Hello Jim,
Nice project as always. Sorry if I missed it but do the braces behind the bass drivers not come in contact with the front baffle and if so why? Not questioning the logic, just trying to learn. Thanks.
Paul
Nice project as always. Sorry if I missed it but do the braces behind the bass drivers not come in contact with the front baffle and if so why? Not questioning the logic, just trying to learn. Thanks.
Paul
That is a good question.
Those braces do not come in contact with the front baffle. To do so would have slightly complicated the assembly because it would have been one more glue surface and one more opportunity for tolerance build up. If the brace had extended all the way to the baffle, it would not have added any additional stiffness. The brace would have to get quite thin in order to clear the woofer frame. The side walls have significant stiffness as a shear panel, so a thin section of bracing would have added negligible stiffness to the already stiff wall. The unsupported span of side wall from the baffle to the start of the brace is only about 1.5 inch.
j.
I will probably measure distortion when the project is complete.
The adjustment of PR mass will be driven by frequency response and tonal balance. This is the primary performance aspect, and everything else is secondary. It is possible that varying levels of PR mass might affect distortion levels to a small degree, but this is less important than getting the tonal balance correct. I feel that optimizing for distortion performance at the expense of frequency response performance would be a mistake.
At this point in the design, with the drivers selected and installed in a cabinet of fixed volume, the distortion performance is mostly unchangeable... Very little could be done to change the performance. So whatever I get, I get.
I have never actually measured the excursion of a woofer.
j.
Please let us know what amp you get. My old Dayton APA150 started acting up a few months ago so I got a used Audiosource AMP100. I found a good quality one on eBay for $62 shipped which was quite the steal as they aren't usually that cheap so I'm content with it. But the new Fosi V3 mono (with PFFB) recently came out and I've looked at them for a stereo music setup but it occurred to me that a single one might make a good measurement amp.
I am leaning towards the Icepower 50ASX2BTL. 50 wpc, flat frequency response to 20k independent of load, able to handle 2 Ohm loads, and THD+n is less than 0.02% from 100 mW up to 20 W. At 40 W the THD+n rises to 0.1%, which is still good performance... For a test amp, I think it is pretty good, especially for $130 including the integrated power supply. It is a board only, so it will need a small case and some connectors.
I have that same 50ASX2. installed on an aluminium plate without cover, just edges turned up for connectors. Ugly but works fine. I can also use an old stereo integrated amp Vincent SV-129
Hello,
Yes one variable will effect another.
Adjusting the mass of the Passive Radiators will effect the excursion of the the woofers. Bl(x) and IMD will be effected as a result.
Placing the woofer in a sealed enclosure will also effect the excursion of the woofer. This time around you are not doing that.
I know you get what you get and the goal is not to change or control the resulting IMD but to adjust frequency response and tonal balance.
However, along the way I would like to measure it.
Thanks DT
DT, How does one measure woofer excursion? I see Klippel have an excursion module with various laser displacement sensors that can measure it, but all are in the multi thousand dollar range. Is there a DIY budget method?
Thanks!
Thanks!
One of the first set of measurements/evaluations I made was an assessment of cabinet radiation. I played pink noise through the woofer at a level of about 85 dBA and used a mechanics stethoscope to probe all surfaces of the cabinet. There is a slight wrinkle in the woofer impedance curve at 330 Hz, so I was listening for that tone. I found a region on the sidewall where I could hear a slight emphasis in the 300 - 350 Hz range, but a nearfield scan of that area showed the cabinet radiation was more than 20 dB below the driver radiation.
I repeated the process playing pink noise through the midrange driver. The mid has a slight wrinkle in the impedance at 435 Hz so again I was listening for that tone. I could find no area of the cabinet that produced a noise signature in the 400 - 500 Hz region. From this evaluation, I determined that the cabinet needs no further bracing or damping.
Initial bass evaluation
The next thing I did is make a preliminary evaluation of the bass tuning. Just to review, I am using a Dayton RS225-8 woofer in a 42 liter (net) volume, with a Dayton DS270-PR Passive Radiator. According to simulation in the VituixCad enclosure tool, I would need 30 g of added mass on the PR to achieve the desired response. The PR peaks at 35 Hz, the woofer has a minimum at 34.5 Hz, and the impedance has a minimum at 34.5 Hz.
Using near field scans of the woofer and PR, I found my tuning is slightly low. The PR peaks at 33.5 Hz, and the woofer has its minimum at 33 Hz. This is a difference of 1.5 Hz, or 5%. I was pleased at how close the measurements match the simulation. For now I am not going to make any changes to the PR, it is close enough. I will wait until the subjective voicing process to make any changes.
j.
I repeated the process playing pink noise through the midrange driver. The mid has a slight wrinkle in the impedance at 435 Hz so again I was listening for that tone. I could find no area of the cabinet that produced a noise signature in the 400 - 500 Hz region. From this evaluation, I determined that the cabinet needs no further bracing or damping.
Initial bass evaluation
The next thing I did is make a preliminary evaluation of the bass tuning. Just to review, I am using a Dayton RS225-8 woofer in a 42 liter (net) volume, with a Dayton DS270-PR Passive Radiator. According to simulation in the VituixCad enclosure tool, I would need 30 g of added mass on the PR to achieve the desired response. The PR peaks at 35 Hz, the woofer has a minimum at 34.5 Hz, and the impedance has a minimum at 34.5 Hz.
Using near field scans of the woofer and PR, I found my tuning is slightly low. The PR peaks at 33.5 Hz, and the woofer has its minimum at 33 Hz. This is a difference of 1.5 Hz, or 5%. I was pleased at how close the measurements match the simulation. For now I am not going to make any changes to the PR, it is close enough. I will wait until the subjective voicing process to make any changes.
j.
One thing to consider with the ASX modules is that the input impedance is relatively low, if the output impedance of the previous stage is high a buffer might be useful.
https://theslowdiyer.wordpress.com/2013/08/25/past-projects-6-channel-icepower50asx-amp/
I bought my IcePower amp modules ten years ago for active 4-way speakers, and ordered two extra units for reserve.
A stereo unit is neat, because you can measure two drivers with and without xo easily
Now I would choose a ready-made Chinese class D amp like Fosi Audio V3 Mono. It has both XLR and RCA input.
https://www.audiosciencereview.com/...ds/fosi-audio-v3-mono-amplifier-review.53474/
A stereo unit is neat, because you can measure two drivers with and without xo easily
Now I would choose a ready-made Chinese class D amp like Fosi Audio V3 Mono. It has both XLR and RCA input.
https://www.audiosciencereview.com/...ds/fosi-audio-v3-mono-amplifier-review.53474/
Last edited:
That is a good thought. I checked the specs of my Motu M2 and the output impedance is 100 Ohm. The input impedance for the 50ASX is 8 kOhm, a factor of 80. That should be OK ? yes?
The Fosi mono amp is a great recommendation. @tktran303 pointed it out to me a few days ago (I was unaware of it until then)...
Here is the fully merged woofer response. Near field scans of both woofer and passive radiator were merged with the far field woofer scans, in accordance with VituixCad instructions. Voltage level was about 0.7 Vrms. The distance for the FF scans was 39 inch (1 m). I made scans ever 15 degrees, from 0 to 180.
I made similar merged responses for all three drivers, all at the same voltage level and distance. When measuring the woofer, I elevated the speaker so the woofer was 44 inch high. When I scanned the mid and tweeter, the mic was 47 inch high. I use a common mic position for mid and tweeter, halfway between the two. The CtC distance from mid to tweeter is 137 mm. Relative to the design axis / measurement axis, the tweeter is +66 mm (up), the midrange is -77 mm (down), and the woofer is -344 mm (down).
When applying a passive crossover to a woofer, there of course will be interactions with the woofer impedance. I never had to worry about such things when doing active crossovers, but this is "unknown territory" for me.
If a 400 Hz LR2 low pass active filter is applied to the woofer+PR, this would be the response, also showing the original.
Now if I apply passive filter, one which achieves an acoustical LR2 rolloff at 400 Hz, something interesting happens.
This passive filter creates about 2 dB of gain from 70 Hz to 110 Hz. While this is certainly welcome, it has the effect of reducing the bass extension. It now looks like the woofer+PR begin a gentle rolloff from 70 Hz down to 40 Hz, then a rapid drop below 40. I knew this sort of thing was possible, but it is interesting to see it in effect. This is another reason why the PR will need to be adjusted during the subjective voicing process.
j.
I made similar merged responses for all three drivers, all at the same voltage level and distance. When measuring the woofer, I elevated the speaker so the woofer was 44 inch high. When I scanned the mid and tweeter, the mic was 47 inch high. I use a common mic position for mid and tweeter, halfway between the two. The CtC distance from mid to tweeter is 137 mm. Relative to the design axis / measurement axis, the tweeter is +66 mm (up), the midrange is -77 mm (down), and the woofer is -344 mm (down).
When applying a passive crossover to a woofer, there of course will be interactions with the woofer impedance. I never had to worry about such things when doing active crossovers, but this is "unknown territory" for me.
If a 400 Hz LR2 low pass active filter is applied to the woofer+PR, this would be the response, also showing the original.
Now if I apply passive filter, one which achieves an acoustical LR2 rolloff at 400 Hz, something interesting happens.
This passive filter creates about 2 dB of gain from 70 Hz to 110 Hz. While this is certainly welcome, it has the effect of reducing the bass extension. It now looks like the woofer+PR begin a gentle rolloff from 70 Hz down to 40 Hz, then a rapid drop below 40. I knew this sort of thing was possible, but it is interesting to see it in effect. This is another reason why the PR will need to be adjusted during the subjective voicing process.
j.
I can't understand why they chose to make it peak in the ultrasonic region instead of dip.
The 3e Audio is the implementation that looks best to me, just not available as a ready to go option.
https://www.audiosciencereview.com/...dio-tpa3255-260-2-29a-amplifier-review.50208/
Here is the fully merged tweeter response. Distance for the far field polar scans was 1 m. As I described in the previous post
This is a waveguide tweeter, although the guide is small and does not really become effective until about 3k. Like most waveguide tweeters, there is a lot of gain at the lower end, this means that practical crossovers will typically have a shallow slope over the whole range of tweeter operation.
For example here is one reasonable passive filter that achieves a 3k LR2 acoustical response
If we look at the filter electrical response, we see that there is no region that we would consider the pass band. The filter rolloff starts at 20k and operates over the whole range of the tweeter.
This is very different than a typical flat flange dome tweeter. One big difference I have found is the method to adjust the gain level of the tweeter. With a traditional dome tweeter, I typically apply some amp-side resistance. But with this tweeter+filter, adding resistance changes the response more than it decreases gain. Surprisingly, changing the series cap has the effect of changing the gain across the whole range of response. Another thing that works is an L-pad resister network on the tweeter side.
Here is the tweeter response with 2.2 and 3.9 Ohm amp side resistor
Here is response with no added resistance, but changing the series cap from 2.7 to 3.3 to 3.9 uF.
This to me seems like a much more effective way of managing the tweeter level over the intended range. I am sure this is old news to many people, but it is certainly new, and interesting, to me.
j.
This is a waveguide tweeter, although the guide is small and does not really become effective until about 3k. Like most waveguide tweeters, there is a lot of gain at the lower end, this means that practical crossovers will typically have a shallow slope over the whole range of tweeter operation.
For example here is one reasonable passive filter that achieves a 3k LR2 acoustical response
If we look at the filter electrical response, we see that there is no region that we would consider the pass band. The filter rolloff starts at 20k and operates over the whole range of the tweeter.
This is very different than a typical flat flange dome tweeter. One big difference I have found is the method to adjust the gain level of the tweeter. With a traditional dome tweeter, I typically apply some amp-side resistance. But with this tweeter+filter, adding resistance changes the response more than it decreases gain. Surprisingly, changing the series cap has the effect of changing the gain across the whole range of response. Another thing that works is an L-pad resister network on the tweeter side.
Here is the tweeter response with 2.2 and 3.9 Ohm amp side resistor
Here is response with no added resistance, but changing the series cap from 2.7 to 3.3 to 3.9 uF.
This to me seems like a much more effective way of managing the tweeter level over the intended range. I am sure this is old news to many people, but it is certainly new, and interesting, to me.
j.
Keep an eye on the DCR that VituixCAD assumes for the inductor. Parts Express has 6.0 and 7.0 mH laminate core inductors. The 7.0 mH has a DCR of 0.61 ohms whereas your screenshot shows an assumed 0.28 ohms. I know the difference will be minor, but I also know you pay attention to the smallest details. (I mean that in a good way.)
Your tweeter response looks pretty smooth, so it likely won't matter, but putting the series resistor on the amp side, the driver side, or an L-pad configuration will have minor differences. (Same for resistors on the midrange.) I have had tweeters that were a challenge to get what I wanted and split the difference and used two resistors, one on the amp side and one on the driver side to fine tune what I wanted to achieve. You can view the power dissipation in VituixCAD also, so I try to pay attention to those resistors on the amp side.