Paul,
Good progress!
The speakers play music nicely, so well done!
How do they compare to your other speakers in real life? Do you plan to take measurements?
Hello Zman
Each new prototype gets directly compared to the Dayton PS95-8 in a 3.3L vented cabinet. Both cabinets are 3D Printed, with same internal volume and port tuning, but the FD61 version has curved internal side walls and thick walls, where the PS95-8 version is thinner walls but has an internal brace built in. Always changing things to determine what is better!
Compared to the PS95-8 I would personally say on midrange they are about the same. FD61 has more low end due to its lower sensitivity and higher Q. Highs the PS95-8 wins easily as the inductance off my driver is too high for the cone profile I have in place.
Earlier models had sharp and nasty break up areas around 2.5Khz to 4Khz depending on the material the cone was made off. The newer builds which use newer cone profiles have lost this.
Adding in ridges disrupts the ability for the cone to have a nasty breakup due to different parts of the cone being able to flex at different rates. You can definitely understand Fostex and there shaped cones on more expensive models, and Scan Speak with there cut and glued cones. It is a super time consuming area though. Each prototype since the very start has had a different cone profile, so the next one will be cone profile 29. I have done over 100 actually on the PC, but I narrow these down from previous findings.
The other driver I compare to is the Mark Audio Alpair 7 but these I made 7 years ago, and are fitted into 20+L TL cabinets I designed back then. The Dayton PS95-8 high frequency response I believe is better than the Alpair but they are a older generation of Marks work and running them in this larger cabinets, it's no competition comparing them with the little 3.3L cabinets.
Maybe I should make a front fascia to match the alpair 7 and run them in my TL cabinets...have to look at that in hornresp!
Measurements wise I need to obtain a new flat response microphone as my old dayton seems to have issues. Money is very strapped right now, so I am making do with what I have in this area.
I analyse each driver with my WT3 analyser and use the mass added method for VAS currently. I do this by using a cup shaped to match the cone and then add weight into this. This way the weight is spread across the cone evenly and always in contact.
So in a quick summary, my opinion of FD61-P28
Lows: Good
Mids: Good
Highs: Bad
Clarity: Surprisingly impressive
Current Stage: Good as a midwoofer up to 8<10Khz. Not good as full range...yet
Paul
Hello Ghemml
Thanks for the comment. Yeah I have attempted a Faraday ring in the form of a copper washer, along with aluminium washer and a entire copper tube which all yielded varying results. At the moment I am looking at each option available to be to make a decision on the best step forward in this area. Have even debated the dual voice coil and opposite direction magnetic fields the likes of JBL and eighteen sound have done on subwoofers but concerned on potential patent infringement issues there.
Any other ways of lowering inductance, let me know. Possibly expand my winding to be taller and sacrifice some linear excursion for an inductance drop?
Paul
Thanks for the comment. Yeah I have attempted a Faraday ring in the form of a copper washer, along with aluminium washer and a entire copper tube which all yielded varying results. At the moment I am looking at each option available to be to make a decision on the best step forward in this area. Have even debated the dual voice coil and opposite direction magnetic fields the likes of JBL and eighteen sound have done on subwoofers but concerned on potential patent infringement issues there.
Any other ways of lowering inductance, let me know. Possibly expand my winding to be taller and sacrifice some linear excursion for an inductance drop?
Paul
Hello everyone. Small update
FD61 will be on show at the Advanced Engineering Show 2019!!!
3D Filaprint will be having them on display at there booth over both days, and I will be making sure a pair of them are there running in the fastidious cabinets. Im very excited knowing they will be seen by so many people, and really appreciate the opportunity to showcase the work so far.
All the parts I require for my next build are finally here so this week I will be working on getting those parts made and tested. Hopefully this will drastically reduce inductance so we can improve that high frequency response.
The new speaker cones will be using a more advanced carbon fiber filament which has a young's modulus in excess of 5,000. This will hopefully improve the high frequency response as well as the other improvements. Previous cones made in a weaker carbon fiber composite yielded a very nice tone.
As always, will keep you all posted. Next post will be with results of the new driver, including that crucial inductance test
FD61 will be on show at the Advanced Engineering Show 2019!!!
3D Filaprint will be having them on display at there booth over both days, and I will be making sure a pair of them are there running in the fastidious cabinets. Im very excited knowing they will be seen by so many people, and really appreciate the opportunity to showcase the work so far.
All the parts I require for my next build are finally here so this week I will be working on getting those parts made and tested. Hopefully this will drastically reduce inductance so we can improve that high frequency response.
The new speaker cones will be using a more advanced carbon fiber filament which has a young's modulus in excess of 5,000. This will hopefully improve the high frequency response as well as the other improvements. Previous cones made in a weaker carbon fiber composite yielded a very nice tone.
As always, will keep you all posted. Next post will be with results of the new driver, including that crucial inductance test
Hello everyone. Results are in on reducing inductance
The test is using an older drive train (Prototype 27) on the brand new copper tube assisted magnet motor. The magnets in use here are ring magnets instead of discs like before. The new motor is 100% compatible with the previous one. P27 was used instead of P28 as the pair of P28 are currently under a long testing phase and some abuse. The new driver looks like this (minus the magnets)
I haven't seen nice, clear and concise comparisons of the effect of a copper shorting ring, so I will be as clear as possible here if it helps anyone else.
The copper tube is a 0.8mm thick tube with a outer diameter of 22mm. The tube is 50mm long, and the test is the exact same driver with the tube fitted and removed. Here are the graphs!
Without:
and then with the copper tube fitted:
The reason for the inductance drop was to improve high frequency response. Adding the copper tube resulted in a drop from 23 Ohm @ 20Khz to 14 Ohm @ 20Khz. 60.8% of what it was. This should be enough to get the high frequency performance where I want with the new cone profile tweaks as well. That will be for the full P29 prototype which should be the next step.
For the engineering show a modified version of P28 will be sent. This will have a new surround and new cone profile, but will use the motor and chassis which has had a lot hours of use with no issues.
As always. Any questions, concerns, doubts, ideas or things you would like to see, let me know. Would people like more information on the design process?
Paul
The test is using an older drive train (Prototype 27) on the brand new copper tube assisted magnet motor. The magnets in use here are ring magnets instead of discs like before. The new motor is 100% compatible with the previous one. P27 was used instead of P28 as the pair of P28 are currently under a long testing phase and some abuse. The new driver looks like this (minus the magnets)
An externally hosted image should be here but it was not working when we last tested it.
I haven't seen nice, clear and concise comparisons of the effect of a copper shorting ring, so I will be as clear as possible here if it helps anyone else.
The copper tube is a 0.8mm thick tube with a outer diameter of 22mm. The tube is 50mm long, and the test is the exact same driver with the tube fitted and removed. Here are the graphs!
Without:
An externally hosted image should be here but it was not working when we last tested it.
and then with the copper tube fitted:
An externally hosted image should be here but it was not working when we last tested it.
The reason for the inductance drop was to improve high frequency response. Adding the copper tube resulted in a drop from 23 Ohm @ 20Khz to 14 Ohm @ 20Khz. 60.8% of what it was. This should be enough to get the high frequency performance where I want with the new cone profile tweaks as well. That will be for the full P29 prototype which should be the next step.
For the engineering show a modified version of P28 will be sent. This will have a new surround and new cone profile, but will use the motor and chassis which has had a lot hours of use with no issues.
As always. Any questions, concerns, doubts, ideas or things you would like to see, let me know. Would people like more information on the design process?
Paul
Hello zman01
Yeah the results are definitely moving in the right direction. Listening with and without the copper tube there is a distinct difference in the highs that you cannot ignore. I am hoping to get the new drive train components done this week. Here is a picture of the new cone profile printed in 3DXTech CarbonX. See what you think
Paul
Yeah the results are definitely moving in the right direction. Listening with and without the copper tube there is a distinct difference in the highs that you cannot ignore. I am hoping to get the new drive train components done this week. Here is a picture of the new cone profile printed in 3DXTech CarbonX. See what you think
Paul
Hello again. Today I am releasing frequency response graphs of FD61 (P27) with the new magnet motor against the other drivers again.
This time the results are from a emm-6 flat response omnidirectional microphone
The setup is crude, with the drivers simply sat on top of a box and measured face on, but it works for the midrange and high end frequencies I am looking at. Picture of the setup:
The comparison has always been against the PS95, So lets look at that one first:
As we can see, FD61 still cannot extend as far, but is a drastic improvement on the previous drivers. I think the PS95 spike in the high frequencies is due to measuring this close.
Next up is the Tectonic Elements driver and boy does is ring at around 12khz, but the fall off from around 15khz is similar to FD61.
Next one is the Bose driver and damn are these two measuring very close to one another. FD61 before was below the Bose driver in frequency response from 6Khz but now tracks it. This means I must of gained as much as 15dB @ 15Khz since P25!
The last comparison is the Samsung which extends better but is the only driver showing less SPL at the lower frequencies and has that dip at 400Hz.
So what to take from all this?
Firstly, frequency response is confirmed to of drastically improved. There is a dip at around 6.5Khz to be resolved, and this is what the ridges are for. I believe this is the frequency at which P27's cone flexes at the edges and creates a dip as it gets to this point. If I am correct, then the ridges on P29's cone along with it being a more rigid material should help reduce this dip. Time will tell.
I am still working on my surround and spider work for P29. When it is ready, I will off course let you all know. Hope the update was helpful or interesting at least.
Paul
This time the results are from a emm-6 flat response omnidirectional microphone
The setup is crude, with the drivers simply sat on top of a box and measured face on, but it works for the midrange and high end frequencies I am looking at. Picture of the setup:
An externally hosted image should be here but it was not working when we last tested it.
The comparison has always been against the PS95, So lets look at that one first:
An externally hosted image should be here but it was not working when we last tested it.
As we can see, FD61 still cannot extend as far, but is a drastic improvement on the previous drivers. I think the PS95 spike in the high frequencies is due to measuring this close.
An externally hosted image should be here but it was not working when we last tested it.
Next up is the Tectonic Elements driver and boy does is ring at around 12khz, but the fall off from around 15khz is similar to FD61.
An externally hosted image should be here but it was not working when we last tested it.
Next one is the Bose driver and damn are these two measuring very close to one another. FD61 before was below the Bose driver in frequency response from 6Khz but now tracks it. This means I must of gained as much as 15dB @ 15Khz since P25!
An externally hosted image should be here but it was not working when we last tested it.
The last comparison is the Samsung which extends better but is the only driver showing less SPL at the lower frequencies and has that dip at 400Hz.
So what to take from all this?
Firstly, frequency response is confirmed to of drastically improved. There is a dip at around 6.5Khz to be resolved, and this is what the ridges are for. I believe this is the frequency at which P27's cone flexes at the edges and creates a dip as it gets to this point. If I am correct, then the ridges on P29's cone along with it being a more rigid material should help reduce this dip. Time will tell.
I am still working on my surround and spider work for P29. When it is ready, I will off course let you all know. Hope the update was helpful or interesting at least.
Paul
Nice project! Very cool to see a deliberate study with multiple cases of development one after another. The PS95-8 is a nice driver to compare against - I am in fact listening to it now and use it high passed in a FAST with a woofer below.
Have you heard of MarkForged 3d printers that can laydown continuous carbon filament? They sent me some samples and they are stronger than aluminum - I could not bend the part with my bare hands. They are able to make aluminum equivalent fixtures and tooling.
Might be perfect for the basket and the cone.
Metal and Carbon Fiber 3D printers for Manufacturing | Markforged
Have you heard of MarkForged 3d printers that can laydown continuous carbon filament? They sent me some samples and they are stronger than aluminum - I could not bend the part with my bare hands. They are able to make aluminum equivalent fixtures and tooling.
Might be perfect for the basket and the cone.
Metal and Carbon Fiber 3D printers for Manufacturing | Markforged
Paul,
Very good stuff and the developments look promising !
Btw, the new cone printed with 3DXTech CarbonX looks nicely done and with the ridges and more rigid material, should have much less flex. It will be interesting to see how it performs in real life; from what I have read, cone material ideally should be a good balance between rigid and damped, and yes, keeping it relatively light at the same time.
Very good stuff and the developments look promising !
Btw, the new cone printed with 3DXTech CarbonX looks nicely done and with the ridges and more rigid material, should have much less flex. It will be interesting to see how it performs in real life; from what I have read, cone material ideally should be a good balance between rigid and damped, and yes, keeping it relatively light at the same time.
@ Xrk971
Thanks for the kind words. It was the kind of research I was looking for prior to starting and during this project but all you can find is the principles. Not the reality and direct comparisons. Along with it helping me to make a driver that I compete, it hopefully helps others in learning and hopefully clarify things as well.
PS95-8 is a very nice driver I agree. A little weak on the bass side but its a small driver so wouldn't expect anything more. I have some alpair 6M's as well I could compare with but the 6M is a larger cone. What kind of frequency to you crossover at if you don't mind me asking?
Yeah I have seen there printers along with some other areas I am interested in but money is always tight so I make do with what I can get. Very british approach I guess!
@ zman
Thanks zman. Due to the magnet motor flux density being weaker than a standard driver, I have to compensate with a longer winding which has put even more focus on keeping weight down. This latest cone on it's own weighs just 1.68g. A massive improvement over the 9g the original P1 prototype had. I have done ridges once or twice before and results of a blue snowball mic showed that it helped but those ridges were straight. This one is curved enough that very little of the cone has a simple arc to follow when looked at from a side view. This also effects the distance the sound has to travel to reach the cone edge which in turn will hopefully control cone breakup in a cleaner fashion.
As always though, the ears and the data will lead me on my path.
Paul
Thanks for the kind words. It was the kind of research I was looking for prior to starting and during this project but all you can find is the principles. Not the reality and direct comparisons. Along with it helping me to make a driver that I compete, it hopefully helps others in learning and hopefully clarify things as well.
PS95-8 is a very nice driver I agree. A little weak on the bass side but its a small driver so wouldn't expect anything more. I have some alpair 6M's as well I could compare with but the 6M is a larger cone. What kind of frequency to you crossover at if you don't mind me asking?
Yeah I have seen there printers along with some other areas I am interested in but money is always tight so I make do with what I can get. Very british approach I guess!
@ zman
Thanks zman. Due to the magnet motor flux density being weaker than a standard driver, I have to compensate with a longer winding which has put even more focus on keeping weight down. This latest cone on it's own weighs just 1.68g. A massive improvement over the 9g the original P1 prototype had. I have done ridges once or twice before and results of a blue snowball mic showed that it helped but those ridges were straight. This one is curved enough that very little of the cone has a simple arc to follow when looked at from a side view. This also effects the distance the sound has to travel to reach the cone edge which in turn will hopefully control cone breakup in a cleaner fashion.
As always though, the ears and the data will lead me on my path.
Paul
3.3L Vented enclosure Test!!!
3.3L Vented 3-Way Shootout
PS95 vs P27 vs P29
Both P27 and P29 using the new copper magnet motor
Hello everyone. Big one today. Testing 2 cone profiles against the Dayton PS95, but in 3.3L vented cabinets. Same size drivers, same size cabinets (Different designs I know but helpful still), same SPL, same distance, same room and equipment. A straight apples to apples comparison!
So for anyone reading along and waiting for some direct comparisons, here you go:
Blue = Dayton PS95-8
Green = Polymate3D P27
Red = Polymate3D P29
As you can see from this, things are looking good. I didn't like how hard it is to distinguish this graph those so I narrowed the limits for a closer look:
So what have a gathered from this test and analysis today?
1) P27 is really doing well up to around 15-16Khz which is fine by me
2) P29 is weaker in the upper range, and sounds more laid back thanks to this
3) The Dayton is showing the bump up above a flat response like before and like it is in the published datasheet.
But wait, there is more today than just this!
P29 was also testing a different flexible filament to see if I could increase compliance and this was a success as well. Testing with low frequencies showed it was moving more freely than P27 is, and it's compliance come in around 0.7mm/N vs the other materials 0.5mm/N. This is vital to help lower the Q value of my driver.
So we have a good frequency response. We have 80dB@1W/1M in reach as well. It fits nicely into printed and sealed cabinets, and it is competing well too in day to day use. Whats next?
The developments from here are tweaking the design to lower mass further, and increase compliance a bit more if I can. Getting the Qts value below 1 is the final goal.
How does P27 sound compared to the Dayton PS95 in the same size cabinet?
P27 sounds to be a little more laid back than the PS95, but not by much. If you have listened to a Mark Audio Alpair 6 or 7, and the PS95 you'll know what I mean. I listen to P27 for hours at a time with no fatigue from the driver or my ears.
P27 due to it's lower 80dB @ 1W vs PS95 85dB @ 1W, and the high Q value, the bass response is significantly higher and in doing so, feels like a bigger driver than the PS95. Listening to something like the MGS2 intro on P27 sounds more full bodied than the PS95 does.
What kind of power handling and XmaX are we looking at with P27/P29?
My design uses a underhung voice coil design and a radically different magnet motor. The underhung design was chosen so I did not need to wind the voice coil cleanly. This is important to me so other people can make the voice coil themselves without a need for panic or stress. The space is sufficient for a messy voice coil and the glue used works perfectly at filling any voids. Inductance tests of multiple voice coils has come out very close. Well within industry tolerances
To the actual question, current designs have a Xmax of 1.25mm, or 2.5mm Peak to Peak. This is not the full story however for a few reasons.
1) The steel tube means that beyond this the magnetic flux density does not fall off a cliff, and is fairly linear, so distortion levels will not jump suddenly.
2) The magnet motor is modular and you can adjust the magnets and other parts TO YOUR TASTE!
If you want you can drop that Xmax down 0.5mm, but gain another 6% power. Possibly your using it in a FAST system and want to extract what you can.
If you want, you can increase Xmax up to 2.2mm, and lose 11% power but you really want to extract as much low end or SPL as you can from the driver
What other driver can do that!? (seriously if you know one, please let me know!)
Actual total power before heat issues varies on the cone material. At the moment my testing has been up to 10W without a issue. A typical cabinet 3 to 6L vented or sealed will result in 3-5W before Xmax is reach on the default design I am using. I want to have some headroom for FAST based setups however
How low is this driver tuning down to?
The design is constantly altering, but the current design in a 3.3L enclosure is producing an F3 of around 48-52Hz, and a F10 of around 42-46Hz. Staying around 50Hz is the target. The Dayton PS95 in comparison has a F3 of around 65-68Hz.
How far are you from getting to a complete product?
Answer is very close! It will be launching as a kickstarter campaign to gauge interest and will go from there. I won't launch it however until I have added my mass reduction tweaks and improved compliance to lower that Qts value.
Oh, and it will launch with 2 drivetrain choices, which will expand further.
What kind of pricing is expected from this driver?
Current target is the same as when I started which is £35. This is a tad on the higher side, but with its interchangeable drivetrain and phase plug, along with a modular magnet motor, I feel it is the right place.
Where did the idea of making your own driver come from?
The idea came from the fustration of not being able to tweak this part of the system, and building countless numbers of speaker cabinets to learn from. I want FD61 to not just be a competitive driver, but be the learners and tweakers driver of choice! Also who else allows you to make your own driver!? Your making your own crossovers and cabinets. Wouldn't it be refreshing to get this hands on!?
and....thats the update for now. Tweaks have already been made to my design whilst these parts have been made and assembled. I very much appreciate all the support, questions, queries, doubts and considerations you all bring to the project, and hope the update has been helpful.
Paul
3.3L Vented 3-Way Shootout
PS95 vs P27 vs P29
Both P27 and P29 using the new copper magnet motor
An externally hosted image should be here but it was not working when we last tested it.
Hello everyone. Big one today. Testing 2 cone profiles against the Dayton PS95, but in 3.3L vented cabinets. Same size drivers, same size cabinets (Different designs I know but helpful still), same SPL, same distance, same room and equipment. A straight apples to apples comparison!
So for anyone reading along and waiting for some direct comparisons, here you go:
An externally hosted image should be here but it was not working when we last tested it.
Blue = Dayton PS95-8
Green = Polymate3D P27
Red = Polymate3D P29
As you can see from this, things are looking good. I didn't like how hard it is to distinguish this graph those so I narrowed the limits for a closer look:
An externally hosted image should be here but it was not working when we last tested it.
So what have a gathered from this test and analysis today?
1) P27 is really doing well up to around 15-16Khz which is fine by me
2) P29 is weaker in the upper range, and sounds more laid back thanks to this
3) The Dayton is showing the bump up above a flat response like before and like it is in the published datasheet.
But wait, there is more today than just this!
P29 was also testing a different flexible filament to see if I could increase compliance and this was a success as well. Testing with low frequencies showed it was moving more freely than P27 is, and it's compliance come in around 0.7mm/N vs the other materials 0.5mm/N. This is vital to help lower the Q value of my driver.
So we have a good frequency response. We have 80dB@1W/1M in reach as well. It fits nicely into printed and sealed cabinets, and it is competing well too in day to day use. Whats next?
The developments from here are tweaking the design to lower mass further, and increase compliance a bit more if I can. Getting the Qts value below 1 is the final goal.
How does P27 sound compared to the Dayton PS95 in the same size cabinet?
P27 sounds to be a little more laid back than the PS95, but not by much. If you have listened to a Mark Audio Alpair 6 or 7, and the PS95 you'll know what I mean. I listen to P27 for hours at a time with no fatigue from the driver or my ears.
P27 due to it's lower 80dB @ 1W vs PS95 85dB @ 1W, and the high Q value, the bass response is significantly higher and in doing so, feels like a bigger driver than the PS95. Listening to something like the MGS2 intro on P27 sounds more full bodied than the PS95 does.
What kind of power handling and XmaX are we looking at with P27/P29?
My design uses a underhung voice coil design and a radically different magnet motor. The underhung design was chosen so I did not need to wind the voice coil cleanly. This is important to me so other people can make the voice coil themselves without a need for panic or stress. The space is sufficient for a messy voice coil and the glue used works perfectly at filling any voids. Inductance tests of multiple voice coils has come out very close. Well within industry tolerances
To the actual question, current designs have a Xmax of 1.25mm, or 2.5mm Peak to Peak. This is not the full story however for a few reasons.
1) The steel tube means that beyond this the magnetic flux density does not fall off a cliff, and is fairly linear, so distortion levels will not jump suddenly.
2) The magnet motor is modular and you can adjust the magnets and other parts TO YOUR TASTE!
If you want you can drop that Xmax down 0.5mm, but gain another 6% power. Possibly your using it in a FAST system and want to extract what you can.
If you want, you can increase Xmax up to 2.2mm, and lose 11% power but you really want to extract as much low end or SPL as you can from the driver
What other driver can do that!? (seriously if you know one, please let me know!)
Actual total power before heat issues varies on the cone material. At the moment my testing has been up to 10W without a issue. A typical cabinet 3 to 6L vented or sealed will result in 3-5W before Xmax is reach on the default design I am using. I want to have some headroom for FAST based setups however
How low is this driver tuning down to?
The design is constantly altering, but the current design in a 3.3L enclosure is producing an F3 of around 48-52Hz, and a F10 of around 42-46Hz. Staying around 50Hz is the target. The Dayton PS95 in comparison has a F3 of around 65-68Hz.
How far are you from getting to a complete product?
Answer is very close! It will be launching as a kickstarter campaign to gauge interest and will go from there. I won't launch it however until I have added my mass reduction tweaks and improved compliance to lower that Qts value.
Oh, and it will launch with 2 drivetrain choices, which will expand further.
What kind of pricing is expected from this driver?
Current target is the same as when I started which is £35. This is a tad on the higher side, but with its interchangeable drivetrain and phase plug, along with a modular magnet motor, I feel it is the right place.
Where did the idea of making your own driver come from?
The idea came from the fustration of not being able to tweak this part of the system, and building countless numbers of speaker cabinets to learn from. I want FD61 to not just be a competitive driver, but be the learners and tweakers driver of choice! Also who else allows you to make your own driver!? Your making your own crossovers and cabinets. Wouldn't it be refreshing to get this hands on!?
and....thats the update for now. Tweaks have already been made to my design whilst these parts have been made and assembled. I very much appreciate all the support, questions, queries, doubts and considerations you all bring to the project, and hope the update has been helpful.
Paul
Phase Plug Test
Today I made 3 different different phase plugs and fitted them to my current design to look at how much it effects the frequency response head on. The difference should be more effected at an angle, but still I wanted to see the results.
To make it simpler for anyone wondering, the differences was very small. Still if you want to see the result, feel free to see this below. I am looking for input on what people prefer style wise whilst I am doing more tests in this area.
So first up is the indented dome:
The blue line is the indented design, whilst the red is with no phase plug. Only real notable mention is the slight extension of 2-3dB at around 15-16khz.
Next one is the big dome:
The dip at around 12Khz is improved a bit and the same improvement at 15-16khz i also there
Finally the Radial design:
The least impact of the designs. The improvement at 15-16khz continues so that means this frequency must of been cancelling itself out.
I will look at doing more tests soon at an angle to see what difference there is there too. I am working on a revised motor design. Not effecting the work I have so far. More moving bits to help reduce moving mass.
Paul
Today I made 3 different different phase plugs and fitted them to my current design to look at how much it effects the frequency response head on. The difference should be more effected at an angle, but still I wanted to see the results.
To make it simpler for anyone wondering, the differences was very small. Still if you want to see the result, feel free to see this below. I am looking for input on what people prefer style wise whilst I am doing more tests in this area.
So first up is the indented dome:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
The blue line is the indented design, whilst the red is with no phase plug. Only real notable mention is the slight extension of 2-3dB at around 15-16khz.
Next one is the big dome:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
The dip at around 12Khz is improved a bit and the same improvement at 15-16khz i also there
Finally the Radial design:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
The least impact of the designs. The improvement at 15-16khz continues so that means this frequency must of been cancelling itself out.
I will look at doing more tests soon at an angle to see what difference there is there too. I am working on a revised motor design. Not effecting the work I have so far. More moving bits to help reduce moving mass.
Paul
P30 - First version might finally be here!
Hello Everyone
For P30 I decided to take my very aggressive shallow cone profile and reduce this in a bit in the hope to extend the high frequency performance just a little more.
Colours were just to stop me from running out of specific colours. Would you all prefer a black basket, or something different like silver?
The white line is FD61
The Blue line is PS95
As you can see, performance all the way to 9Khz is very close. Then the Dayton goes up above the flat response where as FD61 does so with less exaggeration, rolling off from about 14Khz but where P29 @ 20hz was 17dB off the PS95 where as this new design is only 8dB off and PS95 at 20khz is a little above flat SPL,
Using the PS95 drive for comparison, I am very happy that the frequency response is there. SPL is on the lower end of what I would like, but stronger magnets will gain some of this. I will also be testing with a lighter material and check the frequency performance of it as well.
How does it sound in person?
P30 is sounding very good and not harsh on the ear over longer use. Output is clear and crisp, as have the last few iterations and it doesn't feel lacking across a range of genres.
I have also created my first attempt at a datasheet for this driver. People following along know this will vary on drivetrain too, so I will create one of these for each. Feedback is greatly appreciated!
So this development thread is now coming to an end and kickstarter side will be worked on. If your reading this, interested but would like to see something added or altered, now is the time to mention it for consideration.
Paul
Hello Everyone
For P30 I decided to take my very aggressive shallow cone profile and reduce this in a bit in the hope to extend the high frequency performance just a little more.
An externally hosted image should be here but it was not working when we last tested it.
Colours were just to stop me from running out of specific colours. Would you all prefer a black basket, or something different like silver?
An externally hosted image should be here but it was not working when we last tested it.
The white line is FD61
The Blue line is PS95
As you can see, performance all the way to 9Khz is very close. Then the Dayton goes up above the flat response where as FD61 does so with less exaggeration, rolling off from about 14Khz but where P29 @ 20hz was 17dB off the PS95 where as this new design is only 8dB off and PS95 at 20khz is a little above flat SPL,
Using the PS95 drive for comparison, I am very happy that the frequency response is there. SPL is on the lower end of what I would like, but stronger magnets will gain some of this. I will also be testing with a lighter material and check the frequency performance of it as well.
How does it sound in person?
P30 is sounding very good and not harsh on the ear over longer use. Output is clear and crisp, as have the last few iterations and it doesn't feel lacking across a range of genres.
I have also created my first attempt at a datasheet for this driver. People following along know this will vary on drivetrain too, so I will create one of these for each. Feedback is greatly appreciated!
An externally hosted image should be here but it was not working when we last tested it.
So this development thread is now coming to an end and kickstarter side will be worked on. If your reading this, interested but would like to see something added or altered, now is the time to mention it for consideration.
Paul
Hello CraigSu
Thank you for the message and yes I actually did miss this whilst being so eager to get the next build done and testing each stage. If the driver can make a impact once it launches I would love to get into building and designing an amplifier for it and other drivers out there, and had that question in mind when it comes to designing an amplifier. One stage at a time.
Im not really sure how many other people out there are attempting what I am doing. Most DIY drivers are larger for larger applications and anything with 3D printed components I have found doesn't have any data to go with it or sound particularly loud / clear. I would be very interested to see others as I really believe there is a lot of potential by going this route.
Paul
Gale vs Dayton Audio vs Polymate 3D
Hello everyone again
Today is not data and graphs and what is next. Today is my best attempt to allow you all to see what this 3D printed speaker is capable off against competition.
In this video, FD61 P30 is running up against a pair of Gale 3010S bookshelf speakers and the Dayton Audio PS95. Each driver runs for approximately 1.5 minutes. I have my own opinions on the comparison, but I want to see what you all think.
The recording was done off a DH-500 amplifier and Blue Snowball, where the other recordings and video was done off my Pixel smartphone.
YouTube
The final part for today was to get some recording done of vocals as I know this has been lacking until now. Hope you enjoy.
YouTube
I will keep my opinion back to not effect any results. I value all of your opinions.
Paul
Hello everyone again
Today is not data and graphs and what is next. Today is my best attempt to allow you all to see what this 3D printed speaker is capable off against competition.
In this video, FD61 P30 is running up against a pair of Gale 3010S bookshelf speakers and the Dayton Audio PS95. Each driver runs for approximately 1.5 minutes. I have my own opinions on the comparison, but I want to see what you all think.
The recording was done off a DH-500 amplifier and Blue Snowball, where the other recordings and video was done off my Pixel smartphone.
YouTube
The final part for today was to get some recording done of vocals as I know this has been lacking until now. Hope you enjoy.
YouTube
I will keep my opinion back to not effect any results. I value all of your opinions.
Paul
Hello zman
Thanks for listening. I know it's not the best but what did you think?
In relation to the question, no it has not been covered. Mainly due to how much the T&S parameters have been changing. I have some results for you to make an opinion from. I would love to have a lower Q than the current designs, but if I can get close to 1.0 I will be happy.
https://scontent-lhr3-1.xx.fbcdn.net/v/t1.0-9/71199899_498176897404496_4500382082685468672_n.jpg?_nc_cat=107&_nc_eui2=AeFdBegTNjbcjKenxx_uOBqcTtYyO6p21l8DfbHFXhpI4BxCETif8_n9__8Wzi0tpv97zR56-Z6LBbbwI_wQzCE_qFgVr78HWlYdvGHZscIKHA&_nc_oc=AQm3vsCLdUI89yLHubqyR9iFQ1MELM6hZO8xfywZnbjlRfUME4E0aPoWYz0ewj4M50I&_nc_ht=scontent-lhr3-1.xx&oh=846c8dba2a3b349903bd2d226b5347a8&oe=5E2CF19E
This is how the current design works in the cabinet I have made. I picked 3.3L for my enclosure size as it allows it to be 3D printed nicely and give me some space to style my design.
The power output off the motor should increase when I move from my current N35 magnets to N50. My FEMM simuation predicted 2.17Tm for N35 and I measured 2.16Tm, so I am confident with this side of things now. The final power should be more like 2.50Tm, which looks like this:
https://scontent-lhr3-1.xx.fbcdn.net/v/t1.0-9/71384939_498176847404501_8864096347507130368_n.jpg?_nc_cat=109&_nc_eui2=AeGKYGwIGALPbETCmxfdSJqHja8Me-wpbhCa0si1oAOFwrCSbLWZEFHFYZDkdBVbUrn0ZKU9pjNuHhu17sGkMVyQQwbSbjO2XdI-eMxxS3jLeg&_nc_oc=AQn2oe_dtDIKyccOleqY4pUvpuTSzQGXSncne0poTP36fne2OHuYetp0MMEs9mm0mM8&_nc_ht=scontent-lhr3-1.xx&oh=e0b48a8f0768194d5e0b9f6d93d6ca21&oe=5E3221F9
If I was building with wood I would usually build such drivers into a cabinet 6 to 9L in size, and at 9L the result is this:
https://scontent-lhr3-1.xx.fbcdn.net/v/t1.0-9/71388290_498176844071168_1398150582291660800_n.jpg?_nc_cat=107&_nc_eui2=AeGPrC7y2-elI6wLN4EGYw4teAZK0azitUZwy10QSuqjB9Pmp7JivzV_mdi-mYGsX_PRyky42395xbA3D7qfybsRB89yMJfa8DXin76WDrdBQw&_nc_oc=AQkbaXnRMk-qt6TWa7joUR54UrhrQpke-szY3VrdK28fauI_IyOcFLPkcYv8P8BdgDQ&_nc_ht=scontent-lhr3-1.xx&oh=fa9ec979f7810608de7a9f20a0635a54&oe=5E2FE7A9
And then finally how it performs in a little 0.8L sealed cabinet with heavy infill. This is more to allow people not in the diy audio field to have a quick and easy entry point where even a 120x120x120mm 3D printer will allow them to make the cabinet and buy the driver as a kit and get there toes wet. This simulates like this:
https://scontent-lhr3-1.xx.fbcdn.net/v/t1.0-9/71762114_498176840737835_2438099268599283712_n.jpg?_nc_cat=100&_nc_eui2=AeHxAQOEwzPALG6MYt529k9Mjqqt8-6KVWUjAzTXllEVViwVcvuRg9o01BZdKMvgEQXjaV1KyaySk8MfMhXbbI8GTPUXdL1Fc-6wTY-MUrF2tA&_nc_oc=AQkajlt3TjcEOxO6DOhMP9JIIAWPKlyT0YQo4DNxE3FYlilC7-iNByVJhZEvJFiyto4&_nc_ht=scontent-lhr3-1.xx&oh=a0c77116a9be0c65aaa9c0f9ff37356a&oe=5E35B683
So I have made FD61 to potentially work from 1L up to around 12L, enabling it to be possible in:
Bluetooth Speakers
Smart Sound Speakers
Soundbars
Small Bookshelf Speakers
The market has a lot of focus on bigger driver units, but here in the UK room sizes are on the smaller side, and these product ranges are where the most people have an interest in buying something. I want FD61 to cater to many types of people, from new to people like myself who would love to be able to work directly with the driver, and not have it be a fixed component within speaker projects.
Cost wise I am still targetting £35 for a single driver kit, and want to launch with 2 different drivetrain choices to allow you to pick which one you prefer both look and sound wise. I would make another video with both running in the same cabinet for comparison
Any other questions, please don't hesitate to ask away. I by no means know everything....far far from it. I want to do everything I can to make FD61 a success.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.