Beginner DIY DSP XO speaker measurement

[rthorntn]

Hi,

Please help, there is quite a bit of info out there on room correction but not so much on measurement during speaker "design".

I'll start my journey with a "spares bin" pair of Wharfedale 9.0 speakers to simplify things. So crossovers are removed and separate amp channels for each of the four drivers, protection caps on the tweeters.

Windows OS, MiniDSP U-DAC8 and UMIK-1.

CamillaDSP for the crossover.

Which measurements should I do and why?

So I'm trying to come up with a basic PC DSP "speaker build" test checklist, does such a thing exist? Paid software product agnostic would be best, where a free package is available to do the task it would be great if that's mentioned?

I think I would run some measurement tests to find the best HF crossover frequency, correct any channel volume imbalances, remove any channel delays, flatten its response?

So in plain English, ordered steps, the what and the why?

I'll add a separate sub later on, what does that add to the checklist, find best LF crossover frequency and measure and correct phase?

Thanks!

 

[Sokel]

Your strongest barrier for such a built will be anechoic data.
Have you thought of a way to do it?

You strongest tool to your built (as you have to deal with it as a whole new speaker from the ground up) is VituixCAD.
There are ways to get quasi-anechoic too,a great thread about it is here:

How to make quasi-anechoic speaker measurements/spinoramas with REW and VituixCAD

Quasi-anechoic measurements are a way to capture the 'real' frequency response of a speaker without the influence of a room. They are called 'quasi-anechoic' because they have the potential to get you really close to the results one might achieve with a non-reflective environment like an...

www.audiosciencereview.com

 

[rthorntn]

Thanks, I hadn't really thought about it.

Tbh I was just thinking I would put the UMIK-1 a metre away and that would be "good enough".

Reading a bit of the link, I basically want to "gate" the results?

 

[DWPress]

The free software Room EQ Wizard (REW) will get you close enough to where you want to go. Many of us build speakers and use REW to test results without the benefit of any anechoic data. Go Sokel's route if you really want to go off the deep end right away.

You already have all the tools you need, get familiar with REW and its functions, watch some videos and read some guides it's got pretty much everything you need but it's important to calibrate your SPL to get meaningful results for things like distortion. Ultimately you will isolate each channel (L woofer, L tweeter) in the miniDSP app and do a measurement sweep within an appropriate frequency range (don't send anything less than 1khz to most tweeters) to see how the woofer and tweeter respond in frequency and amplitude. Typically tweeters are more sensitive than larger drivers and will need to be reduced in SPL to match more closely to the woofer. If using different amps for tweeters and woofers be aware of the possibility they may have different output voltage which will also contribute to driver matching. Create EQ filters in REW to for any necessary corrections and export them as FIR or IIR for implementation in the XO.

Once you have a sweep of your woofer and tweeter (and yes - about 1 meter away will do BUT don't move the microphone or make any physical volume changes on amps between sweeps) you will have a good idea where the ideal location for the XO and what sort of slope may suit it best. The time alignment and any phase issues is another step. I'd suggest you take baby steps and keep reading, there is literally tons of info out there and what you ask for in simple "plain English" is beyond the scope of a single forum post.

 

[rthorntn]

Amazing, thanks, I purchased six channels of amplification, TPA3255, way too powerful for the tweeters:

200WPC @ 8Ω
315WPC @ 4Ω

But I picked up some 20K ALPS pots and I have a bench power supply so I can find a voltage and pot position that works for both and set them in stone.

I also got two giant "coke can" 80uf radial film capacitors to protect the tweeters.

 

[DWPress]

Might check the math on caps, mine are small polys but can't recall the value - there are online calculators for that.

You might be better off setting VC for the individual drivers in Camilla.