You missed my question. Is it for PP or SE? For PP is easier to figure out the turns and assessing the inductance (with the right permeability). The magnetization curve is not very helpful. One needs to take into account the unvaidable natural air gap which is generally a bit larger than test samples (used to derive magnetization). A practical value for permarbiliy of GO interleaved Es and Is 1x1, goes from 4000 to 6000. So if you use 5000 as an average value you will get a good estimate ( I assume you are keeping B within 1T at lowest frequency of interest).
Apologies, thought I put that in post #1, it's triode PPP Class AB2. So four 2C34 triodes PP or PPP.
Been trying to find a design method taking info from Pat Turner, Wolpert, Lee, Popovitch etc.
Andy.
Been trying to find a design method taking info from Pat Turner, Wolpert, Lee, Popovitch etc.
Andy.
So if P= V^2/R is your target output power, say at 20Hz, then you can figure out the V and derive the turns from:
N=V*10^8/(4.44*Ae*f*B),
where Ae=0.95A is the effective core area (takes into account 0.95 stacking factor for EI cores), f is the lowest frequency, B is the AC induction.
The other thing to consider is copper resistance. Generally you should keep total Rdc=Rprim+ Rsec*n^2 within 10% of the nominal load. Better if 5%. That will tell you if the core size is good enough or need a bigger (o smaller) one. Can also rise the frequency for max Pout to 25-30Hz without practical disadvantages in order to require less turns. The resultant inductance will be easily more than good enough....
Keep the max B at 0.85T (with EI cores) because that's where the magnetization curve starts to visibly bend....With HWR C-cores can work at higher B.
N=V*10^8/(4.44*Ae*f*B),
where Ae=0.95A is the effective core area (takes into account 0.95 stacking factor for EI cores), f is the lowest frequency, B is the AC induction.
The other thing to consider is copper resistance. Generally you should keep total Rdc=Rprim+ Rsec*n^2 within 10% of the nominal load. Better if 5%. That will tell you if the core size is good enough or need a bigger (o smaller) one. Can also rise the frequency for max Pout to 25-30Hz without practical disadvantages in order to require less turns. The resultant inductance will be easily more than good enough....
Keep the max B at 0.85T (with EI cores) because that's where the magnetization curve starts to visibly bend....With HWR C-cores can work at higher B.
P.S.
In the formula above the typical units are: cm^2 for the area, Gauss for B (0.85T = 8500 Gauss), Hertz and Vrms.
In the formula above the typical units are: cm^2 for the area, Gauss for B (0.85T = 8500 Gauss), Hertz and Vrms.
Thank you. I presume the HF response is mainly dictated by inter winding C and other stray C in it's construction as in insulation between windings etc.
Lastly what interleaving do you use 5 to 3 or more?
Andy.
The interleaving depends on the actual size and impedance of the transformer.
HF response depends on stray C and leakage inductance. The interleaving is best used for finding the best balance as one grows or decreases at expense of the other one. If you increase the interleaving leakage inductance is reduced but capacitance grows and also efficiency decreases (because you end up using more space for insulators and likely smaller wire). What one can do is maximize the product (which means highest possible HF resonance for the given case).
Being PPP of triodes I assume it's not high impedance. I would start with 1/2P-S-P-S-P-S-1/2P. Most Sowters are done like that....
Also the PP connection results in halved stray C respect to SE case with same turns and geometry.
HF response depends on stray C and leakage inductance. The interleaving is best used for finding the best balance as one grows or decreases at expense of the other one. If you increase the interleaving leakage inductance is reduced but capacitance grows and also efficiency decreases (because you end up using more space for insulators and likely smaller wire). What one can do is maximize the product (which means highest possible HF resonance for the given case).
Being PPP of triodes I assume it's not high impedance. I would start with 1/2P-S-P-S-P-S-1/2P. Most Sowters are done like that....
Also the PP connection results in halved stray C respect to SE case with same turns and geometry.
Last edited:
If you want to keep it as simple as possible you can avoid the double cave and wind it like if were a SE transformer. Then to get fairly balanced DC resistance out of the 2 halves, you make one half with the inner 1/2P + the second P and the other half with the first P and the outer 1/2P.
With 5 primaries (always half turns for the first and last) and 4 secondaries this connection actually gets near perfect balance of DC resistance of the 2 halves. However also in the simpler case of 4P+3S it should be within 1-2% with no practical consequences.
With 5 primaries (always half turns for the first and last) and 4 secondaries this connection actually gets near perfect balance of DC resistance of the 2 halves. However also in the simpler case of 4P+3S it should be within 1-2% with no practical consequences.