My first REW Measurement - DIY Speakers - Did I get speaker polarity wrong?

[kemmler3D]

Is there any alternative to measuring outside?

Look into doing gated near field measurements, you can somewhat remove the room from the measurement without going outside. It doesn't work on low frequencies but it should at least give you an idea of whether the room or the speaker is ringing. Nope, see below.

 

[Keith_W]

Only by looking at the print. Its all stil in circuit atm, and I only have a multimeter and cannot measure capacitance or inductance.

If you have definitively proven that your tweeters are OK, then the problem has to be in your crossover. Even if the component values are correct, you might still have a faulty component. Unfortunately, I can't give you instructions on how to electrically measure your crossover with REW since I have never done it - but that is something you might want to consider googling. Maybe someone here has that knowledge, or maybe you could head over to DIYAudio.com and ask the folks there.

 

[Keith_W]

Look into doing gated near field measurements, you can somewhat remove the room from the measurement without going outside. It doesn't work on low frequencies but it should at least give you an idea of whether the room or the speaker is ringing.

Sorry, but this is incorrect. If you gate the measurement, you will remove ALL ringing, whether it is speaker or the room. It won't differentiate between one or the other.

 

[kemmler3D]

Sorry, but this is incorrect. If you gate the measurement, you will remove ALL ringing, whether it is speaker or the room. It won't differentiate between one or the other.

Oh, good point, that makes sense. Thanks for the correction. Still useful for diagnosing other problems, although not sure OP has identified any that would call for it.

 

[AnalogSteph]

Even now the surround of tthe new woofer feels much more flexible and compliant than the "good" old woofer. Maybe I should change the second original woofer, too?

Seems like a good idea to me. (Or replace the surround yourself, at least, assuming that matching replacement are available - usually a task reserved for decades-old drivers with foam surrounds.) If you're so inclined, you may want to rig up a TSP measurement setup and observe the differences. I would expect the old driver to have a higher fs and lower Qms.

I do find it disturbing that the surrounds in a name-brand driver would have degraded that much in what can't be much more than a decade, assuming the drivers weren't super old stock to begin with. The driver datasheet does mention a "natural rubber surround", and if I can tell you anything about natural rubber, it's that the stuff doesn't last and tends to go rock-hard or turn to goo after a few decades. Still, this seems a bit extreme.

Side note, the driver datasheet also mentions

Long, high temperature voice coil wound on an aluminium voice coil former gives low distortion and high power handling capacity.

High power handling, yes, but low distortion, not quite. It's a good thing that it's only being used as a woofer and crossed over around 600 Hz, as Al voice coil formers tend to result in major electrical nonlinearity and IM distortion due to eddy currents.

It may be a good idea to check the surrounds on the midranges as well, as they are of similar construction.
That, incidentally, includes the Al voice coil former. Oh crap. That's not a good design. The glass fiber variety would be much preferred. Who does that in a midrange? (Well... with fs = 68 Hz, it's arguably more of a midwoofer.) It would be salvageable with current driving, but that's not happening in an existing passive-XO design. Now again, being in a 3-way would help out quite a bit in practice, but is it ideal? No.

 

[nikb]

Short update: issue ist still there. I am totally overlooking something.

What I did today:
- desoldered all components of the tweeter crossover
- measured them (all resistors within 0.3 of specified value, cap seems OK as does coil)
- reassembled tweeter part of crossover from scratch (using the schematic)
- measured w. REW again - no change
- swapped midrange drivers between speakers (just out of paranoia), still no change

What I will do next (tomorrow):
- revisit the left speaker crossover - maybe I have overlooked something there. But then the left speaker shows the nicer fr in the measurements.

 

[nikb]

Finally found the culprit. As was to be expected, it was something silly.

Today i decided to do one final test. I swapped the crossovers between the speaker cabinets, something I have not done before. To my surprise, the crossover that I thought was "bad" played fine in the other cabinet.

At that point I began to suspect bad wiring. I decided to pull all cables from the cabinet and replace them. To my suprise, when I tugged on the tweeter, what moved down below at the crossover was the cable clearly labeled "midrange"

So now I will listen to some music - I think I have earned it, despite it clearly being my own fault. So glad I finally found the issue.

Next steps:
- try to identify any remaining issues
- attach the crossovers inside the cabinets
- after all is done: apply some sealant where the chassis are screwed into the cabinets

- then: decide on how to best improve my listening room (perhaps a carpet would improve the echo)

Here's todays measurements.
The left speaker still looks better than the one on the right. What do you think, is this normal or does it still need some looking into?

 

[digitalfrost]

While it looks good with the zoomed out Y-scale that you shared, if you compare left and right on a 50dB scale, you will see that the right speaker is quite a bit louder in the midrange and quite a bit quieter in the high frequency. Also generally speaking the frequency response is not very linear. How did you measure? Mic at tweeter height? And was what was the distance?

 

[nikb]

While it looks good with the zoomed out Y-scale that you shared, if you compare left and right on a 50dB scale, you will see that the right speaker is quite a bit louder in the midrange and quite a bit quieter in the high frequency. Also generally speaking the frequency response is not very linear. How did you measure? Mic at tweeter height? And was what was the distance?

The mic was at ear height at the listening position about 15 cm higher than tweeter (or ear-) height, about 2.9m from the speakers. Its not a perfect triangle and the mic was probably about 20cm closer to the speaker on the right but not much more.

I can hear that something is still off - while they sound better than ever before, the right side speaker dominates and sound can clearly be heard coming from the speaker while the left speaker seems to "disappear" (as it should). When listening it seems that everything is happening on the right side and there is no real stereo illusion.
I'm thinking about what I could check.

 

[Keith_W]

While it looks good with the zoomed out Y-scale that you shared, if you compare left and right on a 50dB scale

This has been pointed out over and over to the OP in this thread. I spent a good chunk in Book 1 of the REW guide expounding on this point. It's frustrating that this very basic point isn't coming across.

 

[nikb]

This has been pointed out over and over to the OP in this thread. I spent a good chunk in Book 1 of the REW guide expounding on this point. It's frustrating that this very basic point isn't coming across.

Thanks for the heads-up @Keith_W and @digitalfrost . I learned to properly zoom in the graph today. No nice looking zoomed out graphs from me any more.
Also the midrange looks awful atm - I will tackle that tomorrow.

 

[MaxwellsEq]

Thanks for the heads-up @Keith_W and @digitalfrost . I learned to properly zoom in the graph today. No nice looking zoomed out graphs from me any more.
Also the midrange looks awful atm - I will tackle that tomorrow.

View attachment 491248

I know you built these speakers and are invested, but I really think you should give up on them and buy something that measures well.

 

[Keith_W]

I know you built these speakers and are invested, but I really think you should give up on them and buy something that measures well.

Nah, he should continue. Never give up. He is learning a lot from this exercise, and IMO the learning is far more important than the result.

I bought terrible speakers. In fact I paid a lot of money for them. They did some things right, but some things were really off. It took years of investigating to find out why. If I had never embarked on that journey, I would still be stuck in subjectivist land. So I have a strong belief that the struggle is as important, if not more important, than the result.

 

[MaxwellsEq]

Nah, he should continue. Never give up. He is learning a lot from this exercise, and IMO the learning is far more important than the result.

I bought terrible speakers. In fact I paid a lot of money for them. They did some things right, but some things were really off. It took years of investigating to find out why. If I had never embarked on that journey, I would still be stuck in subjectivist land. So I have a strong belief that the struggle is as important, if not more important, than the result.

Good points!

Sorry @nikb carry on!

 

[Juhazi]

Definitely the right speaker is not OK, midrange plays too loud. Check the serial resistor's value.

I haven't seen independent measurements of TG speakers, but the in-room farfield response of L speaker looks OK. Low end is lean, but there room influence is huge so it is impossible to say anything definitive.

Original version's page has been removed and the new one is quite different http://www.troelsgravesen.dk/SEAS-3-Way-Classic-mkII.htm

 

[nikb]

I know you built these speakers and are invested, but I really think you should give up on them and buy something that measures well.

That was actually my first thought when I started measuring: use it as an excercise to finally learn how to use REW and then decide wheather to buy some current tech speakers (you know: waveguides, controlled dispersion and all that). But: now that I have drawn blood I really want to have them playing as good as I possibly can.
Then I can decide if these are good enough for me for a while.

Thus I will continue.

Also: My money is currently better invested into a carpet than into any fancy audio tech.

 

[nikb]

Original version's page has been removed and the new one is quite different http://www.troelsgravesen.dk/SEAS-3-Way-Classic-mkII.htm

This is the Version that I have, the mkII

 

[kemmler3D]

I know you built these speakers and are invested, but I really think you should give up on them and buy something that measures well.

Given that he is measuring almost 3m from the speaker, I wouldn't say we're looking at measurements of the speaker per se at all. And agree with @Keith_W , progress is being made and this seems like a good crash course in a good handful of audio topics...

 

[nikb]

just a short update:

This is what it currently looks like.
All changes were reflowing solder connections, rewiring and recrimping crim-on connectors only. No out of place components were found in the crossover, all components measured OK.

 

[kemmler3D]

just a short update:

This is what it currently looks like.
All changes were reflowing solder connections, rewiring and recrimping crim-on connectors only. No out of place components were found in the crossover, all components measured OK.

View attachment 491441

Given that the speakers (as I understand it) are not in the same place in the room or relative to the mic, I think this looks OK? The dip on the right speaker between 400-700 is a little concerning but could be due to the room / placement.