Take a look: http://www.tube-amps.net/images/Hashimoto_Specs/HW-100-5_1024.jpg
That is the larger and more expensive transformer they make ($1284 a pair, plus taxes and shipping) and is rated 100W at 30Hz, smaller ones are the same or rated at 40-45 Hz.
You don't need to split the full primaries. Actually if you want to keep it simple you can wind it like if it were a SE transformer and then for one half you connect the first (1/2P) and the 3rd (P) and the other half the 2nd (P) and the 4th (1/2P). If you can make nice layers one on top of the other, the DC resistance of the 2 halves will be matched within less than 1% (maybe 0.5%, it really depends on how good you can make it). I have done it many times....
The resultant capacitance will be half respect to a SE transformer with same geometry and turns thanks to the PP connection. The split bobbin is not really necessary as this can be still considered a relatively small transformer.
Yes UL Tap is referred to turns normally. But do not get stuck on the numbers anything between 40% and 50% will work. So I suggest that you take the UL tap at the end of a layer and do not do strange things in the middle. If you use 0.25 mm copper wire with double insulation (0.301 mm with insulator) you can comfortably wind 130 turns per layer on the EI96 coil former. Then the full primary would be 3+6+6+3 layers for a total of 2340 turns. The height of the primary will then be 5.42 mm. The total height of the winding volume is 12.5 mm (in reality you have 1 mm more and still have enough clearance between the coil and the core). So you have similar height for the secondary plus the space for insulation.
If you take the UL tap after 4 layers from the mid point (you have 9 layers for each half) you will get 4/9=44.44% UL tap. From the beginning of the half primary it will be the 5th layer of course.
I wish it were 50% of the turns for the 4R tap but impendences go with the squared ratio of turns. So, the turns for the 4R tap are the square root of the turns for the 8R. It will never be exact. For 7K and 2340 primary turns, you will need 79 turns for 8R. For 4R you will need to take the tap at 56th turn.
I suggest you make several layers using wire size that fits the width for 79 turns and then put all of them in parallel. The wire size that fills the width with 79 turns is 0.45 mm (0.495 single insulation or 0.516 mm double insulation). With the double insulation you can make 3 or 4 layers for each secondary (it depends on how much insulation you will use) for a total of 9 or 12 layers. If you use double insulation wire for both primary and secondary you don't need to add insulators between layers. Only between primary and secondary. It's more important that you use some tape at the beginning and the end to stop turns falling out. Use thin tape.
If you put 2x4R secondaries in series you get a 16R tap.
If you want to do a better thing and use all secondary turns regardless of 4R or 8R, you can use the Quad/Radford arrangement.
Last edited:
The good news is, with a PP transformer you get away with less capacitance, compared to a SE transformer. Typically for a tube OPT it is equal to 1/4 less secondary to primary capacitance and it varies or can be changed slightly in some sectioning styles. This means you can increase the interleaving amount for PP transformers compared to their SE counterparts. You can check Crowhurst papers and RDH4 diagrams as a start.
Before proceeding with the design, you need to define your criteria:
1. Maximum power for lowest frequency and frequency bandwidth.
2. Amount of losses.
3. Weight, size and cost
4. Define your acceptable Ls roll off
5. Define your acceptable Cp+Cs roll off
6. Define your HF behavior for the specified load.
Before proceeding with the design, you need to define your criteria:
1. Maximum power for lowest frequency and frequency bandwidth.
2. Amount of losses.
3. Weight, size and cost
4. Define your acceptable Ls roll off
5. Define your acceptable Cp+Cs roll off
6. Define your HF behavior for the specified load.
You can but then more space will be occupied by insulators. In my example above, space is tight and I would not do it. With just 4P (18 layers in series) and 3S (9 layers in parallel), 2340 turns, 0.23-0.25 mm Nomex, I estimate leakage inductance will be less than 7 mH. FR will likely be down 3 dB well above 50 KHz.
Last edited:
With regard to winding EI tfmr's you always run out of room so get bigger laminations if you can and a bigger stack, this will give you a better LF response too. Use Nomex paper after every layer then varnish it, make sure the paper is wider than the bobbin width so it goes up the sides a tad, this stops flahover between winding. The paper gives you a stable layer to wind your next layer on. Leave 4mm at each end of the layer and fill with glass fiber tape or what ever. This article gives you a few tips - http://ludens.cl/Electron/trafos/trafos.html
I would take what's in that page with a pinch a salt. There are a couple of things I consider really wrong: 1) coil touching the core is not advisable 2) some of those hand made coil formers waste a lot of space.
4 mm on each side is really a waste of space for standard coil formers. It's 10% less turns on the EI96 coil former. More on smaller ones. Maybe it's ok for home made coil formers that MIGHT maximize space but don't have windows at each side to bring the layer out and lock it. Anyway, quality coil formers are cheap. 2mm on each is rather easy.
4 mm on each side is really a waste of space for standard coil formers. It's 10% less turns on the EI96 coil former. More on smaller ones. Maybe it's ok for home made coil formers that MIGHT maximize space but don't have windows at each side to bring the layer out and lock it. Anyway, quality coil formers are cheap. 2mm on each is rather easy.
If you go to the next size, the EI108, and keep the cross-section the same with the only reason to have more winding space at expense of size and weight you will also get 10-15% more stray capacitance for the same interleaving. If you put more insulation to reduce capacitance you basically get just more size and weight. If you increase the cross-section with less turns you still increase surfaces and thus capacitance.
The best choice for the transformer size is always the smaller core size that can do the job. Setting the low frequency limit for max Pout at 28-30Hz is not a compromise, rather 20Hz limit is more waste than benefits.....
The best choice for the transformer size is always the smaller core size that can do the job. Setting the low frequency limit for max Pout at 28-30Hz is not a compromise, rather 20Hz limit is more waste than benefits.....
Thanks, there is alot of useful info here.
I will go over a few calculations and confirm what I will go with and see what you think in due course.
square root of 2 = 1.414213562373095
100 turns / 1.414213562373095 = 70.7
so would 71 turns thereabouts be correct for the 4 ohm tap?
I dont think Im going to bother with a 16 ohm secondary, im never going to be using such a load, probably ever only 8 ohms, but always havdy to have 4 ohms as an option.
Regarding Nomex paper, I think i might have a roll of this stuff, but is there anywhere i can buy rolls the right width to suit the various bobbin sizes?
I will go over a few calculations and confirm what I will go with and see what you think in due course.
OK, so as an example if the 8 ohm secondary has a total of 100 turns, this is what i calculated for the 4 ohm tap:
square root of 2 = 1.414213562373095
100 turns / 1.414213562373095 = 70.7
so would 71 turns thereabouts be correct for the 4 ohm tap?
I dont think Im going to bother with a 16 ohm secondary, im never going to be using such a load, probably ever only 8 ohms, but always havdy to have 4 ohms as an option.
Regarding Nomex paper, I think i might have a roll of this stuff, but is there anywhere i can buy rolls the right width to suit the various bobbin sizes?
Agree with you there about the homemade bobbins but if you read the whole page there is some good advice there about the practicalities of tfmr winding.
Probably if your willing to pay and buy a 1000 rolls, I've found it's tricky getting it in small amounts full stop here in the UK. Just cut it to size.
The stuff I see available on the likes of aliexpress appears to be a self adhesive tape that comes in various widths. Not too sure if this stuff is suitable.
If I can find a roll of the stuff with no adhesive, that would be the most ideal and I could easily cut it down to various widths.
I've got this other type of paper that the person used that I was supplied with. It's kind of a blue/purplish color on one side and shiny plastic finish on the other side and looks like a brown cardboard kind of colour.
I'm considering using a vacuum chamber for placing the bobbins in the urethane varnish to get all the air out.
If I can find a roll of the stuff with no adhesive, that would be the most ideal and I could easily cut it down to various widths.
I've got this other type of paper that the person used that I was supplied with. It's kind of a blue/purplish color on one side and shiny plastic finish on the other side and looks like a brown cardboard kind of colour.
I'm considering using a vacuum chamber for placing the bobbins in the urethane varnish to get all the air out.
https://www.bernardimotorielettrici...lettrici-flessibili-per-motori-elettrici.html
Last time minimum order was 100 euros. As Diabolical Artificer suggested, I would buy a larger roll and cut it every time for the specific size. If you figure it out before cutting you can use it efficiently with minimal waste.
Pure latheroid paper or latheroid/mylar are also good choices. Not as good as Nomex 410 tough.
Last time minimum order was 100 euros. As Diabolical Artificer suggested, I would buy a larger roll and cut it every time for the specific size. If you figure it out before cutting you can use it efficiently with minimal waste.
Pure latheroid paper or latheroid/mylar are also good choices. Not as good as Nomex 410 tough.