211 IT amp, battery vs fixed bias

[analog_sa]

Originally Posted by Vincent77

LED bias is a bad idea, *unless* the stage is CCS loaded:
- Vf and dynamic impedance change with the signal!
- three green leds in series have rather highish impedance
- does not prevent the tube from running away (an increase in idle current is not throttled by the cathode resistor or anode CCS)

As it happens i have run a 5687 in a preamp in almost exactly the same configuration as shown for years. Briefly tried a CCS and soundwise it presented no benefits compared to transformer or choke load. As for the "highish impedance" of the LEDs or the possibility of the tube running away - well, it didn't run anywhere and it sounded much better than cathode bias. But theorising is great

[kevinkr]

I use fixed bias and batteries extensively in my designs for driver stage duties, but generally avoid Ni-Cad or NiMH batteries in the cathode circuit, preferring battery or negative supply derived grid bias instead. Works well with input transformers or in the case of a power amplifier driven by a line stage with an output transformer which can be floated. (I do this now)

I've used red and infra-red leds for bias and modulation of the small dynamic resistance has been less of a problem IMLE than the colorations imposed by the large electrolytics required for cathode bias with tubes like the 5842/D3A/7788, etc.

On some red leds I measured the vf varied only a few mV from 5 - 20mA forward current, sample to sample variation of vf was well under 20mV for these leds at a fixed current of 10mA.

Using fixed bias without provisions for adjustment may require selection of tubes in order for the circuit to functional optimally. This is a tradeoff I make with some of my designs.

Agreeing with cotdt I prefer fixed bias if for no other reason than the fact it eliminates what is usually a pretty questionable electrolytic capacitor from the signal path. (Small high quality films in the grid circuit IMO do much less damage to the sound) In power stages the ability to easily tune operating point offsets the convenience factor of auto bias (cathode bias) and eliminates that cap I mentioned..

Something I want to mention about battery bias - use alkaline or other non-rechargeable types for grid bias, this will prevent the dead battery zero bias excitement some of my friends have experienced when they did not listen to this bit of advice.. Check the batteries periodically. (Battery life in A1 applications is about shelf life.) Alkalines I generally solder, lithium-iron alkaline AA/AAA and any other lithium types I always use in holders for safety - just check the contacts periodically, probably not a bad idea to apply contact preservative at installation time.

[ColinA]

I used NiCad batteries as bias in the World Audio Design 2A3 amp.
This was a mod which came out on the long gone forum.
My thought was that NiCad cells loose their charge just sitting "on the shelf"
So if the amp had been switched off for some time and the cells were discharged what was happening?

I changed back to standard Cap & Resistor bias just to stop my brain worrying what happened at switch on.

The amp is still fine with 1940 era British 2A3 valves

[tomchr]

Originally Posted by Vinylsavor

1) Is there real advantage to use battery version (lover AC resistance to ground)?

I see no advantage. It's different. In one case you have a cap in the signal path, in the other case a battery. They are different. One is not necessarily better than the other. If I had the choice I'd prefer a cap.

I don't see any advantage of using a battery either. Personally, I wouldn't want the bias to be dependent on a battery that may have discharged since the last use.

As far as the impedance goes, capacitors are very well characterized. At least if you buy one that comes with a data sheet (pretty much any polypropylene cap on Digikey, Solen's caps, etc). You know what you get. With batteries, their internal resistance may be measured, but it's heavily dependent on the state of the battery. Also, I doubt you'll find data for their impedance vs frequency.

Each to their own, but I prefer to know what I'm getting.

Originally Posted by Vinylsavor

As has been mentioned above: No real need to use fixed bias. Cathode bias works extremely well and lets the tubes settle at the op point which is best for their emission which is left and the plate voltage. The downside is the bypass cap which is in the signal path which most often is an electrolytic. At your driver I see an electrolytic in the B+ anyways.

You could eliminate the B+ reservoir cap by connecting the cathode bypass cap from the cathode to B+. This is what Loftin & White did. It reduces the signal path on the output stage to the OPT, output tube, and the bypass cap.

Originally Posted by Vincent77

LED bias is a bad idea, *unless* the stage is CCS loaded:
- Vf and dynamic impedance change with the signal!
- three green leds in series have rather highish impedance
- does not prevent the tube from running away (an increase in idle current is not throttled by the cathode resistor or anode CCS)

Three HLMP-3507 green LEDs in series present a dynamic impedance on the order of 40~50 ohm at 10 mA. That's not that high, actually. Especially when it's up against the impedance of the anode load. It'll have a negligible impact on the gain of the input stage and will, if anything, reduce the THD ever so slightly.

~Tom

[Vinylsavor]

Originally Posted by tomchr

You could eliminate the B+ reservoir cap by connecting the cathode bypass cap from the cathode to B+. This is what Loftin & White did. It reduces the signal path on the output stage to the OPT, output tube, and the bypass cap.

Yes, the so called ultrapath cap, I mentioned that in my post...

Thomas

[kstagger]

I'm at work, so I can find my trusty GEC (and I'm not sure if this is even the right source) tube manual, but there was a discussion of the overload characteristics of fixed versus self-bias.

If I remember correctly, self bias has better overload even though it creates less power, creating the aural illusion of actually making the more power. Again, I could be way off base with this since it has been awhile since I've read the book.

[analog_sa]

Originally Posted by kstagger

If I remember correctly, self bias has better overload even though it creates less power,

And you are probably right in the context of an output tube. The current discussion is about the driver...

[kstagger]

Originally Posted by analog_sa

And you are probably right in the context of an output tube. The current discussion is about the driver...

ah, true... reading too fast as usual.

[tomchr]

Originally Posted by Vinylsavor

Yes, the so called ultrapath cap, I mentioned that in my post...

Oh. I guess I'm not up on the latest GeekSpeek. My bad...

~Tom

[cotdt]

what is ultrapath? i never heard of such a thing. anyone have link?