Q: DuPont Mylar Film for ESLs...

[wrl]

Wow, most of those transformers are well over $100 dollars.

I was hoping to spend no more than $75-$80 on each of them. Also, I don't need it to go lower than about 500-600 hz.

Anyone have any suggestions for places in the US?

Thanks,
Wes

[Alexander Rice]

Using a mains transformer is probably not a good idea because thay are often made of the poorest grade of steel they can get away with. If you try feeding a mains transformer form a signal generator and measure the power in and out you find, unsurprisingly that they are most efficient around 50 Hz and drop off rapidly either side - i only tested to 1khz and already that was down by about 10dB from the 50hz peak - i dread to think what sort of thing happens at 20kHz. As frequency goes up you need thinner and thinner laminations to reduce eddy currents in the laminations. I would have thought you would be better re-winding an old output transformer.

[316a]

Originally posted by Alexander Rice
Using a mains transformer is probably not a good idea because thay are often made of the poorest grade of steel they can get away with. If you try feeding a mains transformer form a signal generator and measure the power in and out you find, unsurprisingly that they are most efficient around 50 Hz and drop off rapidly either side - i only tested to 1khz and already that was down by about 10dB from the 50hz peak - i dread to think what sort of thing happens at 20kHz. As frequency goes up you need thinner and thinner laminations to reduce eddy currents in the laminations. I would have thought you would be better re-winding an old output transformer.

I've built my ESL panels now and have already tried a mains toroid as a step-up transformer . Even though I need to do more work with raising the polarising voltage to increase sensitivity , there appears to be nothing wrong with using one , in fact the top-end in comparable , if not better than with an 8k anode to anode valve output unit . Due to the construction of mains toroids , wide bandwidth is achieved , of course with EI transformers you will run into problems ! Why not repeat your tests with a mains toroid , you'll probably get at least 30k bandwidth

316a

[Alexander Rice]

i did indeed test an EI type transformer - although the tests were not in connection with building speakers so i had no reason to retest a better trannie. I naturally assumed that the core geometry plays a relatively small part in the frequency response of the transformer, this may well not be true, althought i am sure it depends alot on the 'quality' of the particular transformer - ie. the type of steel used and the thickness of the laminations. That said, modern transformers with amorphous silicon steel laminations are already FAR superior to what was being used even 20 yrs ago. I'll run some tests tomorrow if i get the chance beacuse i have always been put off experimenting with valves due to the high cost of a suitable output trannie, if i can buy one 'off the shelf' in the form of an ordinary mains transformer or rewind one of the same to suit my needs i'll definitley be paying more attention to the possibility of home brewing a little valve amp (hell, i've got a 1000W trantmitting trode in the basement - that could make an 'interesting' amp...)

[316a]

Go for it , you may be surprised
This evening I discarded the stepup transformers altogether and drove the ESL straight from the anodes of my amp via teflon capacitors This is definitely the way to go ! Luckily I only need to drive a 5"x20" treble panel . My current workshop amp only runs from a 220V rail though , time to build something with higher B+ and drive these things active . Interestingly the sensitivity of the ESL can be matched to the main unit (Jordan JX92 TL) by tweaking the polaring bias . ESL's are fun

316a