-
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!
converting revel f208 to active crossover with 6 amps
- Thread starter NJ coffee roaster
- Start date
I've converted several speakers from passive to muti-channel active, one amp channel per driver section. The better the passive design to start, the less the difference from conversion...which only makes sense. The F208, being a nice speaker, isn't going to leave a lot of room for uptick in terms of frequency response improvement imo. There's room no doubt, as DSP simply allows for finer EQ and crossover work, but audibility might not be as apparent as better measurements.
Where I feel rather certain you will be able to hear improvement is with SPL and dynamics. I've never encountered a speaker with more than 2-ways, that didn't improve when taken fully active multi-way. I don't care what folks say here if they disagree ... ..I've done too much of this.
The best way I've found so far to convert passive to active, is to measure each driver section's transfer function raw; and to also measure each driver sections' electrical transfer function. It's good to see what the manufacturer did electrically with the passive crossover, and first match that with DSP. But then I always end up doing a multi-way tuning from scratch too, based on the set of raw measurements...just to compare.
At any rate, I've always enjoyed making such conversions. It's been a risk free way to learn for me, other than a relatively low investment in amps and DSP.
There's way too much it depends going on, to say brick walls can't be used.
Like any complementary linear-phase crossover, they have potential pre-ring when the two sides summing together aren't acoustically complementary.
Which means, only when lobing is occurring. So the first question is what is the lobing potential.
If it's severe due to drivers' physical separation, I've measured cases where a brick wall helps reduce lobing so much, it's more than worth any potential pre-ring audibility. (Potential needs to be stressed.)
It also depends on how much usable driver overlap there is, section to section. If there's ample overlaps, brick walls or other high order lin-phase crossovers simply work great.
Last edited:
By distortion, do you mean harmonic? The best high SPL measurements I've made have been with steep filters.
BUT, my designs have more usable frequency range overlap between driver sections, that most designs I see. Due to higher multiway counts (more driver sections).
This lets me use each driver section in the meat of its passband (and only their due to steep). I figure that can't help but reduce distortion.
Probably no improvement, very possible a degradation unless you have a plan and ability to measure. All of the following assumes you have good measurements of your speakers.
The first step would probably be to determine the passive filters to see what Revel did. This is useful since they clearly know how to design a good speaker. From the component values and diagramming the circuit, recreate the filters in a simulator, VituixCAD for example. As @Ageve pointed out the speaker uses 24dB/octave acoustic crossover. This will be the sum of the electrical response of the crossover, plus the acoustic response of the drivers. Revel undoubtedly took into account the acoustic rolloff of each drivers along with the electrical response of the passive crossover. It would also be nice to see what Revel did for the low-frequency compensation filter.
Next you need to the measure electrical and acoustical characteristics of each driver. REW has a tutorial on how to measure T/S parameters and impedance. Or use a DATS. And measurements of the individual drivers as Ageve provided, with and without the passive crossover. Of course being very careful with your tweeter measurements without the filter. The results can be loaded into VituixCAD along with the electrical measurements you made of the drivers.
It should be easy at this point to recreate the same filters as Revel did. They likely have some notches and/or impedance flattening circuitry. You can judge if they had to make compromises with the passive implementation and maybe take advantage of the DSP filters to make small improvements. Some of the small details are the result of the cabinet and lenses, maybe you can find small gains in playing around with the tweeter and midrange filters to smooth the directivity error beyond what Revel accomplished. You can evaluate this with a 360 degree near-field scan as described in VituixCAD's documentation.
Along the way, avoiding blowing your tweeters is something to keep in mind. Using a DC blocking capacitor is probably wise, even using a capacitor as part of the final tweeter filter. In some cases DSP notches don't perform as well as passive notches for controlling driver resonances. May not be applicable to your speakers, but of course measurements would tell that.
It's an interesting project. I doubt if you end up with a magically better speaker, perhaps small gains and the ability to tune some of them to your room better. There are some drawbacks to directly connecting amps to drivers, I compared that in this post. The active version will be sensitive to the amps' noise floor since the passive crossover does a good job of reducing hiss. I typically combine the active tweeter filter with a 6 dB/octave passive to deal with this, which allows me to use a Niles 12-channel amp without having to worry about getting the lowest noise amp. Which is a bit ironic, a pure active implementation becomes sensitive to the amp, where the passive implementation makes the amp inaudible.
Last edited:
Digital_Thor
Addicted to Fun and Learning
I made a copy of the KEF R900, the last affordable good KEF speaker with bigger woofers - IMO, but used the 2018 model from the R3 - did it fully active with LR24dB -but at 2200Hz rather than 2900... Because... I saw that it worked in newer models, proven by all the data from this site , among others.
Being able to tailor bass and the coax, definitely made a difference IME.
Woofers are Satori WO24P, so that there's actually some 'weight' to voices and instruments + but still very detailed and perfectly smooth and also dynamically 'fun'.
I bought the KEF R3 Meta to compare directly... And I only noticed the real difference in bass 'weight'... Volume if you will... The R3 sounds 'smaller'.. which is pretty logical IMO.
I measured everything, and the difference is surprisingly small, except the bass performance.
Just look at the difference between the Meta vs non-Meta ..
So my point, is, that you have to have a specific goal and both the gear and abilities, to pull off this trick. Both Revel and KEF are very good... but, consider your goal and means.
Attachments
terryforsythe
Major Contributor
REW, VituixCAD and a UMIK.
OP is getting a Flex HTx, which is what I use. It is great for this type of project. It includes Dirac Live.
Looking at the chart in post #16, I tend to agree. But, the final arbitrators should be a VituixCAD model and the measurements.
I also turned my 3-way passive speakers into all active using a Flex HTx. By moving my woofer/midrange crossover point up to 350 or 400Hz from around 180-200Hz, I was able to reduce THD quite a bit - the midrange was being pushed a little hard reaching down below 350Hz. But, my speakers only have a single woofer, whereas each F208 has two.
It depends on how much effort you put into it and the shortcomings you percieve on your speakers.
I was unsatisfied with the bass response of my Elac UBR62s, even when using a subwoofer. I added a miniDSP SHD and messed around with it for about a year, tuning, listening, playing with Dirac Live, more tuning and listening, etc. I just could not get a tight bass response out of my woofers. For aethetic reasons and wife approval factor, changing the speakers was not really an option.
So, I want all active all active using a Flex HTx. Almost immediately I noticed the bass issue was gone. The main culprit probably was the high DC resistance of the two inductors in series with the woofer. The woofer is 4.15 ohms, but one inductor was 1.2 ohms and the other was 0.4 ohms, for a total of 1.6 ohms DCR. That tanked the damping factor. To me it was noticable enough to become, over time, irritating. My wife, on the other hand, only listens to background music and never noticed the issue.
In your case, considering the price of the speakers, I would hope the DCR of the woofer's passive filters is not anywhere near as high as mine. It might be worth measuring yours to check. The lower the DCR, the less negative impact they have on damping factor, and thus the less improvement over stock you will experience.
The other benefit will be tackling the dip at 2kHz pointed out by staticV3. That is a slam dunk with the Flex HTx given its options for driver time delay adjustment, crossover frequency selection, crossover topology, crossover slope and PEQ.
Word of caution, be very attentive to making sure your connections are correct, and when tuning start off with the volume low, bringing it up slowly. If you accidentally leave the high pass crossover turned off on your tweater and then hit it with high power, it is very easy to fry a tweater. I have seen some people recommending putting a capacitor in series with the tweater as a safety net. I did not do that, but it may not be a bad idea, though it may complicate the tweater tuning a bit.
Luvchampagne
Active Member
Dirac Live is a $249 add on option for HTX.
terryforsythe
Major Contributor
In an all active setup, choosing very low noise amplifiers has a benefit, though I don't know how audible it is at the listening position.
With my KEF LS60 speakers, which are active, I can hear a small bit of hiss if I put my ear inside the midrange and right next to the tweeter, but I cannot hear it at distances of 0.5m or greater.
I converted my Elacs, which also use concentric drivers, to be all active without any passive filter components whatsoever. I cannot hear any hiss at all, even when putting my ear inside the midrange and right next to the tweeter. I am using a Topping LA90D amplifier for the midrange and a Topping B100 amplifier for the tweeter. I even have them set for high gain, which provides their highest noise floor. Still, the noise floor is too low to be audible with my midranges and tweeters.
I have a Hypex Nilai, set in high gain, powering the woofers. I don't hear any noise coming out when putting my ear right next to the woofer. Recently, in response to someone asking about the Nilai's noise level, I connected the Nilai to the tweeter. Then I could hear just a small amount of noise when I put my ear right next to the tweeter, but it was lower than the noise level of my KEFs.
For reference, when set to high gain, their respective THD+N ratios as tested by Amir are as follows:
Topping B100: 117 dB
Topping LA90D: 117 dB
Hypex Nilai: 104 dB
Of course, their THD+N is even lower (higher THD+N ratio) when set to low gain.
Last edited:
Hello again OP @NJ coffee roaster and other multi-amp colleagues,
On this thread, even though I repeatedly encourage OP to go into his mentioned DSP-based multichannel multi-amplifier exploration, my positive/preferable suggestions are based on my own experiences on my audio setup with my own SP drivers (3-way in sealed cabinet plus subwoofers and super-tweeters) (ref. #931 and #1,006 on my project thread) which maybe not always relevant/valid to OP's specific excellent SP system i.e. Revel F208.
Nevertheless, I still would like to suggest OP to at least try/evaluate to see/find merits/pros (or not?) of DSP-based multichannel multi-amplifier configuration with Revel F208 as some (many?) of ASR friends/colleagues kindly pointed yesterday and today.
Furthermore, prior to starting your exploration, I highly recommend OP implementing and being careful enough about the two points below;
1. At least during your "tests" (and hopefully thereafter too, as I am still doing), place HiFi-grade protecting series capacitors (I recommend non-polar film caps) for your treasure midrange and tweeter. I always have 68 microF cap for midrange (covering 500 Hz - 6 kHz), 10 microF cap for tweeter (covering 6 kHz - 14 kHz), and 10 microF cap for super-tweeter (covering ca. 8 kHz to 22 kHz). I very much carefully measured/confirmed that these protection caps are essentially transparent in these Fq coverage zones (ref. #403, and again recently ref. #1,009).
2. You should not change/control the "DSP gain values" for each SP driver on-the-fly (while listening to music or test signals) especially by numeric-value key-board inputs. After you would change the gain value, you should double check it before actual sound measurements/listening. For example, in case if you would mis-input "+55 dB" instead of your intending "+5.5 dB", you have high possibility of damage/destroy your treasure SP driver(s). If you would have pre-amplifier functionality in analog domain, start measurements/listening at minimal gain, say at -40 dB, and carefully/gradually gain-up to your preferred and/or target gains.
Just for your reference, my DSP software "EKIO" has capabilities of gain control by "numeric key-board input" and "mouse-wheel-rotation up-and-down of 0.1 dB granularity"; I always use, therefore, "mouse-wheel-rotation up-and-down" gain control in DSP "EKIO" which is much safer, as you can easily understand. "EKIO" also has "mouse-wheel-rotation up-and-down" controls for all of the DSP "value change" features.
On this thread, even though I repeatedly encourage OP to go into his mentioned DSP-based multichannel multi-amplifier exploration, my positive/preferable suggestions are based on my own experiences on my audio setup with my own SP drivers (3-way in sealed cabinet plus subwoofers and super-tweeters) (ref. #931 and #1,006 on my project thread) which maybe not always relevant/valid to OP's specific excellent SP system i.e. Revel F208.
Nevertheless, I still would like to suggest OP to at least try/evaluate to see/find merits/pros (or not?) of DSP-based multichannel multi-amplifier configuration with Revel F208 as some (many?) of ASR friends/colleagues kindly pointed yesterday and today.
Furthermore, prior to starting your exploration, I highly recommend OP implementing and being careful enough about the two points below;
1. At least during your "tests" (and hopefully thereafter too, as I am still doing), place HiFi-grade protecting series capacitors (I recommend non-polar film caps) for your treasure midrange and tweeter. I always have 68 microF cap for midrange (covering 500 Hz - 6 kHz), 10 microF cap for tweeter (covering 6 kHz - 14 kHz), and 10 microF cap for super-tweeter (covering ca. 8 kHz to 22 kHz). I very much carefully measured/confirmed that these protection caps are essentially transparent in these Fq coverage zones (ref. #403, and again recently ref. #1,009).
2. You should not change/control the "DSP gain values" for each SP driver on-the-fly (while listening to music or test signals) especially by numeric-value key-board inputs. After you would change the gain value, you should double check it before actual sound measurements/listening. For example, in case if you would mis-input "+55 dB" instead of your intending "+5.5 dB", you have high possibility of damage/destroy your treasure SP driver(s). If you would have pre-amplifier functionality in analog domain, start measurements/listening at minimal gain, say at -40 dB, and carefully/gradually gain-up to your preferred and/or target gains.
Just for your reference, my DSP software "EKIO" has capabilities of gain control by "numeric key-board input" and "mouse-wheel-rotation up-and-down of 0.1 dB granularity"; I always use, therefore, "mouse-wheel-rotation up-and-down" gain control in DSP "EKIO" which is much safer, as you can easily understand. "EKIO" also has "mouse-wheel-rotation up-and-down" controls for all of the DSP "value change" features.
Last edited:
Yes, I essentially agree with you.
I too found it should be very important and highly worthwhile to objectively measure/confirm the Fq-SPL at various "stages" in digital domain, in analog domain, and of course in actual room sound at your/our listening position.
- Where in my multichannel multi-driver (multi-way) multi-amplifier stereo system should I measure/check frequency (Fq) Responses? #393
- Frequency (Fq) responses in the completed system measured by using “cumulative white noise averaging method” under the present standard crossover configurations and relative gains
Part-1_Fq Responses in EKIO’s digital output level: #394
Part-2_Fq Responses in DAC8PRO’s analog output level: #396
Part-3_Fq Responses in amplifiers’ SP output level before protection capacitors: #401
Part-4_Fq Responses in amplifiers’ SP output level after protection capacitors: #402
Part-5_Fq Responses in actual SP room sound at listening position using one measurement microphone: #403
Part-6_Summary, discussions, and a little step forward: #404, #405-#409
- Frequency response in the new (latest) configuration with four amplifiers plus active sub-woofers: #410
- Frequency response measured under subjectively best fine-tuned gain/volume settings with the four amplifiers plus active sub-woofers: #411
- The latest Fq-SPL (re-confirmation) of multiple amplifiers SP high-level output signals and that of room air sound at listening position: all measured by “FFT averaging of recorded cumulative DSP-processed flat white noise” (as of June 8, 2025): #1,009
- Frequency (Fq) responses in the completed system measured by using “cumulative white noise averaging method” under the present standard crossover configurations and relative gains
Part-1_Fq Responses in EKIO’s digital output level: #394
Part-2_Fq Responses in DAC8PRO’s analog output level: #396
Part-3_Fq Responses in amplifiers’ SP output level before protection capacitors: #401
Part-4_Fq Responses in amplifiers’ SP output level after protection capacitors: #402
Part-5_Fq Responses in actual SP room sound at listening position using one measurement microphone: #403
Part-6_Summary, discussions, and a little step forward: #404, #405-#409
- Frequency response in the new (latest) configuration with four amplifiers plus active sub-woofers: #410
- Frequency response measured under subjectively best fine-tuned gain/volume settings with the four amplifiers plus active sub-woofers: #411
- The latest Fq-SPL (re-confirmation) of multiple amplifiers SP high-level output signals and that of room air sound at listening position: all measured by “FFT averaging of recorded cumulative DSP-processed flat white noise” (as of June 8, 2025): #1,009
As long as you’re protective of the original crossover ( disconnect, leave it be) and you’re up for the adventure why not? There are benefits in having an amp connected directly to a driver with nothing but speaker wire in between. FYI I connect my tweeters directly to amp with no capacitor in between. Safe as long as you refrain from stupid actions! Most that poop on your idea have never done this. Good luck and enjoy the ride!
Safe if the amp have adequate DC protection for the task.
If the amp is DC coupled, then there are always some DC at its output, that 'pre-tensiones' a direct coupled tweeter (and/or the midrange) enough which typically leads to more compression (and distortion) compared to series capacitor connection for the drivers.
But with series capacitors, you get the parasitic character of the capacitor to the sound, so one needs to use the type of capacitor which don't bothers the listener sound wise.
A good example of capacitor coupled active crossed speaker is the JBL M2. The passive filter for its high frequency driver protects it from DC and reduces the noise floor (because the passive filter is not just a simple protection capacitor here).
