Audio, high quality like a good ass old tube amp, they had them 5 - 100,000 Hz with less than a percent THD. That level of quality, however the specs are different.
I don't need a primary. "Primarily" what I need is a center tapped choke, but if I can just get a 1:1 audio transformer it would work given one thing - the primary and secondary must be identical.
In fact is it has to be a custom wind I want them wound bifilar but of course not twisted. An E/I core is fine as long as there's enough of it.
I know it is not going to be cheap. I think off the shelf I might be looking at a hundred or more, especially when they hear what wattage I want to pump through it. At least 100 watts.
Anyone had to do this, actually make them make you a transformer ? Anyone got an idea what shit like this would cost ?
Or if you know of something off the shelf I would have to grease you somehow. Anything but firearms, this is Usenet after all.
I like Edcor, if you can't find what you need from a distributor I believe they can custom wind what you want. Made in da USA and prices are pretty reasonable might not cost as much as you expect. downside is long delivery times, custom winding looks like 8 weeks.
Thats interesting they have steel laminations availble. Seems resonable too, last time I had gone to Temple Steel to get a small quantity of laminations.
100 W at 5 Hz? Yikes, that is going to be a real problem! Or, 100 W at
100 KHz, and same voltage at 5 Hz - maybe that is doable with a big pot core. Bifilar winding should be quite easy to do on a pot core bobbin. But, that bifilar winding has HUGE capacitance, as I found out some time ago. I had a 1/2" pot core with 80 turns of quadrifilar, ENTIRELY for ease of winding, and ended up with a staggering 500 pF between wires.
There certainly are outfits that will design and make such things. I just bought the cores, bobbins and clamp, and made my own, but it was a much smaller unit.
I used to use 6V6's driving old filament transformers from the 1940's. They worked great with the pre-war speakers I had at the time. This was back in the early 50's. I soon learned there was a limit to how much amplification you could use. The tube microphonics would make the amplifier howl when you turned the gain up.
The old transformers are hard to find now. Everyone considers them collectors items and wants a fortune. If you are lucky, you might find a scrap wallwart with the right characteristics.
For what purpose? Literally just audio? Refine your spec -- what impedance and (3dB) bandwidth? How much balance either side of CT? DCR?
Does it in fact need to be so precisely balanced that a mere CT choke won't cut it? Do you need a CMC afterwards, or a Guanella style TLT, or a special balanced and shielded design (probably some combination of all of these)?
Can you not solve the problem without a transformer? etc.
Curious, how would you even know that it's bifilar vs. twisted pair? Say it's been potted in tar. X-rays? Actually digging it out? How committed are you to knowing/verifying it..?
5Hz and 100VA, it's going to be a foot-breaker if you drop it and miss the floor not quite microwave oven transfromer size (because single winding), but that ballpark.
- When I tried casting out nines I made a hash of it.
- show quoted text - I like Edcor, if you can't find what you need from a distributor I believe they can custom wind what you want."
They look pretty good. Closed for the holidays but then I wasn't in a hurry much anyway.
Dec 20
- show quoted text - last time I had gone to Temple Steel to get a small quantity of laminations. "
I thought they were still used in a whole buncha stuff.
I might have to settle for less, err more.
ago. I had a 1/2" pot core with 80 turns of quadrifilar, ENTIRELY for ease of winding, and ended up with a staggering 500 pF between wires. "
Wouldn't a heavier gauge wire with thicker insulation help a bit ?
Dec 20
- show quoted text - Hammond supplies high quality tube style transformers and chokes, not at al l cheap."
Is that the organ people ? Audiophiles do have those amps rebuilt.
Dec 20
- show quoted text - I used to use 6V6's driving old filament transformers from the 1940's"
I used the secondary of a power transformer once and while it worked it did n't sound all that good.
Dec 20 For what purpose? Literally just audio? Refine your spec -- what impedanc e and (3dB) bandwidth? "
I could be happy with 8HZ - 30KHz @ -3dB. The low end is slightly more impo rtant that the y high end, I THINK.
Bias on each side would cancel, so whatever minimal heat, which is for clas s AB audio amps would not be much. Fuck class A. Given that there is no hin t of saturation except at high levels, and there is just heat because of th e DCR.
't cut it? "
A CT choke would be fine but every time I look for one I come up with a bl ank. And for this application, precise balance is important.
ced and shielded design (probably some combination of all of these)?"
Not sure what CMC is except for a type of connector, it is not a balun or g oing to drive anything you would call a transmission line because this is a udio.
No, the main principles depend on it.
Wouldn't there be some strange inductive effects when NOT used strictly in parallel ? The inductances would add no ? The main inductances will add of course, why not the inductance of the twist when phased (polaritied ?) that way ?
itted are you to knowing/verifying it..? "
Well by ordering it mainly. I don't have an Xray machine handy. I think it would probably be more expensive to twist than not to twist no ? As such th ey would want more money and call it a feature. Right now I am not even sur e what benefits there are to twisting them.
Dec 20 (21 hours ago)
- show quoted text - So it's an auto-former then ?
Any DC bias or just the 5Hz --- 100 kHz AC ? "
Yes, auto or CT, whichever. Voltage equal. Impedance of the whole thing is
4X each half, accurately. The bias is only in the mA and is applied opposit e to each half of the whole thing. Nothing residual at quiescence.
5:01 AM (16 hours ago)
- show quoted text -
5Hz and 100VA, it's going to be a foot-breaker if you drop it and miss the floor"
Yup. Even if I said 10Hz it will not be small. I can even go higher actuall y but not too much.
---- Cost wise, a hundred bucks each would be great. I only need four, for two p rototypes. After that, if it is a go I'll burn that bridge when I come to i t.
On Saturday, December 22, 2018 at 1:28:26 PM UTC+11, snipped-for-privacy@gmail.com wrot e:
n parallel ? The inductances would add no ? The main inductances will add o f course, why not the inductance of the twist when phased (polaritied ?) th at way?
mmitted are you to knowing/verifying it..? "
t would probably be more expensive to twist than not to twist no ? As such they would want more money and call it a feature. Right now I am not even s ure what benefits there are to twisting them.
Probably not more expensive. Bifilar windings are normally twisted - you do have to twist the wire before you wind it on, but that's never been a prob lem.
The main benefit is that the two twisted wires see almost exactly the same magnetic field, and - because of the twisting - any tiny differences are eq ual and opposite and cancel out almost perfectly.
formatting link
Rayner and Kibble point out that carefully wound multi-wire ratio transform ers give successive taps that are equal to one part in about ten million, b ut in a bifilar wound transformers the two windings can be equal to one par t in a billion.
The Kibble in the author list is the same Kibble remembered in the "Kibble balance" that now replaces the kilogram standard mass.
The impedance level and low-frequency limit are absolutely crucial, the required inductance of the windings depends on those. The upper end of the usable frequency range depends on the leakage inductances (L1-M and L2-M) and windong capacitances, together with the impedance level (again).
In 1:1 transformer equivalent circuit there is a series branch consisting of primary resistance, primary inductance minus mutual inductance, secondary inductance minus mutual inductance and secondary resistance. From the connection between the series inductances, there is mutual inductance to the common voltage. See the LTSpice model below.
If a transformer has some other transformation ratio, it is modeled as an ideal impedance transformation connected to a 1:1 model.
LTSpice:
Version 4 SHEET 1 880 680 WIRE 64 240 -32 240 WIRE 208 240 144 240 WIRE 336 240 288 240 WIRE 384 240 336 240 WIRE 544 240 464 240 WIRE 736 240 624 240 WIRE 336 304 336 240 WIRE 336 432 336 384 WIRE 336 432 -48 432 WIRE 752 432 336 432 WIRE -48 448 -48 432 WIRE 752 448 752 432 FLAG -48 448 0 FLAG 752 448 0 FLAG -32 240 In IOPIN -32 240 In FLAG 736 240 Out IOPIN 736 240 Out SYMBOL res 160 224 R90 WINDOW 0 0 56 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName R1 SYMATTR Value "" SYMBOL res 640 224 R90 WINDOW 0 -3 55 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName R2 SYMATTR Value "" SYMBOL ind 192 256 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 SYMATTR InstName L1-M SYMATTR Value "" SYMBOL ind 368 256 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 SYMATTR InstName L2-M SYMATTR Value "" SYMBOL ind 320 288 R0 SYMATTR InstName M SYMATTR Value ""
Again, depends on specs. With high capacitance, you also get staggeringly low leakage. More generally speaking: the impedance is lower. Which gives you more bandwidth at a given system bandwidth.
I'm guessing your impedance is quite low, but again, without specs...
?
Okay but there were two questions there?
Okay, so, a push-pull amplifier maybe? Transistors? Is that going to be, like, an ampere? Two? A hundred? Or one volt, ten or a hundred? You've said 100VA but that's missing one more number. What's wrong with complementary push-pull?
Common mode choke -- a real CT choke can't give perfectly differential balance, but a CMC can clean it up.
You've said about nothing about the application so I have no clue where you're actually going with this. :-\
Which inductances? The winding inductance is the same, it's going around the core the same number of times. (If it weren't, it wouldn't be a 1:1 transformer!)
The leakage inductance is more or less the same, because the distance between wires, and wire length, is more or less the same. Or, more accurately, inseperable from differences in wire/insulation diameter, or variation in winding length.
Remember, if you can't measure it -- it doesn't matter!
Some time ago the OP posted a transistor amplifier with two emitter followers in push-pull and the autotransformer as the emitter feed choke. The speaker load was to be connected between the emitters. The expected frequency range is simply not realizable with audio transformers. Five decades is a heavy challenge even with RF transmission line and magnetics transformers.
I just do not understand him wanting all the troubles of the transformer, as there is no such pressing need for it as in tube amplifiers. But more than 50 years in professional electronics has not given me an understanding of the audio things ...
emitter followers in push-pull and the autotransformer as the emitter feed choke. The speaker load was to be connected between the emitters. "
Well hell, emitter followers, tubes wish they could have an output impedance so low -- even in class AB, neither transistor is really let go of controlling that load. As long as you're driving the bases with complementary signals and no weird predistortion business (which would only make things worse, anyway), the transformer really doesn't matter.
Get a 120/240V 50Hz toroid, probably uh, 500VA or so. More is better. If it's isolation (120:120V), wire it together in CT. If it's just whatever primary, hook the two halves together in CT and ignore the secondary.
You won't know the difference. I guarantee it. Full money back, pending double blind testing, offer expires non transferrable etc etc...
Honestly I'd be more worried that you're going to blow your spec (such that it is) on output impedance alone. Even a fairly high gNFB tube amp does worse than an emitter follower with none. Zo is the dominant part of tube sound.
There is one point the OP may not be thought about: The emitter followers need bipolar drive, so he has to provide a negative supply out of the originally intended unipolar car voltage supply.
He'd better off with a DC-coupled H-bridge and a suitable diode function generator to bring the tube 3/2-power distortion. The DC coupling will get rid of the lower frequency limit, but it brings along possible problems with DC offsets.
I'm assuming it's going to be an old-fashioned "resistance coupled" interstage, so the negative voltage comes for free as long as you aren't expecting it down to 0Hz. But, as I've said many times in this thread -- who knows.
I will post up the design. It doesn't matter. If I even get a provisional on it,l that gives the Chinese, the Japanese, and whoever else-ese the design. If I sell two of them, them people will make sure to have one at least.
My thing now is f*ck it,l you think you can build it better - DO IT. I think people will buy mine for a couple of reasons,l and I will of course tout them in the ads.
Fuck a patent, copying me is better flattery than any money could buy.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.