Thylantyr - Sounds like you've done this before!
I will definately use almost all your ideas on
the final implementation. I like the clipping
indicator LED. Please send link if you have one.
Cocolino - I agree with your implementation.
Raven is -5 dB, fixed voltage divider.
Scan speak will be adjustable from +2.5 dB to
+7.5 dB
Dave/Steve - Yes, I think i need a baffle step
block. Thanks for the link/info.
Much reading/work to do, but AFTER breakfast!
-herm
I will definately use almost all your ideas on
the final implementation. I like the clipping
indicator LED. Please send link if you have one.
Cocolino - I agree with your implementation.
Raven is -5 dB, fixed voltage divider.
Scan speak will be adjustable from +2.5 dB to
+7.5 dB
Dave/Steve - Yes, I think i need a baffle step
block. Thanks for the link/info.
Much reading/work to do, but AFTER breakfast!
-herm
That brings up another question:
How high can I cross? If you look at that freq response in the
previous post, it looks like the scan speak is good up to 5 kHz
on-axis. But off axis, it starts dropping off pretty strongly after
3 kHz.
How high can I go before that off axis drop comes in to play?
I realize there is no straight answer, but what do you guys
think? Could I get away with 4 kHz? How about 5 kHz?
I have suggested 24dB slopes only because I am crossing
so close to the limit of the Raven. 4-5 kHz will allow me to
use a flatter slope.
I'm open for opinions, suggestions, or facts.
How high can I cross? If you look at that freq response in the
previous post, it looks like the scan speak is good up to 5 kHz
on-axis. But off axis, it starts dropping off pretty strongly after
3 kHz.
How high can I go before that off axis drop comes in to play?
I realize there is no straight answer, but what do you guys
think? Could I get away with 4 kHz? How about 5 kHz?
I have suggested 24dB slopes only because I am crossing
so close to the limit of the Raven. 4-5 kHz will allow me to
use a flatter slope.
I'm open for opinions, suggestions, or facts.
Wait until You`re in the phase of actually doing the alignments to know what`s real brain spinning
Beginning with Your last sentence:
Yes that`s what I would do in this case as well.
As I said, depending of the x-over point and the final lowpass slope the low-pass shelving might not (likely is not) be necessary.
Even in the case of a hump, I do not believe a notch filter is very suitable here.
You`d better (and more simply) tame that modifying the Q of the LR-filter lowpass circuit.
power amplifier clipping indicator;
http://sound.westhost.com/project23.htm
but, for pre-amp only, you can strip out some of the circuits... or design something simple with a LM339
comparator.
pre-amp projects.
http://sound.westhost.com/projects-2.htm
linkwitz transform
http://sound.westhost.com/project71.htm
*************************
To confuse you even more
1. Balanced input stage (differential).
2. Subsonic filter
3. Linkwitz Transform
4. Selectable crossover frequency
knob, or crossover frequency
plug in modules.
Item #2 and #3 can be "switch in/out" of the circuit if you need it
or not, user selectable.
http://sound.westhost.com/project23.htm
but, for pre-amp only, you can strip out some of the circuits... or design something simple with a LM339
comparator.
pre-amp projects.
http://sound.westhost.com/projects-2.htm
linkwitz transform
http://sound.westhost.com/project71.htm
*************************
To confuse you even more 1. Balanced input stage (differential).
2. Subsonic filter
3. Linkwitz Transform
4. Selectable crossover frequency
knob, or crossover frequency
plug in modules.
Item #2 and #3 can be "switch in/out" of the circuit if you need it
or not, user selectable.
herm said:
An electronic crossover with a
knob to select from pre-determined
crossover frequencies does wonders.
built at least a crossover with
a 4 position selector, then build
a module so you can plug in
different frequencies, ie
generic idea
high pass -> 4 pos switch
2khz, 3khz, 4khz, 5khz
replace pluggable filter module;
1.5khz, 2khz, 2.5khz, 3khz
Worry about final system tweaking
when you are ready to listen to music.
herm said:
I would create a bypass switch for
"EQ" type of circuit, let the user
determine if they need it or not.
Here is a cosmetic concept for
EQ/crossover that is 15 years old.
http://dreadlordpk.50megs.com/cgi-bin/i/ElectronicStuff/ParametricEQConcept/EQ-X.gif
Master volume (optional)
Input Gain
EQ section, 3 band parametric or equivalent with bypass push button switch.
Active crossover section with
level controls (volume)
Full range output that bypasses
the active crossover just in case
you find a need to use it.
Clipping indicators on each output.
Input clipping indicator.
*****************************
All you really need to do is
build the basic blocks then "switch in"
the extra goodies as needed.
basic block;
input -> crossover --> output
need the subsonic filter?
flip switch
input --> subsonic --> crossover --> output
Need EQ ?
flip switch
input --> EQ --> crossover --> output
Need LT circuit, subsonic and EQ ?
input --> EQ --> subsonic --> LT --> crossover --> output
I'm not giving this much thought,
so order of blocks needs to be
analyzed.
Throwing generic ideas at you
Hi,
Be aware of the fact that the transformer of the R1 is a short at DC. This can give you big problems when directly coupling the R1 to an amplifier. Either the amp or the transformer can become smoking with even small amounts of DC, not to mention the large distortion it can cause. It is always wise to couple this kind of ribbons through a coupling capacitor (high quality PP). In your case Herm, you can split the 24 dB XO in a 1st order (the coupling cap) and a 3rd order.
With the Scanspeak’s I don’t have experience but they look excellent ones. Also have a look at the Audax HM130-Z12 These have a smooth roll off at the high end.
Long time ago I combined a Decca DK30 ribbon with a KEF B110 with an XO at 2.5 kHz. These matched very well. Although a good sounding speaker, the B110 is not up to nowadays standards anymore, so when I have time I will replace them for the Audax HM130-Z12’s
Be aware of the fact that the transformer of the R1 is a short at DC. This can give you big problems when directly coupling the R1 to an amplifier. Either the amp or the transformer can become smoking with even small amounts of DC, not to mention the large distortion it can cause. It is always wise to couple this kind of ribbons through a coupling capacitor (high quality PP). In your case Herm, you can split the 24 dB XO in a 1st order (the coupling cap) and a 3rd order.
With the Scanspeak’s I don’t have experience but they look excellent ones. Also have a look at the Audax HM130-Z12 These have a smooth roll off at the high end.
Long time ago I combined a Decca DK30 ribbon with a KEF B110 with an XO at 2.5 kHz. These matched very well. Although a good sounding speaker, the B110 is not up to nowadays standards anymore, so when I have time I will replace them for the Audax HM130-Z12’s
Well, I think it`s hard to predict wether higher crossing/flatter slope or lower crossing/steeper slope would be the best solution.
There are just too many variables and finally You have to find the best compromise by trying in reality.
I fear without breadbording the thing and actually trying different things You won`t find the right answer.
3kHz certainly is the lowest and 5kHz the highest reasonable x-over point. Why not to begin in the middle.
The idea of thylantyr switching in and out blocks is very good and for doing comparisons I would also suggest eventually to switch different parts values inbetween the blocks (concerning switching: take care and watch the x-over`s outputs first without connecting the power amps & speakers - the switching might introduce transient spikes which Your amps & speakers might not like so much!!)
For the beginning I would breadboard the x-over according Your last block-diagram.
All the other goodies thylantyr suggested for the moment IMO are not necessary in order getting the most important parts of the thing started but if desired of course You can extend the circuit and break Your head on this accessories later.
I've been playing with Circuit Maker, and here's what I came
up with for the L-R crossover blocks. I will post more blocks
as I build them. This is built for -6dB at 3.5 kHz.
For my experimentation, I plan on using panasonic polypropolene
caps (Digikey P3333-ND), Yageo metal film resistors (Digikey
976XBK-ND), and NE5532 opamps like John has in his kit.
I will buy parts for several different crossover frequencies, and
experiment like you guys have suggested.
Just wanted to add that I use a Marantz preamp and Carver
AV-806 amps. I have 2 channels of Zen V4 and will build two
more eventually to power these speakers.
Edit: Hey Pjotr, do you think a 2 Hz highpass like Linkwitz uses
will take care of the DC? Or does it need to be done at the
speaker (to protect it from the amp as well)?
up with for the L-R crossover blocks. I will post more blocks
as I build them. This is built for -6dB at 3.5 kHz.
For my experimentation, I plan on using panasonic polypropolene
caps (Digikey P3333-ND), Yageo metal film resistors (Digikey
976XBK-ND), and NE5532 opamps like John has in his kit.
I will buy parts for several different crossover frequencies, and
experiment like you guys have suggested.
Just wanted to add that I use a Marantz preamp and Carver
AV-806 amps. I have 2 channels of Zen V4 and will build two
more eventually to power these speakers.
An externally hosted image should be here but it was not working when we last tested it.
Edit: Hey Pjotr, do you think a 2 Hz highpass like Linkwitz uses
will take care of the DC? Or does it need to be done at the
speaker (to protect it from the amp as well)?
interesting reading material.
http://www.fortunecity.com/marina/pontoon/3/audio/elliptic/elliptic.html
http://www.fortunecity.com/marina/pontoon/3/audio/elliptic/elliptic.html
Tweeter Delay circuit.
I have done some reading on the tweeter delay circuit.
Although I probably don't understand most of it, here
is a possible solution. Please let me know if I have
missed the facts ....
To determine the amount of delay necessary, I will
mount the drivers in the baffle. The tweeter should
be hooked up reversed polarity.
While measuring the frequency response, delay should
be varied until there is a MINIMUM seen at the crossover
frequency.
This should take into account the differences in baffle
mounting AND the effect of the crossover.
How did I do?
P.S. Thylantyr, I will get more info on the transformer & get back with you, I just don't have time today ....
I have done some reading on the tweeter delay circuit.
Although I probably don't understand most of it, here
is a possible solution. Please let me know if I have
missed the facts ....
To determine the amount of delay necessary, I will
mount the drivers in the baffle. The tweeter should
be hooked up reversed polarity.
While measuring the frequency response, delay should
be varied until there is a MINIMUM seen at the crossover
frequency.
This should take into account the differences in baffle
mounting AND the effect of the crossover.
How did I do?
P.S. Thylantyr, I will get more info on the transformer & get back with you, I just don't have time today ....
herm said:
Desigining is a game of compromises,
trying to put the important stuff
on top of the list.
Unless you plan to mount the tweeters
with a huge offset from the midrange, perhaps a few feet, I would
place this esoteric topic on the bottom of my list......
If you are not sure if this is important
for your sonic tastes, then do a
crude experiment. Place a tweeter
next to a midrange on your test
bench, move the placement of the tweeter and determine if the 1, 2, or
3 inches of offset is audible.
thylantyr said:
Hi,
The transformer is designed and intended to work above 1 kHz AC only. Below 1kHz is will quickly go into saturation with substantial signal. What do you think what will happen if you apply small amounts of DC? It will magnetise the core substantial. That can cause severe distortion. Not only that, it will also limit its “headroom” one way at frequencies it is intended for. Such transformer can handle only minor amounts of DC.
Only transformers designed for DC current (like SE output transformers) can handle DC current.
In the manual of audio line transformers I found once a note that it should not be measured with an ohmmeter. Even the small amount of current of such a meter will cause magnetisation and rise the distortion.
Whether or not the transformer is an isolation transformer does not matter. It is an impedance transformer. It transforms the low impedance of the ribbon to manageable levels, in the case of the R1 to 4 ohms. But this is only effective above a certain frequency.
Thylantyr;
Thanks for asking this question on Madisound Forum!
For those of you who have not read it yet, here's the discussion
going on over there:
http://www.madisound.com/cgi-bin/discuss.cgi?read=242561
And here's one response from Andre @ E-speakers:
>Greetings,
>
>No problem at all with connecting the Raven's directly to an
>amplifier. Good amps have extremely low DC at their output,
>typically between 1 and 5mv. This will in no way disturb the
>Raven's transformer. If DC was to be substancially higher, the
>ribbon would probably stretch at turn on/turn off and would
>require a new ribbon elament much earlier than the transformer
>would saturate.
>
>Active Raven based systems are quite popular and I have never
>had a report of a problem.
>
>Andre
Looks like it would be a good idea to measure DC before hooking
up the Ravens...
Thanks for asking this question on Madisound Forum!
For those of you who have not read it yet, here's the discussion
going on over there:
http://www.madisound.com/cgi-bin/discuss.cgi?read=242561
And here's one response from Andre @ E-speakers:
>Greetings,
>
>No problem at all with connecting the Raven's directly to an
>amplifier. Good amps have extremely low DC at their output,
>typically between 1 and 5mv. This will in no way disturb the
>Raven's transformer. If DC was to be substancially higher, the
>ribbon would probably stretch at turn on/turn off and would
>require a new ribbon elament much earlier than the transformer
>would saturate.
>
>Active Raven based systems are quite popular and I have never
>had a report of a problem.
>
>Andre
Looks like it would be a good idea to measure DC before hooking
up the Ravens...
I knew Andre would reply since he has alot of experience
with Raven and Esg ribbons. There seems to be two
issues. Theoretical and real world. According to his reply,
no problems have been reported with a direct connect.
Adding that capacitor to play it safe is probably ok too,
but now you introduce the phasing gremlin. Which
gremlin is worse, let the end user decide I guess.......
with Raven and Esg ribbons. There seems to be two
issues. Theoretical and real world. According to his reply,
no problems have been reported with a direct connect.
Adding that capacitor to play it safe is probably ok too,
but now you introduce the phasing gremlin. Which
gremlin is worse, let the end user decide I guess.......
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.