Zaph Audio ZA-SR71 Review!

[gornir]

Hi everyone,

I have written a small review on the Zaph Audio ZA-SR71 loudspeaker design. The ZA-SR71 is a popular and a well-regarded design among the DIY community and rightfully so.

The ZA-SR71 is both technically well-designed and a good performer sonically, paired with a high performance/price ratio.

I have personally built and used the ZA-SR71 for some years now and I listen to them from time to time and I still enjoy them. I’ve used the ZA-SR71 as a sort of DIY reference loudspeaker and sometimes used it to benchmark other loudspeaker designs against.

I’ve made a small modification to the original ZA-SR71 regarding the enclosure volume. I found the 14 liter enclosure too small for the Seas ER18RNX mid-woofer and I increased the volume to 18 liters. Besides that, I’ve used exactly the same cross-over and the same cross-over components sold by Madisound in the ZA-SR71 kit.

In my review I’ve measured the driver units individually and the measurements includes, T-S parameters, distortion, impedance and frequency measurements as well as measurements for the finished design.

I’ve also made some cross-over simulation based on my own measurements to show some effect of changing the listening distance to the loudspeaker as well as showing the driver units cross-over slopes and phase behavior etc.

I think John Krutke (Zaph Audio) made a very nice DIY loudspeaker design with the ZA-SR71 and I just want to say that this review is my personal opinion and is based on my own sound preferences, measurements and design views and shouldn’t be consider as criticism to another loudspeaker designers design choices.

Please visit Zaph Audio website for the full ZA-SR71 design documentation regarding cross-over schematics, tuning options and the designer’s comments to the ZA-SR71.

For the full review see:
ZaphAudio ZA-SR71 Review!
SEAS H1456 ER18RNX Measurements!
SEAS H1189 27TDFC Measurements!

Regards

/Göran

 

[samadhi]

Hi Goran,

Ive just been browsing your website and looking at all the nicely documented designs. Good job!

Interesting thoughts on the ER18 enclosure volume, something i'll need to consider when I finally get mine in a proper cabinet.

 

[gornir]

Hi Goran,

Ive just been browsing your website and looking at all the nicely documented designs. Good job!

Interesting thoughts on the ER18 enclosure volume, something i'll need to consider when I finally get mine in a proper cabinet.

Thank you samadhi!

I did try the 14 liter option first and didn't like it. I've seen other people using a 16 liter enclosure for the ZA-SR71, but 18 liter and a lower port-tuning fitted my preferences better.

You can always build the enclosure larger and stuff it with something solid if you want to try a smaller enclosure tuning.

Regards

/Göran

 

[ODougbo]

Not much to add, here's a pic though - plans came with.

 

[sreten]

Hi,

I guess a sensible question is have madisound missed something by
not offering a MD20 box version as well as the MD14 box version.
The way things work the same port should work in both boxes.

rgds, sreten.

 

[PeteMcK]

Gornir, how long had the ER18RNX drivers been 'burnt in' before you took the measurements? I ask because mine don't measure with such a low Fs (I'm about to measure them again next weekend). I suspect that SEAS drivers take a bit longer to break in than some cheaper drivers.

 

[Michael Chua]

ER18RNX T/S measured with Delta Mass Method.

 

[gornir]

Gornir, how long had the ER18RNX drivers been 'burnt in' before you took the measurements? I ask because mine don't measure with such a low Fs (I'm about to measure them again next weekend). I suspect that SEAS drivers take a bit longer to break in than some cheaper drivers.

Hi PeteMcK,

I have done measurements several times on these mid-woofers. The last measurements are on my website, when they had about 50+ hours total break-in.

Fs rarely match the manufacturer specs. In my experience Seas is slightly better than others to match the spec sheets, but not for all their drivers.

I usually first measure the woofers brand new without a burn-in. Then I massage them with a medium-low signal for 24 hours and then measure them again. After that some drivers change the Fs, others don’t. Then I stress them for about 1 hour with a signal at approximately 80% of the drivers excursion limit and then measure them both warm and cold.

My two ER18 drivers seem to be from two different production batches and they differ about 3-4Hz between the samples. They have almost the same Fs after the burn-in as brand new.

You can also carefully move the cone back and forth 5-6 times to soften them up before measuring them, this often lowers the Fs a couple of Hz.

Regards

/Göran

 

[gornir]

ER18RNX T/S measured with Delta Mass Method.

Hi Michael Chua,

I see that you got a Mms of about 5.7grams, which seems to be very low. The Mms for the ER18 is around 14grams +/- 1gram.

I think you should try to measure them again with a higher "Delta Mass". A "Delta Mass" of 12gram is to small as added weight. It should be at least 125-150% of the actual Mms, sometimes more.

Regards

/Göran

 

[gornir]

Hi,

I guess a sensible question is have madisound missed something by
not offering a MD20 box version as well as the MD14 box version.
The way things work the same port should work in both boxes.

rgds, sreten.

Hi sreten,

Yes, that would work and you would have plenty of room for the cross-over filter. 🙂

I'm shure many think it sounds ok in the original 14 liter box, but I simply think it became a better speaker with the 18 liter option.

Regards

/Göran

 

[PeteMcK]

Thanks Gornir.

Michael, there is a wide discrepancy between your Vas measurement and everyone else's. Gornir get between ~ 27 - 36; Zaph, 26.3, my new ones were 36L, the data sheet has 32L

 

[Michael Chua]

Pete, that's what LMS calculated. I was shocked by the discrepancy. VAS of 61 Liters is almost double the manufacturer's specs.

This is not uncommon. I've measured other drivers with vastly different VAS from specs. Maybe the manufacturers decided to make changes (like a more compliant surround) and forgot to redo their measurements.

I see that you got a Mms of about 5.7grams, which seems to be very low. The Mms for the ER18 is around 14grams +/- 1gram.

I think you should try to measure them again with a higher "Delta Mass". A "Delta Mass" of 12gram is to small as added weight. It should be at least 125-150% of the actual Mms, sometimes more.

I usually start with 30 grams but with the ER18, it was too much. With 12 grams, the resonance frq already shifted from 44Hz to 25Hz. LMS recommends a one octave shift for accuracy. 12 grams is slightly more than 100% of 5.7 grams.

 

[PeteMcK]

re:'This is not uncommon' - well, not in my experience, that degree of difference would make me check the method / calibration. But we are using different methods, I'm surprised you guys use such high added mass, using ARTA / LIMP, I use an Aussie 10c coin, 5.65 grams, I'd use 20c (11.3 grams) for larger drivers, 8" and 10". (LIMP recommends 20-50% freq shift)

 

[5th element]

I haven't had much experience with measuring T/S parameters, I usually just trust the manufacturer, but some parameters, imo, should not be open to misinterpretation. Some perhaps more then others mind you, but the one I'd definitely expect not to change is the moving mass.

I too, Michael, find your measurements of the T/S parameters somewhat questionable, mainly because your arrival at 5.7g is not slightly off, but way off the mark. I would expect perhaps a couple of grams either side of SEAS measurement to appear due to experimental error, but not something as large as that.

 

[gornir]

Pete, that's what LMS calculated. I was shocked by the discrepancy. VAS of 61 Liters is almost double the manufacturer's specs.

This is not uncommon. I've measured other drivers with vastly different VAS from specs. Maybe the manufacturers decided to make changes (like a more compliant surround) and forgot to redo their measurements.



I usually start with 30 grams but with the ER18, it was too much. With 12 grams, the resonance frq already shifted from 44Hz to 25Hz. LMS recommends a one octave shift for accuracy. 12 grams is slightly more than 100% of 5.7 grams.

Hi Michael,

I've used both 19gr and 28gr added weight in the different measurements I've done on the ER18 and they both worked fine.

When I look at your impedance plots I can see that when you add the 12gr weight you get a wobbly impedance curve. This often indicates that the added weight has bounced around the cone or/and the driver is placed on a soft surface. Ideally the driver should be clamped to a solid surface, but that's just me guessing since I'm not familiar with your measurement setup.😉

I have not used LMS for T/S measurements, but I’ve used e.g. Clio and SoundEasy but none of them is as accurate as the WT3 tester which is the one I use today. Even though it’s a cheap tool it’s very fast, very accurate and consistent and I’ve measured tons of drivers with it without issues.

Seas 12cm drivers have larger Mms than 5.7gr so I think you should try to re-test them again.

Regards

/Göran

 

[tvrgeek]

If you change the box, it is no longer John's design. It is then a variant. Period. Now, on that variant, I was planning the exact same thing, adding depth so as not to change the baffle geometry. My other plan was to try out my big hooking 1 inch radius bit. So, that means some additional wood in the front corners and bottom loading of the parts. I was thinking about how to play off the SR-71 theme other than the obvious black paint. Red pinstripe or something. Sounds like an excuse to get out to Udvar-Hazy to see one again. The traffic should have backed off from the Discovery

 

[tvrgeek]

Strong second on the WT unit. I have an old WT-2. I also have a audio generator and VTVM for the old way, and Sound Easy for the new way. WT-2 is far more consistent. BUT. it really does not matter that much. You are looking at 20% maybe. If you have SE, you are just that much closer to a model. I measure with both mass displacement and sealed box. I do it until I am within 5% per method and 10% between them. Then I believe it except for Le which is calculated. I find I usually double it as close enough. This I confirm by impedance measurements with a Zobel in place. I discussed this with Smith & Larson and they confirmed, it is a very tricky problem even if you glue the vc in place. So, measure and double. Close enough. I measure all my drivers with it, not just woofers. This gives me something to compare to the SE model.

 

[tvrgeek]

VAS will vary with drive level BTW. I have never had an issue with Seas specs.

 

[PeteMcK]

re:'This often indicates that the added weight has bounced around the cone' - I've noticed this too...

 

[Michael Chua]

Göran

I will redo the measurements again later. I'm very busy working on a major landscaping project right now.

I haven't had much experience with measuring T/S parameters....

It's pretty straightforward with LMS. No calculation required. Basically, the steps are:

1) Impedance Sweep (Free Air)
2) Impedance Sweep with weight added. Weight should shift Fs about one octave
3) Input Added Weight
4) Input Sd

and LMS will generate T/S.

I too, Michael, find your measurements of the T/S parameters somewhat questionable, mainly because your arrival at 5.7g is not slightly off, but way off the mark. I would expect perhaps a couple of grams either side of SEAS measurement to appear due to experimental error, but not something as large as that.

There's no Input for Mms. All the information is generated by LMS based on the 2 graphs. It'll be interesting to see what comes out when I redo the measurements.

I've attached measurements for the VIFA NE265W-4 for comparison. With this model, my measurements are very close to Vifa's.

The first graph is of the two units (free air). Looks match to me.

Second graph is with 140 grams added.

Third is T/S (LMS)

Fourth is from Vifa datasheet.