Well, since we sorta hi-jacked Wrenchone's "Killer" screen drive thread, I'll start this new one for discussing DIY toroid winders for making OTs.
The Potthoff patent being the starting point, but any new ideas are welcome too.
http://www.pat2pdf.org/patents/pat4127238.pdf
Hopefully the moderators can move our previous toroid related posts over to this thread? Thankyou.
Don
The Potthoff patent being the starting point, but any new ideas are welcome too.
http://www.pat2pdf.org/patents/pat4127238.pdf
Hopefully the moderators can move our previous toroid related posts over to this thread? Thankyou.
Don
I'll mention some other issues related to the toroid winder for cogitating on. Some of these are probably dealt with in the Potthoff patent write-up, I need to go back and read it again. Been a while since I looked at it.
1) Turns counting on the toroid. The winding of turns on the toroid is not quite synchronous with the hand crank rotation since the wire spill off point moves around the edge of the C during operation. Some kind of optical interrupter is likely called for.
2) Rotation of the toroid core. The Potthoff design shows a magnetic chuck holding the toroid core, which is in turn mounted on a rotary table with a screw drive. This can only move the core thru some limited rotation angle before the core has to be re-positioned in the chuck. Not very convenient if you want to automate this whole thing.
Other winders use multiple concave rubber rollers, spring loaded against the outside of the core/winding assembly. These are driven to rotate the core then. This solves the limited rotation angle problem, but introduces some inaccuracy in the rotation angle control due to winding thickness on the core. Perhaps some combination of the two schemes, with an electromagnetically controlled chuck for turning control and spring loaded passive rollers for temporary holding during chuck repositionings. Still, there may be a centering problem during repositioning. Accuracy of core rotation is desirable for accurate progressive winding with its fine core dithering.
3) The hand crank operation shown in the patent does have a certain iconic appearance for the result of trillions of our tax dollars in development cost (apple peeler?). But for automating this, I would use a smaller cogwheel sprocket in the center of the C assembly for the belt drive and a hefty stepper motor to drive it.
4) Automation control. It would seem almost de-rigour to propose a programmed PC control of a bunch of stepper motors for operating this. But for DIY purposes, I think this level of automation and development is un-necessary. I think a foot pedal or "throttle-lever" operated pulse rate generator pot, routed thru a bunch of TTL counter chips with binary present control switch banks (setting counter divide ratios) would be sufficient to control the steppers. Then one can derive tables for typical counter settings for progressive wind rates, and dither angle etc. which can be interpolated for new designs. But if some PC expert out there would like to write up a Basic software control module with fancy visual screen interface, that's fine too.
5) During the pre-wind of wire onto the shuttle or "shuttleless C", it might be helpful it the wire guide between master spool and the shuttle or C could dither back and forth some to evenly distribute the wire build up. Either some flex/spring in the support for manual control or some stepper controlled contrivance for automated operation.
6) ?
Don
1) Turns counting on the toroid. The winding of turns on the toroid is not quite synchronous with the hand crank rotation since the wire spill off point moves around the edge of the C during operation. Some kind of optical interrupter is likely called for.
2) Rotation of the toroid core. The Potthoff design shows a magnetic chuck holding the toroid core, which is in turn mounted on a rotary table with a screw drive. This can only move the core thru some limited rotation angle before the core has to be re-positioned in the chuck. Not very convenient if you want to automate this whole thing.
Other winders use multiple concave rubber rollers, spring loaded against the outside of the core/winding assembly. These are driven to rotate the core then. This solves the limited rotation angle problem, but introduces some inaccuracy in the rotation angle control due to winding thickness on the core. Perhaps some combination of the two schemes, with an electromagnetically controlled chuck for turning control and spring loaded passive rollers for temporary holding during chuck repositionings. Still, there may be a centering problem during repositioning. Accuracy of core rotation is desirable for accurate progressive winding with its fine core dithering.
3) The hand crank operation shown in the patent does have a certain iconic appearance for the result of trillions of our tax dollars in development cost (apple peeler?). But for automating this, I would use a smaller cogwheel sprocket in the center of the C assembly for the belt drive and a hefty stepper motor to drive it.
4) Automation control. It would seem almost de-rigour to propose a programmed PC control of a bunch of stepper motors for operating this. But for DIY purposes, I think this level of automation and development is un-necessary. I think a foot pedal or "throttle-lever" operated pulse rate generator pot, routed thru a bunch of TTL counter chips with binary present control switch banks (setting counter divide ratios) would be sufficient to control the steppers. Then one can derive tables for typical counter settings for progressive wind rates, and dither angle etc. which can be interpolated for new designs. But if some PC expert out there would like to write up a Basic software control module with fancy visual screen interface, that's fine too.
5) During the pre-wind of wire onto the shuttle or "shuttleless C", it might be helpful it the wire guide between master spool and the shuttle or C could dither back and forth some to evenly distribute the wire build up. Either some flex/spring in the support for manual control or some stepper controlled contrivance for automated operation.
6) ?
Don
Here's a whopper - a 3-4 foot toroid wind on a Jovil winder. http://video.google.com/videoplay?docid=-660662384802328397
But these videos are not shuttleless as per Don's scheme!
Another thread on a DIY approach but again with shuttle: http://www.physicsforums.com/showthread.php?t=130135&page=3
Perhaps there is some nuggets in these?
But these videos are not shuttleless as per Don's scheme!
Another thread on a DIY approach but again with shuttle: http://www.physicsforums.com/showthread.php?t=130135&page=3
Perhaps there is some nuggets in these?
Here' another DIY effort but again not shutterless: http://waterfuelcell.org/phpBB2/viewtopic.php?t=1034
Unfortunately I can't see any of the videos on my computer. Reading thru the threads was interesting. But again, no pictures here.
The physics thread seems to discover the wire going slack each turn just before it quits with a vague approach of re-winding the wire back on the shuttle. The upside down shuttle slot approach is interesting though, and used by several commercial designs. The other discussion using the bicycle rim comes up with some guide plates to constrain the slack wire. Sort of precursor thinking to a slack accumulator.
Looks like using rubber drive wheels for rotating the toroid disturbs the windings, and wouldn't be very accurate once windings are on the core. I like the magnetic chuck approach used in the Potthoff design. I would improve it to use guide wheels during repositioning of the toroid on it. (the wheels would statically hold the toroid while the de-activated magnetic chuck rotates incrementally to a new angle then re-energizes)
If anyone noticed any technique used in the videos for handling the slack wire each turn, please mention it.
I have come to the conclusion that it is actually a plus to have slack wire available off the "shuttle" prewinding when handled correctly. In the Potthoff design, most of the actual winding onto the toroid comes from wire accumulated in the loop accumulator slot. This makes for smooth control of wire tension when winding around toroids with a square cross section. I would use a velvet covered, metal backed sheet pressed against the side of the Teflon C to control the tension in the slack loop accumulator.
Then I would construct the slotted C from three round Teflon sheets. This way, the accumulator edge sheet can be mounted slightly off center from the cog belt track/slot. This would allow the height of the rounded edge rim, that the wire spills over, to be varied versus rotation angle around the C. Where wire is already accumulated in the slack loop, the barrier could be higher to resist feeding from the prewound "shuttle".
From where the slack loop has pulled down tight on the core (meaning the position of the spillover point then), out until the spillover point reaches the furthest point from the core, the edge rim could be lower to allow easy spillover from the pre-wind to form the next slack loop.
Could it be that only the Potthoff design has the accurate control of wire position needed for progressive winds? Seems that the shuttle type designs loose tension on the wire every rotation and suffer from a slack uncontrolled wire loop. This might explain why most commercial toroids are wound "randomly".
Don
The physics thread seems to discover the wire going slack each turn just before it quits with a vague approach of re-winding the wire back on the shuttle. The upside down shuttle slot approach is interesting though, and used by several commercial designs. The other discussion using the bicycle rim comes up with some guide plates to constrain the slack wire. Sort of precursor thinking to a slack accumulator.
Looks like using rubber drive wheels for rotating the toroid disturbs the windings, and wouldn't be very accurate once windings are on the core. I like the magnetic chuck approach used in the Potthoff design. I would improve it to use guide wheels during repositioning of the toroid on it. (the wheels would statically hold the toroid while the de-activated magnetic chuck rotates incrementally to a new angle then re-energizes)
If anyone noticed any technique used in the videos for handling the slack wire each turn, please mention it.
I have come to the conclusion that it is actually a plus to have slack wire available off the "shuttle" prewinding when handled correctly. In the Potthoff design, most of the actual winding onto the toroid comes from wire accumulated in the loop accumulator slot. This makes for smooth control of wire tension when winding around toroids with a square cross section. I would use a velvet covered, metal backed sheet pressed against the side of the Teflon C to control the tension in the slack loop accumulator.
Then I would construct the slotted C from three round Teflon sheets. This way, the accumulator edge sheet can be mounted slightly off center from the cog belt track/slot. This would allow the height of the rounded edge rim, that the wire spills over, to be varied versus rotation angle around the C. Where wire is already accumulated in the slack loop, the barrier could be higher to resist feeding from the prewound "shuttle".
From where the slack loop has pulled down tight on the core (meaning the position of the spillover point then), out until the spillover point reaches the furthest point from the core, the edge rim could be lower to allow easy spillover from the pre-wind to form the next slack loop.
Could it be that only the Potthoff design has the accurate control of wire position needed for progressive winds? Seems that the shuttle type designs loose tension on the wire every rotation and suffer from a slack uncontrolled wire loop. This might explain why most commercial toroids are wound "randomly".
Don
Don,
You may have addressed this already & I missed it but how do you envisage the spill-off point moving across the gap in the C?
With the shuttle versions, the spill-off point is continuously controlled by a guide on the outside of the rim but this is not feasible in the Pothoff design as the rim no longer passes through the toroid but bypasses the toroid. This would mean the spill-off guide would have to drop the wire as it enters the C gap & pick it back up after coming out of the C gap.
Does this make sense?
Here's some other shuttleless toroid winder patent application from 1957 re US3050266 & from 1955 US2978193 & 1962 US3191878. I haven't had a chance to examine them but there may be something of interest in them.
You may have addressed this already & I missed it but how do you envisage the spill-off point moving across the gap in the C?
With the shuttle versions, the spill-off point is continuously controlled by a guide on the outside of the rim but this is not feasible in the Pothoff design as the rim no longer passes through the toroid but bypasses the toroid. This would mean the spill-off guide would have to drop the wire as it enters the C gap & pick it back up after coming out of the C gap.
Does this make sense?
Here's some other shuttleless toroid winder patent application from 1957 re US3050266 & from 1955 US2978193 & 1962 US3191878. I haven't had a chance to examine them but there may be something of interest in them.
"how do you envisage the spill-off point moving across the gap in the C?"
Yes, I mentioned this issue earlier (in the earlier thread section). The spill-off has to be captured again after crossing the C's gap. I would imagine that the leading edge of the Teflon C needs some judicious filing to get a long tapered knife like edge that expands back out and upward to the rounded spillover edge. It needs to radially expand upward from the belt's outer radius (bottom of the wire pile) up to the normal edge radius in order to capture the wire as it moves by.
It needs to first capture the wire, then bring it up to the normal rounded C edge level. There should be some V shape to the wire gap (relative to the belt) as the spill-off comes up thru the gap since the slack wire loop slot is feeding in the wire from maybe a 1/4 inch outward from the belt. So it needs to get the leading edge of the C inserted into that gap as it passes by. I need to read the patent details to see if that is mentioned. It could be one of those critical details intentionally left out though. One might be able to clip some kind of guide piece in place after inserting the toroid in the C's gap too.
Here is a picture of a Jovil head. Its using a shuttle still. But notice the toothbrush like appendage on the side (spring loaded) for holding back slack wire. Also, it is curiously using a cog belt to drive the shuttle. From their write up, this is necessary for heavy wire winding. Apparently the usual rubber drive wheels can't handle the torque for heavy stuff.
http://www.infantron.com.sg/manufact/jovil/6fb-pic.htm
http://www.infantron.com.sg/manufact/jovil/2fb-pic.htm
this one has got a much bigger slack wire dragger more similar to the Potthoff friction slot:
http://www.infantron.com.sg/manufact/jovil/7qa-pic.htm
Don
Yes, I mentioned this issue earlier (in the earlier thread section). The spill-off has to be captured again after crossing the C's gap. I would imagine that the leading edge of the Teflon C needs some judicious filing to get a long tapered knife like edge that expands back out and upward to the rounded spillover edge. It needs to radially expand upward from the belt's outer radius (bottom of the wire pile) up to the normal edge radius in order to capture the wire as it moves by.
It needs to first capture the wire, then bring it up to the normal rounded C edge level. There should be some V shape to the wire gap (relative to the belt) as the spill-off comes up thru the gap since the slack wire loop slot is feeding in the wire from maybe a 1/4 inch outward from the belt. So it needs to get the leading edge of the C inserted into that gap as it passes by. I need to read the patent details to see if that is mentioned. It could be one of those critical details intentionally left out though. One might be able to clip some kind of guide piece in place after inserting the toroid in the C's gap too.
Here is a picture of a Jovil head. Its using a shuttle still. But notice the toothbrush like appendage on the side (spring loaded) for holding back slack wire. Also, it is curiously using a cog belt to drive the shuttle. From their write up, this is necessary for heavy wire winding. Apparently the usual rubber drive wheels can't handle the torque for heavy stuff.
http://www.infantron.com.sg/manufact/jovil/6fb-pic.htm
http://www.infantron.com.sg/manufact/jovil/2fb-pic.htm
this one has got a much bigger slack wire dragger more similar to the Potthoff friction slot:
http://www.infantron.com.sg/manufact/jovil/7qa-pic.htm
Don
This is apparently a Jovil head for putting mylar insulating tape on the toroid core. Makes no sense to me how it could work. Any ideas? Appears to have a shuttle with rollers imbedded in it. Then a leather belt wrapped around it that turns the rollers as the shuttle rotates by. The belt appears to be tied down at the ends.
http://www.infantron.com.sg/manufact/jovil/7th-pic.htm
Don
edit:
Maybe something like this: The mylar is wound up on the shuttle in a single stack and fed to the toroid from the bottom (inside) side as the shuttle rotates around the toroid. The leather strap is just a friction device to keep the mylar tight. As the mylar stacked winding is used up from the inside it rotates with respect to the shuttle on the rollers embedded in the shuttle.
This is clearly a new technique. I can't help but wonder if it could be adapted to winding wire as well.
On further analysis, this only works for very flexible stuff like thin mylar film, since it makes a 180 degree turn around the roller feeding the tape from the bottom. This would be too hard on wire without making the shuttle rather thick.
http://www.infantron.com.sg/manufact/jovil/7th-pic.htm
Don
edit:
Maybe something like this: The mylar is wound up on the shuttle in a single stack and fed to the toroid from the bottom (inside) side as the shuttle rotates around the toroid. The leather strap is just a friction device to keep the mylar tight. As the mylar stacked winding is used up from the inside it rotates with respect to the shuttle on the rollers embedded in the shuttle.
This is clearly a new technique. I can't help but wonder if it could be adapted to winding wire as well.
On further analysis, this only works for very flexible stuff like thin mylar film, since it makes a 180 degree turn around the roller feeding the tape from the bottom. This would be too hard on wire without making the shuttle rather thick.
"some other shuttleless toroid winder patent application from 1957....."
The two early patents mentioned use two external shuttles with the wire bundle rounted around both of them like a belt on two pulleys (with the toroid core over one of the straight sections). Wire spills over the edges there too. The 1st one using double spherical surfaces to route the spilled wire loop, like I had earlier suggested. The second one reverts to just flat friction loop accumulators. Both of these double shuttle designs would severely wear the wire/insulation with all that dynaminc bending of the wire bundle.
The third one is in the Rube-Goldberg category in my opinion. Might as well program a robot to hand wind it.
Back to the spill-off in the C gap issue (Potthoff design). I read thru the Potthoff patent and there's no mention of this issue. But on thinking about this, I think it is easy to solve. The leading edge of the Teflon gap needs to be knife edge like and immediately adjacent to the belt, and a little lower than the belt's top surface (ie, just below the bottom of the wire bundle).
That way the wire can spill over the belt edge onto the lower Teflon edge (likely at the position of the leading C edge roller bearing below the belt, which will be curving it) just like normal spill-over throughout the remaining Teflon C track. Once the wire has dropped onto the Teflon knife edge it is captured, and the Teflon can now rise and widen as the rotation angle proceeds.
Eventually (smoothly) it rises back up and widens to the normal Teflon rounded spill edge. So judicious filing of the Teflon outer C leading edge is called for. This could be a mechanical wear point however (for the Teflon), unless the wire naturally wears the same contour, which it might. (a self sharpening knife?)
Don
The two early patents mentioned use two external shuttles with the wire bundle rounted around both of them like a belt on two pulleys (with the toroid core over one of the straight sections). Wire spills over the edges there too. The 1st one using double spherical surfaces to route the spilled wire loop, like I had earlier suggested. The second one reverts to just flat friction loop accumulators. Both of these double shuttle designs would severely wear the wire/insulation with all that dynaminc bending of the wire bundle.
The third one is in the Rube-Goldberg category in my opinion. Might as well program a robot to hand wind it.
Back to the spill-off in the C gap issue (Potthoff design). I read thru the Potthoff patent and there's no mention of this issue. But on thinking about this, I think it is easy to solve. The leading edge of the Teflon gap needs to be knife edge like and immediately adjacent to the belt, and a little lower than the belt's top surface (ie, just below the bottom of the wire bundle).
That way the wire can spill over the belt edge onto the lower Teflon edge (likely at the position of the leading C edge roller bearing below the belt, which will be curving it) just like normal spill-over throughout the remaining Teflon C track. Once the wire has dropped onto the Teflon knife edge it is captured, and the Teflon can now rise and widen as the rotation angle proceeds.
Eventually (smoothly) it rises back up and widens to the normal Teflon rounded spill edge. So judicious filing of the Teflon outer C leading edge is called for. This could be a mechanical wear point however (for the Teflon), unless the wire naturally wears the same contour, which it might. (a self sharpening knife?)
Don
Potthoff says (col 3, lines 15 and next)
"... Belt 46 is now turned in the same direction as it was to place the wire onto [C shaped] belt loop ..."
This means, I believe, that the wire is unwound off the belt the inner turn first.
Seems feasible if there are only few layers stored on the belt loop.
Confused !
Yves.
"... Belt 46 is now turned in the same direction as it was to place the wire onto [C shaped] belt loop ..."
This means, I believe, that the wire is unwound off the belt the inner turn first.
Seems feasible if there are only few layers stored on the belt loop.
Confused !
Yves.
Looking at figure 2 in the Potthoff patent, I notice that the wire bundle is shown as a straight section thru the gap of the C. This would cause undesireable bending of the wire bundle for each turn of the "shuttle". (which could be 1000s of times for a Hi Z primary) A few fixes for this come to mind. First it depends on how much tension is applied to the wire during the initial windup on the "shuttle" as to whether the wire actually straightens there, so maybe it is not really a problem.
If it does prove to be a problem, a couple of fixes:
A thin metal section, shaped with the same curvature could be inserted into the C gap after insertion of the core. It would have a covering of Teflon adhesive tape over the outside curvature (this does exist, I have a roll, but it's expensive) to minimize friction. More elaborate, would be a machined piece of Teflon to fill the gap, but it could include a rounded edge for transporting the spill-off wire across with the same form factor as the rest of the C's spill-off edge.
Alternatively, a curved section of spring metal could be fitted to the surface of the cog belt, just like the bottom of a normal shuttle. It would have a gap in it for fitting thru the toroid initially, that could be taped shut (fiberglass based adhesive tape). This would rotate around with the wire bundle just like a normal shuttle, but forming just the bottom constraint.
On upper windings on the core, one might just use Teflon insulating tape between the layers, and let the existing lower winding shape provide the sliding support for the shuttle wire bundle passing by.
This would eliminate space consuming inserts as the winding window gets cramped. Although for high bandwidth OT work, one does not want to fill the core window, a long thin winding gives low leakage L.
------------------------------
Some undeveloped (and deranged?) ideas for making a toroid winder:
Seeing the Jovil mylar strip winding setup, gets me thinking that maybe something could be done using a fixed Teflon bobbin or shuttleless C configuration with a Mylar belt sliding along under (or over) the wire bundle for rotation of the wire. A continuous belt of this could easily be formed by taping together the ends after passing thru the toroid. Since large spools of Mylar are available, this may not even require a continuous loop belt, but could just de-spool and re-spool the mylar film continuously during a wind, capstan driven like a tape recorder.
The large 12 inch Jovil gear driven shuttles appear to bottom feed wire off the shuttle thru an inner slot (maybe a V bottom inside), and have the shuttle split in concentric sections somehow. One section appears to be gear driven, the other appears to be either friction dragged by a brake shoe in one case, or in the other case there is a second gear surface with maybe a drag or variable speed control motor connected. I think this scheme is driving the outside of the wire bundle and re-winding the slack wire during each turn back into the shuttle instead of letting it spill out into a slack loop. I notice that the specifications still allow progressive winding, like the smaller shuttles using friction plates (which this 12 inch spool does not have, nor need apparently)
The mylar tape winding shuttle scheme with rollers in the shuttle also gives me some ideas. If the shuttle is made of Teflon, no rollers are required (the Mylar stack has to rotate with respect to the shuttle and its bottom feed slot), so the shuttle can still be thin. Next, instead of a leather friction drag belt on the outside of the shuttle rubbing the top of the mylar wound stack (the shuttle itself is driven in their setup), this could be a driven belt around the whole loop.
This way the mylar does not have to perform a 180 degree turn coming out of the bottom feed slot, and rewinds on the shuttle when slack wire results. Now, replacing the mylar wound stack with wire, we have the equivalent of the Jovil 12 inch design. Bottom feeding wire from a V slot on the shuttle bottom thru an angled feed hole. Belt drive of the wire bundle itself on the top surface, the shuttle itself free wheeling or controlled by a computer controlled drive motor (variable speed during each shuttle rotation as it dispenses wire from its slot onto the toroid surface or rewinds slack wire back into the shuttle). No friction loop accumulator or edge spill-off needed here. Wire position is controlled continuously/accurately by the bottom feed hole.
Just for fun, I'm going to try getting a quote on a useable Jovil SMC-2 system (the minimum setup that can handle a 5 inch max OD winding), and for parts like the shuttle head or just a shuttle spool.
Don
If it does prove to be a problem, a couple of fixes:
A thin metal section, shaped with the same curvature could be inserted into the C gap after insertion of the core. It would have a covering of Teflon adhesive tape over the outside curvature (this does exist, I have a roll, but it's expensive) to minimize friction. More elaborate, would be a machined piece of Teflon to fill the gap, but it could include a rounded edge for transporting the spill-off wire across with the same form factor as the rest of the C's spill-off edge.
Alternatively, a curved section of spring metal could be fitted to the surface of the cog belt, just like the bottom of a normal shuttle. It would have a gap in it for fitting thru the toroid initially, that could be taped shut (fiberglass based adhesive tape). This would rotate around with the wire bundle just like a normal shuttle, but forming just the bottom constraint.
On upper windings on the core, one might just use Teflon insulating tape between the layers, and let the existing lower winding shape provide the sliding support for the shuttle wire bundle passing by.
This would eliminate space consuming inserts as the winding window gets cramped. Although for high bandwidth OT work, one does not want to fill the core window, a long thin winding gives low leakage L.
------------------------------
Some undeveloped (and deranged?) ideas for making a toroid winder:
Seeing the Jovil mylar strip winding setup, gets me thinking that maybe something could be done using a fixed Teflon bobbin or shuttleless C configuration with a Mylar belt sliding along under (or over) the wire bundle for rotation of the wire. A continuous belt of this could easily be formed by taping together the ends after passing thru the toroid. Since large spools of Mylar are available, this may not even require a continuous loop belt, but could just de-spool and re-spool the mylar film continuously during a wind, capstan driven like a tape recorder.
The large 12 inch Jovil gear driven shuttles appear to bottom feed wire off the shuttle thru an inner slot (maybe a V bottom inside), and have the shuttle split in concentric sections somehow. One section appears to be gear driven, the other appears to be either friction dragged by a brake shoe in one case, or in the other case there is a second gear surface with maybe a drag or variable speed control motor connected. I think this scheme is driving the outside of the wire bundle and re-winding the slack wire during each turn back into the shuttle instead of letting it spill out into a slack loop. I notice that the specifications still allow progressive winding, like the smaller shuttles using friction plates (which this 12 inch spool does not have, nor need apparently)
The mylar tape winding shuttle scheme with rollers in the shuttle also gives me some ideas. If the shuttle is made of Teflon, no rollers are required (the Mylar stack has to rotate with respect to the shuttle and its bottom feed slot), so the shuttle can still be thin. Next, instead of a leather friction drag belt on the outside of the shuttle rubbing the top of the mylar wound stack (the shuttle itself is driven in their setup), this could be a driven belt around the whole loop.
This way the mylar does not have to perform a 180 degree turn coming out of the bottom feed slot, and rewinds on the shuttle when slack wire results. Now, replacing the mylar wound stack with wire, we have the equivalent of the Jovil 12 inch design. Bottom feeding wire from a V slot on the shuttle bottom thru an angled feed hole. Belt drive of the wire bundle itself on the top surface, the shuttle itself free wheeling or controlled by a computer controlled drive motor (variable speed during each shuttle rotation as it dispenses wire from its slot onto the toroid surface or rewinds slack wire back into the shuttle). No friction loop accumulator or edge spill-off needed here. Wire position is controlled continuously/accurately by the bottom feed hole.
Just for fun, I'm going to try getting a quote on a useable Jovil SMC-2 system (the minimum setup that can handle a 5 inch max OD winding), and for parts like the shuttle head or just a shuttle spool.
Don
Don,
I haven't read though your deranged? ideas section yet but some issues have been at the back of my mind:
- The circumference of the shuttle would have to be only slightly greater than the length of a full wind on the toroid otherwise each rotation of the shuttle will cause an accumulation of excess wire in the spill-off area.
- In other words unless most of the spilled-off wire is used up in a wind on the toroid it will simply accumulate because it can't be wound back onto the shuttle.
- Is this different to the continuous non-gapped shuttle which keeps the wire in the guide at all times?
I haven't read though your deranged? ideas section yet but some issues have been at the back of my mind:
- The circumference of the shuttle would have to be only slightly greater than the length of a full wind on the toroid otherwise each rotation of the shuttle will cause an accumulation of excess wire in the spill-off area.
- In other words unless most of the spilled-off wire is used up in a wind on the toroid it will simply accumulate because it can't be wound back onto the shuttle.
- Is this different to the continuous non-gapped shuttle which keeps the wire in the guide at all times?
I suppose you could spool up a toroid of enameled iron wire,
if you really wanted to "Roll your own" from scratch. Aren't
needing frequencies a low-tech iron core wouldn't handle...
Don just need to figure a machine to fill up this "core" bobbin
with iron, while copper primary and secondary windings are
already wrapped over the entire mess...
If the empty core bobbin is a "C", you won't need anything
fancy to wind the copper through the donut hole.
if you really wanted to "Roll your own" from scratch. Aren't
needing frequencies a low-tech iron core wouldn't handle...
Don just need to figure a machine to fill up this "core" bobbin
with iron, while copper primary and secondary windings are
already wrapped over the entire mess...
If the empty core bobbin is a "C", you won't need anything
fancy to wind the copper through the donut hole.
"The circumference of the shuttle would have to be only slightly greater than the length of a full wind on the toroid otherwise each rotation of the shuttle will cause an accumulation of excess wire in the spill-off area."
I wrestled with this issue myself, but finally concluded it is a non issue. 1st, either gapped C or continuous shuttles handle this the same way.
Imagine a really huge shuttle diameter, one with circumference vastly greater than a single turn length on the core. Seems like there will be way too much wire pulled off the shuttle when the spill-off is at the far side. But now imagine the same situation after winding one turn. There is still the same huge length of wire pulled off the shuttle forming the wire spoke to the back.
What happens is the 1 turn wound uses up a little of this each time, and the spill-off feeds just enough more, at some small angle just before reaching the far point again, to make up for the loss. The huge loop formed in the friction pad/loop accumulator simply gets pulled all the way thru to form the next long wire spoke to the back again next time. The spill-off can not (should not anyway, I notice there IS an EMERGENCY OFF switch on the Jovils) pull wire off until the wire has gone tight again (so only after the complete loop is pulled thru).
Obviously, an excessively sized (diameter) shuttle does cause the wire to get dragged around thru the friction/accumulator many times, but only one loop is ever present in it if it is operating correctly. This again highlights the need for correct control of friction in the spill-off and accumulator section and to variations of this with wire size.
Just a comment on the Jovil belt drive shuttle heads. At first I thought the belt wrapped around the shuttle was to obtain sufficient torque for turning the shuttle with large wire. But I now believe this is to increase friction at the spill off point for large wire. Ie, the wire has to pull-off between the shuttle and the tight rubber belt over it.
The angle range of the belt covering can also be put to use for controlling the spill-off force versus shuttle rotation angle, but I'm not sure they made use of this. One would like the required spill-off pulling force to be high while pulling the loop thru the core, but to be low while replenishing the wire in the loop for the next turn once its pulled up tight on the core.
Don
I wrestled with this issue myself, but finally concluded it is a non issue. 1st, either gapped C or continuous shuttles handle this the same way.
Imagine a really huge shuttle diameter, one with circumference vastly greater than a single turn length on the core. Seems like there will be way too much wire pulled off the shuttle when the spill-off is at the far side. But now imagine the same situation after winding one turn. There is still the same huge length of wire pulled off the shuttle forming the wire spoke to the back.
What happens is the 1 turn wound uses up a little of this each time, and the spill-off feeds just enough more, at some small angle just before reaching the far point again, to make up for the loss. The huge loop formed in the friction pad/loop accumulator simply gets pulled all the way thru to form the next long wire spoke to the back again next time. The spill-off can not (should not anyway, I notice there IS an EMERGENCY OFF switch on the Jovils) pull wire off until the wire has gone tight again (so only after the complete loop is pulled thru).
Obviously, an excessively sized (diameter) shuttle does cause the wire to get dragged around thru the friction/accumulator many times, but only one loop is ever present in it if it is operating correctly. This again highlights the need for correct control of friction in the spill-off and accumulator section and to variations of this with wire size.
Just a comment on the Jovil belt drive shuttle heads. At first I thought the belt wrapped around the shuttle was to obtain sufficient torque for turning the shuttle with large wire. But I now believe this is to increase friction at the spill off point for large wire. Ie, the wire has to pull-off between the shuttle and the tight rubber belt over it.
The angle range of the belt covering can also be put to use for controlling the spill-off force versus shuttle rotation angle, but I'm not sure they made use of this. One would like the required spill-off pulling force to be high while pulling the loop thru the core, but to be low while replenishing the wire in the loop for the next turn once its pulled up tight on the core.
Don
kenpeter said:
Yes, but that's another attracting/exciting part of this funny game
Don, be prepared to insert someting else than a mylar tape between windings to control stray capacitances
Yves.
Full manual approach ... for the fun.
http://www.audiyofan.org/forum/viewtopic.php?t=6822
re kenpeter:
"just need to figure a machine to fill up this "core" bobbin with iron, while copper primary and secondary windings are already wrapped over the entire mess..."
How about using Teflon covered steel wire wound in a BIG loop, then held flat on one side to wind the coils on. Then pull on the Teflon covered steel wire end until the whole thing shrivels up to a small round toroid.
"If the empty core bobbin is a "C", you won't need anything
fancy to wind the copper through the donut hole.
The problem with C cores is that they don't have the low leakage L of a toroid to get the huge bandwidths. They can do a little better than an EI if the windings are split (both primary and secondary split, this means ALL sections ie, the P-P halves too) between the two sides of the completed assembly.
To get low leakage L requires very long thin winding profiles so that the leakage paths are very long and comparable to the high Mu core pathes. The toroid is the master at solving this. It not only gets the long profile right, but even makes the leakage paths cancel by symmetry (provided each winding covers the full core). (unless the toroid hole gets significantly filled in with wire, then circular leakage paths form around the inner hole wires, which are shorter in circumference than the high Mu toroid path.)
Notice that toroids do not require interleaved windings either. With a progressive toroid winder, one can wind super performing toroids easier than EI or C cores. Just one automated progressive layer wind for each winding.
Don
"just need to figure a machine to fill up this "core" bobbin with iron, while copper primary and secondary windings are already wrapped over the entire mess..."
How about using Teflon covered steel wire wound in a BIG loop, then held flat on one side to wind the coils on. Then pull on the Teflon covered steel wire end until the whole thing shrivels up to a small round toroid.
"If the empty core bobbin is a "C", you won't need anything
fancy to wind the copper through the donut hole.
The problem with C cores is that they don't have the low leakage L of a toroid to get the huge bandwidths. They can do a little better than an EI if the windings are split (both primary and secondary split, this means ALL sections ie, the P-P halves too) between the two sides of the completed assembly.
To get low leakage L requires very long thin winding profiles so that the leakage paths are very long and comparable to the high Mu core pathes. The toroid is the master at solving this. It not only gets the long profile right, but even makes the leakage paths cancel by symmetry (provided each winding covers the full core). (unless the toroid hole gets significantly filled in with wire, then circular leakage paths form around the inner hole wires, which are shorter in circumference than the high Mu toroid path.)
Notice that toroids do not require interleaved windings either. With a progressive toroid winder, one can wind super performing toroids easier than EI or C cores. Just one automated progressive layer wind for each winding.
Don
smoking-amp said:
This is another thing I also had in mind - based on the above point - we need to have a sufficient toroid hole left after winding so could we revert to the standard circular shuttle and just make the shuttle cross-section small enough that it doesn't take up much more space than the bare wire bundle? I know this imposes a limitation on the smallest size core that can be used.
Just how many times do we need the ability to wind a toroid that has a smaller hole than the shuttle cross-section? Maybe for these rare occasions the final windings could be hand wound?
__________________________________________________
"Imagine a really huge shuttle diameter,....." - here's my logic on this situation - lets say the toroid needs a 1000 turns then the shuttle is loaded with 1000 turns of wire. When we have wound the 1000th turn on the toroid all the wire must be off the shuttle but this is vastly more wire than is needed to wind around the toroid so where is the wire located?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.