• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

The Midlife Crisis - My 833C Amp Build

-My wife, upon seeing the 833C triodes in person - "That's not an amp, that's a Midlife Crisis!"

-My reply, "Hey, it's cheaper than a Corvette..."


So was born my 833C SET monoblock project.

The 833C is a ruggedized version of the 833A, with thick graphite plates capable of 400W dissipation with forced air cooling to prevent melting of the glass envelope. See picture #1. It's a BIG tube, and I'll need some sort of containment vessel due to the exposed ultra-HV plate caps...I'm thinking a tempered glass cylinder on each mono.

This tube loves voltage, so I'm building with a B+ of 2300VDC, using a choke input supply. To do this I need a very high voltage PT, and a custom made first choke that can take the pounding of thousand of volts from the rectifiers; more on that in a minute. I've chosen the Hammond 733A for the 833 supply, bridge rectified with 16 Cree 1200V SiC schottky diodes per bridge, and the Hammond 369KX for the input/driver supply. One of the 733A is on backorder from Digikey until early April, but there's plenty to do before that gets here. Seven Panasonic T-HA and TS-HB caps in series will handle the C2 and C1 duties, respectively, with the appropriate balancing resistors.

The schematic is in picture #2. It borrows a bit from Michael Koster's Meteor, with a few differences, especially in the power supplies. I'll be using the K&K Audio versions of the Salas HV regulator to supply the bias voltages. Rod Coleman filament regulators (2 in parallel per 833C) will handle the 10V@10A filaments on each side. Driver is the 6E5P, triode strapped.

The key to all this is, of course, the 12K:4 series feed OPT. It will be custom made by Monolith Magnetics, with a 15kg amorphous core! They are also building the custom first PS chokes to withstand the pounding of the HT. They have guaranteed 8Hz to 60kHz small signal response and full power saturation at 35Hz max.

Full power, by the way, sims to be 195W at 2.5V input.

All this does not come cheaply. Total cost for parts will be ~$10K...but it's not a Midlife Crisis if it doesn't cost at least 5 figures, right!

Stay tuned.

.
 

Attachments

  • 2013-01-30 22.25.07.jpg
    2013-01-30 22.25.07.jpg
    33.6 KB · Views: 10,441
  • Midlife Crisis Amp.jpg
    Midlife Crisis Amp.jpg
    120.1 KB · Views: 10,346
Last edited:
Here's a snapshot of some of the parts I've gathered so far.

In the back are four Nytronic HV chokes (2260 W.V. at 10,00feet altitude) to be used as second chokes for both the 833C and 6E5P supplies - picked them up cheap at Fairradio.com.

You can also see the Cree Schottky diodes, the assembled Coleman regs (minus pass transformers until I figure out how I want to mount the heat sinks), Lundahl 13H chokes for the 6E5p choke input supply, the T-HA caps, two of the K&K CCS/Salas regs for bias, two huge Claritycap TC 430uF film caps for the 6E5P supply (got to get that grid current from somewhere!), etc.
 

Attachments

  • 833 amp supplies 1.jpg
    833 amp supplies 1.jpg
    395.6 KB · Views: 9,525
Yes, I know that's a possibility, but in reality the custom chokes were "cheap" compared to the OPTs, and I'd rather use a conventional wiring scheme on my first multi-KV amp!

BTW, the chokes will be 35H and <160ohm DCR, so combined with the 8H Nytronics chokes and lots of capacitance in LCLC I should have very low ripple. I'll post the PSUD files once I actually get the Xfrmrs and can measure the true resistances, up to now it's all approximations.

I've tried putting chokes in the - leg of a PS in the past (Lundahl "common-mode" configuration) and I preferred having all the inductance in the + leg. It just sounded punchier to me that way. Could have been my imagination, I suppose, but my perception was that it was significantly punchier in the + leg.
 
Remember no sharp edges . lock up the cats when that thing is on . I like tubelab bit about choke on the ground leg rather than the hv leg . Look at it this way choke inhibit change in current and the current flows in a loop so which side the choke is on does not matter to it's function in that loop. aarl books on hv 3.5k and up in the past have used the choke on the return leg ( at ground ) .
 
Remember no sharp edges . lock up the cats when that thing is on . I like tubelab bit about choke on the ground leg rather than the hv leg . Look at it this way choke inhibit change in current and the current flows in a loop so which side the choke is on does not matter to it's function in that loop. aarl books on hv 3.5k and up in the past have used the choke on the return leg ( at ground ) .

Yup, all solder joints smooth to prevent corona discharges.

Wire for 833 B+ will be rubber insulated high voltage test lead.

I'm even thinking about installing the 833 upside down so that all ultra-HV is completely contained in the chassis.

I'm allergic to cats.

Another trick is to isolate the choke core from ground, but the benefit of having Monolith do a custom job is that they guarantee it will hold up just fine under the pounding. Based on my discussions with Ward, I have faith...he's even crazier than I am! Of course, the chokes and the OPT will be gigantic, so fitting them plus three large power transformers onto a single chassis per mono that doesn't collapse under its own weight will be a trick.

Approximate OPT dimensions: 10"x10"x8" :D


.
 
Last edited:
  • Like
Reactions: 1 user
An online search didn't turn up any warnings about this phenomenon that I could find, so I sent an inquiry to Cree about it. Let's see what they say.

A definitive (and very quick!) answer from Cree:

Due to a negative temperature coefficient on forward current these devices parallel very nicely to multiply current.

This negative temperature coefficient of forward current allows us to parallel more than one die in a package without any unequal current sharing issues.

This behavior is unlike high voltage Si PiN diodes.

However if clients plan to use the SiC Schottky diodes in series to multiply voltage our engineers would suggest adding balancing resistors of a suitable value.

You do want to force voltage sharing among components . You have identified the issue.



So it appears that using them in series is fine if you encourage voltage sharing. I've seen that done in various references using just caps (10-47nF) and caps plus resistors (470K), so I asked them a followup question about which is optimum. I suspect that anything that evens out the voltage distribution will be OK.
 
If I may make a suggestion...why not put the CHOKE in the Negative Leg of you Bridge Rectifier...that way you will not subject it to any High Voltage...it does exactly the same job using it this way without the expense of having a Choke wound to take the pounding