Maybe something to consider. Maybe I am wrong.
Your box is heavily stuffed. What we are achieving with stuffing closed boxes? We get apparently bigger volume due to lovered speed of sound. As you wrote Q fell from 0.90 to 0.74, what is not small step. So, as box is apparently biger, first standing wave is lower.
Try to model the box which dimensions giving Q=0.74 without any stuffing and calculate standing wave resonance. Could be close to your measured resonance.
Last edited:
@Zvu I have, and yes, it looks like it's at the top of the pile.
My only concern is the high Fs (95 Hz) and the rising distortion from 250 Hz upward shown in the datasheet. I know it's measured at 110 dB, but I’m not sure where distortion will start creeping in at lower levels.
I came across this set of measurements:
https://www.robsandiy.com/html/mid_bass/bms_5s117_8.html
Distortion appears to sit between 0.1–0.2% at 85 dB across the passband I’m interested in, which is great. The Fs there shows 100 Hz, slightly higher, but it's still a lot of money to commit to something that's unproven for my application. I'd love to see some polar measurements too.
What are your thoughts?
I did A-B the single speaker against one of my Genelecs and I didn't like it but then I was probabbly looking for problems.
My only concern is the high Fs (95 Hz) and the rising distortion from 250 Hz upward shown in the datasheet. I know it's measured at 110 dB, but I’m not sure where distortion will start creeping in at lower levels.
I came across this set of measurements:
https://www.robsandiy.com/html/mid_bass/bms_5s117_8.html
Distortion appears to sit between 0.1–0.2% at 85 dB across the passband I’m interested in, which is great. The Fs there shows 100 Hz, slightly higher, but it's still a lot of money to commit to something that's unproven for my application. I'd love to see some polar measurements too.
What are your thoughts?
I did A-B the single speaker against one of my Genelecs and I didn't like it but then I was probabbly looking for problems.
Callooh callay, data confirmation day!
I had to get to the bottom of the cube crisis. Or is it a driver crisis? Crisis of confidence? Either way, I ran some tests.
I measured the Faital Pro 5FE120 (Drivers A and B) and used a Peerless P830870 as a control and 3D-printed an adapter that let it fit in the same box (adorable). All three were measured in:
Turns out the oddity around 1kHz in the 5FE120 made the box resonance tough to spot. It was there all along, just buried in the chaos. What are the chances that the 5FE120 response throws a fit right on top of the resoannce? The P830870 makes it easier to see - the resonance shows up around 1.2kHz and is completely killed by damping.
As @danibosn rightly pointed out, the resonance shifts lower when the 5FE120 is in the box. My guess is the big magnet and general driver bulk change the physics.
I had to get to the bottom of the cube crisis. Or is it a driver crisis? Crisis of confidence? Either way, I ran some tests.
I measured the Faital Pro 5FE120 (Drivers A and B) and used a Peerless P830870 as a control and 3D-printed an adapter that let it fit in the same box (adorable). All three were measured in:
- Free Air
- Un-damped Box
- Damped Box (rockwool + melamine foam)
Turns out the oddity around 1kHz in the 5FE120 made the box resonance tough to spot. It was there all along, just buried in the chaos. What are the chances that the 5FE120 response throws a fit right on top of the resoannce? The P830870 makes it easier to see - the resonance shows up around 1.2kHz and is completely killed by damping.
As @danibosn rightly pointed out, the resonance shifts lower when the 5FE120 is in the box. My guess is the big magnet and general driver bulk change the physics.
- Faital Pro 5FE120
- Peerless P830870
Also...
I did a couple more distortion tests on the 5FE120 alone in the cube at 80cm.
One where the SPL peaks at 85dB and another at 95dB.
I did a couple more distortion tests on the 5FE120 alone in the cube at 80cm.
One where the SPL peaks at 85dB and another at 95dB.
Good for you in persevering and doing some more experiments.
Printing the holder for the Peerless driver was a great idea to help you see what was really going on.
From what I remember for distortion for the 5fe120 that looks right. I have a Clio pocket and it simply provides THD, 2nd and 3rd distortion measurements which is handy for a sanity check.
Printing the holder for the Peerless driver was a great idea to help you see what was really going on.
From what I remember for distortion for the 5fe120 that looks right. I have a Clio pocket and it simply provides THD, 2nd and 3rd distortion measurements which is handy for a sanity check.
Attachments
I purchased 5fe100 and 5FE120's to try out an idea approximately two and a half years ago, so quite some difference in build dates I imagine.
My cabinets are slightly larger than your cube, being an approx 8Litre reflex and my measurements were at a lower level than yours. So there is likely to be some differences. Your point about spikiness could be simply differences in chosen resolution. I usually try to aim for 12th octave smoothing.
My cabinets are slightly larger than your cube, being an approx 8Litre reflex and my measurements were at a lower level than yours. So there is likely to be some differences. Your point about spikiness could be simply differences in chosen resolution. I usually try to aim for 12th octave smoothing.
Measuring Sanity (and Distortion) at 20cm
I'm human. I'm weak. I can't control myself.
The 5FE120 is OK, but I'd always be left wondering what it could’ve been like with something better. The BMS 5S117 gets recommended all over the place but I haven’t seen anyone actually use them in a project. So... here we are.
I ran some measurements to keep myself grounded and promptly found something that ungrounded me further. Distortion at 80 cm is clearly better than the 5FE120, but is it 400% more expensive better? At 80 cm, probably not.
But then...
My room’s noise floor is too high to measure distortion accurately at 75 dB, which is my actual listening level. So I set levels for 75 dB and 85 dB at 80 cm, then moved the mic to 20 cm. The measurements were enormously better.
Which got me thinking:
Anyway, I now have the best of the best (on paper, anyway). What’s clear is that most - if not all - 5" drivers have some weirdness around 2 kHz, probably down to inherent geometry. I’m hoping to tame that in the crossover. Either way, it's nowhere near as offensive-looking as the 5FE120.
Anyway, I can get on and finish the final cube cabs now...
I'm human. I'm weak. I can't control myself.
The 5FE120 is OK, but I'd always be left wondering what it could’ve been like with something better. The BMS 5S117 gets recommended all over the place but I haven’t seen anyone actually use them in a project. So... here we are.
I ran some measurements to keep myself grounded and promptly found something that ungrounded me further. Distortion at 80 cm is clearly better than the 5FE120, but is it 400% more expensive better? At 80 cm, probably not.
But then...
My room’s noise floor is too high to measure distortion accurately at 75 dB, which is my actual listening level. So I set levels for 75 dB and 85 dB at 80 cm, then moved the mic to 20 cm. The measurements were enormously better.
Which got me thinking:
- Is the room the reason distortion looks worse than it is? Maybe.
- Would the 5FE120 have looked better using this method? Probably.
- Was all this fuss for nothing? Possibly.
Anyway, I now have the best of the best (on paper, anyway). What’s clear is that most - if not all - 5" drivers have some weirdness around 2 kHz, probably down to inherent geometry. I’m hoping to tame that in the crossover. Either way, it's nowhere near as offensive-looking as the 5FE120.
Anyway, I can get on and finish the final cube cabs now...
Last edited:
Building a cab with Fibracolour feels like getting back with a troublesome ex - I know you have problems, but I’m sure I can fix them this time...
If you’re mixing Valchromat/Fibracolour with anything clean-looking, treat them separately:
Now you’ve got a dust-proof barrier. When the black dust returns (and it will), you can just wipe it off.
On a massive positive: CAD and 3D printing saved me so much time, effort, and rage on the baffle recess.
If you’re mixing Valchromat/Fibracolour with anything clean-looking, treat them separately:
- Sand everything to 120g
- Then the birch with 180g to get rid of the black
- Mask it off
- Apply liberal coats of finish
Now you’ve got a dust-proof barrier. When the black dust returns (and it will), you can just wipe it off.
On a massive positive: CAD and 3D printing saved me so much time, effort, and rage on the baffle recess.