The Mad Science Challange -DIY MIC IPT

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Thanks, Abbey

Yeah, I see and agree your points, but (much) more experimentation definitely needed..

The unit we're going for is one of the simpler ones - the GW 9020-B6 linked below - but with additional "S6" magazine/head so that it can work both either as a Belt-system (for wire >0,2) or as Slider-system (for 0.05-0.2mm wire):

https://www.coilwindingmachinechina...ital-stepping-motor-coil-winding-machine.html
/Jakob E.
 
Im sad to say that my 1:8 IPT past away… the tiny wire got two breaks that I could not fix☹️

Its just to much work so I wont try again… I think 0,1 mm is probably the thinnest wire that is manageble, doing it ny hand… R.I.P
 
New plan... no more 0,05mm wire gage...
A layered cake with double electrostatic screens: 1:4+4

SIngel layer 0,1mm 845T (sec)
Electrostatic screen
Singel layer 0,4mm 210T (pri)
Electrostatic screen
Sectioned layer 0,1mm 845T (sec)
 
And we got a Winner 1:8 toroid IPT that beats Lundahl 1538 1:5 IPT in hi freq. response. The Lundahl specs -0,3dB at 100Khz with a 200 Ohm source and that is very hard to beat. Here are both Lundahl in black 1:8 toroid in red with a 2K source to provoce rolloff

my soundcard also starts to roll off above 20Khz so I cannot do any accurate measurment in the very hi end. But I have a clear relative winner…
 

Attachments

  • 8C416D6D-8C2D-461C-84CE-9535955B267E.jpeg
    8C416D6D-8C2D-461C-84CE-9535955B267E.jpeg
    200.5 KB · Views: 65
Last edited:
I polished the core so it would not cut the 0,05mm wire. Singel layer secondary + screen + singel layer pri 0,4mm👍
Primary inductance ca 4H5 dcr1,6 Ohm. Secondary dcr 720ohm
The mad science challange is completed😳
 

Attachments

  • C92F8C59-884F-49FF-80A6-E405D90F09A0.jpeg
    C92F8C59-884F-49FF-80A6-E405D90F09A0.jpeg
    138.7 KB · Views: 74
Last edited:
…its always hard to review your own stuff, but I like the sound of this toroid, it sounds detailed, clear and very natural… I connected the IPT directly to the tube grid without secondary load resistor an their is no hi freq ringing… the self resonance point must be well above the audio band… this one is a keeper…👍
 
Last edited:
Im start to wonder if inductance really is flat vs freq in this cores. Measured inductance at 250Hz is 4H. Calculated from 2K freq plot in REW is 25H. I will measure the trafos with freq generator + DVM and see what the results are.
 
worth investigating

The amorf cores I've seen until now have been rather even vs. frequency
If the inductance was 4H independent of freq. the -3dB in my 2K measurment should be around 80Hz, but its actually around 15Hz:unsure:
 
Last edited:
Lessons learned a couple of turns later:

Goal: Toroidal design that has better hi. freq. response than LL1538.

High step up design:
I think this singel layer aproach really is the way to go with toroids. 1 . you dont have to count turns when doing a complete layer (manual winding). 2. max. noise rejections. 3. Min. capacintance beteween turns. 4. Electrostatic screen min. capacitance beteween pri and. sec. 5. More than one screen becomes messy on the toroids, the insulated copper tejp never fitts perfect on a toroidal shape, Leakage inductance is always low on toroids, so you are ok with singel section pri and sec. 6. 2000 turns in total is pretty manageble for a 1:8 design.

High step up design: secondary first, then screen, then primary. You need a big core to get descent primary inductance with this aproach.

Expanding the concept:
0,25 mm pri/ 0,05mm sec =1:5
0,4mm pri/ 0,05mm sec =1:8
0,5 mm pri/ 0,05 mm sec 1:10
and so on...
lower step up higher pri inductance

Lower step up on smaller cores 1:4-5 range (much less critical design). Primary first (singel layer), then screen, then wind the secondary in section to min. capacitance between turns.


Wire gauge: 0,05mm is hard to handle and breaks easy. For newbies: 0,1mm wire is much more easy and robust (could be prefered at lower step up)

If anyone can handle 0,025mm wire you could do a 1:20 - 1 layer sec./screen/1 layer pri. super transformers... Anyone up for the challange? ;)
 

Attachments

  • ACF5D7FD-B72D-4118-966E-15FAF75D27E9.jpeg
    ACF5D7FD-B72D-4118-966E-15FAF75D27E9.jpeg
    157.4 KB · Views: 29
Last edited:

Latest posts

Back
Top