so who do you trust more? it takes extreme care to do a setup that can measure at this level of performance. we suspect that the high frequency distortion of the 94dB normalised plot is dominated by distortion from the mic pre in the GRAS 46AC capsule driving the 5m cable.
-
Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
PURIFI finally did a fully purified passive speaker design! The SPK 16 prototype is here - with a PTT tweeter
- Thread starter DTTOM37
- Start date
Although it doesn't matter whom I trust, I really want to believe in the measurements in your datasheets just because I want this level of performance to exist and be available to me.
Based on your comment I checked a few other tweeter measurements at their site. And indeed some harmonic distortion plots look suspicious although others do show 6dB lower H2 at 2kHz. Depending on the measurement setup -60dB may be hard to measure but they measured -40dB at 2kHz. This level of distortion should be well above noise/distortion of any GRAS 1/2" mic even on a 5m cable. Otherwise they could use any cheap hobby mic.
I also found multiple Purifi speakers measure worse on Klippel at audioxpress than in your datasheets. BL curves are often de-centered. Worst example is the PTT5.25X-NFA-01. Datasheet performance is impressive, 3rd party measurement not so much. So I'm asking myself if this is a matter of production or measurement tolerances.
Edit: Looking at the BL graphs again, part of the difference is the range of both axes. Still audioxpress result is considerably de-centered.
Based on your comment I checked a few other tweeter measurements at their site. And indeed some harmonic distortion plots look suspicious although others do show 6dB lower H2 at 2kHz. Depending on the measurement setup -60dB may be hard to measure but they measured -40dB at 2kHz. This level of distortion should be well above noise/distortion of any GRAS 1/2" mic even on a 5m cable. Otherwise they could use any cheap hobby mic.
I also found multiple Purifi speakers measure worse on Klippel at audioxpress than in your datasheets. BL curves are often de-centered. Worst example is the PTT5.25X-NFA-01. Datasheet performance is impressive, 3rd party measurement not so much. So I'm asking myself if this is a matter of production or measurement tolerances.
Edit: Looking at the BL graphs again, part of the difference is the range of both axes. Still audioxpress result is considerably de-centered.
Last edited:
HifiCompass does a really good job measuring:
hificompass.com
Purifi PTT1.3T04-HAG-01 33mm Alu dome tweeter review | HiFiCompass
1. What and why am I testing? 2. The datasheet 3. The package and construction 4. Impedance frequency response 5. On-axis frequency response (at 315 mm) 6. Off-axis frequency responses (at 315 mm) 7. Harmonic distortion (at 315 mm) 8. Voice coil current harmonic distortion 9. Step response 10...
Ok, that looks much better indeed. Also love that spiral below the pole piece vent.
Is the BL curve audioxpress measured for the PTT5.25X-NFA-01 correct? I considered this speaker as a replacement for defective Scan Speak 15W9530K02 but was put off by that BL measurement. It seems it could benefit from some adjustment of VC position although it may still be better than the Scan Speak.
Is the BL curve audioxpress measured for the PTT5.25X-NFA-01 correct? I considered this speaker as a replacement for defective Scan Speak 15W9530K02 but was put off by that BL measurement. It seems it could benefit from some adjustment of VC position although it may still be better than the Scan Speak.
Klippel LSI has trouble measuring Bl(x) and Kms(x) for some drivers. We developed our own analyser. in addition, we have seen that Voice Coil used a Bl(x) Klippel curve from another driver in a review of one of ours. the graph had the same time stamp as in a published review of a Satori 6.5.
Paul7777x
Addicted to Fun and Learning
- Joined
- Feb 11, 2026
- Messages
- 646
- Likes
- 912
That alone is enough to quash any doubt.
Checking your info before publishing is essential for your own self respect, never mind simply doing a proper job. And such a massive clanger is mockable.
So I just did.
Would be interesting to see proof for that. Klippel likely also wants to fix any issues that may exist with LSI.
If they publish wrong data of your speakers you should have them correct it. This includes excessive harmonic distortion compared to your own measurements. Otherwise they become a source of doubt regarding price to performance ratio.
we informed both.
A few questions for Lars Risbo?
Is the original box design as published in the Build Description V1.00 still a valid choice or has the move to cast components resulted in significant improvements that now supersede the original mdf version? From a visual perspective alone the two boxes differ considerably.
Is the crossover design still valid as well or have there been changes since the build description was released?
Is the 50mm radius on the front panel extremely important or would a 38mm (1 1/2") radius suffice? Are the 50mm upper and lower radii as important as the sides? I would really like to do these boxes in a nice veneer. If just the sides of the front panel were radiused I could fabricate the entire rear of the box as square sided and wrap the veneer without issue. It's the front corners that would negate that if the upper and lower radii are really needed.
The reason I ask is that I've now made the decision to move forward with this build and am ready to order drivers. I'm a retired lifelong cabinet maker / machinist and am well versed and equipped to fabricate the boxes in mdf. Although having the front panels done on a cnc would obviously be easier, a few small changes to the design would result in panels that could be very accurately made with tooling I currently have on hand. In all honesty, once the front panel has been jigged up and test pieces made, the rest of the panels required are simplistic.
I hope all these questions aren't out of bounds... the amount of work you all have done to bring this design to fruition is astonishing. And the results of all that work appear to be equally so!
Thanks!
Is the original box design as published in the Build Description V1.00 still a valid choice or has the move to cast components resulted in significant improvements that now supersede the original mdf version? From a visual perspective alone the two boxes differ considerably.
Is the crossover design still valid as well or have there been changes since the build description was released?
Is the 50mm radius on the front panel extremely important or would a 38mm (1 1/2") radius suffice? Are the 50mm upper and lower radii as important as the sides? I would really like to do these boxes in a nice veneer. If just the sides of the front panel were radiused I could fabricate the entire rear of the box as square sided and wrap the veneer without issue. It's the front corners that would negate that if the upper and lower radii are really needed.
The reason I ask is that I've now made the decision to move forward with this build and am ready to order drivers. I'm a retired lifelong cabinet maker / machinist and am well versed and equipped to fabricate the boxes in mdf. Although having the front panels done on a cnc would obviously be easier, a few small changes to the design would result in panels that could be very accurately made with tooling I currently have on hand. In all honesty, once the front panel has been jigged up and test pieces made, the rest of the panels required are simplistic.
I hope all these questions aren't out of bounds... the amount of work you all have done to bring this design to fruition is astonishing. And the results of all that work appear to be equally so!
Thanks!
Not Lars but with a tweeter that radiates as widely as the Purifi does any change will have some impact on the diffraction signature. 38mm is still a very good roundover size and the difference between that and 50mm is measureable but small. There will be some change from the effective baffle width of the change in roundover size. The tri corners do have an obvious impact on the baffle diffraction and leaving them out will result in a change. Whether this slight reduction/change in measured performance is worth an improved aesthetic look to you is a personal choice.
Thanks for your comments. This thread got awfully quiet after my post... your thoughts regarding the impact these radii have are very well taken. My desire to reduce the radius to 1 1/2" was driven by that being the maximum size bit that's available in a 1/2" shank. My router table is very capable and can handle that with ease but anything above that and it gets scary fast. A shaper is by far the more appropriate machine and I've now been able to locate someone locally that will work with me at the full 50mm dimension.
To cut to the chase, I'm looking for a way to create these boxes in a highly accurate and repeatable manner without a cnc being required. I'm darn close to achieving that goal now and should have completed drawings within a few days. The finished redesign does not modify the external appearance or internal structure in any way and stays true to Purifi's original dimensional values. Modifications to the original design are constrained to increasing ease of production only. I would be more than willing to post the final drawings if that's acceptable.
So, let me ask everyone... is there serious interest in a flat pack of the box as released in V1.00 of the build description?
Luvchampagne
Senior Member
I would be interested. I'm located in California. Where are you located?
South East Georgia.
However, I have fairly frequent reason to be in LA and the Bay Area as well as few other places about the country.
Hoping for a less rounded design. Ive asked around and most woodworking businesses are not doing it due to the difficulty.
Something akin to neumann perhaps, or mofi/cadentia 3 with their angled edges. I think those designs might be more accessible. But I absolutely get it, the design here is one of a kind and a statement.
Something akin to neumann perhaps, or mofi/cadentia 3 with their angled edges. I think those designs might be more accessible. But I absolutely get it, the design here is one of a kind and a statement.
I'd love to put together a pair of these, but not sure about the cabinet.
Lars, is there any word on providing a kit including front baffle? The rest looks "doable" by hand to me
Alternatively:
How does one turn the engineering drawings into a CAD /CAM file for a milling machine? I have never uses a CNC mill before, but I am contemplating buying one and learning how to use it to make the front baffle out of Valchromat
Thanks for any comments
Lars, is there any word on providing a kit including front baffle? The rest looks "doable" by hand to me
Alternatively:
How does one turn the engineering drawings into a CAD /CAM file for a milling machine? I have never uses a CNC mill before, but I am contemplating buying one and learning how to use it to make the front baffle out of Valchromat
Thanks for any comments
- Joined
- Jan 7, 2020
- Messages
- 1,689
- Likes
- 2,423
Did you order a set?
Frankly speaking most of us just need the front. Thinking of having the front 3d printed and the rest done normallyI'd love to put together a pair of these, but not sure about the cabinet.
Lars, is there any word on providing a kit including front baffle? The rest looks "doable" by hand to me
Alternatively:
How does one turn the engineering drawings into a CAD /CAM file for a milling machine? I have never uses a CNC mill before, but I am contemplating buying one and learning how to use it to make the front baffle out of Valchromat
Thanks for any comments
Audionaut
Major Contributor
I've built speaker cabinets many times before, using quarter-round molding to create the curves. It works perfectly.
Yes, a Kit with the Front would be perfekt.
I asked some carpenrty companys to Finish the cabinet with a CNC for me, but must could Not do that or it would Not fit in the proces or only if ordered large amount of pieces, so no Chance in my area and i dont want to get it shipped a Long way, to expensive and Potential to damage. Even a friend of Mine looked at the 3d Plans and could not do the Front because the CNC has to be very modern and accurate, and it would Take a lot of time. If @Lars Risbo would sell the Kit with the Front it would be possible for a acceptable price, otherwise a finish speaker Like Buchardt e50 or the AsciLab a6b would be at Same cost with warranty and less chances to get Something wrong
Similar threads
