• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. 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!

Help me convert old speakers to actives

Hypex FA 123 is 500€ so 1 K€ for stereo and Umik around 130 € (you have to look up prices outside EU ;), tho).

REW for free.
Multimeter should not be missed in any household (10 bucks), but not needed essantially.

Add a solder iron/station or/and crimping tool to your list.
have soldering iron already. Multimeters look cheap on Amazon, so will get one once ready to start.
 
For multichannel multi-SP-driver multi-amplifier active system, the "right-person-in-right-place" selection of amplifiers would be one of the critical issues.

For example, you do not need powerful amplifier for your highly-efficient tweeters, but you need reasonably-small-power high S/N low-distortion amplifier for tweeters (and super-tweeters) resulting better/best total sound quality.

Even Greg Timbers uses "reasonable and budget" Pioneer Elite A-20 for compression drivers (super tweeters) in his extraordinary expensive multichannel stereo system with JBL Everest DD67000 which he himself designed and developed (ref. my post here).
I think if I use the existing drivers, those would be from the very bottom of the manufacturers' range and therefore almost any modern amp would work fine (i.e the drivers will be the limiting factor in performance). And the Hypex plate amp I would think will be good enough for mid-range and even top-of-the-range drivers anyway. There is no lower-powered amp in the Hypex range.
 
Some comments.

1. Don't get a USB mic. If you must get a USB mic, get a UMIK-2. The reason: unreliable timing information due to different clocks between the ADC on the mic and DAC. John Mulcahy (author of REW) tested a few mics here. In addition, the Audiolense user forum is full of complaints from users who use USB mics. A 48V Phantom power mic with an interface is much better. Someone suggested a Focusrite 2i2, but I would suggest that you get an interface with as many DAC channels as drivers that you need to control. In your case, 6. Something like an RME Fireface UC/UCX, Motu Ultralite Mk.5, Merging Anubis, or its Neumann clone (I forget what it's called).

2. Choice of DSP. You can choose either hardware (MiniDSP, DSPNexus, etc) or software. You are limited to software that can actually generate crossovers, which means Acourate, Audiolense, REW, DRC-FIR, Eclipse Audio FIR Designer, etc. PC based DSP is vastly more powerful, both in terms of processing power and the fine corrections that can be done. Hardware DSP is more robust and more convenient.

3. Choice of XO points and slopes. Measure your drivers and look for distortion and anomalies in the output. Decide if you want to deal with it, and how. For e.g. cabinet resonances can be dealt with through physical means, which is the proper way. The lazy way is to notch them out with DSP or remove them with a crossover so you don't have to worry about them :)

Good luck!! It's a fun journey. I had tonnes of fun doing exactly what you are proposing.
 
For some reason, I keep thinking that the existing drivers, once I measure them, will measure crap and I will want to use better drivers. Or is not that likely?
Will I be able to measure distortion (not just frequency response) though? Assuming I just have REW.
 
And the Hypex plate amp I would think will be good enough for mid-range and even top-of-the-range drivers.

Yes for your starting-up, I assume.
Please take intensive care, however, to implement proper safety measures (protection capacitors) for your midrange drivers and tweeters (ref. here).
And also take best care of your ignition/startup sequences and shutdown sequences (ref. here).
 
Will I be able to measure distortion (not just frequency response) though?

I too say "Yes!".

And, the time alignments between the SP drivers would be another real thrill and great benefit in DSP-based multichannel audio setup (summary ref. here and here).
 
Some comments.

1. Don't get a USB mic. If you must get a USB mic, get a UMIK-2. The reason: unreliable timing information due to different clocks between the ADC on the mic and DAC. John Mulcahy (author of REW) tested a few mics here. In addition, the Audiolense user forum is full of complaints from users who use USB mics. A 48V Phantom power mic with an interface is much better. Someone suggested a Focusrite 2i2, but I would suggest that you get an interface with as many DAC channels as drivers that you need to control. In your case, 6. Something like an RME Fireface UC/UCX, Motu Ultralite Mk.5, Merging Anubis, or its Neumann clone (I forget what it's called).

2. Choice of DSP. You can choose either hardware (MiniDSP, DSPNexus, etc) or software. You are limited to software that can actually generate crossovers, which means Acourate, Audiolense, REW, DRC-FIR, Eclipse Audio FIR Designer, etc. PC based DSP is vastly more powerful, both in terms of processing power and the fine corrections that can be done. Hardware DSP is more robust and more convenient.

3. Choice of XO points and slopes. Measure your drivers and look for distortion and anomalies in the output. Decide if you want to deal with it, and how. For e.g. cabinet resonances can be dealt with through physical means, which is the proper way. The lazy way is to notch them out with DSP or remove them with a crossover so you don't have to worry about them :)

Good luck!! It's a fun journey. I had tonnes of fun doing exactly what you are proposing.
Agree in basic aspects, but as it is a "Fun and Learning" project I would lean to Einstein "make it as simple as possible but not simpler" and add "not to expensive for start" to it.;)
 
Sorry, but I'm unfamiliar with Hypex modules, so my advice is out of date. I still use measuring microphones that need phantom power and an XLR input!
Not completely out of date. It just depends on the use case.

When tuning with REW I use a UMIK microphone, which has a USB output that plugs directly into a computer USB port. This probably is all the OP needs for this project.

When making initial driver measurements for my latest speaker project to use for modelling in VituixCAD, I used ARTA, an ARTA box, an XLR microphone and an audio interface that provides power to the XLR microphone. This allowed me to compensate for amplifier non-linearities. In that regard, for convenience, I used an old small class D amplifier for my initial measurements, which probably is not very linear. So, I wanted amplifier compensation to ensure accurate results. If making measurements with the amplifier that will be used with the speakers, though, a UMIK and REW probably are perfectly fine.
 
For some reason, I keep thinking that the existing drivers, once I measure them, will measure crap and I will want to use better drivers. Or is not that likely?

At least in my case, fortunately the results were not always as you are thinking about!
You would please read the"pre-history" background of my project in this post.
 
For some reason, I keep thinking that the existing drivers, once I measure them, will measure crap and I will want to use better drivers. Or is not that likely?
With DSP you will be able to compensate for frequency response non-linearities and use time delays for phase correction. Other issues, though, not much you can do.

Will I be able to measure distortion (not just frequency response) though? Assuming I just have REW.
Yes. REW has a "Distortion" tab that displays the measured harmonic distortion.

On my speakers the THD was a little high for my taste. I added glue fillets around the inside of the enclosures (one of my cabinets was not glued together very well), added an additional internal brace, and added additional damping material (I used fiberglass insulation). This improved the THD.

If adding additional damping material, keep the areas around the back of the woofers and midranges, and the ends of the ports inside the cabinets, clear of any damping material.
 
Some comments.

1. Don't get a USB mic. If you must get a USB mic, get a UMIK-2. The reason: unreliable timing information due to different clocks between the ADC on the mic and DAC. John Mulcahy (author of REW) tested a few mics here. In addition, the Audiolense user forum is full of complaints from users who use USB mics. A 48V Phantom power mic with an interface is much better. Someone suggested a Focusrite 2i2, but I would suggest that you get an interface with as many DAC channels as drivers that you need to control. In your case, 6. Something like an RME Fireface UC/UCX, Motu Ultralite Mk.5, Merging Anubis, or its Neumann clone (I forget what it's called).
Ooooohhhh, it is getting complicated. I would be OK to go with Umik 2 (does it resolve the above issues?), but getting lots of boxes I would like to avoid for now. If I ever move to a place where I can have a workshop, I will start a full-scale speaker and sub design and build project and then get full range of hardware. For now, would Umik 1 (or 2) be enough as a minimalist but fully adequate set?

2. Choice of DSP. You can choose either hardware (MiniDSP, DSPNexus, etc) or software. You are limited to software that can actually generate crossovers, which means Acourate, Audiolense, REW, DRC-FIR, Eclipse Audio FIR Designer, etc. PC based DSP is vastly more powerful, both in terms of processing power and the fine corrections that can be done. Hardware DSP is more robust and more convenient.
Hypes amps have DSP built in and I assume will have all functionality I need. At least for now I am not aware of anything missing there - they can do various crossovers, EQ filtering, power protection and maybe other things.

3. Choice of XO points and slopes. Measure your drivers and look for distortion and anomalies in the output. Decide if you want to deal with it, and how. For e.g. cabinet resonances can be dealt with through physical means, which is the proper way. The lazy way is to notch them out with DSP or remove them with a crossover so you don't have to worry about them :)
My current level of understanding of how I will chose and set crossover points and EQ given the existing drivers:
- get access to inside of the speaker and disconnect drivers
- connect drivers one at a time to an amp output and feed the amp with REW signal
- look at FR curve and choose crossover frequency so the two drivers' performance overlaps before the FR curve falls off. Also note driver SPL levels.
- program the crossover points and slopes (e.g. 24dB or 48dB per octave). Set channel level to match all 3 drivers between themselves
- connect all 3 drivers to amp, then run REW, look at FR and hope it is smooth
- start obsessive tweaking without understanding what to look for and what to do
 
Last edited:
Yes for your starting-up, I assume.
Please take intensive care, however, to implement proper safety measures (protection capacitors) for your midrange drivers and tweeters (ref. here).
And also take best care of your ignition/startup sequences and shutdown sequences (ref. here).
I'll have to read on it. I understand Hypex amp has power limiters (so I do not fry drivers with too high level), and the filters once programmed will mean the amp will not send e.g LF to HF. So, I still need capacitors for protection? Why? And will they not screw the sound?
 
Last edited:
Don't worry too much from the beginning!
The Hypex software gives everything you need: volume, filters (Highpass if recommended before measuring tweeter, as mentioned above), limiters and so on.
Just dig through the Hypex Filter Design's manual.
 
Then you need the mic to measure FR of every driver solely to then calculate the xover, for what you need some basic technological knowledge.
What do you mean by 'calculate'? is there an exact calc or do I just go approx by looking at FR curves and picking a point where both drivers still do not roll off much?
 
Don't worry too much from the beginning!
The Hypex software gives everything you need: volume, filters (Highpass if recommended before measuring tweeter, as mentioned above), limiters and so on.
Just dig through the Hypex Filter Design's manual.
OK.
 
What do you mean by 'calculate'? is there an exact calc or do I just go approx by looking at FR curves and picking a point where both drivers still do not roll off much?
Second one.
Coming from analogue design where trial and error is the main approach and simulation can provide just a frame, it now is a relief with DSP. Instead of hours of switching every part and remeasure again you now switch frequency and mode within seconds and see results instantly. It has a high educational aspect and with overlays every step can be compared to other.
How to find the correct crossover frequency?
First approach could be measuring the speaker as is (every single driver) so you can estimate what the constructors idea was, and then try to imitate this after extraction of the passiv xover.

Second approach could be measuring single drivers directly without xover and decide with given FR and distortions and radiation pattern where to cross.

For the crossover most use LR 24 dB, so that would be a good beginning.

Concerning the electrical part in the project I see no problems ahead.
Problems may pop up if one or more of the drivers are bit worn out. But even if so, when you should arrive at this point you have learned a lot and had a lot of fun.
 
do I just go approx by looking at FR curves and picking a point where both drivers still do not roll off much?
That is a good start.

Also, you may want to measure the off-axis frequency response, e.g., at 15 degrees, 30 degrees, 45 degrees and at 60 degrees. If possible, it is good to have about the same dispersion from the tweeter and the midrange at the crossover frequency. However, given the tweeter is flush mounted and not in a waveguide, it may not be optimal. Nonetheless, you can play with the crossover to try to get a good balance.
 
Back
Top Bottom