What about using iron powder (fiilings) in resin to diy output transformer for single ended 2a3?
Maybe mixing iron powder with copper or some other metal fillings.
The construction of such core seems easy and you could create different shapes and sizes....
Maybe mixing iron powder with copper or some other metal fillings.
The construction of such core seems easy and you could create different shapes and sizes....
If it were so easy, everybody would be doing it that way, What you describe will result in a low quality, low permeability, high loss, core not anywhere near as good as commercially available powdered metal magnetic materials, which have insufficient permeability for use in a good SE transformer. The higher permeability powder cores also tend to vary in permeability with applied AC flux, which would make the transformer nonlinear. The permeability of lamination based SE transformers is to a large extent linearized by the air gap used to define the inductance of the transformer and make it capable of supporting the DC bias current necessary for a SE amp.
I doubt it would be high loss. Likely non-conductive. It is a reasonable idea to put forward. There is no need to get into a transformer rage about it. I say give it a try, you can only learn.
I tried packing powdered ferrite material (Mu 3000 matl. when in solid core form) into a donut mold once, and the best effective Mu I could get was around 7. (commercial powdered cores are extremely compressed with a hydraulic press to make the metallic powder used actually deform) Since most OTs will sorta work without a core above 1000 Hz, this might get you down to 1000/7 = 143 Hz. But with no primary inductance reserve at all.
The problem with DIY packed powdered material is that adjacent particles touch each other in small points, while air gaps of varying width separate the rest. The small contact points magnetically saturate very easily, while the air gapped portions have variable saturation effects depending on how much air gap there is. This makes for a well known non-linear Mu for loose packed powdered cores. Not so much a problem if you have a large inductance reserve, but for audio LF you won't.
How it sounds might be a different thing though, since SE xfmrs depend on non-linear 2nd harmonic distortion (due to the DC biasing tilting the perm. versus the AC swing) for some of their effect.
The problem with DIY packed powdered material is that adjacent particles touch each other in small points, while air gaps of varying width separate the rest. The small contact points magnetically saturate very easily, while the air gapped portions have variable saturation effects depending on how much air gap there is. This makes for a well known non-linear Mu for loose packed powdered cores. Not so much a problem if you have a large inductance reserve, but for audio LF you won't.
How it sounds might be a different thing though, since SE xfmrs depend on non-linear 2nd harmonic distortion (due to the DC biasing tilting the perm. versus the AC swing) for some of their effect.
Last edited:
If you must build an OPT with found objects then look for the steel strap used to hold heavy loads to wood pallets. This is a kind is spring steel. Then you roll it around a form and make a toroid core. You can stack them and make a larger size too.
Another way to get cores is to disassemble old microwave ovens but it is much work to strip them to the bare core, cut the copper off with a hacksaw
Though if you're going to use steel strap, make sure it's annealed mild steel first. That should be passable up to a few kHz (mostly depending on thickness) and 0.6 to 0.8T. Hardened shipping strap makes poor magnets and much poorer transformers!
Tim
Tim
My point was that the already available powdered metal cores are nowhere near good enough for use in an audio transformer. The manufacturers of these cores have access to a much higher grade of source material than available to the casual experimenter (and have done a lot of proprietary research on binders and additives). There's nothing to stop one from trying this approach, but be prepared for disappointment.
On another note, however, it may be interesting for those who have a bi-amplifed system to consider trying a ferrite SE transformer for the HF amplifier to feed the tweeters. Since low frequency response is not required, the primary doesn't need a whole lot of inductance, and the power requirement is modest. This would be squarely in the range of a largish E core like the 42 X42X20 core or the E55.
On a historical note, Indiana General briefly offered what they called "bonded ferrite" cores in the early 80's - ferrite powder with some sort of binder. I never tried them - I was told by others that they were lossy. Indiana General dropped them after a short while
On another note, however, it may be interesting for those who have a bi-amplifed system to consider trying a ferrite SE transformer for the HF amplifier to feed the tweeters. Since low frequency response is not required, the primary doesn't need a whole lot of inductance, and the power requirement is modest. This would be squarely in the range of a largish E core like the 42 X42X20 core or the E55.
On a historical note, Indiana General briefly offered what they called "bonded ferrite" cores in the early 80's - ferrite powder with some sort of binder. I never tried them - I was told by others that they were lossy. Indiana General dropped them after a short while
Ferrite is fine transformer material, it's just the lower saturation (0.4T, vs. 1.2T for silicon steel) means you need that many more turns for the same voltage/frequency capacity.
Although the permeability of ferrite (up to ~15k, though most power materials are ~2k) is generally lower than good steels (GOSS ~20k in toroids; super-perm materials up to about 1M), it's much higher at high frequencies, past where iron rolls off. So you'll get better performance up there (not that it matters, as the inductance will be huge either way, far in excess of your speaker's impedance).
Amorphous cores are a good solution for absolute best efficiency / power density in the intermediate range (about 1-10kHz), above where steel of any thickness is economical, while having higher flux density than ferrite until it gets too hot as well.
Tim
Although the permeability of ferrite (up to ~15k, though most power materials are ~2k) is generally lower than good steels (GOSS ~20k in toroids; super-perm materials up to about 1M), it's much higher at high frequencies, past where iron rolls off. So you'll get better performance up there (not that it matters, as the inductance will be huge either way, far in excess of your speaker's impedance).
Amorphous cores are a good solution for absolute best efficiency / power density in the intermediate range (about 1-10kHz), above where steel of any thickness is economical, while having higher flux density than ferrite until it gets too hot as well.
Tim
I would never think of using ferrite for a full range transformer for all the reasons cited - I'm quite familiar with ferrite and powder material characteristics, having used them in SMPS designs for 30+ years. In the niche application of an HF transformer to drive a tweeter, the lower saturation flux density could be easily worked around, as the power and primary inductance requirements are less severe. The very point of the exercise in my mind would be to not use steel or amorphous materials. I'll put it on my to-do list, as I have a couple of small SE tube designs contemplated. Winding a set of ferrite tweeter drivers won't be much more difficult than some of the byzantine multi-output SMPS designs I've done in the past, though I'll use more interleaving than I'd normally do with an SMPS design. Incorporating the crossover filter into the amp front end will prevent it from being driven with "inappropriate frequencies".
How will it sound? I'll have to try it to find out. Look for a thread some time, though not real soon.
How will it sound? I'll have to try it to find out. Look for a thread some time, though not real soon.
Speaking of Ferrite OTs and Switchmode power supplies, there is that hybrid combination used by D. Berning for a HF carrier Ferrite OT. An old thread or two has discussed variants to avoid the patent. The unmentioned problem with the scheme however is that the OT leakage inductance is suffered at the 200 KHz or so carrier freq., so is a real issue. You'll notice that Bernings OTs have very high impedance ratios to make up for leakage L loss (not really a loss, but acts like a series resistance at the audio freq, spoiling the LF output damping factor without feedback). A switched capacitor impedance converter can avoid the issue though.
Last edited:
I really wouldn't want to heap my plate with a mess like that - it's already overflowing with stuff more likely to work without a bunch of effort. What I have in mind is a simple linear transformer with low limit ~1.5kHz. I did some back-of-the-envelope calculations last night, and want to start with 900-1000 primary turns and about 1-1.5H primary inductance. I have some Ferroxcube 42X33X20 cores in 3F3 material that would be excellent candidates for a first build, as they are long-legged E cores that will help keep the leakage inductance down. Shed of the requirements to handle the low end, the low (relatively speaking) leakage and stray capacitance of a targeted sdesign could make this an excellent HF performer. Of course, one could do a design like this using steel, though one would want to hold off on the flux levels to keep the losses down. As a result, you'll get less of a size/turns bonanza than you'd initially think. This is normally not a problem with a full-range design, as satisfying the low end requirements will automatically keep the flux density down in the mud for the high end. With ferrite, it's less of a problem, as long as you work within the limmitations of the material.
I'm shoving off from this thread, as this is diverging quite a bit from its original intent. When I start this project, I'll start another thread.
I'm shoving off from this thread, as this is diverging quite a bit from its original intent. When I start this project, I'll start another thread.
its the winding the transformer that would be much more difficult than making the transformer core itself.
Its probably a lot cheaper to get a good transformer.
Its probably a lot cheaper to get a good transformer.
- Status
- Not open for further replies.