Thanks for all of the responses. I've been busy this week so haven't had much time to thank you all.
I'm following Patrick Turner's method and looking at cores. He recommends a pretty big core. If I were to build a 100W transformer he would have me use a much larger core than you would see on an Edcor or Hammond 100W transformer. That fine, there is a difference in design philosophy there. I'll probably end up building something with a core similar in size to a 100W Hammond/Edcor (EI-150) and use a pair of 40W plate output tubes.
Turner has me calculate core cross section and I'm a bit drawn to a double C-core configuration since the window size seems to be bigger with these and I'm thinking I can fit larger wire and keep resistive losses down. I can also get much wider C-cores. Seems like this would result in needing a ton of copper, though and then the surface area of the windings would also go up leading to increased capacitances.
If I were to go with C-cores, is it best to just stick with a window that is similar in width to what EI-150 would give me? I get considerably more height for the same width.
EI-150 has a window that is 19mm x 57mm. C-cores are 32mm x 60mm or I could even go with 80mm or 100mm for the width.
If I understand correctly, the longer core lengths would result in lower inductance per turn, but the other concern is that the wider cores would also need a lot of copper to fill.
I'm following Patrick Turner's method and looking at cores. He recommends a pretty big core. If I were to build a 100W transformer he would have me use a much larger core than you would see on an Edcor or Hammond 100W transformer. That fine, there is a difference in design philosophy there. I'll probably end up building something with a core similar in size to a 100W Hammond/Edcor (EI-150) and use a pair of 40W plate output tubes.
Turner has me calculate core cross section and I'm a bit drawn to a double C-core configuration since the window size seems to be bigger with these and I'm thinking I can fit larger wire and keep resistive losses down. I can also get much wider C-cores. Seems like this would result in needing a ton of copper, though and then the surface area of the windings would also go up leading to increased capacitances.
If I were to go with C-cores, is it best to just stick with a window that is similar in width to what EI-150 would give me? I get considerably more height for the same width.
EI-150 has a window that is 19mm x 57mm. C-cores are 32mm x 60mm or I could even go with 80mm or 100mm for the width.
If I understand correctly, the longer core lengths would result in lower inductance per turn, but the other concern is that the wider cores would also need a lot of copper to fill.
Copper is pretty cheap compared to the time put into winding a tfmr, £15 gets a 1kg roll here in the UK enough to wind three big tfmr's ish. Never wound an OPT though I have wound numerous mains tfmrs & I do have some big C cores and toroid cores once I have the courage to start doing the calcs. I have a mate in Oz who used to wind OPT's,he kindly gave me four of the last few he had. He used more laminations than a commercial OPT, they sound ******** lovely, more iron = better low frequency though I know there are caveats to that ethos. You also have to bear in mind that most commercial OPT's are built down to a price, that said comepare the size of a Sowter OPT to an Edcor or Hammond.
Andy.
Andy.
I'd be interested in apprenticing with an experienced winder, I'd sweep his floor, organize his shop, get him coffee, clean his bathroom, unload his deliveries, for free! I see the craft of it as being the big hurdle, tooling, materials, the finesse, the varnishes, making the mistakes, the skill, etc. Until you have all that down knowing all the theory in the world won't help. Its like if I tried to be an expert welder just from a book and not experiencing thousands of welding jobs and making thousands of mistakes. The best winders probably don't have time to make learning videos here, but I'd pay for them if they did. If the forum is advancing the science of tube design, a huge part of that is the iron.
Using a larger core CAN (not always) be a smart idea since you can get more inductance and less capacitance with more windings. You will have greater copper and core losses, but that's not a big deal. You can worry about that stuff when you have a professional winding machine.
Another thought is to wind a parafeed transformer and a matching choke.
You will get significantly better performance for a little bit more effort.
You are vaguely
I'd love to accept such help and I also love giving advice, but.. Keep in mind that few people keep doing this job worldwide and not because they're stupid. I dare to say it's the other way around. It can be mentally backbreaking, especially when making a living out of it. Think about it carefully.
Designing and building a few OPTs, ITs, chokes, power transformers is no big deal, but I had to design over 300 prototypes for a timespan of 5 years to acquire some basic knowledge, such as market and customers needs, customer's way of thinking, own experience. Then at some point the prototype way of thinking comes to an end and production efficiency comes into play, that's a whole new story.
Half a year ago, I started attempting giving the OEM production to other reputable winder factories. This included tutoring from my part. It was a disaster.
I'd love to accept such help and I also love giving advice, but.. Keep in mind that few people keep doing this job worldwide and not because they're stupid. I dare to say it's the other way around. It can be mentally backbreaking, especially when making a living out of it. Think about it carefully.
Designing and building a few OPTs, ITs, chokes, power transformers is no big deal, but I had to design over 300 prototypes for a timespan of 5 years to acquire some basic knowledge, such as market and customers needs, customer's way of thinking, own experience. Then at some point the prototype way of thinking comes to an end and production efficiency comes into play, that's a whole new story.
Half a year ago, I started attempting giving the OEM production to other reputable winder factories. This included tutoring from my part. It was a disaster.
I wish Gerry at Transcendar had an apprentice! All his knowledge is gone now! There are so few winders and designers with real practical knowledge left on earth. We need more people to get this knowledge and keep it going. So much has been lost in just the past 20 years because the old timers have died and not passed it along. It is truly a crime to let it all go without sharing it with someone to keep it going for the sake of the industry. I’d think people would want to learn a talent with a built in clientele.
Reminds me of the gentleman from the US who as an accomplished knifesmith went to Japan to study knifesmithing as an apprentice to a master smith. Spent the first 6 months cutting wood up into tiny cubes to make charcoal for the forge. Thankfully for him he stayed i think it was 2 years and learned a lot.
Usually if we show interest and listen the experienced folks will teach us.
Usually if we show interest and listen the experienced folks will teach us.
Transformer winding is part science, part art, as is building valve amps and the world is changing fast, the old apprenticeship system has gone; what kid nowadays wants to spend 10 years learning a craft,then another 50 honing that craft? Craftsmen like guru's are thin on the ground and as someone noted a lot of the old boys who tinkered with cats whisker radios back in the day have gone silent key, sad, but inevitable. Lastly, a boffin who used to design test gear at Marconi who lives near me said transformer designers are all a bit mad, he said the tfmr designer at Marconi used to do his calculations for a new tfmr lying down under his desk with his shoes off. Most of the old boys I've come across who wind OPT's are lets say eccentric, come to think of it I talk to myself and have been known to burst into hysterical laughing on occasion when working at the bench.
Having said all that a tfmr is just a lump of iron with copper wire wrapped round it, very simple really. My approach is to largely ignore all the complicated maths and calculations, take an old mains tfmr apart & use the lams to wind a new tfmr. They often don't work as designed 100%, I then work out why and try a few simple calculations to get a better tfmr as well as testing it, often I have to pull it apart and start again several times, each time they get better. Have a go, don't worry if the result aint right, forget all the waffle, get winding and have fun.
Andy.
Having said all that a tfmr is just a lump of iron with copper wire wrapped round it, very simple really. My approach is to largely ignore all the complicated maths and calculations, take an old mains tfmr apart & use the lams to wind a new tfmr. They often don't work as designed 100%, I then work out why and try a few simple calculations to get a better tfmr as well as testing it, often I have to pull it apart and start again several times, each time they get better. Have a go, don't worry if the result aint right, forget all the waffle, get winding and have fun.
Andy.
Last edited:
I would say it is normal. How many of us have a plethora of friends with the same passion? How many of us would be laying under their own desks without shoes if we had no forums to talk about it with people all around the world? 😆
Operator self improvising - misuse of glues as instructed, crossed turns in layers, poorly constructed bobbins, falling wires from insulation edge, hammering the coil (guys never heard of arch bulge compensation during winding). All of this happened despite instructions and mutual agreements it SHOULD NOT.
One company however took it seriously and is doing good now. It took them almost a year to learn to build the first set of coils.
Not far from true, it takes a specific mindset to design an OPT or any audio transformer, especially a prototype. Do it everyday and your mentality will definitely shift from "common normal". A beginner OPT designer has a TON of simultaneous problems. Imagine doing this unique activity for a long timespan, which rarely someone of your friends or family understands. When doing an activity with little social support, it's not unusual to start questioning is worthiness. That combined with extreme concentration, patience, confidence, handskill, study, takes years of buildup. You'll become geeky. But with some good self-awareness, you can remain social and nice to the rest of society.
Last edited:
Ouch, man.
I was more asking from the point of view of whether I could buy parts to only do a couple or if I would have to order an absurd amount of materials, making it a dumb thing (financially) to attempt to do for a guy who only ever intends on doing a couple, just for the fun of it and the learning.
I have no doubt that I could tackle the job and that my first cut at it would be adequate and usable, and that I'd learn a ton. There's nothing dumb about that...
Getting a non-absurd amount of material to experiment with might still cost an absurd amount of money. Not really a ‘Dumb’ thing if the knowledge you gain is worth the extra you spent. Personally I wouldn’t spend $1000 to make something I could have someone else make for 400, could buy for 200 and goes for 100 wholesale. Some people might. Or you may be able to do better.
Some additional comments may be appropriate.
You could ask around for dead OPT's - I certainly have kept about 5-10 vintage types, and maybe repair services may have kept more modern ones (especially guitar amp origins). The value is in doing the tear-down (including a suite of testing if some windings still allow), and then having the core (and possibly former) to do some trials (including a simple sample winding to try and measure inductance versus signal and other parameters).
There is also a learning curve with testing any construction sample, including impedance sweep to identify internal parameter values (or an LCR for snapshot frequency measurements) and basic HF resonances. Of course there is also the method to use any existing amp (and knowing how it performs for frequency response and stability) and then swap in the diy OPT and cross-compare.
If you luck upon an ok first up diy construction then you may not feel inclined to pursue the more onerous path of gaining more experience and working through the design and testing of how different aspects change things. As you may have gleaned, Patrick started with no formal education, but had the drive to seek out references, and the intellect to hammer them into submission over many years, and the projects to work through a complete design and test loop many times to end up knowing what made a difference and what didn't.
You could ask around for dead OPT's - I certainly have kept about 5-10 vintage types, and maybe repair services may have kept more modern ones (especially guitar amp origins). The value is in doing the tear-down (including a suite of testing if some windings still allow), and then having the core (and possibly former) to do some trials (including a simple sample winding to try and measure inductance versus signal and other parameters).
There is also a learning curve with testing any construction sample, including impedance sweep to identify internal parameter values (or an LCR for snapshot frequency measurements) and basic HF resonances. Of course there is also the method to use any existing amp (and knowing how it performs for frequency response and stability) and then swap in the diy OPT and cross-compare.
If you luck upon an ok first up diy construction then you may not feel inclined to pursue the more onerous path of gaining more experience and working through the design and testing of how different aspects change things. As you may have gleaned, Patrick started with no formal education, but had the drive to seek out references, and the intellect to hammer them into submission over many years, and the projects to work through a complete design and test loop many times to end up knowing what made a difference and what didn't.
Patrick Turner gave us all a great gift in his website. I downloaded it all before it went offline, but it seems others have put it up so it can live on. His practical examples that run you through the process are great but sometimes leave me wondering why, and so far I've been able to run to other references and figure things out when I have a question.
Right now I'm leaning toward not doing this now but maybe later after my kids fly from the nest in the next few years. It seems like $1000 is probably in the ballpark of the bare minimum I'd spend getting all set up to wind two transformers, but the next two would be much cheaper.
But I'll admit that I've spent ~$1000 on a pair of output transformers before. It's still a lot cheaper than a car or a boat hobby.
Right now I'm leaning toward not doing this now but maybe later after my kids fly from the nest in the next few years. It seems like $1000 is probably in the ballpark of the bare minimum I'd spend getting all set up to wind two transformers, but the next two would be much cheaper.
But I'll admit that I've spent ~$1000 on a pair of output transformers before. It's still a lot cheaper than a car or a boat hobby.
For over a hundred transformers built, you should be already designing them shoeless under your desk! Nice work btw!
Here's some tips for beginner DIY winders (non commercial).
1. Think of your target first. And start small. I consider a max of 5W@25Hz for SE, 20W for PP is a good beginning spot for first attempt OPTs. Such range can also work with some household materials for insulation.
2. A universal design will save you a lot of costs. To work on a budget, try working your set of products around the bulk of your materials (wire spool, core format).
3. Preferably start with power supply chokes and transformers. They have more forgiving mistakes.
4. Arm yourself with extreme patience and expect a lot of "beginner's luck". The later especially failing you when working on different models and conceptions. If pursuing further into this activity, you'll eventually find out EVERY thickness of wire and insulation type needs time to be empirically studied and accustomed to.
5. Studying your equipment is mandatory and can take a lot of time, especially between different models. You can experience a lot of failures in the beginning and only a trained winding equipment engineer can solve your problems fast. Early winder's problems such as wire snaps, entanglement, breaks, mandrel and bobbin imperfections can occur. And they all contribute and work as a whole.
6. You'll need at least an frequency generator and oscilloscope to verify your final result via square waves. Identifying square wave shapes takes some training. Preferably you'll also need a bode plotter.
7. Also an ohmmeter, various resistive loads and inductance meter. The later is tricky, you'll need low frequency with good voltage excitation.
1. Think of your target first. And start small. I consider a max of 5W@25Hz for SE, 20W for PP is a good beginning spot for first attempt OPTs. Such range can also work with some household materials for insulation.
2. A universal design will save you a lot of costs. To work on a budget, try working your set of products around the bulk of your materials (wire spool, core format).
3. Preferably start with power supply chokes and transformers. They have more forgiving mistakes.
4. Arm yourself with extreme patience and expect a lot of "beginner's luck". The later especially failing you when working on different models and conceptions. If pursuing further into this activity, you'll eventually find out EVERY thickness of wire and insulation type needs time to be empirically studied and accustomed to.
5. Studying your equipment is mandatory and can take a lot of time, especially between different models. You can experience a lot of failures in the beginning and only a trained winding equipment engineer can solve your problems fast. Early winder's problems such as wire snaps, entanglement, breaks, mandrel and bobbin imperfections can occur. And they all contribute and work as a whole.
6. You'll need at least an frequency generator and oscilloscope to verify your final result via square waves. Identifying square wave shapes takes some training. Preferably you'll also need a bode plotter.
7. Also an ohmmeter, various resistive loads and inductance meter. The later is tricky, you'll need low frequency with good voltage excitation.
If it is an interstage 1:1 that would be a peace of cake. Give some more information, turn ratio, copper loss etc etc