Hi all.
I'm using a ZK-TB21amp @ 12V and aiming for 8W. I have a sealed .1 ft^3 box and some songs play well but for part of songs it sounds like a detuned AM radio.
I get the same sound with another smaller box and also with no box (low volume).
When adding a basic filter with what I had at hand, 0.33mH and 10 Ohm in parallel it improved, but I still get the AM sound.
I wonder what I am doing wrong. I better get it to work because I got a few of them.
Regards.
I'm using a ZK-TB21amp @ 12V and aiming for 8W. I have a sealed .1 ft^3 box and some songs play well but for part of songs it sounds like a detuned AM radio.
I get the same sound with another smaller box and also with no box (low volume).
When adding a basic filter with what I had at hand, 0.33mH and 10 Ohm in parallel it improved, but I still get the AM sound.
I wonder what I am doing wrong. I better get it to work because I got a few of them.
Regards.
The PS95 has a massive dip ~10dB in the middle of the vocal range. Driver has to be crossed to a tweeter, because the response begins to flatten about 15 degrees off-axis. Best to work the dip into the crossover slope, so you need a tweeter that can go down to 2kHz comfortably. I like the SB26STCN, nice small faceplate that takes very little box volume and baffle space.
Thanks Sangram.
A lot of people on the Internet say that it sounds great at a single-driver speaker. Wondering what they did... Cannot be listening to the same music I'm listening to 🙂
I'll try to get a tweeter soon. What could I do in the mean time if I want to listen to them?
I have an umik-1 and can I use REW in case it helps.
A lot of people on the Internet say that it sounds great at a single-driver speaker. Wondering what they did... Cannot be listening to the same music I'm listening to 🙂
I'll try to get a tweeter soon. What could I do in the mean time if I want to listen to them?
I have an umik-1 and can I use REW in case it helps.
That dip is indeed not good, but for the rest it's specs look ok.
But AM is mainly the loss of high frequency. And that should not be the case with this driver, so it must be something else. What kind of filtering do you use if any?
But in general, that is not the driver i would use. The Fountek FR88EX or the Peerless TG9FD10-04 would be a much better choice i think in that size and price bracket. And the best would be the exellent but way overpriced Scanspeak 10F/8424G00. Only a pitty that that price is so ridiculous high.
But AM is mainly the loss of high frequency. And that should not be the case with this driver, so it must be something else. What kind of filtering do you use if any?
But in general, that is not the driver i would use. The Fountek FR88EX or the Peerless TG9FD10-04 would be a much better choice i think in that size and price bracket. And the best would be the exellent but way overpriced Scanspeak 10F/8424G00. Only a pitty that that price is so ridiculous high.
lene, my experience matches yours. The driver is unusable in any form of one-driver array because of the dip - though it gets better off axis the top end just falls off completely at anything above 5k once you go that far off either axis.
Many people like the driver as is, as it recesses the midrange and makes for a relaxed presentation. You can work with a very narrow front baffle and try use a filter design that shelves the higher frequencies a bit, then use EQ to make up the top end if you want. This way you can probably get the dip to integrate with the baffle response (the baffle should be just as wide as the driver, but even then there will be stuff missing).
I tried valiantly to get a linear response fro the driver, and have not managed something I'm happy with (+/-2dB FR target) so I opted to add the tweeter and a bottom woofer to get it to sound sort of OK.
Many people like the driver as is, as it recesses the midrange and makes for a relaxed presentation. You can work with a very narrow front baffle and try use a filter design that shelves the higher frequencies a bit, then use EQ to make up the top end if you want. This way you can probably get the dip to integrate with the baffle response (the baffle should be just as wide as the driver, but even then there will be stuff missing).
I tried valiantly to get a linear response fro the driver, and have not managed something I'm happy with (+/-2dB FR target) so I opted to add the tweeter and a bottom woofer to get it to sound sort of OK.
I measured the Ps95-8 against many other well respected full range drivers and it’s not bad in the general evenness or high end output. In fact, it has the most extended top end. There is a small peak at 8kHz. But I would not complain that the driver lacks top end. These are gated measurements so only look above 800Hz.
For bottom end to be apparent, you will need to use 1mH and circa 6.8ohm in parallel for a baffle step correction to attenuate the highs to allow the bass to be heard.
0.1cu ft won’t get more than 100 Hz or so bottom.
The ZK-TB21 amp you are using a TPA3116 based Class D 2.1 with lots of tone controls and a low pass subwoofer output. Make sure you have all your tone controls set to defeated or neutral. And make sure you are not listening to the subwoofer output, or it would sound like a cheap AM radio under water. 🙂
I suspect your problem is amp related. Try a plain non-tone control amp if you have one. Or set all knobs at 12 o’clock and make sure you connect speaker to either left or right output and not sub.
For bottom end to be apparent, you will need to use 1mH and circa 6.8ohm in parallel for a baffle step correction to attenuate the highs to allow the bass to be heard.
0.1cu ft won’t get more than 100 Hz or so bottom.
The ZK-TB21 amp you are using a TPA3116 based Class D 2.1 with lots of tone controls and a low pass subwoofer output. Make sure you have all your tone controls set to defeated or neutral. And make sure you are not listening to the subwoofer output, or it would sound like a cheap AM radio under water. 🙂
I suspect your problem is amp related. Try a plain non-tone control amp if you have one. Or set all knobs at 12 o’clock and make sure you connect speaker to either left or right output and not sub.
Last edited:
Every single driver in that graph has a dip @2k. Is there a baffle issue?
Also, the scale is 10dB, so the peak to dip delta in the PS95 is about 8dB. That is egregious an not really high fidelity. It might be acceptable in some very marginal applications.
Also, the scale is 10dB, so the peak to dip delta in the PS95 is about 8dB. That is egregious an not really high fidelity. It might be acceptable in some very marginal applications.
2kHz is 3.36in distance from driver to some edge/lip - so probably yes as these were all measured on a relatively small trapezidal baffle of a Nautaloss speaker cabinet sitting on top of a ported XKi bass unit for a FAST setup. More info here:
A Subjective Blind Comparison of 3in to 5in Full Range Drivers
Here is photo of the baffle test setup:
Go sligthly off axis and the PS95-8 sounds quite nice. Maybe not HiFi - but the issues the OP has is far from HiFi - more like basic setup error because it "it sounds like a detuned AM radio" comment. I am saying that the PS95-8, properly implemented sounds far from a cheap AM radio.
A Subjective Blind Comparison of 3in to 5in Full Range Drivers
Here is photo of the baffle test setup:
Go sligthly off axis and the PS95-8 sounds quite nice. Maybe not HiFi - but the issues the OP has is far from HiFi - more like basic setup error because it "it sounds like a detuned AM radio" comment. I am saying that the PS95-8, properly implemented sounds far from a cheap AM radio.
Thanks a lot for the answers. The noise (that sounds like detuned AM ratio) happens only in parts of songs. So, you're listening to songs that sound nice but then there's the noise that ruins the experience. So, I like the songs but there's this weird noise that gets in the middle. English is not my first language so I don't really know how to describe it.
At some point I thought I had burned a speaker and replaced it with a new one and kept the volume low but I got the same behavior.
I just noticed that if EQ is completely neutral it sounds better. Even neutral for bass, I had it dampened a bit because I have a sub. It's a LW150-4 6" and it sounds great with this AMP.
At the beginning I thought I had issues with the power source but now I'm running with a 12V LIPO batt and I can rule this out.
I'll get some measurements (REW). Would recording a bit of a known song or sample help?
Getting measurements now...
At some point I thought I had burned a speaker and replaced it with a new one and kept the volume low but I got the same behavior.
I just noticed that if EQ is completely neutral it sounds better. Even neutral for bass, I had it dampened a bit because I have a sub. It's a LW150-4 6" and it sounds great with this AMP.
At the beginning I thought I had issues with the power source but now I'm running with a 12V LIPO batt and I can rule this out.
I'll get some measurements (REW). Would recording a bit of a known song or sample help?
Getting measurements now...
Made the measurements. Using calibrated UMIK-1, listening position. Let me know what you think. Keeping the EQ knobs in neutral.
Box internal volume aims for .1cf, about 3lt.
Internal volume(cm^3): 13W x 23H x 10W.
Box internal volume aims for .1cf, about 3lt.
Internal volume(cm^3): 13W x 23H x 10W.
Place speaker on a stool (at least 1m high) in middle of room, put microphone 0.5m away. This will help eliminate some of the room floor bounce. Use IR Window and set to 6 cycles frequency dependent window so that we can reduce the effect of the room/floor.
It looks like there is a big dip at 1000Hz and 2000Hz and that causes lack of presence. There is a deep 250Hz dip probably floor bounce cancellation.
It looks like there is a big dip at 1000Hz and 2000Hz and that causes lack of presence. There is a deep 250Hz dip probably floor bounce cancellation.
Done.
This is with the IR window, one per speaker:
And without the IR window (default settings):
This is with the IR window, one per speaker:
And without the IR window (default settings):
Something is very wrong. Your response should look somewhat like the red curve at the top:
I don’t know why you have such a big peak at 1.5kHz or a big dip near 2kHz.
Try to plot your graphs on a similar scale from 40dB to 90dB with 5dB major tick mark increments.
I don’t know why you have such a big peak at 1.5kHz or a big dip near 2kHz.
Try to plot your graphs on a similar scale from 40dB to 90dB with 5dB major tick mark increments.
Last edited:
Hmm, looked up 'detuned AM radio', but still don't know how it affects its 100-5kHz frequency response or kind of noise it makes, so no way to really help beyond maybe a cavity resonance in the bullet phase plug comb filtering with the diaphragm's breakup modes if it's thin enough since the ~ 4-7 kHz sibilance dip normally damped by a dust cap is dimensionally inside the 2 cm diameter voice coil [VC].
This is the image I got.
I don't know how to describe what I am hearing. I recorded a sample (umik+mac+audacity). See second 10 for the noise I get. It sounds as saturated audio but it's not the recording, that's what I hear...
Let me know! Wondering if I need to use a different box.
I don't know how to describe what I am hearing. I recorded a sample (umik+mac+audacity). See second 10 for the noise I get. It sounds as saturated audio but it's not the recording, that's what I hear...
Let me know! Wondering if I need to use a different box.
Attachments
I'm not that sure if it matches your situation but I had a rather similar experience with a tiny 3" fullrange on a Icepower amp.
After some experiments I found out that chipamps perform very well on nearly constant loads.
After a simple impedance linearization of the driver the sound improved alot.
Next improvement was the implementation of a bafflestep and the little setup spread joy 🙂
After some experiments I found out that chipamps perform very well on nearly constant loads.
After a simple impedance linearization of the driver the sound improved alot.
Next improvement was the implementation of a bafflestep and the little setup spread joy 🙂
Thanks a lot. I noticed something yesterday, I don't really know how a 3" driver should sound like. I'm using the same test songs I use in the car and in a KRK ROKIT 8 G4 and yesterday I listened to them with monitoring headphones and some of these songs are bass heavy or don't sound good in the headphones.
As I said before some songs sound good, even if there's something there that gets in the middle or something missing.
Unfortunately I don't have another amp to test with. I have more identical ones and I already checked. I ordered a small Pam8610 just to see if I see different behavior.
Is it important that the borders of the baffle are sanded? They're a bit sharp at the moment.
I'll check impedance linearization. Any tips are welcome.
As I said before some songs sound good, even if there's something there that gets in the middle or something missing.
Unfortunately I don't have another amp to test with. I have more identical ones and I already checked. I ordered a small Pam8610 just to see if I see different behavior.
Is it important that the borders of the baffle are sanded? They're a bit sharp at the moment.
I'll check impedance linearization. Any tips are welcome.
Hi there.
I switched back to a power supply and using previous instructions I did more measurements and tweaked the amp EQ a bit. I need to do another build with these speakers so I'll look around for proven designs. Red line is with the sub. I'll listen to them like this unless there's a suggestion for a filter that I can use.
I switched back to a power supply and using previous instructions I did more measurements and tweaked the amp EQ a bit. I need to do another build with these speakers so I'll look around for proven designs. Red line is with the sub. I'll listen to them like this unless there's a suggestion for a filter that I can use.
Active speakers (ROKIT-8-G4), but not baffles to hook to the amp. If this is useful I can measure tomorrow in the same position.