Operating Point of 12AT7

[rlyach]

While I am patiently waiting for my amp to arrive I have been trying to understand the circuit. Keep in mind I am use to dealing with solid state circuits and have only recently ventured into the world of tubes. I have downloaded the anode current curves for both the 12AT7 and the 2A3 (RCA datasheets). From reading the schematic and manual, I have calculated the operating point of the 2A3 at Vgk~-60V, Va~290V and Ia~50mA. The 12AT7 is biased at Vgk~-2.5V, Va~227V and Ia~3.4mA. The 2A3 looks to be biased in the sweet spot but I am a little confused about the 12AT7. Using the constant current circuit and the fixed bias of 2.5V, it appears that the operating point could be moved down in voltage to Vgk~-3.0V. Then I noticed that the C4S circuit board has a place for bias resistors to set the LM431 to whatever voltage is needed. When I looked at the gain of the first stage, and assuming the mu of the 12AT7 is 60, there may be too much gain to move the bias point down to -3.0V. I was wondering if this was originally tried and found to clip on full signal. Then, the jumper was added to short the LM431 to 2.5v. Just curious if I should even bother playing with the bias point.

Also, I have been reading a lot of stuff on this board and have purchased a set of Jantzen Audio Superior Z-Caps for the coupling and parafeed caps using the same values as the original design. The Mundorf silver in oil were a little too pricey. I was toying with adding power supply chokes but after calculating a nominal ripple of <7V I think I will skip the chokes for now. Does that make sense or will the chokes add a marked improvement in performance. I plan on driving a pair of Orca speakers and no sub-woofer with this amp.

Thanks,

 

[Paul Birkeland]

Biasing at 3V vs. 2.5V will move the plate voltage up a little bit (maybe to 250V). 

There is some required compliance for the operation of the C4S, which can become a limitation for the peak voltage swing of the driver tube if it is not addressed, and on the other end if the plate voltage is very low, you will run out of room on the other peak.

If you look at the image of the plate curves, the red dot is the approximate quiescent operating point of the 12AT7 in the Stereomour, and the red line is a practical limit to the available B+ (actual B+ minus the compliance of the load).  What you'll notice is a lovely balance of the ratio of the two red line segments when divided by the operating point.

When you move to 3V of bias, you bump up a bit more on the high voltage limit, while having greater compliance below. 

Now, if you did bias the amp at 3V, I think it would still work just fine.  In fact, I have experimented with some other tubes in the Steremour that took more bias voltage, while at the same time having a lower resulting plate voltage, and everything is still balanced.

If you'd like to mess around with operating points, the PC board in the Steremour offers some flexibility, and I'm sure PJ could suggest some fun things to try!

-PB

 

[rlyach]

Thanks for the post. I guess my anode current plot was not accurate. The one I downloaded showed a much wider spread between -1 and -2 volts, leading to a possible non-linearity issue and which is Why I wanted to move the bias point up. The curves you supplied look very reasonable. Thanks for the clarification. I can't wait to get building.

 

[Paul Birkeland]

Yeah, plate curves are a funny thing, it's important to remember that they were hand drawn.  If you look at the GE vs. RCA datasheets, they can suggest differences in linearity, especially in the lower plate current regions. 

I actually went to the RCA datasheet first to pull the curves from, and they looked so nasty at low currents that I went looking for a more reasonable representation. 

 

[Paul Joppa]

With the current source as a plate load, the variation in plate voltage from tube to tube is fairly large, maybe +/-30% or so. We've had difficulty with that in the early days of the Paramour, so for Stereomour I aimed at making sure there was plenty of voltage margin, both positive and negative.

For those who care:

On negative excursions, I prefer to not let the grid get within less than 1v of the cathode potential, because the grid will begin to conduct some current and introduce nonlinearity.

On positive excursions, as the tube ages it will demand more and more plate voltage in order to maintain the current. Having a healthy margin means the tube will continue to work for a longer time. (In the old days, there were also some C4S LEDs that had excess voltage drop, causing increased plate current and hence too much plate voltage.)

I believe the 12AY7 or 6072 will drop in without change; that's a premium tube designed specifically for audio and has a pretty good reputation, with a mu of 40.

By adjusting the bias you can use other lower-mu tubes, as long as they have 12-volt heaters - 12AU7 with mu=17, 12FQ7 (a 9-pin version of the 12SN7, mu=20) or 7861 (12-v version of the 5670, which is used in Paramount, with mu of 35) for example.

 

[rlyach]

Thanks Paul. I guess there are some drawbacks to a constant current load. One being the very high output voltage swing (flatter load line) and the other is a lower dropout voltage. I think you need about 13 volts to make the C4S load work, two 6 volt diodes and the resistor drop on the PNP, but it stays more linear all the way up! There are a lot of neat things about this circuit. One more that I can't wrap my mind around is the output transformer primary connected to the cathode bias of the 2A3 instead of Gnd.

 

[jrebman]

Randy,

Another benefit of  the c4s is better isolation from power supply noise as well, and the other benefits will become far more apparent when you hook up those Orcas .

Actually, the more you look at and study the BH products, the more you will find and and appreciate the incredible level of engineering that has gone into them.

I choose to stick with bottlehead, not only because I can build, customize and voice them to my taste, but also because they are simply top notch sonic peerformers -- and this is coming from somebody who has Owned Shindo and Audio Note, etc. in the past.

BTW, glad to see Paul mention the 6072 -- a really nice tube and one I had planned on trying in my stereomour once I get it completed.

Welcome, but be forewarned, this stuff is highly addictive!

-- Jim

 

[Paul Birkeland]

Thanks Paul. I guess there are some drawbacks to a constant current load. One being the very high output voltage swing (flatter load line) and the other is a lower dropout voltage [...] One more that I can't wrap my mind around is the output transformer primary connected to the cathode bias of the 2A3 instead of Gnd.

I can't see how a little extra gain and a lot less distortion (approximately horizontal load line) are drawbacks.  In the sense of messing with different tubes, the CCS makes things very, very easy.

You could connect the bottom of the primary of the output transformer to ground if you wanted to, this is a bit more of a traditional approach.  One advantage of going to the cathode is that you shave off some DC voltage from the parallel feed capacitor, increasing the safety margin. 

 

[Grainger49]

There are getting to be a lot of guys in the Orca club too.

 

[rlyach]

I think it is too late.... I am already hooked. It all started when my son and I built a guitar amp. It sounded so good that we built another out of a 1957 Gilbranson organ. My son did the design changes. That turned out even better and both amps are now in the recording studio he works at. You can see the head here (http://www.myrmidonstudios.com/?page_id=1251). Now I want a SET stereo amp. I looked all over the internet for designs and parts and came across bottlehead. I was very impressed with what I saw. Then I came across the orcas, and I also ordered a matching base. Then I purchased a DAC.... Then upgraded capacitors... You can see where this is going. Now if my amp would only get here I could stop spending and start building. I can't wait to hear this thing. Wish me luck.

 

[Paul Joppa]

The reason for the cathode connection of the output transformer is to keep the signal current out of the cathode bypass capacitor. In the musical range the impedance of the plate choke is high, so the signal current generated at the plate goes through the parafeed capacitor, then through the output transformer primary, back to the cathode and through the tube to the plate. The bias resistor current is the same as the plate choke current, nearly constant. All the signal current gets confined to the 3-element loop.

I really love it when customers ask intelligent questions!

 

[Grainger49]

Paul,

Does this mean that with the cathode return on the primary of the output transformer a better cathode bypass capacitor will not make a difference?

 

[rlyach]

Now I get the OT connection. The signal current returns through the tube and not the bypass cap. That should keep the cathode voltage more stable. Boy, for such a simple circuit there are sure a lot of nuances. I think then that the only reason to use a better bypass cap would be for leakage and noise. If it were to keep the voltage stable, a larger cap would be more useful than a better cap. Here a higher quality cap will help but it may be only a second or third order improvement. I will keep staring at the schematic and see what else pops out.

 

[jrebman]

Randy,

Don't just stare at the schematic -- also look at the iron, iron arrangement, grounding scheme, etc. .

-- Jim

 

[Paul Joppa]

Paul,

Does this mean that with the cathode return on the primary of the output transformer a better cathode bypass capacitor will not make a difference?

Well, it means it shouldn't. But I'm pretty sure it does anyhow, at least a little. Everything seems to matter to some extent.

Things like that make the engineer in me crazy, so I tell myself that a better bypass cap is also a more reliable, long-lived cap - an actual good engineering reason to do it.  :^)

 

[Grainger49]

It's 3:50 AM and I stifled the laugh when I read "It shouldn't.  But I'm pretty sure it does anyhow,"  Yes, everything seems to matter. 

 

[rlyach]

The spending continues... I just ordered a pair of matched JJ 2A3's and a gold pin JJ 12AT7.

 

[rlyach]

OK... I just bought a JJ gold pin ECC81 for the pre-amp stage. The curves for this tube are attached. Basically it shows good linearity with a 2V p-p input. The output should swing -57V to +58V relative to the operating point, give or take drawing error. Since the mu of the tube is 60 everything seems in order. the only difference is that my operating anode voltage will be 192.5 instead of the manual's 230V or the schematics 200V. I think this should work well. Also, this operating point is well away from the low and high limits of -1.0v Vgk and the dropout of 330V (with safety). There seems to be room to move the operating point up a bit, but I think I will leave it at the designed -2.5V. Is my analysis correct?

 

[Paul Birkeland]

Yes, but you have to factor in that the curves are for a bogey tube; your specimens may show 192V on the plate, or they may be up at 220V (220V is somewhat typical of what I've seen).

From what I know of the Soviet tubes, I think you may find that they are closer to 220V for 100 hours or so, then there might be a little drop in the plate voltage, as I'd imagine that JJ would not use a super fresh tube to draw the curves from. There's a lot of learning to be done with all of this, and much of it transfers to other kits pretty well.

-PB

 

[earwaxxer]

Hey Granger - I like the new 'avatar'. My pic was taken on a pirate ship. We weren't much in the way of pirates, but I love the feel of the old sailing vessels.