First 300b Design, Please comment

[SGregory]

I think it is important to understand what the aim is. This amp is to be my building block of understanding for the nuances of tube audio. Start with KISS (Keep It Simple Scott) and build complexity refinements in later once I understand what is going on. I will have two designs to immediatley compare it to. Tubelab Simple SE (not sure what tube yet) and a Tubelab SE which will be runnning 300b's as well. Hopefully at some point in the near future I can contribute to this community more than I suck away from it. Thank you very much for your help so far.

Thanks zigzagflux, I didn't think it should be bypassed but wasn't 100% sure. Your explination make perfect sense.

dsavitsk. Don't pick on Pinto's, they are cars of my era. (That and Disco). You are right, in its current design it won't be ultimate. Thats okay for today and will give me something to do tommorrow. Although not yet in good enough form to present yet, the P.S. and heater circuits will not be insignificant. I will probably use silicon in the heater circuits to obtain the goals. I'll post them shortly. And yes coupling cap is too large. For the present I will just put in 0.47uF until final output stage design is completed.

As far as the 300b operating point, my suggested operating points are based on being a newbie and using SE Cad. While I considered the recomended operating points by W.E. they were all less than 3.5k @ 350V plate to get any kind of power output.
Initially was trying to get power, then to balance power vs. THD. Maybe SECad isn't the tool I should use as it flags the 5.0k and 4.0k primaries as too high of an impedance. THD calculations do agree with above comments.

Attached is a load line curve were I bumped up Vplate to about 395v using a 4.0k primary. From my inexperienced perspective, this looks good. I am driving the tube to 80% of its max power, producing 10.1W with 2.5% 2nd and 0.3% 3rd order harmonics. Vin is +/-66 V. There is quite a bit of headroom before hitting A2. At the limit of A1 it is showing close to 15W with 3.3%2nd and .3%3rd. I know I must be missing something in this process so please help me. I probably will have the OPT made by Electraprint, so we can make it pretty much anything. This is an area where I want to get right so I am all ears.

Miles, As I am learning I am opening up to the use of mosfets. I certainly like my alephs. I have plenty of 12b4's laying around and am quite happy with their sound at this point as well. After running kids around tonight I will read-up on those ideas as well and try to come up with a circuit.

My ultimate goal is to have my system tri-amped. Using SE for Upper and Mids which are are 93 and 91 db/watt and and PP for the lower as it is only 89db. I don't need or want to blow the windows out. I know really good h.e. fullrange and horns exist, but I have never been exposed to them other than what I read on the boards.

[45]

Originally Posted by SGregory

Attached is a load line curve were I bumped up Vplate to about 395v using a 4.0k primary. From my inexperienced perspective, this looks good. I am driving the tube to 80% of its max power, producing 10.1W with 2.5% 2nd and 0.3% 3rd order harmonics. Vin is +/-66 V.

From the picture, with +/-66V it looks like you get approx. 5.5W. Rather poor efficiency. You get 10W with +/- 82V input.

The original WE recommended loads for 350V/60mA are 4K, 3K, 2K. If you go for 4K you get 7W (THD is a bit more than 3%), if the load from your speaker drops at 6 or 4 ohms the primary impedances will be 3 and 2K respectively. The Pout for 2K is more than 10W. I don't think you can find more info for other tubes.

Cheers,
45

[SGregory]

Thanks 45, Calculated by hand and your right 5.5W. I came up with 5.8W before transformer loss. Went back to SECad and found the I had not checked RMS out. I checked it and the results came back at 5.18W with the transformer I used.

Proves the danger of sims in that they are only as good as those who operate it. Unless you can determine what is reasonable, then something as simple as a check box will throw everything off.

I am now back to Standard operating points, designing for better A2 operation, and hand calculating.
Thanks Again

[SGregory]

After looking at a FET follower, the complexity involved in setting up a negative power supply violates my KISS at this point. I will get to hear that design with the tubelab SE. I am sure it sounds great.

Back to simplicity, the interstage xfmr.
Is the Lundahl 1660 @ 1:2.25 considered to be a "good" interstage? Another possibility would be to have Electraprint wind one. Don't hear much about his stuff however. Another one I saw was a Sowter @1:1.

I guess if I go this route then a DRD should be considered. I am far away from my original concept at this point but oh well.

[JPeitzman]

the complexity involved in setting up a negative power supply

Huh? It is true you will hear it with the Tubelab SE so it may be moot, but all it takes is a couple of diodes and a cap or two.
All three of the listed transformer companies make top notch iron. You may not hear to much about Electra-Print because people are to busy listening As a bonus Jack is one of the nicest people I have chatted with. The inter-stage with fixed bias is just about as simple as it gets, DRD is right up there with it. Here is something to chew on a bit, keep in mind that it uses an all out cost-no-object power supply. You could take it another step even and lose the R/C on the driver cathode, replacing it with an LED.
http://www.users.globalnet.co.uk/~va...s/SE108IT2.GIF


James

[Yvesm]

Originally Posted by Miles Prower

. . .
Music programs have fast rising transients that can be way above the average value, and these will likely drive the grid positive . . .

Agreed 200% !

Using a cathode follower (or source follower) is a good way to recover quickly after an overload if, and only if it is DC coupled to the 300B grid.
And to do so, you'll inevitably have to stack some negative and positive supplies.
If the link cap remains, you loose all the profit !

In short, problems arise when this link cap receive charging current thru the direct biased diode constitued by the G1 to K in the 300B.
This shifts bias wich recover slowly while the cap discharges thru grid leak resistor.

Yves.

[SGregory]

I retract my statement of complexity after looking at SJS schema. My first impression of negative supply is from tubelab SE, which when boiled down is equally as simple in concept.
I was hoping Jack at electraprint made a good interstage. He was good to deal with although he would only talk through email. I don't have much data on his IT so I will use the LL1660 as a reference for the time being.

The SJS uses a 1:1 IT so it will require a driver tube with a mu greater than 30 in my case. I have plenty of gain in my pre to drive the amp. What is an equiv to a 437A?

If the IT has a ratio 1:2.25 (LL1660 T) I can use a lower mu tube, possibly with higher bias current. This would raise the impedance driving the 300b, which I would assume ( hate to use that word but in this case a good choice ;-) ) compresses the dynamics by reducing the grid recovery. My basis for this statement is that the grid current would have to fight secondary impedance and not the DCR. The secondary DCR remains the same for the both cases using a LL1660.

I will post a schema later for your consideration. Again I would like to thank everyone that has contributed so far. This has been a great learning experience.

[45]

I would prefer the LL1671/30mA. With 3K source and open secondary you get 30Hz-30KHz +/-1dB which is very good for a SE. The ratio is 1:1, the inductance 35H. The max output at 30Hz is 130 Vrms!
Using a 1:2.25 step-up the driver will see a capacitive load 5 times higher.
Could be used with a triode-strapped 6F6GT (which is a bit better than the 6V6 in this application) if you want just a slight swing (few volts) into positive grid.
If you want more drive you could go for a 45 which is more linear with lower loads.
For example, at 240V/30 mA you can get 200V P-P into 4K, i.e. 1.2W, with 3% THD approx.
The input stage can be a good medium/high gain triode.

[Allen Wright]

Using a CF will NOT add anything to the sonic quality - more like worsen it greatly.

A 300B is not hard to drive, use a decent tube like a 6H30 and run it at a lot of current (say 20mA) and you'll get all the drive you need, and a far simplier signal path.

Regards, Allen (Vacuum State)

[45]

P.S.
Don't worry about that 3% THD from the 45 it is practically 2nd harmonic only at 200V P-P into 4K!
The curves of the 45 and 300B are extremely similar in terms of linearity and very likely you get the best condition for perfect harmonic cancellation. That is, you get low overall THD from your amp if the input triode has very low distortion!