Hello all,
I'm backing with some updates and asking for additional help.
Now I have all a need to perform some measurement, including "time", so the idea is to measure properly the frequency response and the impedance till Sunday 3th.Sep.
Unfortunately the cabs, after built, were storage for a long time and I didn't had time to used them till now so, the measurement below was done with the cab "cold" and after "long storage time" without plug them to the amplifier. Probably isn't the right way for measurements.
Hornresp simulation
Direct link https://s26.postimg.org/cvalkt46h/Hornresp_Z.png
REW comparison (Blue = Hornresp / Green = measured)
Note: Once the REW measurement is offset I subtracted the ~R-sense to align the graphics
Direct link: https://s26.postimg.org/6nz7kn8fd/Z_comparison.png
Devices
Direct link: https://s26.postimg.org/mdexylf8p/Devices.jpg
What is the proper way to measure impedance? warm? cold? after 30min playing 50% of the max load?
What is the proper way to measure Frequency Response? warm? cold? after 30min playing 50% of the max load?
For Impedance measurement:
For Frequency Response measurement:
I need to measure again in a proper way to understand if the first peak will be proper align with the simulation or not before conclusions.
I'm backing with some updates and asking for additional help.
Now I have all a need to perform some measurement, including "time", so the idea is to measure properly the frequency response and the impedance till Sunday 3th.Sep.
Unfortunately the cabs, after built, were storage for a long time and I didn't had time to used them till now so, the measurement below was done with the cab "cold" and after "long storage time" without plug them to the amplifier. Probably isn't the right way for measurements.
Hornresp simulation
An externally hosted image should be here but it was not working when we last tested it.
Direct link https://s26.postimg.org/cvalkt46h/Hornresp_Z.png
REW comparison (Blue = Hornresp / Green = measured)
Note: Once the REW measurement is offset I subtracted the ~R-sense to align the graphics
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/6nz7kn8fd/Z_comparison.png
Devices
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/mdexylf8p/Devices.jpg
What is the proper way to measure impedance? warm? cold? after 30min playing 50% of the max load?
What is the proper way to measure Frequency Response? warm? cold? after 30min playing 50% of the max load?
For Impedance measurement:
- R-sense device = 100ohm
- Software: Room Eq Wizard 5.18
- Audio device: Behringer UCA 222
For Frequency Response measurement:
- Dbx RTA-M Mic (there is no calibration file so i found one at Internet that will present some error)
- Audio device: Behringer Mixer XENYX Q502USB
- Software: Room Eq Wizard 5.18 and Smaart 7
I need to measure again in a proper way to understand if the first peak will be proper align with the simulation or not before conclusions.
That would depend on what you're using the impedance measurement for. If you're just checking to see how close your build is to the sim, and if you're checking for leaks and if panel flex is affecting the build, then measuring while its cold is fine.
From your measurement and its comparison to the Hornresp sim, I suspect you may have a leak somewhere, likely around the driver. The impedance peaks are never EXACTLY as simmed, but they should be pretty close. The only time I've seen the first impedance peak shifted that much is when the TH had a leak around the driver because of a problem with the cutout - see this thread for details: - When things don't go quite as planned...
The presence of a leak can be confirmed by running a very low frequency tone through the TH and listening for whistling or gushing noises around the edge of the driver, particularly close to where it's screwed to the baffle. BTW, did you use the manufacturer's specifications for the diameter of the cutout for the driver? If so, that could be part of the problem, as the spec is usually for those cases when you're mounting the driver through the baffle, and not on top of it, and using the manufacturer's specification could lead to leaks around where the driver is screwed to the baffle.
The second thing I noticed is that the impedance curve has some minor ripples in it, between 60-70 Hz and just above the impedance peak at the right. The middle impedance peak is also somewhat malformed. This could be due to panel flex. A quick way to find out is to put a sandbag on top of the cabinet and redo the impedance measurement (or, sit on top of it and do the same), to see if the ripples disappear and the middle impedance peak is better formed. Unfortunately panel flex is not that easy to cure once the cabinet is built (that's why I prefer to over-brace my builds), but if the cabinet performs to your satisfaction, that may be a non-issue.
Thanks for your feedback Brain,
For leakage it could be around the screws places or through driver rubber, is there any glue or something special you do in those interfaces?
The hole spec is lower then manufactures specification just to guarantee proper rubber seat, but the driver rubber maybe isn't that perfect to avoid leakage so maybe additional work need to be done.
Probably the frequency response confirm if there is leakage or not as you suggest on the linked post.
For leakage it could be around the screws places or through driver rubber, is there any glue or something special you do in those interfaces?
The hole spec is lower then manufactures specification just to guarantee proper rubber seat, but the driver rubber maybe isn't that perfect to avoid leakage so maybe additional work need to be done.
Probably the frequency response confirm if there is leakage or not as you suggest on the linked post.
1/8" weatherstripping tape is usually what I use to ensure a good seal. If the driver has a good front sealing gasket though, this shouldn't be required.
Yup, I suggest doing the FR test, and also the "leak" test that I suggested. The latter will help to pinpoint exactly where the leak is occurring, if in fact there's a leak. Run a low tone at a decent volume through the TH (start below 40 Hz) and decrease the frequency slowly. If you start to hear chuffing or whistling noises (that are not coming from the driver's pole vent), then you've got a leak, and the noises can help to locate where the leak is.
ok I will try to do it.
In this meantime as you can see on the image below (highlighted in red), the driver has it rubber flange cut around the screw places, so that places for sure is the leakage source through the screw because there is also lash between screw and the hole on polywood. Probably I will also need to inject some silicon there.
Direct link: https://s26.postimg.org/ae8hoefzd/HPX2150_front2.jpg
In this meantime as you can see on the image below (highlighted in red), the driver has it rubber flange cut around the screw places, so that places for sure is the leakage source through the screw because there is also lash between screw and the hole on polywood. Probably I will also need to inject some silicon there.
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/ae8hoefzd/HPX2150_front2.jpg
In addition I'm posting an schema comparing the fastening type, T-nuts vs Wood Screw.
I saw so many people using T-nuts o their projects that i didn't pay much attention regarding the leakage and probably is the main source.
If someone can share solution on this interface I would be very grateful.
Direct link: https://s26.postimg.org/wp5jfus15/Fastening_Comparison.png
I saw so many people using T-nuts o their projects that i didn't pay much attention regarding the leakage and probably is the main source.
If someone can share solution on this interface I would be very grateful.
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/wp5jfus15/Fastening_Comparison.png
Once the cutout in the baffle sits inside those cutouts for the screws in the driver's front seal, it should not be leaking through that point if the driver is properly screwed down. An easy way to find out is to simply unscrew the driver and examine the front seal. If it's evenly compressed all around, then the cutout in the baffle is good enough to ensure a seal. However, if its not, then cutout might be too big.
Attachments
If the T-nut is firmly fastened into the wood, then there shouldn't be a path for air to get into that gap. If you're not sure, you could seal up the edges of the T-nut on the inside of the box with some silicone sealant.
I suggest performing the test I mentioned to identify if there are actually any leaks before doing any modifications to your build.
Hi Brain,
Until now you're hitting everything. I double checked the baffle cutout and I used manufactures' specification, so probably there are some small gaps there that I need to fix with gasket as you did. What thickness did you used?
Once I build 4 cabs, 2 of them don't have driver for a while.
Direct link: https://s26.postimg.org/6hwposf49/Baffle_cutoff-cab.jpg
Direct link: https://s26.postimg.org/hu4pstyfd/Baffle_cutoff.png
Today I performed some SPL measurements, The REW worked but for some reason the Smaart 7 didn't.
I didn't understand why the amplifier was enabling the Limiter while REW was always under 75dBA, I'm trying to find the answer but probably it was because I didn't calibrate the REW SPL Meter. Considering this, the output voltage controlled as 2,9V (Sine Wave @ 60Hz) was changed once I tried to increased the gain to reach simulation value without success.
Finally I hope the backyard I have available is at least OK for the SPL measurement.
Drive condition = Cold
Results: The Frequency looks really good but probably some differences from the simulation are linked to air leakage as you demonstrate in your post.
Mic placement
Direct link: https://s26.postimg.org/3so4tj5hl/image.jpg
Setup - Perspective View
Direct link: https://s26.postimg.org/uk8y5x6dl/image.jpg
Setup - Front View
Direct Link: https://s26.postimg.org/m739b9761/image.jpg
Setup - Back View
Direct link: https://s26.postimg.org/id9t23pu1/image.jpg
Setup - Output Voltage (it can be disregarded once it was modified)
Direct link: https://s26.postimg.org/4v2wptdop/image.jpg
Setup - Audio Interface
Direct link: https://s26.postimg.org/5zwyv705l/image.jpg
REW - RTA Measurement (no offset)
Direct link: https://s26.postimg.org/fa92yqavd/SPL_RTA.png
REW - RTA Measurement (with offset) vs Hornresp Simulation
Direct link: https://s26.postimg.org/cevzrv6vd/SPL_Hornresp_vs_RTA.png
Until now you're hitting everything. I double checked the baffle cutout and I used manufactures' specification, so probably there are some small gaps there that I need to fix with gasket as you did. What thickness did you used?
Once I build 4 cabs, 2 of them don't have driver for a while.
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/6hwposf49/Baffle_cutoff-cab.jpg
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/hu4pstyfd/Baffle_cutoff.png
Today I performed some SPL measurements, The REW worked but for some reason the Smaart 7 didn't.
I didn't understand why the amplifier was enabling the Limiter while REW was always under 75dBA, I'm trying to find the answer but probably it was because I didn't calibrate the REW SPL Meter. Considering this, the output voltage controlled as 2,9V (Sine Wave @ 60Hz) was changed once I tried to increased the gain to reach simulation value without success.
Finally I hope the backyard I have available is at least OK for the SPL measurement.
Drive condition = Cold
Results: The Frequency looks really good but probably some differences from the simulation are linked to air leakage as you demonstrate in your post.
Mic placement
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/3so4tj5hl/image.jpg
Setup - Perspective View
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/uk8y5x6dl/image.jpg
Setup - Front View
An externally hosted image should be here but it was not working when we last tested it.
Direct Link: https://s26.postimg.org/m739b9761/image.jpg
Setup - Back View
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/id9t23pu1/image.jpg
Setup - Output Voltage (it can be disregarded once it was modified)
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/4v2wptdop/image.jpg
Setup - Audio Interface
Channel 1 (L) = Mic
Channel 2 (R) = Sound Source + Reference
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/5zwyv705l/image.jpg
REW - RTA Measurement (no offset)
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/fa92yqavd/SPL_RTA.png
REW - RTA Measurement (with offset) vs Hornresp Simulation
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/cevzrv6vd/SPL_Hornresp_vs_RTA.png
I used a gasket made of matt board with an internal diameter that was a few mm less than the manufacturer's suggested driver cutout. See this post for more details: - http://www.diyaudio.com/forums/subwoofers/275969-dont-planned-2.html#post4371645
It worked, and the intent was to replace it with a gasket made of 1/8" ply, but work on that particular TH stopped at that point.
Measurements will be impacted by anything that's less than one wavelength of the lowest frequency of importance away from the driver under test. So, if you're looking for accurate measurements down to 20 Hz, there should be nothing within about 50 feet of the speaker and the mike. Of course the *size* of any nearby items that may impact the measurement needs to be taken into consideration too, and the lower the frequency, the larger such items have to be to make any appreciable impact. e,g, a mike stand or even a chair that's within 50 feet of the mike and speaker should not make any difference to the measurement at 20 Hz).
Amazing to see there are no big dips to see in RTA. Did you add cone compensation? I notices the mic is not on the ground. Can you do measurement with mic on ground?
Thanks, I'm learning a lot on this forum and trying to leave a trace for the others.
Yes, I added cone compensation, it's published somewhere in the thread, not too far.
Sure, I'm working on the leakage issue suggested from Brain, so as long things are ok i will try to perform new measurement, next time i will also check SPL Meter, but, it may take time for new measurement.
You can put the db(c) meter at same distance as microphone (alongside the mic) and then put in the measured db into rew SPL Reading calibration so you can accurately get the right db levels in the graph. You can also do it through the SPL meter popup window.
I'm backing with more updates.
I've been working theses last two days on the leakage stuffs and here is some feedbacks.
The driver was so fasten that we can see some print mark from driver's basket on the cab.
Direct link: https://s26.postimg.org/iaw5n7gqx/Fasten_Marks.jpg
I bought some pieces of Ethylene-vinyl acetate (EVA) to prepare the gasket and other small peaces, this is exactly the same material of drivers original gasket.
Thickness 2mm = to prepare gasket
Thickness 4mm = to prepare small pieces
In addition I bought a silicon to try to seal all T-nuts perimeters (no images available).
Drivers gasket holes detail
Direct link: https://s26.postimg.org/5nbswycg9/Holes.jpg
Small pieces
Direct link: https://s26.postimg.org/51rfel6rt/Small_Parts.jpg
Holes filled
Direct link: https://s26.postimg.org/cry7jq93d/Hole_fill.jpg
Drivers secondary gasket (inner diameter little shorter and outer diameter little larger).
Direct link: https://s26.postimg.org/xso94ncll/Gasket.jpg
Secondary gasket assembly detail 1
Direct link: https://s26.postimg.org/mhllga5qh/Gasket_Assembled_1.jpg
Secondary gasket assembly detail 2
Direct link: https://s26.postimg.org/4g2gihbpl/Gasket_Assembled_2.jpg
So, I made some measurements, with fastening more tight, less tight, with secondary gasket, without it, and see no big difference in the result, just some small improvement due to small pieces. The secondary Gasket brings more osculation the impedance response and make it worst.
The tests were performed at two cabs and both always present the same results.
Here are some graphics:
First measurement (as first built)
Direct link: https://s26.postimg.org/us7eo41ah/Z_A.png
Secondary Gasket + Small pieces
Direct link: https://s26.postimg.org/duczg0hah/Z_A-_Gasket.png
Small pieces only
Direct link: https://s26.postimg.org/kyusp1ojt/Z_A-_Holefill-1.png
Small pieces + fastening ajustment
Direct link: https://s26.postimg.org/5eneyiefd/Z_A-_Holefill-2.png
All graphics overlay
Direct link: https://s26.postimg.org/8a0i5difd/Z_All.png
Before my conclusion I prepare the graphic below to easy show what I'm thinking:
Direct link: https://s26.postimg.org/oz1y1af0p/Comparison.png
The impedance measured and simulated are very close till 35Hz, and the frequency response too, this frequency is close to the crossover one, so it will not bring much impact (The yellow area) once the slew rate is very high to constrain diaphragm displacement.
In addition, when the driver is brand new and been storage for long time it affects Thiele small parameter and them affect impedance response and frequency response, so other them leakage issue probably the system has just some T/S parameter different from the simulation.
I wish I could measure it now but I hadn't controlled mass to do it and I need to learn how to proper do it.
Finnaly, the result is good and maybe there is no need to put much effort on it, it's time to have fun 😀
Note: At the end I also discover that some wind and background noise was affecting my impedance curve adding more oscillation to the graphic.
I've been working theses last two days on the leakage stuffs and here is some feedbacks.
The driver was so fasten that we can see some print mark from driver's basket on the cab.
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/iaw5n7gqx/Fasten_Marks.jpg
I bought some pieces of Ethylene-vinyl acetate (EVA) to prepare the gasket and other small peaces, this is exactly the same material of drivers original gasket.
Thickness 2mm = to prepare gasket
Thickness 4mm = to prepare small pieces
In addition I bought a silicon to try to seal all T-nuts perimeters (no images available).
Drivers gasket holes detail
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/5nbswycg9/Holes.jpg
Small pieces
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/51rfel6rt/Small_Parts.jpg
Holes filled
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/cry7jq93d/Hole_fill.jpg
Drivers secondary gasket (inner diameter little shorter and outer diameter little larger).
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/xso94ncll/Gasket.jpg
Secondary gasket assembly detail 1
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/mhllga5qh/Gasket_Assembled_1.jpg
Secondary gasket assembly detail 2
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/4g2gihbpl/Gasket_Assembled_2.jpg
So, I made some measurements, with fastening more tight, less tight, with secondary gasket, without it, and see no big difference in the result, just some small improvement due to small pieces. The secondary Gasket brings more osculation the impedance response and make it worst.
The tests were performed at two cabs and both always present the same results.
Here are some graphics:
First measurement (as first built)
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/us7eo41ah/Z_A.png
Secondary Gasket + Small pieces
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/duczg0hah/Z_A-_Gasket.png
Small pieces only
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/kyusp1ojt/Z_A-_Holefill-1.png
Small pieces + fastening ajustment
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/5eneyiefd/Z_A-_Holefill-2.png
All graphics overlay
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/8a0i5difd/Z_All.png
Before my conclusion I prepare the graphic below to easy show what I'm thinking:
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/oz1y1af0p/Comparison.png
The impedance measured and simulated are very close till 35Hz, and the frequency response too, this frequency is close to the crossover one, so it will not bring much impact (The yellow area) once the slew rate is very high to constrain diaphragm displacement.
In addition, when the driver is brand new and been storage for long time it affects Thiele small parameter and them affect impedance response and frequency response, so other them leakage issue probably the system has just some T/S parameter different from the simulation.
I wish I could measure it now but I hadn't controlled mass to do it and I need to learn how to proper do it.
Finnaly, the result is good and maybe there is no need to put much effort on it, it's time to have fun 😀
Note: At the end I also discover that some wind and background noise was affecting my impedance curve adding more oscillation to the graphic.
very nice work in measurements and thumps up for actually testing different approaches for a 'problem'.
Related to cabs stack, may I ask for more support?
maybe someone will ask me to test it myself but I like to learn with people experience and sometimes the theory didn't cover all real life behaviors so we save a lot of time sharing experience.
Note: The below face is always touching the floor.
When using two cabs, what would be the best stack option below and why?
Once I have just a few gears probably stereo set up isn't the best option so I didn't consider it. Putting cabs closer is better to get more SPL 😀
2A - Vertical - Sibe-by-side
Direct link: https://s26.postimg.org/3t0juqnbt/SS15_Stacks_-_2_-_SBS_-_Vertical.png
2B - Horizontal - Sibe-by-side
direct link: https://s26.postimg.org/5jjkw84ux/SS15_Stacks_-_2_-_SBS_-_Horizontal.png
2C - Horizontal - Over
Direct link: https://s26.postimg.org/vcjfsl315/SS15_Stacks_-_2_-_Over_-_Horizontal.png
and when using four cabs?
4A - Vertical side-by-side
Direct link: https://s26.postimg.org/xspuk6ahl/SS15_Stacks_-_4_-_SBS_-_Vertical.png
4B - Horizontal Over
Direct link: https://s26.postimg.org/wjxdkst5l/SS15_Stacks_-_4_-_Over_-_Horizontal.png
4C - Horizontal Hybrid
Direct link: https://s26.postimg.org/wmluldm6x/SS15_Stacks_-_Cluster_-_Horizontal.png
maybe someone will ask me to test it myself but I like to learn with people experience and sometimes the theory didn't cover all real life behaviors so we save a lot of time sharing experience.
Note: The below face is always touching the floor.
When using two cabs, what would be the best stack option below and why?
Once I have just a few gears probably stereo set up isn't the best option so I didn't consider it. Putting cabs closer is better to get more SPL 😀
2A - Vertical - Sibe-by-side
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/3t0juqnbt/SS15_Stacks_-_2_-_SBS_-_Vertical.png
2B - Horizontal - Sibe-by-side
An externally hosted image should be here but it was not working when we last tested it.
direct link: https://s26.postimg.org/5jjkw84ux/SS15_Stacks_-_2_-_SBS_-_Horizontal.png
2C - Horizontal - Over
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/vcjfsl315/SS15_Stacks_-_2_-_Over_-_Horizontal.png
and when using four cabs?
4A - Vertical side-by-side
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/xspuk6ahl/SS15_Stacks_-_4_-_SBS_-_Vertical.png
4B - Horizontal Over
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/wjxdkst5l/SS15_Stacks_-_4_-_Over_-_Horizontal.png
4C - Horizontal Hybrid
An externally hosted image should be here but it was not working when we last tested it.
Direct link: https://s26.postimg.org/wmluldm6x/SS15_Stacks_-_Cluster_-_Horizontal.png
The graphs show some improvement when you used the "small pieces" to fill in the holes in the speaker's gasket, so it looks like that was a step in the right direction. Strange that the use of a secondary gasket didn't show similar improvement however.
The position of the lowest impedance peak can also be impacted by the driver's t/s parameters, but they'd have to be quite different from published specs to move the peak so much. You don't need to measure ALL of the parameters to confirm this - just use DATS to measure Fs, Qes, Qms. If Fs is significantly higher than published, that could account for the shift in the lowest impedance peak.
Anyway, well done with the testing! It's nice to see someone take a pretty analytical approach to their creation, rather than just building it and powering it up.
in my own experience with th18, best results is horizontale placement, mouth to mouth in both stack or row.