813 SE build

[santa]

Hello,
I am just starting to breadboard an 813 SE amp and I thought I would post my progress here.
This is my third or fourth big SE amp. Previous amps used the Svetlana 572-10 and 6BM8 mu follower as a driver.
This time I want to go with some lovely 813's that I have.
I'm planning to have them at the SF DIY show so I gotta get busy.
I have a pretty good handle on what it takes to make this happen but as I go I want to hear from others on what I can do to improve my build as my skills are still developing.
My speakers are very efficient so I dont need to squeeze every watt from the tubes.

Power supply is 1500 VCT to 6D22s rectifier, 8H 300mA choke, 220uF cap.
Driver is an Aikido 6SL7/6SN7@300V B+ (derived from the 813 plate supply)
I'm a bit concerned about overvoltage voltage on the driver tubes. At first I considered just using resistors to drop the voltage to the driver but what if I lost a filament? The 6D22s starts up plenty slow and smooth but I can see a potential problem if I lost a filament in a driver tube.

I'm starting to think that a seperate transformer for the driver b+ would be the safest way to go.

What options do I have here?

[Ryssen]

I'm starting to think that a seperate transformer for the driver b+ would be the safest way to go.

+1

[Fuling]

Safer and might sound better.

[costis_n]

Consider that at startup, there well be still some time before the 813's conduct fully, which could easily mean the poor little Aikido will see 1KV on the supply.

[santa]

I will be using the 6D22s damper diode for the B+. The 813 will be heated by the time the diodes start to conduct and if I lost an 813 filament the b+ would rise to something like 1060V I think.
Not much would live very long at that voltage level so I'll use a 230 VAC tranny and CLC filtering for the front end power supply.

Previously I used AC for the filaments but now my speakers are about 97dB@1W so I am concerned that I may not be able to null the hum enough for it to be acceptable.

Considering the sensitivity of my speakers should I persue the DC heating approach for the 813 or just go with the AC on the 813 filaments?
I can accept a heatsink if I must

Starting to layout the top plate.

[radiostar]

I think you need DC for heating. Which type is 813 on the photo?

[the_manta]

Looks good
I have a couple of questions:

• What power transformer and choke is that ? I like the finishing of the end bells. What kind of Hammond OPT is that ? 1638SEA ?
• You're using a brass chassis, I assume. Are you gonna coat it or something ? Those are getting brown/black doe to oxidation. After some time you see every fingerprint due to the acidic sweat which catalizes this reaction.
• Since your Giant-7 sockets have round edges, they aren't chinese. Are those the russian ones or USA-Johnson ? Are those worth the extra money ? (One should concern that an 813 draws 5A heater current, so the contact material has to be of high quality. But till now I didn't have problems with chinese sockets)
• The round miliampere meter looks really good. Is that a chinese one ? The Simpson meters are really expensive and I'm still looking for cheap old-fashioned meters.
• Is the 6D22S save to use at Voltages up to 1kV ? I've got a full box of EY500A, the european version. There are only values for short voltage peaks in the datasheet (as occuring in TV sets) but no continous ratings.
• Speakers with 97dB, nice. Do you really need the Power of a 813 ?

[santa]

Radiostar, the 813 in the photo is RCA and has an open filament.
I have a pair of VT-144/813 new in box dated 1944 to use.

Manta, The PT is a custom wound Edcor with hammertone spray paint on the bells. Choke is a Stancor 8H 300mA, OPT is Hammond 1628 SE 5K. Not the best choice but it is what I had on hand.
We'll see how my HF response looks.

I will have to have the top plate clear coated or possibly antiqued as it will corrode quickly
It is 4 mm thick and easily able to support the transformers.

The sockets are USA johnson NOS,milliameter is chinese and accurate to about 3%.

The 6D22S is just loafing at 1K. I have been using them for 10 yrs at 1KV and have has zero failures. (I have new ones for this build.)

I really dont need the power afforded by the 813 but I have had the tubes for over a decade and I want to get them in service. I have a large room and like my jazz at somewhat realistic listening levels

[tubelab.com]

I built an 845 SE and I needed to use DC on the filament to get rid of 60 Hz intermodulation products. If you use AC and play a 1KHz tone through the amp you can see 940 Hz and 1060 Hz IMD products in the output. They are masked by the level of the fundamental tone so you dont hear them in musical applications, but they cause listener fatigue. All my DHT amps have used DC heating, but the DC must be clean. Sawtooth ripple on DC is worse than clean AC because of the higher order harmonics.

[Rod Coleman]

Originally Posted by radiostar

I think you need DC for heating. Which type is 813 on the photo?

Here's a pic of my 5A solution for 813 heating.

It's buffered on both sides of the filament to keep the rectifier and mains noise right out. It's CCS controlled, using a transistor Vbe reference - no reference diodes or chips to put the noise back in.

If you have a chassis with 1/8" [3mm+] thickness or more, the controller can usually use that as a heatsink, or you can add a piece of extrusion to cool it further.

There's quite a few of these - as self-assembly kits - out there heating Transmitters now!