3k vs. 5k transformer for PP 2A3

[isaacc7]

I took a look at the 2a3 data over at the NJ7P database and saw something I wasn't expecting. It seems as though all of the plans I've seen for pp 2a3 amps use 5k output transformers, but the database shows some other operating points that look better when using a 3k transformer. It looks like you can get higher output power AND lower THD, but I don't hear about this too often. I'd assume that in a pp design you'd be left with mostly 3d harmonics in any case, so why wouldn't you want the lower distortion? Is the distortion spectrum really that different between the two?

One other question, is there any harm in using a transformer "rated" (whatever that means) for 100 watts when you're only going to pump 15 through it? Thanks for any info..


Isaac

[cerrem]

Optimum plate loading depends on operating points...Plate voltage...bias ..ect..ect... WHat Class of operation are you wanting to use???? If you go into grid current +region, then you need a smaller load to extend up there, for example..
So first thing is to decide on what your main objective is with this PP amp... Are you looking to squeeze out the most power output ?? This will mean operating the valves in such a way that they will wear out quicker than a conservative design...
Also..with the 100W OPT... OPT's don't have standards anymore for ratting them..100W may be correct for your case...SInce for Hi-Fi you need 15W at 20 Hz........ That OPT you have may be 100W at 50HZ for example... ..
I will do some analysis and get back to you...

Chris

[isaacc7]

Originally posted by cerrem
Optimum plate loading depends on operating points...Plate voltage...bias ..ect..ect... WHat Class of operation are you wanting to use???? If you go into grid current +region, then you need a smaller load to extend up there, for example..

OK, sure I figured that, just wondering why I hadn't seen the other operating points being used before...

So first thing is to decide on what your main objective is with this PP amp... Are you looking to squeeze out the most power output ?? This will mean operating the valves in such a way that they will wear out quicker than a conservative design...

Well, I'm not really sure just yet, but it would seem to me that IF I were to go with AB1 output for the most power that I'd want to set it up for lowest distortion too. I'm just looking into possibilities at this point.

Also..with the 100W OPT... OPT's don't have standards anymore for ratting them..100W may be correct for your case...SInce for Hi-Fi you need 15W at 20 Hz........ That OPT you have may be 100W at 50HZ for example... ..
I will do some analysis and get back to you...

Chris

Thanks! Yeah, I figured the 100w thing was more or less marketing speak as opposed to a really useful spec. The output transformer is rated at 20hz to 20 khz at 100w and is a 5k load. This is the Edcor pp transformer line. I assumed that those numbers were a tad optimistic at 100w, or at least with a considerable amount of distortion at that output. I'm wondering if I'm only putting 15w through it if I could get some decent numbers out of it. Seem like it would be a handy transformer to have...

Isaac

[EC8010]

I measured output power vs distortion for PP 2A3 with various loadings and found that the best load for Sovtek 2A3 really was the textbook setting of 5ka-a, biassed at 50mA per valve with 250V across them (grid bias).

[SemperFi]

Before you buy an OT rated to 100watts for an amp that'll give 20watts: I'm not sure who said this, Lundahl perhaps?, but an oversize OT is not optimum sonically. The reason is something with the magnetic field in a tranny being optimized when used at the rated current. In other words, if the idle current is 50mA, don't use an OT that is rated to 200mA. This may only apply to SE OTs, and even then it may not make an audiable difference to most of us

[isaacc7]

Originally posted by EC8010
I measured output power vs distortion for PP 2A3 with various loadings and found that the best load for Sovtek 2A3 really was the textbook setting of 5ka-a, biassed at 50mA per valve with 250V across them (grid bias).

Yeah, but the "textbook" also says this:
Push Pull Class AB1 Amplifier
Plate Voltage ................................. 300 V
Grid No. 1 Voltage ............................ -62 V
Peak Grid No. 1 Voltage........................ 124 V
Plate Current (Zero Signal) ................... 80 mA
Plate Current (Maximum Signal) ................ 147 mA
Load Resistance ............................... 3K ?
Power Output (approx) ......................... 15 W
Total Harmonic Distortion ..................... 2.5 %

That's what was confusing me, anyone ever see this used in an actual design? Thanks for running the sims though, it confirms the decisions made by most of the designers out there:-) Odds are I'll probably go with them, who am I to buck the trend, plus a 5k transformer will be more useful for other experiments than a 3k:-)

Isaac

[tubelab.com]

It is true that an SE transformer has an optimum operating current that is determined primarily by the magnetic gap. This does not generally apply to a P-P transformer since the DC fields should cancel. If you use a transformer that is too big for the job, you will get better bass response than a "correct" transformer. On the other hand you will give up some efficiency and possibly some HF response due to the fact that you have more wire losses, and more iron to magnetize.

[kevinkr]

Another thing to consider is your effective damping factor particularly if you are not going to use global negative feedback. Pure class A operation with a 5K P-P impedance might get you a damping factor as high as 2.5 once you take into account the transformer primary and secondary dcr. Using a 3K transformer you will be lucky to get anything over half that. Class AB operation worsens the picture and also implies a modulation of the output source impedance depending on the power level..

[cerrem]

Hold the horses!!!
You left out a crucial detail about the 5K load spec....
The 5K load spec is for CATHODE BIAS while the 3K spec is for FIXED BIAS....
With the Cathode bias the spec states a 780 ohm resistor in the cathodes..this brings the effective Plate voltage down to 237V..that explains the larger plate load..ect..ect...
A P-P plate load can only go so small...your limited by the plate dissipation curve..which you sketch your plate-load/load-line from the plate voltage operating point up till your are tangent with the plate dissipation curve...if you go any further past that curve then you run the risk of the plates getting red durring signal operation of the amp, regardless of how cool your idle current/dissipation is...SO that is the point you do your number crunching and see your power output and % odd distortions....If you are not happy with them in a TRIODE Class A P-P amp..then you make the plate load bigger to reduce the odd distortions while reducing power output.. It's a trade-off and as simple as that...
As for frequency response of the OPT...you need to calculate the transformer specs and then make or buy a transformer with those specs.. These specs are SPECIFIC to the 2A3 at 3K load and at that operating point....
OK..I did the OPT numbers for you for a respectable Hi-Fi OPT..
For -3dB points at 3.3Hz and 60kHz on a 3K plate load you would need a minimum of 50H Inductance, and a maximum of 12mH of leakage inductance...you then need to keep your TOTAL capacitance under 900pF, since at 900pF you just start to get rise at the Q point...
You should have a P-P OPT that hits no higher than 12,000 Gauss at 15W @20Hz... Personally I would prefer 10,000 Gauss limit even with M6 core...

Chris

[i_should_coco]

Originally posted by cerrem
Hold the horses!!!
...if you go any further past that curve then you run the risk of the plates getting red durring signal operation of the amp, regardless of how cool your idle current/dissipation is...S

Actually, for class-A operation, as long as the static dissipation is within the plate rating, going above the max dissipation curve for one half of the swing won't matter. The tube's anode has enough thermal inertia for this not to be a problem, because during the other half of the swing, the dissipation will be much less, balancing things out.