ok long time member under old name and lost logins and all my data in flood so back on here and starting back in finishing several many yrs old on hold projects
1) Revel F36 blew my speaker budget for the year ( $1200/ pair, lol) and yes I have read , reread, all the reveiews , data , spinorama on all the revel/harman speakers I could find.
the only thaing that bugs me abotu the F 36 is that 4 k broadband hump, esp when compared to the F 35 and others that hit the target curve a bit better,
I have a very large room and high ceilings and sit about 13' away but that hump is still noticible, not enought to make me not love the seakers,But..
There has to be a easy way to DIY those factory crossovers to get rid of that hump. I know Revel had to cut costs somewhere.. I have seens some pics in a vid of the x.o
and it doesn't look like too many components and maybe even bi-polar caps>??
So anybody up for a project? I have Holm RTA REW a nice focusrite DA , calibrated mikes old 50mhz scope ,etc.
I'm thinking, just pull the board, put a R that matches the DCR of the tweet and measure it then work to get a shelf filter in and upgrade the parts while I am at it?
Few hundred$ would be worththe investment to really amke these beauties sing?
1) Revel F36 blew my speaker budget for the year ( $1200/ pair, lol) and yes I have read , reread, all the reveiews , data , spinorama on all the revel/harman speakers I could find.
the only thaing that bugs me abotu the F 36 is that 4 k broadband hump, esp when compared to the F 35 and others that hit the target curve a bit better,
I have a very large room and high ceilings and sit about 13' away but that hump is still noticible, not enought to make me not love the seakers,But..
There has to be a easy way to DIY those factory crossovers to get rid of that hump. I know Revel had to cut costs somewhere.. I have seens some pics in a vid of the x.o
and it doesn't look like too many components and maybe even bi-polar caps>??
So anybody up for a project? I have Holm RTA REW a nice focusrite DA , calibrated mikes old 50mhz scope ,etc.
I'm thinking, just pull the board, put a R that matches the DCR of the tweet and measure it then work to get a shelf filter in and upgrade the parts while I am at it?
Few hundred$ would be worththe investment to really amke these beauties sing?
You might want to check out da Spinoramas to see if there is an axis or listening height where this bump goes away and the response is flatter. You can make this your usual listening axis by tilting the speaker back (if the good axis is below) or toeing in the speakers a lot if it's better on some horizontal axis. Doing the later will also improve the stereo image.
Revel 2 ways are likely to have an acoustic AS3 crossover alignment and the best axis is NOT on treble axis bla bla
Revel 2 ways are likely to have an acoustic AS3 crossover alignment and the best axis is NOT on treble axis bla bla
That would give you rubbish. The xover needs the unit(s) and box to behave anywhere near what it does in real life. (ie in the speaker) Your bump is almost certainly an artifact of the interaction between the box, bass unit, treble unit AND the xover.
Last edited:
yes and modding the crossover cand fix all those interactions which a good crossover design takes into account... but it might add 4-5 componenets which cost harman $$$ on their "lower budget" seaker that,,, i can do it with dsp but id rather not add that to signal chain.
also no to the above positioning hassels (WAF) AND I have arady been thur that .. if you look the bump causes the "curve" to have a slower tilt...
check the f35 vs f36 or f 206 and 208 for even the 228 266 version and you will see what I mean.
check the f35 vs f36 or f 206 and 208 for even the 228 266 version and you will see what I mean.
Why don't you try my suggestions which will cost you nothing and see if this gives you the sound you want.
It's unlikely Revel will release a speaker with a bad sound 4kHz bump. You might just not be listening to it as they intended.
ALL xovers, including acoustic LR & AS xovers, have an optimum axis and it's often NOT on treble axis bla bla
I'm sorry your WAF problems won't allow you to position your speakers properly but this means, a new xover or fancy DSP might only help the yuckiness but won't cure it.
You can try your DSP and see if it improves the sound If not, a new xover is unlikely to work either.
It's unlikely Revel will release a speaker with a bad sound 4kHz bump. You might just not be listening to it as they intended.
ALL xovers, including acoustic LR & AS xovers, have an optimum axis and it's often NOT on treble axis bla bla
I'm sorry your WAF problems won't allow you to position your speakers properly but this means, a new xover or fancy DSP might only help the yuckiness but won't cure it.
You can try your DSP and see if it improves the sound If not, a new xover is unlikely to work either.
Last edited:
I would add DSP and equalize it. With regard to the passive crossovers, at most I would replace any laminated steel core inductors with heavy gauge air core inductors with the same inductance values. Together that should make a noticable improvement.
Here is why: If you just add a tuned resonant circuit to the tweeter filter to reduce the 4k bump, it likely will change the group delay of the tweeter, which in turn can mess with the phase alignment of the tweeter and the midrange. This could create a dip at the crossover frequency. To do the crossover properly, you may end up needing to change more than merely adding the tuned resonant circuit. To get it right, plan on taking a lot of frequency response measurements of the individual drivers, as well as impedance measurements, and spend a lot of time modeling the system (e.g., in VituixCAD).
With regard to adding DSP, there are a lot great options, some very inexpensive and some more costly. I use a Raspberry Pi running CamillaDSP (free) in one of my systems. MiniDSP also has a lot of DSP options for various budgets. I use one in my other system.
I should rephrase that. What you are asking for is a complete re-design of the Revel F36 xover. The success of that is totally dependent on the skill & experience of whoever does it and the gear you have to do the required measurements.Why don't you try my suggestions which will cost you nothing and see if this gives you the sound you want.
....
You can try your DSP and see if it improves the sound If not, a new xover is unlikely to work either.
I'll withdraw my comments and just add $0.02 Modify only ONE of your speakers and keep the other original so you can compare the sound with what you dream up. Making it measure 'better' doesn't always make it 'sound' better.
no digital add on,,, no room no desire,,,, I was one fo the first with the didden and other dc2496 mods ,,, it still sounds better thans any of the dsp been there done that,,,
do not want nor have a place for any and one is this system, analog only hence the desire to tweak the crossover.. for the spinorama link been there seen it no compare it to the F 35 and other highter end revel speakers. I still use the dcx for testing purposes and have already tried it and eq'd out the hump and it does sound a touch more like the higher end Revels but it nor any digital is goin in my system, as far as Totally redesigning the crossover.. not.. just the tweeter section has to be touched, unless it s series crossover... but i'll fin out soon enuff
do not want nor have a place for any and one is this system, analog only hence the desire to tweak the crossover.. for the spinorama link been there seen it no compare it to the F 35 and other highter end revel speakers. I still use the dcx for testing purposes and have already tried it and eq'd out the hump and it does sound a touch more like the higher end Revels but it nor any digital is goin in my system, as far as Totally redesigning the crossover.. not.. just the tweeter section has to be touched, unless it s series crossover... but i'll fin out soon enuff
Attachments
Last edited:
It's not difficult to shape the response to your liking. Draw the complete x/o filter of the tweeter unit.
Prepare to measure tweeter only at desired angle/angles and distance.
Load frd and zma of a raw tweeter to a simulator and introduce passive parts you found in original x/o.
Some sort of rlc network should do the job.
Prepare to measure tweeter only at desired angle/angles and distance.
Load frd and zma of a raw tweeter to a simulator and introduce passive parts you found in original x/o.
Some sort of rlc network should do the job.
yea sim suggestions? my best freind , who passed aways 2 yrs ago, had leap and all the goodies , so i beenn out of the loops for a while.
getting the data is not an issue. was going to measure the tweeter with crossover installed and tweak from there but might be better to do raw tweeter measurments? thanks
getting the data is not an issue. was going to measure the tweeter with crossover installed and tweak from there but might be better to do raw tweeter measurments? thanks
VituixCAD.
Take frequency response measurements and impedance measurements of both the midrange and tweeter, and import those into VituixCAD. Use an IR window that closes before the first reflection (probably the floor reflection) to keep room reflections from interfering with the measurements. The more off axis angles you measure, the more accurate will be the prediction of the estimated in-room response, if that is important to you. Assuming you only are tweaking the tweeter's filter and do not change anything in the Woofer's or midrange's filters, that should be sufficient.
Take frequency response measurements and impedance measurements of both the midrange and tweeter, and import those into VituixCAD. Use an IR window that closes before the first reflection (probably the floor reflection) to keep room reflections from interfering with the measurements. The more off axis angles you measure, the more accurate will be the prediction of the estimated in-room response, if that is important to you. Assuming you only are tweaking the tweeter's filter and do not change anything in the Woofer's or midrange's filters, that should be sufficient.
The reason I recommend also measuring the midrange is so that you will see how changes to the tweeter's filter affects its integration with the midrange. If your tweeter filter design changes the tweeter's group delay, it could worsen the phase alignment between the midrange and tweeter, which could cause a dip in the frequency response at the crossover frequency. By having the midrange in the model, you will be able to see how the tweeter's filter affects the crossover frequencies.
I would try a DSP or analog EQ first. Having an option to dial in adjustments freely and precisely - and instantly bypass it - will show well what kind of FR will be most pleasant. With a crossover mod, you won't have the option of instant A/B testing.
By apologies. I thought 'spinorama' meant a collection of off-axis curves in all directions rather than a single room response estimate. This is (possibly) a crude 'measure' of my Room Interface Profile
AbsoluteListening Tests-Further Progress but there is at least a PhD's worth of research into what the RIP actually is.
Then you are only likely to achieve somewhat less improvement than the touch you found .. if you are good at xover juju.
If you believe that, you are probably not good at xover juju ... but I might be wrong 😲
That you will 😊
But I didn't ask the obvious question which I should have done at first.
What don't you like about the standard Revel F36 in your listening environment?
Last edited:
Not having the midrange and its xover in the model is essential to the final result. Its output affects stuff well above the 'xover frequency'
This could get interesting. The first thing I'd look at is the amount of tweeter padding. If there is none, it may be difficult. Also check the speaker impedance above 1k. Adding a notch may lower it some.
This should be close. It shaved 2dB off the response at 4k in my sim.
After looking again I'd probably shift the notch slightly higher in frequency. Changing the cap to a smaller values shifts it higher.
After looking again I'd probably shift the notch slightly higher in frequency. Changing the cap to a smaller values shifts it higher.
Attachments
Last edited:
temp 25 thanks you very much .. on the right track with my thought process and for the naysayers yes the mid might effect it some but its x/o is one ovtave below where we need the correction so,,,,, hopefully just the tweeter and that's where we start... TEMP 25 thanks! be a few weeks before I can tear into it.. got to finish house painting first i'll post pics valuse and measurements soon as I can get them
Terry yes and thanks for tip for VituixCAD i'll look it up. i'll do enuff tweet measurements to see if anything get wonky..Like i said infortunately my filter guru passed away a few years ago.. so Im having to reboot my memory banks