Burntcoil.
Throughout this video there is a lot of hi frequency distortion .
I think this is down to the down sampling used,
I have found the same problem when converting wav files to upload to this site.
The rest as you say , sounds like overloading.
Recordings can only give an idea of the sound, not the actual in room experience.
I was amused at the musical box and bending the paper demonstration, as that is not actually what they are doing .
Steve.
Throughout this video there is a lot of hi frequency distortion .
I think this is down to the down sampling used,
I have found the same problem when converting wav files to upload to this site.
The rest as you say , sounds like overloading.
Recordings can only give an idea of the sound, not the actual in room experience.
I was amused at the musical box and bending the paper demonstration, as that is not actually what they are doing .
Steve.
This sounds exactly like what I imagine if one took a spruce bodied cello and mounted an exciter where the bridge is and played cello music. 
Curved spruce panels with ribs and multiple exciters may be an extension to get different modes going and be able to reach wider bandwidth.
Curved spruce panels with ribs and multiple exciters may be an extension to get different modes going and be able to reach wider bandwidth.
All the machined components arrive on Monday so this is a first rough draft of a build and test approach.
Build and Test Stages for Option 1
Problem- IF HF performance out of design spec
Solution - Increase B by increasing motor size
- Wind lower impedance VC
Problem - IF efficiency is out of design spec
Solution- increase BL
I will be using REW with a calibrated Umic for the testing
The standard test panel will use a Dayton Audio DAEX32EP-4 on a plywood panel because it is a well known exciter design with published performance data and is an exciter I am very familiar with.
All helpful thoughts and criticisms are very welcome.
Burnt
Build and Test Stages for Option 1
- Calibrate magnetometer. Test ring and rectangular magnets fields using magnetometer and compare with published figures.
- Test and note field from two stacked ring magnets and two stacked rectangular magnets. compare with single test magnets to establish field increase as % of single magnet ( published figures suggest X 1.8 increase.)
- From 2 calculate potential field for Option 1 design
- Test assemble Option 1 magnet motor without VC pole pieces motor and test VC gap field strength.
- Compare results with calculated predicted field
- Apply VC pole pieces and retest VC gap field strength.
- Review results and decide on motor design. The options are, two layer magnet structure, four layer magnet structure, with VC pole pieces or without VC pole pieces
- Assemble motor design with adhesives
- Measure assembly and adjust motor casing and print out motor casing
- Calculate wire winding to give target BL, Rdc and Impedance
- Build VC former and wind VC
- Assemble VC to test panel and apply position shims
- Assemble motor to VC providing motor support.
- Remove position shims.
- Manually test motor and VC assembly to ensure free movement of VC in gap, adjust if required
- Run test regime using REW
- Run listening tests against standard panel
- Run high power test ( potential for ‘magic smoke’)
- Publish test results with test recordings and commentary.
Problem- IF HF performance out of design spec
Solution - Increase B by increasing motor size
- Wind lower impedance VC
Problem - IF efficiency is out of design spec
Solution- increase BL
I will be using REW with a calibrated Umic for the testing
The standard test panel will use a Dayton Audio DAEX32EP-4 on a plywood panel because it is a well known exciter design with published performance data and is an exciter I am very familiar with.
All helpful thoughts and criticisms are very welcome.
Burnt
Step 1- Calibration and Measurement
I am going to report on every step in this project in case someone wants to replicate the build and also to help me keep notes of my progress. First step is to calibrate the measurement gear I have against the published measurements.
The published data for the ring magnet shown in the photo below is
OD 40mm
ID 25mm
T 5mm
and these measure pretty much spot on.
The field strength is quoted as 3400 Gauss
Measuring with the magnetometer I bought gives a variable field which I will comment on in more detail in the next post.
The measurements I get for one ring magnet are 0-16 depending on the position around the circumference
The objective of this test is to calibrate my tester against the published data so I am going to take 16 as 3400 Gauss as the reference. Please note I am not claiming either my magnetometer is more accurate that the tester used by the manufacturer nor am I claiming the 3400 figure is correct.
Both magnetometer and magnets are consistent- a second sample measure the same. This gives me the calibration I need on a ‘good enough’ basis.
As ever, If anyone can give constructive guidance I am happy to receive it.
Burnt
I am going to report on every step in this project in case someone wants to replicate the build and also to help me keep notes of my progress. First step is to calibrate the measurement gear I have against the published measurements.
The published data for the ring magnet shown in the photo below is
OD 40mm
ID 25mm
T 5mm
and these measure pretty much spot on.
The field strength is quoted as 3400 Gauss
Measuring with the magnetometer I bought gives a variable field which I will comment on in more detail in the next post.
The measurements I get for one ring magnet are 0-16 depending on the position around the circumference
The objective of this test is to calibrate my tester against the published data so I am going to take 16 as 3400 Gauss as the reference. Please note I am not claiming either my magnetometer is more accurate that the tester used by the manufacturer nor am I claiming the 3400 figure is correct.
Both magnetometer and magnets are consistent- a second sample measure the same. This gives me the calibration I need on a ‘good enough’ basis.
As ever, If anyone can give constructive guidance I am happy to receive it.
Burnt
Step 2- Magnet Measurement
The first big surprise for me is the field of the ring magnet is nothing like I expected.
The ring magnet has an opposing pole on each face, positive on one side, negative on the other. What I had anticipated is that the field would be donut shaped with the field lines curving round from pole to pole. What was unexpected was that the field is not evenly distributed around the circumference, it is strongly lobed as shown in the sketch of the measurements below. The full set of measurements vary smoothly from the minimum to the maximum figure.
This lobed field is repeated when two rings are stacked. Hmmm, magnets are weird!
So I took the dog for a walk for a bit and realised, just like Spaniel hunting down a squirrel, the field lines are bunching up into two lobes rather than being evenly distributed because thy are seeking the minimum path between poles, which makes perfect sense. Thank you squirrel.
The other notable effect is that the field strength is increasing more or less linearly. The literature suggested filed strength would not double but would increase by 80% so I was expecting to see a maximum of circa 6120 Gauss. This result is pleasing but as you can see one lobe is more than twice the original figure which I expect is due to this field line bunching I am seeing and can’t be relied upon in a conservative motor design.
Magnets are not weird but they are very counterintuitive. Big learn for the future.
Burnt.
The first big surprise for me is the field of the ring magnet is nothing like I expected.
The ring magnet has an opposing pole on each face, positive on one side, negative on the other. What I had anticipated is that the field would be donut shaped with the field lines curving round from pole to pole. What was unexpected was that the field is not evenly distributed around the circumference, it is strongly lobed as shown in the sketch of the measurements below. The full set of measurements vary smoothly from the minimum to the maximum figure.
This lobed field is repeated when two rings are stacked. Hmmm, magnets are weird!
So I took the dog for a walk for a bit and realised, just like Spaniel hunting down a squirrel, the field lines are bunching up into two lobes rather than being evenly distributed because thy are seeking the minimum path between poles, which makes perfect sense. Thank you squirrel.
The other notable effect is that the field strength is increasing more or less linearly. The literature suggested filed strength would not double but would increase by 80% so I was expecting to see a maximum of circa 6120 Gauss. This result is pleasing but as you can see one lobe is more than twice the original figure which I expect is due to this field line bunching I am seeing and can’t be relied upon in a conservative motor design.
Magnets are not weird but they are very counterintuitive. Big learn for the future.
Burnt.
Step 3- Calculate Potential VC Gap B
So this is not in the right sequence and I will have to save establishing the VC B figure until I have an assembly as lobeing effect is making this impossible. My hope is that when assembled with the central magnetic pole the filed distribution evens out so I can measure the VC gap B directly. Lets see.
So this is not in the right sequence and I will have to save establishing the VC B figure until I have an assembly as lobeing effect is making this impossible. My hope is that when assembled with the central magnetic pole the filed distribution evens out so I can measure the VC gap B directly. Lets see.
Step 4A - Test Assemble Option 1
Why has Step 4 become Step 4A?
Good question. In rocketry there is a well known event known as an R.U.D. - a Rapid Unplanned Disassembly. When working with high strength magnets there is a similarly destructive event which I will call R.U.A. A Rapid Unplanned Assembly can result in magnets shattering and the remains of the magnet sticking to the rest of the assembly rendering it as useful as rocket wreckage. I knew this, but this morning I ‘just wanted to see” so of course the magnet took over and reminded me I need to build some jigs to help handle these magnets-there are non-trivial forces involved.
Also, there will be no test assembly, its not practical as keeping sufficient control of the magnets without using epoxy to permanently fix at least half of the assembly is not going to work. So its now a case of building a ‘complete’ magnetic assembly, using jigs and epoxy, and then testing each assembly.
I have managed to assemble part of the Option 1 motor, images below, and can report that applying the rear pole piece has led to an increase in the lobeing effect.
Burnt
Why has Step 4 become Step 4A?
Good question. In rocketry there is a well known event known as an R.U.D. - a Rapid Unplanned Disassembly. When working with high strength magnets there is a similarly destructive event which I will call R.U.A. A Rapid Unplanned Assembly can result in magnets shattering and the remains of the magnet sticking to the rest of the assembly rendering it as useful as rocket wreckage. I knew this, but this morning I ‘just wanted to see” so of course the magnet took over and reminded me I need to build some jigs to help handle these magnets-there are non-trivial forces involved.
Also, there will be no test assembly, its not practical as keeping sufficient control of the magnets without using epoxy to permanently fix at least half of the assembly is not going to work. So its now a case of building a ‘complete’ magnetic assembly, using jigs and epoxy, and then testing each assembly.
I have managed to assemble part of the Option 1 motor, images below, and can report that applying the rear pole piece has led to an increase in the lobeing effect.
Burnt
Attachments
Step 4B- Magnet Assembly
And with the use of a very simple printed jig the assembly of the central magnet was straightforward. You definitely need to handle Neodymium magnets with great care.
The effect on the field remains interesting. The external field reduced significantly, which is what you would expect with a 'closed' magnetic circuit. The lobeing effect is still present with measurements running from a minimum of 0 Gauss at two places on the circumference at 180 degrees from one another to a maximum of 2550 Gauss at two points 90 degrees from the zero poles.
It is proving impossible to measure the VC gap direct with the magnetometer so I will just have to see what sort of BL I get with the VC in place.
Burnt
And with the use of a very simple printed jig the assembly of the central magnet was straightforward. You definitely need to handle Neodymium magnets with great care.
The effect on the field remains interesting. The external field reduced significantly, which is what you would expect with a 'closed' magnetic circuit. The lobeing effect is still present with measurements running from a minimum of 0 Gauss at two places on the circumference at 180 degrees from one another to a maximum of 2550 Gauss at two points 90 degrees from the zero poles.
It is proving impossible to measure the VC gap direct with the magnetometer so I will just have to see what sort of BL I get with the VC in place.
Burnt
Attachments
Last edited:
Update on very slow progress
Progress is slow because Covid has turned up to Burnt Ranch and laid us all up but I managed to do some further assembly today. The end result is good but it was an ‘exciting’ process with the components all actively involved in ‘helping’. Pictures of assembly attached and for scale I have included a couple of my spare exciters for comparison.
I have yet to print out the mandrel to help form the Aluminium VC bobbin but I can at least model this over the next day or two. ( Covid wrecks your ability to think.)
From this point on I am still looking at two routes forward. The first is the most conventional using a suspension/spider attached to the VC bobbin and attached to the printed casing holding the magnet assembly. The second and more interesting route is to go the zero suspension/spider route where the VC is bonded directly to the panel. This raises a whole stack of positional issues but has the advantage of eliminating the distortion mechanisms inherent with the spider.
More news as it happens.
Burnt.
Progress is slow because Covid has turned up to Burnt Ranch and laid us all up but I managed to do some further assembly today. The end result is good but it was an ‘exciting’ process with the components all actively involved in ‘helping’. Pictures of assembly attached and for scale I have included a couple of my spare exciters for comparison.
I have yet to print out the mandrel to help form the Aluminium VC bobbin but I can at least model this over the next day or two. ( Covid wrecks your ability to think.)
From this point on I am still looking at two routes forward. The first is the most conventional using a suspension/spider attached to the VC bobbin and attached to the printed casing holding the magnet assembly. The second and more interesting route is to go the zero suspension/spider route where the VC is bonded directly to the panel. This raises a whole stack of positional issues but has the advantage of eliminating the distortion mechanisms inherent with the spider.
More news as it happens.
Burnt.
Attachments
Another ‘exciting’ assembly completed. It’s a very interesting structure magnetically in that there is no external field and everything is concentrated in the voice coil gap which is, as you might anticipate, very strong.
Attachments
