diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Tubes / Valves (http://www.diyaudio.com/forums/tubes-valves/)
-   -   Eliminating Cathode Bypass Caps In SET Amplifier (http://www.diyaudio.com/forums/tubes-valves/216543-eliminating-cathode-bypass-caps-set-amplifier.html)

Davec113 20th July 2012 09:35 PM

Eliminating Cathode Bypass Caps In SET Amplifier
 
I would like to eliminate the electrolytic cathode bypass caps and try a different way to bias my amplifier, it's a Tubelab SSE with 12BH7A driver tube and EL34 power tubes running in triode mode with no feedback. Currently, the cathode bias is set up as shown in the schematic, linked below:

Simple SE schematic

I would appreciate any suggestions! A loss of 6-9 dB in gain is ok, and would be welcome.

soundguruman 20th July 2012 10:01 PM

Quote:

Originally Posted by Davec113 (Post 3099168)
I would like to eliminate the electrolytic cathode bypass caps and try a different way to bias my amplifier, it's a Tubelab SSE with 12BH7A driver tube and EL34 power tubes running in triode mode with no feedback. Currently, the cathode bias is set up as shown in the schematic, linked below:

Simple SE schematic

I would appreciate any suggestions! A loss of 6-9 dB in gain is ok, and would be welcome.

I think yo might lose some low frequency too, if you remove the cathode bias cap.
I think you can reduce the value "some." The existing is 1500uF and that can be reduced. But you may not be able to eliminate a bypass entirely, just from the standpoint of low frequency roll off.
I think you can make the gain less by changing the plate resistor.
(link)
Cathode Bypass Capacitor Calculator
You see with a 1500 uF bypass, you are well down to below 20 Hz,
With a 22 uF bypass, you are rolling off at about 100 Hz, and 20 Hz is gone.

In this set up, a voltage regulator is used to supply voltage to the plate
resistor, and I am thinking that by adjusting the gate voltage from R19,
you might be able to adjust the gain of the thing...maybe I'm wrong but
it looks that way.

TheGimp 20th July 2012 10:13 PM

1) LED Bias
2) Fixed Bias

andyjevans 20th July 2012 11:14 PM

The best way to do this in the driver stage - and what I've done myself - is to use a DHT in filament bias. This requires some knowledge, but isn't complex though it does require a DC filament supply to each DHT. You'd have to change your amp - add stuff and make an additional hole for a second tube in the driver stage.

My driver stage is a 4P1L (mu=10) which would give you about the gain you desire. I run it into a Hammond 126C interstage but a 126B may be better. Cathode resistor is 20 ohms so no need to bypass. 165v on the plate, 15v supply to the filaments in filament bias, so this feeds the filament and cathode resistor in series and the resistor sets the bias.

I loathe cathode bypass caps - they kill the sound and make it bland. Electrolytics are worst but polypropylene aren't that good either, in fact I don't want caps anywhere near the signal path. I tolerate a cathode bypass of the output stage, but these days I use PP amps anyway so use CCS and pentode sinks under the diff pairs.

andy

kevinkr 20th July 2012 11:14 PM

Quote:

Originally Posted by TheGimp (Post 3099211)
1) LED Bias
2) Fixed Bias

:up:

Davec113 20th July 2012 11:25 PM

Quote:

Originally Posted by andyjevans (Post 3099262)
The best way to do this in the driver stage - and what I've done myself - is to use a DHT in filament bias. This requires some knowledge, but isn't complex though it does require a DC filament supply to each DHT. You'd have to change your amp - add stuff and make an additional hole for a second tube in the driver stage.

My driver stage is a 4P1L (mu=10) which would give you about the gain you desire. I run it into a Hammond 126C interstage but a 126B may be better. Cathode resistor is 20 ohms so no need to bypass. 165v on the plate, 15v supply to the filaments in filament bias, so this feeds the filament and cathode resistor in series and the resistor sets the bias.

I loathe cathode bypass caps - they kill the sound and make it bland. Electrolytics are worst but polypropylene aren't that good either, in fact I don't want caps anywhere near the signal path. I tolerate a cathode bypass of the output stage, but these days I use PP amps anyway so use CCS and pentode sinks under the diff pairs.

andy

Thanks, if you have a schematic for the driver stage you are talking about I'd love to see it. Also for the PP amp, I am already thinking of an amp that will top the SSE, but I'm also interested in modifying the SSE... so far by bypassing the el34's cathode caps and replacing the last electrolytic in the ps with a 100 uF film cap. Next big thing in my system is looking like a hirez usb dac, my current dac is 20 years old.

Davec113 21st July 2012 12:03 AM

Quote:

Originally Posted by TheGimp (Post 3099211)
1) LED Bias
2) Fixed Bias

Thanks, is the sound of LED biasing better than a resistor/cap?

Also, the driver stage has a CCS loaded anode, so if the cathode cap isn't there the CCS will just push the ac through the resistor, would the driver stage work without any cathode cap at all, that is without rolling off bass freqs?

Wavebourn 21st July 2012 12:22 AM

Cathode bias with resistor/cap provides feedback by DC current, so current through the tube stays stable. If you eliminate it you need to take care of bias. No free lunch. The best way of course is to ground cathode and provide a negative bias voltage on the first grid from additional voltage source. Adjusting this voltage you can adjust tube current. However you can achieve similar result using Zeners, or diodes, or LEDs, in cathode, but you can't adjust bias such a way.

Davec113 21st July 2012 01:19 AM

Quote:

Originally Posted by Wavebourn (Post 3099301)
Cathode bias with resistor/cap provides feedback by DC current, so current through the tube stays stable. If you eliminate it you need to take care of bias. No free lunch.

I was wondering, since the driver tube's anode is ccs loaded, if it would be possible to get rid of the cathode cap in the driver stage as the ccs would stabilize the current?

andyjevans 21st July 2012 09:17 AM

Going back to filament bias - I think this is the "free lunch". No input cap that you'd have with battery grid bias, and a cathode resistor that doesn't need to be bypassed. I'd prefer the sound of a good 20 ohm wirewound resistor (needs to be as good as possible - I use 20W DALE surplus stock off ebay). I've compared all these options - including LEDs etc in the cathode - and to my ears a properly executed filament bias is my preferred sound.

It's unconventional and there are only a few tubes you can use comfortably at lower voltages but these include the 26 and 4P1L, which is enough for me. Basically for a driver stage you need a bias voltage of up to 15v and a low filament voltage ( 26 = 1.5v, 4P1L = 2.1v) to get away with a supply voltage of around 20-24v into a filament board such as Rod Coleman's. For a preamp, you can use a supply of as little as 12v and a bias voltage of more like 4-5v. Easy stuff - I've even used a standard DC 12v mains supply! Choke input is even better - the supply really needs to be uber-clean. But satisfy these conditions, and in my view you have your "free lunch".

Andy


All times are GMT. The time now is 12:32 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2