Hy Guys,
I had a "semi" successfull shot building some wire panels earlier this year (http://www.diyaudio.com/forums/planars-exotics/238622-first-esl-attempt-3.html)
But I'm currently constructing some new, and hopefully, better one's. I'll use some sheets of wood and route all the openings, will give me a much flatter panel this way.
Before I found myself constantly checking out every wooden slat at the diy store and no matter how "straight" they were when constructing panels this way I never could get them perfectly straight, there was always a little warp or twist in such a panel and they didn't align 100% either...
I'm playing a bit with the esl_seg_ui program and I have a few questions.
When I use a resistor on the (+) input of 4.7 Ohm, can I put this in the program as 0.0047 kOhm ? It "seems" that this value hasn't got any effect, is this correct ?
I use this resistor to protect my amp but putting in 0.0047kOhm or 0.008kOhm apparently has no effect, so why isn't everybody doing this, must be something I'm missing here ...
When using this resistor, can I calculate a passive crossover based on this (~300Hz) since the impedance won't drop below 4.7Ohm ?
Or can I discard this resistor when segmenting this way ?
I use 2 230V/6V toroids for this project.
Greetings, Kevin.
I had a "semi" successfull shot building some wire panels earlier this year (http://www.diyaudio.com/forums/planars-exotics/238622-first-esl-attempt-3.html)
But I'm currently constructing some new, and hopefully, better one's. I'll use some sheets of wood and route all the openings, will give me a much flatter panel this way.
Before I found myself constantly checking out every wooden slat at the diy store and no matter how "straight" they were when constructing panels this way I never could get them perfectly straight, there was always a little warp or twist in such a panel and they didn't align 100% either...
I'm playing a bit with the esl_seg_ui program and I have a few questions.
When I use a resistor on the (+) input of 4.7 Ohm, can I put this in the program as 0.0047 kOhm ? It "seems" that this value hasn't got any effect, is this correct ?
I use this resistor to protect my amp but putting in 0.0047kOhm or 0.008kOhm apparently has no effect, so why isn't everybody doing this, must be something I'm missing here ...
When using this resistor, can I calculate a passive crossover based on this (~300Hz) since the impedance won't drop below 4.7Ohm ?
Or can I discard this resistor when segmenting this way ?
I use 2 230V/6V toroids for this project.
Greetings, Kevin.
Attachments
When adding a resistance to the primary side you must mutiply it by the square of the transformation ratio to see the refelcted effect, IE "Pri R" * Ratio^2= reflected "Pri R".
The program assumes that all of the resistances are on the secondary side of the transformer, so therefore if you have a 1:100 ratio, the 4.7 ohms resistor on the primary is reflected as a 47k to the secondary side.
jer 🙂
The program assumes that all of the resistances are on the secondary side of the transformer, so therefore if you have a 1:100 ratio, the 4.7 ohms resistor on the primary is reflected as a 47k to the secondary side.
jer 🙂
I see, I will have a look at it tomorrow afternoon. In my case I have a stepup of 1:70, so I have to put in 23kOhm, ... Hope it doesn't, affect the graph too much ...
Do you by any chance now if I can use the fix 4.7Ohm resistor as a starting point to calculate a passive crossover ?
Thank you for helping me out here.
Greetings, Kevin.
Do you by any chance now if I can use the fix 4.7Ohm resistor as a starting point to calculate a passive crossover ?
Thank you for helping me out here.
Greetings, Kevin.
The First series resistance will only effect the very high end roll of just as it does for EQ'ing the peaking effect caused by the transformers leakage inductance and the total capacitance.
I plan on omitting this resistor and using a shelving EQ to make this adjustable in real time as needed.
I started a passive design once and it got quite complicated with very large capacitors.
But a simple one can be made but you must know the primary inductance of your step-up transformer.
I did a series type of crossover design using Circuitmaker 2000, you can use LTspice just the same only I am more fluent with CM2K.
You also have to represent the transformers leakage inductance as well.
The results looked promising but like I said the parts became much more expensive than what a simple opamp circuit could do with a spare amplifier for the woofer.
For instance in my design it required some 200uf and in some cases 2000uf non-polar caps at high voltages for the voltage range I was designing it for.
Since it was a simulation there was no way for me to know if some of the added networks would actually be of any use.
They did make some changes in phase shift at different frequency's that otherwise would not be there in a straight through system.
So I gave up on the idea.
But even one of my simplest designs might have worked well I just never got as far as testing it.
I had to build new panels First.
I found that you can add some more inductance in series with the Primary to form the filter for the woofer.
This works well for a series type of crossover.
I may explore this some more once I get my desktop setup running again.
But, Mainly I am just shootin' four amplifier Bi-Amped design as it is much more cost effective.
If I can find those sim files again I will certainly post them for you to mess with and get an idea of how it would work.
jer 🙂
I plan on omitting this resistor and using a shelving EQ to make this adjustable in real time as needed.
I started a passive design once and it got quite complicated with very large capacitors.
But a simple one can be made but you must know the primary inductance of your step-up transformer.
I did a series type of crossover design using Circuitmaker 2000, you can use LTspice just the same only I am more fluent with CM2K.
You also have to represent the transformers leakage inductance as well.
The results looked promising but like I said the parts became much more expensive than what a simple opamp circuit could do with a spare amplifier for the woofer.
For instance in my design it required some 200uf and in some cases 2000uf non-polar caps at high voltages for the voltage range I was designing it for.
Since it was a simulation there was no way for me to know if some of the added networks would actually be of any use.
They did make some changes in phase shift at different frequency's that otherwise would not be there in a straight through system.
So I gave up on the idea.
But even one of my simplest designs might have worked well I just never got as far as testing it.
I had to build new panels First.
I found that you can add some more inductance in series with the Primary to form the filter for the woofer.
This works well for a series type of crossover.
I may explore this some more once I get my desktop setup running again.
But, Mainly I am just shootin' four amplifier Bi-Amped design as it is much more cost effective.
If I can find those sim files again I will certainly post them for you to mess with and get an idea of how it would work.
jer 🙂
Your idea to use a resistor in series with the transformer to protect your amplifier is a good one. However, it won’t define the impedance for crossover calculations because it is in series with the transformer primary. Adding a capacitor in series to form a hi-pass doesn't work well since the impedance of the transformers varies widely with frequency. The simplest approach is to add a 10ohm - 15ohm resistor across the primary of the step-up transformer. This goes a long way to stabilizing the impedance magnitude so the crossover network will work properly. Lower values of shunt resistance result in a more constant impedance. The trade off is that lower values of resistance will need to be able to dissipate more power than if you chose a higher value. In my experience, 10 - 15ohm is a good range to choose from. Much lower than 10 ohm and power dissipation becomes a problem. Much higher than 15ohm and the impedance isn't as constant and resistive as you would like for an easy passive crossover.
Both the Acoustat Spectra 11 and Final 0.3 used a 10 ohm shunt resistor in the crossover.
Attachments
What does the single 1k ohm resistor on the secondary leg on the Final do? Some kind of delay front to back stator?
im curious as well 🙂 onlyin to one stator.. maybe its to compensate for the offset? if you measure between the 0 and one secundary and then between 0 and the other secundairy they have different values.
I can see that the resistor in the second picture is "only" 5W, my 4.7 Ohm resistor is rated for 11W, is this enough or does it need to be higher (the Wattage) ?
By putting the resistor across does this change the impedance value that the amp sees ? Eg: a 10 Ohm resistor across means 10 Ohm for the amp ?
Does a wirewound resistor suffice ? I still have a few 4.7 Ohm 11W resistors that I can work with...
How does this relate to the value that I need to put in the ESL_set_ui program ?
Many thanks for your help with this project guys.
Greets, Kevin.
By putting the resistor across does this change the impedance value that the amp sees ? Eg: a 10 Ohm resistor across means 10 Ohm for the amp ?
Does a wirewound resistor suffice ? I still have a few 4.7 Ohm 11W resistors that I can work with...
How does this relate to the value that I need to put in the ESL_set_ui program ?
Many thanks for your help with this project guys.
Greets, Kevin.
Manufacturers like Acoustat make speakers (such ones which include an ESL and a cone driver) that can be driven by one amp channel, like their customers have at home.
But that makes little sense for a DIYer. As the cautionary tale posts previously suggest, black arts have to be invoked trying to create anything half useful in a passive crossover, ESLs or cones, and certainly not paired.
Much more sensible to just bi- or tri-amp with a digital signal processor crossover.
Ben
But that makes little sense for a DIYer. As the cautionary tale posts previously suggest, black arts have to be invoked trying to create anything half useful in a passive crossover, ESLs or cones, and certainly not paired.
Much more sensible to just bi- or tri-amp with a digital signal processor crossover.
Ben
The 10 ohm resistor gives the amp something to work against, but it will not be a 10ohm load to the amp, as it is parallel to the Primary winding of the transformer.
I would use both the 10ohm across the tranny primary, and use a DCX2496 Digital Crossover (or some other digital crossover)
Hy guys,
The reason why I want to build an "easy" passive crossover is if a friend wants me to build him a pair I can do it without having him have to upgrade his entire setup.
I want to cross over at ~300Hz ...
I personally use a Marantz receiver, the SR7007
Greets, Kevin.
The reason why I want to build an "easy" passive crossover is if a friend wants me to build him a pair I can do it without having him have to upgrade his entire setup.
I want to cross over at ~300Hz ...
I personally use a Marantz receiver, the SR7007
Greets, Kevin.
That is a good question…1kOhm on the secondary side is too low a resistance to affect the response. I guess it might have been included in a misguided attempt to equalize DC resistance between the center-tap and the each stator. But when operating in constant charge mode, there is no current flow from the center tap. It is effectively just one series circuit(transformer-stator1-stator2-1k-transformer) so it doesn't matter where the resistance is along the series circuit path.
More details on effect of adding resistance in series with stators here:
http://www.diyaudio.com/forums/planars-exotics/258958-help-esl-simulator.html#post3989632
Perhaps it is there as a safety precaution to limit current in case of stator-to-stator short.
Yeah, I am surprised that the Final crossover used a 10ohm resistor rated at only 5W. The power dissipation in this resistor when playing music will depend on how low the crossover is and how big an amplifier you use. Based on my experience I would recommend something in the 20W to 40W range. Note that testing with steady tones or pink noise will generate more heat in the resistor than playing music because of the high peak/rms nature of most music.
The impedance seen by the amp will depend on transformer parasitics and where you position your series resistance. The most common arrangement for a first order HP filter is shown in Figure (A). The impedance seen by the amplifier will be the parallel combination of the 10 ohm resistor and the transformer in series with the 1 ohm resistor. In practice, it will be something around 8 – 10 ohm in the 200 – 2Khz range, falling to 1 – 2 ohm at 20Khz.
Note that the 1ohm series resistor forms a voltage divider with the 10ohm, so you will lose about -1dB of your amplifier output. If you increased this to 4.7ohm like you are wanting, you would lose a bit over -3dB. This is obviously undesirable. My preference is to move the 10ohm shunt resistor upstream of the series resistor as shown in Figure (B). The impedance seen by the amplifier is similar, but you avoid the -3dB loss.
In either case, to crossover around 300Hz a series capacitor of about 33uF – 47uF would be a good starting point. Measurements of the response are required to tweak the exact value. You will probably find that 4.7ohm series resistance rolls off the top octave transformer response a bit. You can either adjust you preamp tone controls to compensate, or reduce the series resistance. Usually something in the 1.5 – 2 ohm range works well and doesn’t trigger protection circuits in AV receivers as the impedance only falls below 4 ohm in the top octave where there is little musical content.
These should work fine. You can use two in series to form a 9.4ohm 22W resistor for the shunt.
The shunt resistance does not directly affect the response and can be ignored for the ESL_set_ui program calculations.
Any series resistance should be reflected to the secondary and used as part of the first ladder resistance.
(Multiply series resistance on primary side by square of step-up ratio as geraldfryjr mentioned in post#2.
Attachments
- Status
- Not open for further replies.