diytube.com

I've been researching filament supplies for a 811A. Most people say it has to be DC. A few have said you can use AC if you use a balancing pot and maybe try Steve Bench's circuit. But everyone likes Rod Coleman's. I like AC because it's easier, and makes less heat. DC has to have a bigger filament transformer, lots of big caps and will make something like 10 watts of heat. Rod Coleman's circuit needs a bigger transformer (more current and voltage), a circuit board, a bid UGLY heat sink and will make like 24 watts of heat.

A 845 will make a lot of heat itself, but it's the good kind of heat (above the chassis).

hang on, why would DC supply cause "more heat"...? i thought filaments were designed to run at X temperature, and it is up to the designer to provide the correct filament voltage to achieve that temperature.....

DeathRex wrote:I've been researching filament supplies for a 811A. Most people say it has to be DC. A few have said you can use AC if you use a balancing pot and maybe try Steve Bench's circuit. But everyone likes Rod Coleman's. I like AC because it's easier, and makes less heat. DC has to have a bigger filament transformer, lots of big caps and will make something like 10 watts of heat. Rod Coleman's circuit needs a bigger transformer (more current and voltage), a circuit board, a bid UGLY heat sink and will make like 24 watts of heat.

A 845 will make a lot of heat itself, but it's the good kind of heat (above the chassis).

Claudio from audiodesign uses AC on the fils in one of his 845 designs.

http://www.audiodesignguide.com/Fiat/845Amp2011.html

AC on the fil of an 845 may be marginal. You would have to listen for yourself to decide whether the level of hum is acceptable. 845s require 3.25A per filament. 811As require 4A. 4A may be pushing it with AC.

The 813 requires 5A, and the AC hum is not acceptable for hi-fi.

However, the 811a uses 6.3V, the 845 and the 813 use 10V. You may be able to get away with AC on the fils of the 811As.

nyazzip wrote:hang on, why would DC supply cause "more heat"...? i thought filaments were designed to run at X temperature, and it is up to the designer to provide the correct filament voltage to achieve that temperature.....

Just from the rectifiers. With 6.3 volts and 4 amps, you need some TO-220 case rectifiers with heat sinks. Wish there was a 0.2 volt drop rectifier that I could use.

kheper wrote:Claudio from audiodesign uses AC on the fils in one of his 845 designs.

http://www.audiodesignguide.com/Fiat/845Amp2011.html

AC on the fil of an 845 may be marginal. You would have to listen for yourself to decide whether the level of hum is acceptable. 845s require 3.25A per filament. 811As require 4A. 4A may be pushing it with AC.

The 813 requires 5A, and the AC hum is not acceptable for hi-fi.

However, the 811a uses 6.3V, the 845 and the 813 use 10V. You may be able to get away with AC on the fils of the 811As.

I don't have any high sensitivity speakers. I did get a 2 X 6.3 volt 6 amp transformer, military grade.

I've read that DC will give enough voltage variation from one end of the filament to the other to cause shorter tube life.

John

I've heard of people switching the positive and negative of the filaments with a relay every turn on, to help give even wear of the filament.

I ordered a couple of Rod Coleman's regulators, but I already have a big 6.3 volt transformer for the 811s. His regulator requires about 4 volts more than the filament voltage. I also ordered a 12V transformer for the 838s.

DeathRex wrote: I also ordered a 12V transformer for the 838s.

BAH. I was going to add that I have a NOS Thordarson 10V@12A - if you were going to run 845s on AC. I was planning to use it for a low voltage chip amp, but it seems a waste of a nice transformer for that purpose.

After you build it, report back as to how much hum the 811A has.

Well I don't have a transformer to run Rod's regulator with 811s. So I'm going to use either just DC, or a transistor in series to help eliminate ripple. Michimori uses a 2SD2052 pass transistor but no one stocks it. I found some power transistors at mouser, but don't know if they are compatible:
http://www.mouser.com/ProductDetail/Fai ... DRzA%3d%3d
http://www.mouser.com/ProductDetail/Fai ... tSYw%3d%3d
http://www.mouser.com/ProductDetail/Fai ... yOEg%3d%3d

Here's the 2SD2052: http://pdf1.alldatasheet.com/datasheet- ... D2052.html

Here's Michimori's 811A: http://www6.plala.or.jp/Michi/renewal81 ... l811A.html

The transformer I have is a 2 X 6.3 @ 6 amps military, sealed from Triad. I'm betting it would do 8-10 amps, it weighs more than Hammonds' 185G12 doing 2 X 14 amps. It's also selectable from 125, 115, 105 volts in. With 123VAC connected to 105 input I could get 7.3VAC out.

Try a CCS

Geek wrote:Try a CCS

Might be what Rod uses? But a filament regulator seems to need more voltage, like a LD1117 only needs 0.8 volts input above the out to regulate, but as a current regulator it needs 0.8 volts plus 1.25 through the current resistor.

I did put a current regulator on a RTF Sonneberg radio's filaments to replace the NTC when I changed it from 220 to 120 volts. It takes a long time for the filaments to come up now. Maybe a GE NTC in series with the transistor. A CL-150 will do 5 ohms cold and about 0.15 at 4 amps, but will still drop 0.6 volts.

Don't know if the 1000uf cap on the transistor's base in Michimori's makes a delay.

I once built a 811A SE amplifier (15 W) with AC-filaments and hum balancing pot.
The un-weighted hum level was -55 dB (ref. 10 Vrms) without any negative feedback.
When added 16 dB GNFB, hum level was -74 dB.

I got a pile of parts from mouser. I put in a bridge rectifier (4 X MBR1645 17 amp schottkys), 2 X 8200 caps, 0.1ohm resistor, and 2 X 8200 caps and I have 0.5 volts of ripple on the oscope. Would that be good enough?

I tried instead of a resistor I put in a GE NTC CL-150 (5 ohms cold, .014 ohms at 4 amps) It gave me a slower ramp up, well slower than instant. I'm going to time it and see if I can make a chart.

Here's the warm up delay caused by a CL-150 (NTC inrush current limiter) inbetween 2 X 8200uf caps and 2 X 8200uf caps going to the filaments of a 811A.
The left side is volts, the bottom is time.

Note; I took the voltage reading by the second up to 10 seconds, then every five seconds. I couldn't figure how to space it out in the software, so I filled in the times.

It's got to be less jarring to the 811, but still still quick enough to get your Soft Cell fix while you're in the mood.

It takes a little over 3 seconds before the filaments start glowing, maybe 2 or 3 seconds after that the tube should be conducting.



FIRST DOWN .... OLD MISS.