AnalogPackrat
Well-known member
OK, I figured it made more sense to post this in the Lab than over on the 2503 transformer winding thread in the Drawing Board... I got on the winding bandwagon thanks to Mr. CJ. Met up with him before the holidays and got two sets of lams, two bobbins, and other bits and pieces. I had recently acquired some 3x34 trifilar magnet wire at a local surplus shop so I decided to try my hand at winding an output transformer for my APIs using it. I didn't need the fourth winding (VU meter) anyway.
I didn't expect to winding a bunch of things, so I set myself up with the lowest of low tech winding, uh, "equipment." Said equipment consisted of a few old books, some clean rags, some samples and offcuts of stone from my home remodel, a tapered stick of wood (found in garage), flat toothpicks, and yellow tape. Real cave-man stuff here.
I needed a way to allow the spool to smoothly unreel the wire and a way to keep good tension on the wire as I wound the coil. I arranged the books to support the spool so it could spin without rolling around or sliding off the bench. I folded the clean rags over the wire and stacked stone pieces on top to get a the right amount of friction for winding. The two stacks make it easier to back out a turn our two without fouling the spool. The bobbin fit over the stick and the flat toothpicks were wedged under to keep it in place.
The rag didn't produce enough friction so I added a layer of that thin white foamy sheet packing material inside the folded rag. This really worked well and solved the tension problem. Here are the first few turns going on the coil. Since I was using a three strand wire (instead of four as on the real 2503) and of smaller guage (34 instead of 30 or 31) I wasn't sure how many turns I would be putting on. I took CJs advice and just packed on as many as I could, leaving enough clearance for taping up at the end.
I got thirty turns per layer and at nine layers I had plenty of room left, so I kept going until I had 680 turns on the coil. Here is a good "winder's eye view" of the setup.
Measuring the raw coil inductance seemed like an interesting thing to do. The next step is adding the lead breakouts. I found some multicolored 22awg ribbon cable with the correct colors (per the original) and decided to use it.
-Ready to solder the starts.
-The Packrat in action with Mr. Weller (photo coutesy of Mrs. Packrat).
-Starts soldered and taped for strain relief.
-Finishes soldered.
-Last look at the breakouts!
-All taped up and ready to lace.
-E's laced--CJ threw in a couple of Nickel Steel lams which I used for the outer layers.
-Slip those I's into place.
-Add some brass machine screws, tap the lams all 'round to get them snug and she's done! :guinness:
If it works well, I've got another set of parts for a second one. Here are some measurements I made:
DCR of individual coils is 50.5ohms (one is 47ohms--slightly larger wire?).
Raw coil with air core has inductance of 8.12mH per winding measured with cheap-O Bel Merit digital LC meter.
Leakage capacitance between center and outer windings (remember I used flat trifilar wire) is 9.9nF and 6.7nF between the two outer ones.
Inductance of a winding with the core in is 0.82H (it was 0.5H with only the E's).
The lower DCR came from the center wire of the trifilar set which allowed me to use it as the primary to balance both the leakage capacitance and DCR on the secondaries. Not sure if it matters, but hey, it worked out. I noticed that I could squeeze the corners of the core and get the Inductance to read over 1.5H. I can't tighten the screws more because it starts to deform the core. Should I clamp the corners and epoxy/varnish the edges of the lams?
Does anyone have measurement for a real 2503, profile clone, and/or one of CJ's jobs? I'm curious how my bastard son of 2503 compares.
Thanks again to CJ for getting the parts together and for prividing all the guidance. When I finish my API pcbs I'll try the thing out...
A P
I didn't expect to winding a bunch of things, so I set myself up with the lowest of low tech winding, uh, "equipment." Said equipment consisted of a few old books, some clean rags, some samples and offcuts of stone from my home remodel, a tapered stick of wood (found in garage), flat toothpicks, and yellow tape. Real cave-man stuff here.
I needed a way to allow the spool to smoothly unreel the wire and a way to keep good tension on the wire as I wound the coil. I arranged the books to support the spool so it could spin without rolling around or sliding off the bench. I folded the clean rags over the wire and stacked stone pieces on top to get a the right amount of friction for winding. The two stacks make it easier to back out a turn our two without fouling the spool. The bobbin fit over the stick and the flat toothpicks were wedged under to keep it in place.
The rag didn't produce enough friction so I added a layer of that thin white foamy sheet packing material inside the folded rag. This really worked well and solved the tension problem. Here are the first few turns going on the coil. Since I was using a three strand wire (instead of four as on the real 2503) and of smaller guage (34 instead of 30 or 31) I wasn't sure how many turns I would be putting on. I took CJs advice and just packed on as many as I could, leaving enough clearance for taping up at the end.
I got thirty turns per layer and at nine layers I had plenty of room left, so I kept going until I had 680 turns on the coil. Here is a good "winder's eye view" of the setup.
Measuring the raw coil inductance seemed like an interesting thing to do. The next step is adding the lead breakouts. I found some multicolored 22awg ribbon cable with the correct colors (per the original) and decided to use it.
-Ready to solder the starts.
-The Packrat in action with Mr. Weller (photo coutesy of Mrs. Packrat).
-Starts soldered and taped for strain relief.
-Finishes soldered.
-Last look at the breakouts!
-All taped up and ready to lace.
-E's laced--CJ threw in a couple of Nickel Steel lams which I used for the outer layers.
-Slip those I's into place.
-Add some brass machine screws, tap the lams all 'round to get them snug and she's done! :guinness:
If it works well, I've got another set of parts for a second one. Here are some measurements I made:
DCR of individual coils is 50.5ohms (one is 47ohms--slightly larger wire?).
Raw coil with air core has inductance of 8.12mH per winding measured with cheap-O Bel Merit digital LC meter.
Leakage capacitance between center and outer windings (remember I used flat trifilar wire) is 9.9nF and 6.7nF between the two outer ones.
Inductance of a winding with the core in is 0.82H (it was 0.5H with only the E's).
The lower DCR came from the center wire of the trifilar set which allowed me to use it as the primary to balance both the leakage capacitance and DCR on the secondaries. Not sure if it matters, but hey, it worked out. I noticed that I could squeeze the corners of the core and get the Inductance to read over 1.5H. I can't tighten the screws more because it starts to deform the core. Should I clamp the corners and epoxy/varnish the edges of the lams?
Does anyone have measurement for a real 2503, profile clone, and/or one of CJ's jobs? I'm curious how my bastard son of 2503 compares.
Thanks again to CJ for getting the parts together and for prividing all the guidance. When I finish my API pcbs I'll try the thing out...
A P
