Using PP output for SE?

[skip_scratch]

What happens if I take an 8K PP OPT and use half of it (that is center tap to one side) for use as a 4K OPT?

Will it work? Any recommendations on how to wire it?

/.skip

[anatech]

Hi Skip,
No, that will not work.
You have the wrong type of transformer. The core may saturate due to unbalanced plate current.

A proper SE output transformer has a gap designed into the core so there is lower chance of saturation. You need the correct type of output transformer.

In a P-P transformer, the plate current induced flux cancels in the core. Therefore you only have to deal with a small imbalance current. Not the full flux due to plate current as in an SE design.

-Chris

[zobsky]

a parafeed arrangement might work, .. but you still have the issue of 4K vs 8 k impedance

[Dave Cigna]

The short answer is that it should not work because, as anatech said, PP output transformers are normally designed to handle little DC imbalance current in the primary (the current on each side cancels, the difference is the imbalance current.) The way that SE transformers accommodate the DC is by adding a gap in the core. The gap sacrifices inductance, so PP transformers generally don't include it since it theoretically does no good. (Note: at some point both Partridge and Crowhurst advocated the use of a small gap in PP transformers. Lundahl currently includes one in their PP transformers.)

Having said that, some people have reported success using PP transformers for SE. Typically they are using comparatively large PP transformers with low power tubes run at a relatively low DC idle current. Some PP transformers can handle DC better than others. If the laminations are fully interleaved (every E and I is flipped compared to the adjacent one) then it might not work. OTOH, some have the laminations flipped in groups. The chance of success is better with those. You can tell by looking at the long side of the transformer.

One other thing: if you use only 1/2 of the primary (not recommended) then the effective impedance is 1/4 that of the whole thing. Using 1/2 of a PP transformer that is 8k plate-to-plate will result in a 2k load.

[lndm]

I have acquired a largish (7lb) PP OPT and am trying to convert to SE. AFAIK, this xfmr was designed to give >100w PP from 6550's.

It had an inductance of 20H a side, till I separated the E's and I's. Now 10H a side in series for 40H total.

I have tested with between 50-130mA of DC with different devices. It seems to have a limit of 6-7 watts. I should be seeing 2-3 times this power.

I have a sheet of paper between the E's and I's but have tried up to 4 sheets. What else can I do to release the power?

[kevinkr]

This seems like a reasonable amount of power for your modified output transformer.

I assume the problem here is that the core is still saturating.. Probably not quite enough core area and this is inspite of the changes you have made.

I also assume that you have sufficient supply voltage headroom to achieve higher output levels with the tubes you are using to drive this transformer. (Are you certain that the output tube is not saturating or being driven to cut off while you are testing?)

More detail as to what you are using to drive it might help.

In theory widening the gap still further will increase the dc flux density that can be accomodated before saturation occurs, but at the expense of primary L.

[lndm]

Looking at the dummy load one peak will clip, then within a couple of dB, the other peak will clip.

The first peak clips in conjunction with the driver pushing the grid of the output stage to 0V. This clip is fairly flat looking.

The other peak clips but the driver seems unaffected by this. This clip is wedge shaped (chamfering the wave).

I tried removing most of the I's from the xfmr and there was little discernable difference.

The B+ is 400V and the primary sees 450Vp-p at clipping. The signal across the primary clips also but not as badly (not as soon as the secondary).

The behaviour seems fairly uniform across the spectrum.

[kevinkr]

Need to know a few more details, what output tube are you using to drive this transformer? Also what is the primary Z and secondary load impedance you are using?

Once you have driven the grid of your output tube to 0 V unless you are running class A2 you are already well into the nonlinear region of the tube's characteristic curve and no more power is available..
This isn't transformer related at all.

You may be looking at nothing more than your output tube operating beyond its linear operating range.

[lndm]

2 x PSE KT66 triode mode. B+400, Va 370, Ik_TOTAL 110mA. Speaker is a steady 6 ohms. Primary sees 2k, playing with taps this seems the best (least ugly distortion, most power).

I also tried 2 x 6550 in triode mode. At 250V with 180mA and 800 ohms at the primary, and at 320V with 140mA and 1k5 ohms.

The differences in output power seemed marginal.