300B preamp - best operating point
 

Hello,

i am building a single stage 300B preamp and looking for best operating points for the 300B. High voltage supply will be around 440VDC. 300B should be operated with automatic bias and anode resistor.

Possible operating point would be:
Anode voltage: 225VDC
Voltage on kathode resistor: 45VDC
Votage on anode resistor: 440VDC - 225VDC - 45VDC = 170VDC
Anode current: 50mA
Anode resistor: 170VDC / 50mA = 3.4kOhm; Power loss 8.5W
Kathode resistor: 45VDC / 50mA = 900 Ohm

Any comments on this operating point? Does anyone have experience with 300B operated at low voltage/ low current as small signal gain stage?

regards,

Markus

Western Electric does.

300B Performance Graphs & Mechanical Dimensions
http://www.westernelectric.com/products/300b/300B.pdf

I'm not a fan of "automatic bias" and in this topology the quality of that capacitor had better be extremely good (i.e. read as expensive) and will probably still color the sound.

I've been designing with the 300B for about 14yrs (WE, JJ, and others) and strongly recommend using fixed bias. If you make plate current easily measurable it is very worthwhile to make the bias voltage pot adjustable so that you can tweak the operating point to suit.

Also very important is a good filament supply and I would strong recommend you talk to Rod Coleman about his CCS designs targeted at DHTs.

I have not used the 300B in a line stage but have heard 2A3 based line stages and they sound quite good. I would expect with the right 300B this could sound exceptional, and even with RC coupling will have a nice low source impedance.

I don't think there is a 'best' point for everyone. I found some people like lower current than what I like; some others like higher current...


to Kevin,

In a fix bias SE output circuit, how about the last cap of the power supply? It encloses the top of the OPT and ground (now also the cathode), providing AC current for the tube (along with the signal/amplified AC current). So it's still in the 'signal loop' as what I see it.

In an auto (cathode) bias circuit, one can connect this very last cap of the power supply in the same manner - enclosing OPT and cathode - WE connection. Now they should work in the same way. No?

Is there any 'pitfall' I didn't see?

In most designs the last cap in the supply is definitely in the signal path, this is obviously of direct consequence with transformer coupling, and less so in the case of RC coupling which the OP proposed to use here. The additional cathode bypass cap just compounds the problem and because of the low impedances involved may be a much larger value than utilized in the supply. Generally IMHO bigger does not equal better. :D
The WE connection would generally be subject to the same quality of capacitor issues as a supply capacitor - go to fixed bias and that WE cap becomes the equivalent of a standard supply bypass capacitor as well. Note that in some applications the WE connection is a mighty fine way to inject an excessive amount of supply ripple into a circuit node with gain.. It is also possible to add a (cathode bypass) capacitor and form a capacitive divider that cancels the supply ripple - in fact it lends itself to applications where the supply is noisy. Look up some of my old posts on the subject. Basically the cathode cap assuming the WE cap value is known would be as follows: (mu+1) x WEcap value. The value is not too critical as long as the value of the cathode cap is equal to or somewhat greater than the ratio of (mu +1) x WE cap.

In my supplies I usually use fairly small high quality film capacitors, and voltage regulation which is not generally based on the currently trendy shunt regulator topology often discussed here, although it could/should be superior to the series pass stuff I usually use. Ideally shunt regulation would be used with a CCS which would give you very good ripple and mains noise rejection..

I probably haven't really answered your question, but I don't think you missed anything important either. I believe most rational solutions can be optimized to work well, but what ultimately works best for you will have to be the result of experimentation. It's all about the compromises that are right for you, and I have a definite bias towards fixed bias on several levels.

Ah! voltage regulator, that's a top notch solution.

I've heard a commercially available item - JC Verdier 300B amp, which consists of an almost text book tube regulator for its HT supply. It's a really good sounding amp. (The down side is heat. That amp is very compact and full of tubes. HT is 600V before regulator and 400V after. Holy **** lot of heat!! )

I've been using the capacitive divider you mentioned and get good result in noise. My recent 2A3 uses 20uF PIO cap for WE connection and 100uF electrolyte on the cathode, very quiet:D

My last 300BSE also used similar setup in a period before I switched to fixed bias. Both bias methods sounded very similar. I couldn't quite distinguish them actually - especially instant A-B test was impossible for me...

Compared side by side, the JC Verdier wins hands down in the bass authority. Regulator rules in this regard !!

My 300B SE stereo amp is on two chassis with the amp on one and the supply on the other. The amp chassis has six tubes, and the PSU fourteen.. :eek: :D The raw supply is 650V and the regulated outputs are 400V.. The pass tubes are KT88 connected in tetrode mode - it all runs very warm, particularly problematic in the summer. The amp however does go all the way down to 10Hz at full power.. (I built the first three commercially in 1999-2000, and finally a somewhat cheaper one for me in 2004.) Some 22yrs ago I decided voltage regulators would be part of my design philosophy and to date I have only designed four amps that don't use regulated supplies. (Out of the dozens I have designed.) Guess I am pretty sold.. :D

Thanks to everyone for your answers especially Kevin.

I know how important the filament supply will be and designed (built) the following:
1.Power transformer 13VAC/2A
2.Schottky bridge rectifier
3.Choke input filter 60mH/2A
4.Smoothing capacitor 10mF Epcos Sikorel
5.Voltage regulator with LT1085 high performance regulator set at around 10V
6.Another smoothing capacitor 10mF Epcos Sikorel
7.CCS based on LM317
Calculated/ simulated ripple voltage of this supply is exceptional low, actually lower than noise of the LT1085/ LM317.

For the cathode capacitor i purchased cheaply 1000pcs. of good old axial Roederstein film capacitors 2.2uF/ 250VDC. I need to parallel quite a lot of this capacitors but there is room inside the case.

For the high voltage voltage supply i purchased a batch of Epcos MKV high performance film capacitors. Plan is LCLC filter with 2x15H and around 2x20uF, but need to simulate if ripply is acceptable.

Back to the original question: Any suggested operating point?

That's interesting... It's pretty easy to find 30uF or 40uF 200V metallized plastic film capacitors at not too great cost. Since a 300B cathode bias resistor is likely to be about 900 to 1000 ohms, it should be possible to use as low as 100uF of bypass to get the f3 down to a couple of Hertz. Bypass those with some 2uF or so film/foil caps to guild the lily. That should work well, right?


--