Alastair,
Yes I have just started playing with it. However I do not know what core I will be using at the moment. I have just tried the different cores shown within the program until I get a bobbin height larger than the height of the turns. I have used:
RP ohms 22500
Z6600<<<<<<<<<KT66 PP
Watts 16
IP 0.06
core m102b<<<<just a type that fits.
I have 7 sections 4primary and 3 secondary.
What did you do about the core? Did you use the program to give you a core or buy and input the data size etc?
Regards
M. Gregg
Yes I have just started playing with it. However I do not know what core I will be using at the moment. I have just tried the different cores shown within the program until I get a bobbin height larger than the height of the turns. I have used:
RP ohms 22500
Z6600<<<<<<<<<KT66 PP
Watts 16
IP 0.06
core m102b<<<<just a type that fits.
I have 7 sections 4primary and 3 secondary.
What did you do about the core? Did you use the program to give you a core or buy and input the data size etc?
Regards
M. Gregg
Hi M.Gregg.
A word about core size and their availabilty.
The most common form is "scrapless EI" laminations consitutued by a piece looking like an "E" and another like an "I" to close the magnetic path.
They are referenced as "EIxxx"
The "xxx" number indicates the size, but saddly, the same number from US (and, I suppose Brithish) manufacturers has not the same meaning than from German, French or Italian ones !
AFAIR (I was told by BudP), in the US system "xxx" represents the width on the center leg of the "E" part using a curious decimal fraction of inches while here -in froggy land-, it represents the lenght of the "I" part.
Fortunately (yes, it happens !) the lenght of the "I" is ALWAYS 3 times the width of the center leg ! !
From a link given by Sheldon in post #30, look here:
EI
where you can verify that the width of the center leg of the "E" (also named "tongue") is 1 inch multiplied by the tenth of the "xxx", that is 0.87 inches ( or 7/8 !)
Translated in millimiters the tongue is 22mm, and the lenght of the "I" is 3 times this value : 66mm.
Conclusion: The US "EI87" is equivalent to the "metric" EI66.
Putting all constants in a single formula:
EI"metric" = EI"US" * 0.765
Hope this helps !
And even more, each country seems to have its own "preference list" and some sizes may prove to be difficult to obtain at your location . . .
The M102 to wich you point has a different geometry and no so easy to find outside Germany.
Remind you'll have to obtain some additional pieces of hardware like bobbin (coil formers) bells, and so on.
Yves.
A word about core size and their availabilty.
The most common form is "scrapless EI" laminations consitutued by a piece looking like an "E" and another like an "I" to close the magnetic path.
They are referenced as "EIxxx"
The "xxx" number indicates the size, but saddly, the same number from US (and, I suppose Brithish) manufacturers has not the same meaning than from German, French or Italian ones !
AFAIR (I was told by BudP), in the US system "xxx" represents the width on the center leg of the "E" part using a curious decimal fraction of inches while here -in froggy land-, it represents the lenght of the "I" part.
Fortunately (yes, it happens !) the lenght of the "I" is ALWAYS 3 times the width of the center leg ! !
From a link given by Sheldon in post #30, look here:
EI
where you can verify that the width of the center leg of the "E" (also named "tongue") is 1 inch multiplied by the tenth of the "xxx", that is 0.87 inches ( or 7/8 !)
Translated in millimiters the tongue is 22mm, and the lenght of the "I" is 3 times this value : 66mm.
Conclusion: The US "EI87" is equivalent to the "metric" EI66.
Putting all constants in a single formula:
EI"metric" = EI"US" * 0.765
Hope this helps !
And even more, each country seems to have its own "preference list" and some sizes may prove to be difficult to obtain at your location . . .
The M102 to wich you point has a different geometry and no so easy to find outside Germany.
Remind you'll have to obtain some additional pieces of hardware like bobbin (coil formers) bells, and so on.
Yves.
Yves,
2 questions.
Is the best option to go for an oversized core or will that create losses?
Is it better to put brown paper insulation between each layer to use up space on the bobbin (Better high frequency performance)
Or increase the size of the wire used to take up any spare space on the former?(lower DC resistance).
Regards
M. Gregg
2 questions.
Is the best option to go for an oversized core or will that create losses?
Is it better to put brown paper insulation between each layer to use up space on the bobbin (Better high frequency performance)
Or increase the size of the wire used to take up any spare space on the former?(lower DC resistance).
Regards
M. Gregg
JOCKER ? !
M.GREGG: Well, there are no "best" options (althought some are "worst") in designing transformer, specially output ones.
More iron means less iron losses (if properly "filled" with copper) and usually needs less turns of wire to maintain performance at lo end.
Also less turns reduce the leakage inductance that affect hi end.
But various parasitic caps increase since the size on the coil former is larger : more surface.
It'is then that you may need (more) paper between section AND between layer to keep this unwanted caps low.
But . . . this increase the height of the bobbin (the distance btwn the first and the last wound layer) and degrades the leakage inductance . . . and so on !
Play with software and observe the results !
Yves.
This is a great thread! Thank you so much to everyone who posted links to all the resources.
I'm interested in winding my own single ended OPT's myself for guitar amps of 1-5 Watts. I have buckets of very large 5 and 12 volt EI power transformers. I realize that the quality of the iron may not be quite as good. I'd appreciate comments on the likely results of using this kind of iron. Realize, this is not for hi-fi, but for a guitar amp. What would be the likely tradeoff?
Thanks again for an excellent thread!!
I'm interested in winding my own single ended OPT's myself for guitar amps of 1-5 Watts. I have buckets of very large 5 and 12 volt EI power transformers. I realize that the quality of the iron may not be quite as good. I'd appreciate comments on the likely results of using this kind of iron. Realize, this is not for hi-fi, but for a guitar amp. What would be the likely tradeoff?
Thanks again for an excellent thread!!
Actually, using power core for guitar amp outputs is a fine idea. In our line of guitar OPT's (see attachment) and powers we use the lower performance grades because they provide a far more musical response in the extremely dynamic environment of guitar amps.
If you download the RDH4 from Pete Millet's web site
Technical books online
the very first coil construction spoken of, one used in the RCA public address amplifier in the back section of the book (that Leo Fender derived his circuit from), is an extremely good winding configuration to follow. I would recommend using Nomex insulation material, two layers of 0.003" thick material will do nicely. You do need to edge seal the insulation to the bobbin walls with good quality, flexible electrical grade tape, about 3mm wide. Even better is to build up the bobbin side walls with 3mm wide tape, on both sides of each secondary section, and seal the Nomex to the bobbin walls with tape.
This provides a narrower winding space but does provide two benefits. 1.) Full safety agency creepage and clearance dimensions for the voltages involved.
2.) Linear flux transform for the antenna event that is really what is going on between primary and secondary windings, across the dielectric barriers. The core is just along for the ride above 250 Hz, with typical E/I core, control of this antenna event is what will control the predominant character of your amp.
Incidentally, that coil design I mentioned above will give you a Fender sound character. Using other materials as additional dielectric barriers, in between the two layers of Nomex, will allow you to change the response and to a degree tone character of your amp. Tone character is entirely electrostatic in nature. Using different grades of core will affect the leading edge rise time and settling time for signals. Using high dielectric constant plastics, between the Nomex layers will also affect the response characteristics, but only as amplitude increases.
Go explore! And, someday, take one of your concoctions to a transformer company and ask them to vacuum impregnate it with a polyester resin. This will change you forever.
Bud
If you download the RDH4 from Pete Millet's web site
Technical books online
the very first coil construction spoken of, one used in the RCA public address amplifier in the back section of the book (that Leo Fender derived his circuit from), is an extremely good winding configuration to follow. I would recommend using Nomex insulation material, two layers of 0.003" thick material will do nicely. You do need to edge seal the insulation to the bobbin walls with good quality, flexible electrical grade tape, about 3mm wide. Even better is to build up the bobbin side walls with 3mm wide tape, on both sides of each secondary section, and seal the Nomex to the bobbin walls with tape.
This provides a narrower winding space but does provide two benefits. 1.) Full safety agency creepage and clearance dimensions for the voltages involved.
2.) Linear flux transform for the antenna event that is really what is going on between primary and secondary windings, across the dielectric barriers. The core is just along for the ride above 250 Hz, with typical E/I core, control of this antenna event is what will control the predominant character of your amp.
Incidentally, that coil design I mentioned above will give you a Fender sound character. Using other materials as additional dielectric barriers, in between the two layers of Nomex, will allow you to change the response and to a degree tone character of your amp. Tone character is entirely electrostatic in nature. Using different grades of core will affect the leading edge rise time and settling time for signals. Using high dielectric constant plastics, between the Nomex layers will also affect the response characteristics, but only as amplitude increases.
Go explore! And, someday, take one of your concoctions to a transformer company and ask them to vacuum impregnate it with a polyester resin. This will change you forever.
Bud
Last edited:
Miles is right, winding good output transformers really takes something. I have been down this road, successfully and it was the most difficult thing I have done in this hobby to date. The basic math is the easy bit. Transformers are never perfect and the art is in balancing issues such as leakage inductance, distributed capacitance and shunt capacitance not to mention speaker interface and (for SE and sometimes PP), gap for a given tube. This takes the willingness to wind, and wind again and play with the section interconnection configurations. The variables are endless and there is no one text to cull this knowledge from, such knowledge tends to be protected, with reason, it takes serious commitment to put it all together.My first essay was a 11k job for 845 SE. It took three goes together with much testing of insulating materials to get the lowest capacitance and a stable surface to wind onto. Then there is the issue of obtaining good cores. I lucked out on 8, 2 mil hypersil C cores at a London junk emporium, one of my cases was HEAVY!
If you have the patience, the end result is extremely rewarding.
Richard.
If you have the patience, the end result is extremely rewarding.
Richard.
This text is in image format, it cannot be translated by online tools. Do you have them in written text?
Beware that there´s two main kinds of PDF.
One was straight "written" in PDF or converted from a "text based" source (even if containing images) such as a Word or HTML document
But the other is basically a page scan, so an image, just saved as PDF; this one does not contain google translatable text, not even text which can be copypasted, except as an image block.
That said, you can OCR each page, one by one, and translate that ... only it will be much longer and inevitably contain some errors.
There are some "automatic" tools to do that, even online, but they require proof reading and correction later.
One was straight "written" in PDF or converted from a "text based" source (even if containing images) such as a Word or HTML document
But the other is basically a page scan, so an image, just saved as PDF; this one does not contain google translatable text, not even text which can be copypasted, except as an image block.
That said, you can OCR each page, one by one, and translate that ... only it will be much longer and inevitably contain some errors.
There are some "automatic" tools to do that, even online, but they require proof reading and correction later.