the power tubes.  It's just the screens this time around so it may be ok. 


I think I'm gonna try the totem idea.  I was already biased around 60% when it failed.
And I will check the socket when I get in there. 


Anybody have any info on how this lower screen supply will effect OT primary spec.  Math welcome!

[goldstache]

Yeah the PT is 190-0-190v. I'm running it FWB rectifier, with the PT CT tied to that totem center point.  I was thinking that I could since its bridge rectified, though, I've been wrong before.

[92Volts]

Yeah the PT is 190-0-190v. I'm running it FWB rectifier, with the PT CT tied to that totem center point.  I was thinking that I could since its bridge rectified, though, I've been wrong before.

Yep, you can do this because it's bridge rectified with the CT connected between the caps. The tubes should last forever with only 250-something on the screens!

If you hadn't been able to use the half voltage I was going to suggest ultralinear connection if your OT supports it. If you need to have the screens at the same voltage as the plates it's actually easier on them. Lowering the screen voltage should affect the sound less, though, and if you have that option I'd go with it.

[goldstache]

Thanks for confirmation!  Yeah, when I built it sounded so good I just boxed it up thinking, you know that screen voltage is gonna decimate those EL's. 


I was looking at a mig 60 schematic some time after and was like "yeah I should derive a second screen supply.".

[goldstache]

Speaking of UL mode, I was under the impression that the UL tap specs (70%) was reference to an AC figure.  That's why when I use UL the screens read the same potential as the plates from a DC standpoint but receive 70% of the AC of the plates?


Anybody want to chime in on this?


Thanks everyone!

[92Volts]

Speaking of UL mode, I was under the impression that the UL tap specs (70%) was reference to an AC figure.  That's why when I use UL the screens read the same potential as the plates from a DC standpoint but receive 70% of the AC of the plates?


Anybody want to chime in on this?


Thanks everyone!

Exactly.

DC is the same since they're connected to the same winding on the OPT. For example 500v.

At peak current, say a tube draws enough to pull voltage down to 50v at the plate, a downward swing of 450v. The screen sees downward swing of (70% of 450) 315v so it's at (500-315) 185v at that time.

If you take this to an extreme the plate will see 100% of the voltage swing at the plate and is effectively connected to the plate, or might physically be connected to the plate and this is triode-mode operation. Ultralinear is basically partway between pentode and triode mode.

[goldstache]

Brilliant!  Thanks for the explanation.

[pdf64]

Just to note that the tapping point on the primary for UL taps is usually around 40% of the turns (16% of the impedance), when using the HT as the starting point, ie the CT of a p-p OT.

[pdf64]

ditch the standby switch. it's nothing more than small box full or transients waiting to blow shit up. check the failed tube's socket for arc trail from pin 2 to pin 3. if it has arced, you definitely want to replace it.

bias tubes colder - they'll last longer and sound just as good... with 525V on the plates, you'd want to be in the 60% of 28W arena: about 32mA per EL34.

--pete

As much as it sticks in the craw to advocate HT standby, in this scenario (high HT, solid state rectifier and a DCCF), the DCCF tube will be significantly stressed without it
http://www.valvewizard.co.uk/dccf.html
http://www.valvewizard.co.uk/standby.html
So a standby with trickle bypass may be beneficial here.

Re bias, see http://www.aikenamps.com/index.php/idle-current-biasing-why-70-percent
With the intended load setting, 70% is likely to result in significant plate overdissipation with signal.
So it may be best to bias colder, eg 20mA cathode current at idle, or set the load impedance selector to half the actual value and bias at 70%, preferably with a bit more current limiting for the g2s.

[92Volts]

As much as it sticks in the craw to advocate HT standby, in this scenario (high HT, solid state rectifier and a DCCF), the DCCF tube will be significantly stressed without it
http://www.valvewizard.co.uk/dccf.html
http://www.valvewizard.co.uk/standby.html
So a standby with trickle bypass may be beneficial here.

The grid-to-cathode diode shown further down the DCCF page should keep it safe, with or without the limiting resistor. I've used that without the resistor and didn't notice a tone difference, but the way it's shown on Valvewizard is likely best

[DummyLoad]

if you'd be so concerned that the DCCF is even an issue, then use a protection diode. ditch the standby switch, use a MOV on the primary side of the PT to control inrush current.


--pete

[pdf64]

The point is that if ditching standby, there's other stuff to consider.
Re MOVs, they can't help with current surges; I suspect positive temperature coefficient thermistors were what was being thought of.
But even a PTC on the primary, I suspect that tube conduction will inevitably lag the HT; a PTC on the HT winding (or CT 0V return if bi-phase arrangement) may be a better way forward?

However, nickb has discovered a further DCCF failure mode, ie that the initial cathode current surge caused by the positive bias of the CF seems to damage the emission of Russian tubes, see http://music-electronics-forum.com/t46108/

[DummyLoad]

yes, sorry MOV are for OVP. was thinking thermistor, had MOV on the brain. still may be wise to use a MOV since B+ is high(er) than most amps we see these days, and esp, with lower quality tubes currently available.


PTC on primary or secondary: give it try? see if they live long-er? maybe a new PT? see below...
 
interesting discussion on DCCF. don't use russian tubes, but good to know for service work.


i worked on a laney 60 a few years ago that had similar issues + a blown PT. rebuilt it with new sockets and i used NOS GE or sylvania tubes for the preamp. scrounged up a pair of svetlana EL34 that held up ok before it left the shop. client sold it some time after that, so have no idea if the EL34 actually held up long term. amp had 525V B+ to the plates. i spec'd new iron from heyboer with a reduced sec. V rating since the original PT would have had over 600V B+ to plates.

should anyone ever need a PT to fit a laney 60, the heyboer p/n is HTS-12134. secondaries are: 800VCT @ 250mA; 6.3V @ 5A. they are built with hi-watt style endbells.


--pete