• 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

Regarding the transformers, and shielding from line noise, the 369KX has an electrostatic shield between primary and secondary, and the filament transformers are split bobbin design and thus do not need a shield as they will inherently reject line noise.

The 733A has no shield, but it will be fed from a soft start circuit with integral RFI and DC filters. PCX was blowing these out at 50% off a month or two ago and I picked two up specifically for the 733A both to soften the start-up and filter the line. They're similar to this unit:

NewClassD Soft Start
 
That makes sense. I thought about using Xenon rectifiers but decided that SS would be smaller, lighter, and a lot safer.
xenon has (at least c3ja has) ~1200Vmax, xenon is nicer than mercury because you don´t need to wait so long before applying the anode voltage (mercury evaporation) bugger that filament though (2,5V 9A)
edit: you still have to wait, but is shorter than mercury ,AFAIK
 
Last edited:
xenon has (at least c3ja has) ~1200Vmax, xenon is nicer than mercury because you don´t need to wait so long before applying the anode voltage (mercury evaporation) bugger that filament though (2,5V 9A)
edit: you still have to wait, but is shorter than mercury ,AFAIK

3B28 has a PIV of 10KV. I would have used it in a bridge setup. Too much space, too much filament voltage to supply, too much HT topcaps!
 
The filament supply chokes arrived today, along with the two small 12V transformers for the fan supply (if I use a fan). Now just waiting on the one 733A and the custom Monolith OPT and choke.

Time to start exploring layouts and decide on a chassis.
 

Attachments

  • 833 filament chokes.jpg
    833 filament chokes.jpg
    282.8 KB · Views: 2,084
I tried the 3B28, still have a few around here somewhere. I was hoping for some cool tube glow, but they are so dim that you can't see the glow in a lit room...or anywhere near the 833A.

Breadboard and test everything before thinking about the chassis. With 100 watts of heater power in the 833A hum will be a problem. I HAD to use DC for the filament. I would get audible hum in the speakers with ONLY the filament transformer switched on, NO B+. Moving things around did not help. I needed a big fat bridge some big caps, and a choke wound with something resembling welding cable!

The 833A has so much emission that you will measure negative voltage on the plate from heater power alone. Put a voltmeter between the heater and grid or plate, then light the filament with AC. The plate will go a few volts negative from the electrons hitting it.
 
I tried the 3B28, still have a few around here somewhere. I was hoping for some cool tube glow, but they are so dim that you can't see the glow in a lit room...or anywhere near the 833A.

Breadboard and test everything before thinking about the chassis. With 100 watts of heater power in the 833A hum will be a problem. I HAD to use DC for the filament. I would get audible hum in the speakers with ONLY the filament transformer switched on, NO B+. Moving things around did not help. I needed a big fat bridge some big caps, and a choke wound with something resembling welding cable!

The 833A has so much emission that you will measure negative voltage on the plate from heater power alone. Put a voltmeter between the heater and grid or plate, then light the filament with AC. The plate will go a few volts negative from the electrons hitting it.


Yes, definitely DC on the filaments. I'll be using Rod Colemans excellent filament supply boards which use a CCS and Gyrator to control the current and voltage. I have to parallel two per side at 5A each to get to 10A. The chokes in the post above are rated at 2.5mH at 10A, so they should do.

Below is the PSUD simulation for the supply to the Coleman boards. Ripple will be ~40mV going into the boards, and in the uV range at the tube filaments.

I will definitely be breadboarding/testing before assembly, I just want to get an idea of the size of chassis needed - I'd like to keep it around 18"x22"x4.25"deep if I can.
 

Attachments

  • 833 Filament PSUD.jpg
    833 Filament PSUD.jpg
    168.3 KB · Views: 2,069
Tubelab.com: Don't expect a lot of glow from a 3B28. You will get a blue glow from the 866A (the mercury counterpart). If you look carefully (BE CAREFUL, HIGH VOLTAGE!), what glow you see is between the plate and the filament. Preheating the filament before applying HT is mandatory.

There was a comment about the C3J thyratron. If that tube has a PIV of 1200 volts, the RMS AC supply must not exceed 850 volts. Hopefully, you won't be "pushing your luck" using a 800 volt transformer. The grid controlled regulator/control is a nice concept...thyratrons are sort of like SCR's, once they fire, they conduct until the AC supply goes through almost zero crossing. But the grid control can deal with the conduction cycle and other nice things...

Last but not least is the 833A. Nice tube. Still used in some older broadcast gear still in service. The 833A was used (in a pair) as a class B modulator in older Gates, RCA, and other 1 kW am transmitters. Two of these (in parallel) were used as RF final amplifiers at the 1 kw level. The filament transformers (secondary voltage 10 V at 20 A for a pair of 833's) was center tapped and the dc return circuits were taken off the center tap. Fixed bias was normally used with this tube to set the operating point (and protect it).
 
My main worry about the single chassis design is hum from the 10A filament supply, even with the AC part 20"or so away from the signal circuitry. I'm kicking around the idea of a separate small steel chassis just for the filament trannie, chokes and caps, with an umbilical connection to the main chassis. It would also give me a lot more room in the main chassis for all the rest of the parts.

Stay tuned.
 
Magz: Suggest using a cooling fan and the heat radiating connectors on the plate and grid connections to the 833A. The ceramic filament "socket" supports the tube. Avoid mechanical stress to any of the connections to the 833A to avoid damaging the glass to metal seals. The 833A is installed filament connections down, with air directed to the top of the tube to cool it in CCS (continuous commercial service) use. This tube is sensitive to filament voltage and in broadcast use, the tube wants 10 volts on the filament, measured at the tube. Since audio output use is essentially equivalent to continuous broadcast operation, use the CCS ratings to maximize tube life.

In the older tube broadcast transmitters using the 833A, the heavy iron components are located at the bottom of the cabinet (plate transformer, HT choke, modulation transformer, modulation choke, etc.) and the power components (modulator, RF final) are located near the top of the cabinet (on separate chassis assemblies). The cooling fan is located on top of the cabinet to force air on the power tubes. Filament transformers for the power tubes were located on the cabinet side walls, and a variac typically was used to fine adjust the filament voltage.

If you want a tube power supply with a bridge rectifier, four 3B28 or 866A will work with plate current on the order of 250 mA. In broadcast use, the 8008XE or the 575A was used. I liked the 575A with its higher peak inverse voltage and less prone to arc back. Virtually all the tube broadcast transmitters in the 1 kW range were powered off standard 220 volt single phase AC (same power that runs your electric stove, central air conditioning, well pump, etc.).

Did you think of locating a metal cabinet (approx. 36-48 in high) equipped with standard 19 in rack rails that accommodate the power supply and audio subassemblies "under one roof"? The older Bud cabinets had the rack rails in front and a door in the back, which could be closed and keep people and pets away from the HT...
 
Breadboard and test everything before thinking about the chassis. With 100 watts of heater power in the 833A hum will be a problem. I HAD to use DC for the filament. I would get audible hum in the speakers with ONLY the filament transformer switched on, NO B+. Moving things around did not help. I needed a big fat bridge some big caps, and a choke wound with something resembling welding cable!

Why not use these instead of heavy iron? exists also types with active pfc
gen_Q-120.jpg
 
Magz: Suggest using a cooling fan and the heat radiating connectors on the plate and grid connections to the 833A. The ceramic filament "socket" supports the tube. Avoid mechanical stress to any of the connections to the 833A to avoid damaging the glass to metal seals. The 833A is installed filament connections down, with air directed to the top of the tube to cool it in CCS (continuous commercial service) use. This tube is sensitive to filament voltage and in broadcast use, the tube wants 10 volts on the filament, measured at the tube. Since audio output use is essentially equivalent to continuous broadcast operation, use the CCS ratings to maximize tube life.

In the older tube broadcast transmitters using the 833A, the heavy iron components are located at the bottom of the cabinet (plate transformer, HT choke, modulation transformer, modulation choke, etc.) and the power components (modulator, RF final) are located near the top of the cabinet (on separate chassis assemblies). The cooling fan is located on top of the cabinet to force air on the power tubes. Filament transformers for the power tubes were located on the cabinet side walls, and a variac typically was used to fine adjust the filament voltage.

If you want a tube power supply with a bridge rectifier, four 3B28 or 866A will work with plate current on the order of 250 mA. In broadcast use, the 8008XE or the 575A was used. I liked the 575A with its higher peak inverse voltage and less prone to arc back. Virtually all the tube broadcast transmitters in the 1 kW range were powered off standard 220 volt single phase AC (same power that runs your electric stove, central air conditioning, well pump, etc.).

Did you think of locating a metal cabinet (approx. 36-48 in high) equipped with standard 19 in rack rails that accommodate the power supply and audio subassemblies "under one roof"? The older Bud cabinets had the rack rails in front and a door in the back, which could be closed and keep people and pets away from the HT...

I have finned top caps for the 833, and plan to have a small fan (33cfm DC brushless) for cooling. I have read that too much cooling is detrimental to performance, and the Penta 833C has a maximum cfm spec of 40. What they used to do in some cases was to use the warmed air from the power supply chassis to cool the 833 to prevent over-cooling. I can do that by placing a fan under the 833 and sucking the warmed air from the chassis and directing it up and around the 833, which will be surrounded by a borosilicate glass tube.

For rectification I'm going with SiC schottky diodes, four in series per leg of the bridge, 4800PIV. My supply is choke input with a soft start circuit, so there will be a relatively smooth current draw from the supply.

A 3-4 foot tall cabinet will not meet WAF. I need to keep it to max 19"x23"x4.25".
 
Last edited:
There was a comment about the C3J thyratron. If that tube has a PIV of 1200 volts, the RMS AC supply must not exceed 850 volts. Hopefully, you won't be "pushing your luck" using a 800 volt transformer. The grid controlled regulator/control is a nice concept...thyratrons are sort of like SCR's, once they fire, they conduct until the AC supply goes through almost zero crossing. But the grid control can deal with the conduction cycle and other nice things...
yes i did mistake in concept, thanks to the zero cross circuit it would never fire thyratrons.
i like idea of having a full regulated anode supply (with tubes)