Hi
I'm building 4 inverting gain clones which are to be used to drive my 2 way setup. (Fostex FE206E OB and jbl2225 sealed)
I was planing to do something like Pedja (http://www.geocities.com/react_71/audio/gainclone/gainclone.htm)
For high/midrange I planing to use a 6db hp filter and some lp filter to remove hf garbage and a 6 or 12 db filter for the bass.
The best documentation I have found so far is:
http://www.t-linespeakers.org/tech/filters/passiveHLxo.html
http://www.silcom.com/~aludwig/Sysdes/Crossove_Design.htm#Bi_amping
It have given me some clues, but also some questions.
And now for the questions:
For the lowpass, 2 resistors in series will affect the gain right? Should I just add the value of all the resistors in series and use that to calculate the gain?
For the highpass how do I calculate the values when I need to combine a hp and a lp filter?
Or better yet, where can I read about this (see some examples (LLXO for dummies))?
I'm building 4 inverting gain clones which are to be used to drive my 2 way setup. (Fostex FE206E OB and jbl2225 sealed)
I was planing to do something like Pedja (http://www.geocities.com/react_71/audio/gainclone/gainclone.htm)
For high/midrange I planing to use a 6db hp filter and some lp filter to remove hf garbage and a 6 or 12 db filter for the bass.
The best documentation I have found so far is:
http://www.t-linespeakers.org/tech/filters/passiveHLxo.html
http://www.silcom.com/~aludwig/Sysdes/Crossove_Design.htm#Bi_amping
It have given me some clues, but also some questions.
And now for the questions:
For the lowpass, 2 resistors in series will affect the gain right? Should I just add the value of all the resistors in series and use that to calculate the gain?
For the highpass how do I calculate the values when I need to combine a hp and a lp filter?
Or better yet, where can I read about this (see some examples (LLXO for dummies))?
If you are looking at the Pedja’s circuit those HF filters are kind of weird. The higher the frequency, the higher the gain of the amp. Starts at F=6.3*(Rout +R1)*C1). C2 makes it even more complicated. The gain goes up but the filter suppresses the HF more, so it’s pretty tricky. Mind you the effect becomes smaller the smaller the Rout+R1 is compare to the 9k resistor. So if Rout+R1 has similar value to 9kOhm, than the gain will increase 2 times for HF. So if Rout+R1<<<9k, then F=1/6.3*(Rout+R1)*C1 if C2 is not installed and the gain will change very litle.
For the LF, F=1/4.7uF*6.3*(R1+9k).
I hope I’m not missing anything.
Greg
For the LF, F=1/4.7uF*6.3*(R1+9k).
I hope I’m not missing anything.
Greg
I haven't got a schematic yet. I wanted to do an XO without using opamps.
Wow it is that simple (joke).
If I add a buffer more after the XO I will get a more "normal" filter response. So I end up with: buffer -> hp -> hp -> buffer -> gainclone.
How about combining a 6db lp and a 6 db hp after each other? How will they affect each other? Should they also be separated by a buffer to get some normal filter responses?
Wow it is that simple (joke).
If I add a buffer more after the XO I will get a more "normal" filter response. So I end up with: buffer -> hp -> hp -> buffer -> gainclone.
How about combining a 6db lp and a 6 db hp after each other? How will they affect each other? Should they also be separated by a buffer to get some normal filter responses?
Konnichiwa,
Yes.
Yes and no. This is a little complex. You may be better off using fully active filters placed around the Chip Amplifiers. AD (IIRC) has a suitable on-line calculation tool that handles inverting filters up to high order, HPF and LPF.
Here it is, I did RC....
http://www.analog.com/Analog_Root/static/techsupport/designtools/interactiveTools/filter/filter.html
Simply select MFB LP (or MFB HP) and fill in the blanks and it gives you a schematic for the whole filter. If you like, tell me what Order X-Over you want and what particular frequency and I run the Filter Designer and post the results here.
Sayonara
Flemming J P said:
Yes.
Flemming J P said:
Yes and no. This is a little complex. You may be better off using fully active filters placed around the Chip Amplifiers. AD (IIRC) has a suitable on-line calculation tool that handles inverting filters up to high order, HPF and LPF.
Here it is, I did RC....
http://www.analog.com/Analog_Root/static/techsupport/designtools/interactiveTools/filter/filter.html
Simply select MFB LP (or MFB HP) and fill in the blanks and it gives you a schematic for the whole filter. If you like, tell me what Order X-Over you want and what particular frequency and I run the Filter Designer and post the results here.
Sayonara
Konnichiwa,
There is a way to simplify even more. As long as you use even numbers of filter sections you can alternate NPN Transistors as follower with the J-Fet CCS to the negative rail with PNP Transistors as follower with the J-Fet CCS to the positive rail for the LPF. If the NPN and PNP transistor are thermally coupled you remove the output offset completely and you can dispense with coupling capacitors entirely.
Equally, if you include an added PNP Follower after the final HPF section (maybe implementing a mild LPF while at it?) you can limit the cap's used again strictly to those in the filter.
Sayonara
Nuuk said:
There is a way to simplify even more. As long as you use even numbers of filter sections you can alternate NPN Transistors as follower with the J-Fet CCS to the negative rail with PNP Transistors as follower with the J-Fet CCS to the positive rail for the LPF. If the NPN and PNP transistor are thermally coupled you remove the output offset completely and you can dispense with coupling capacitors entirely.
Equally, if you include an added PNP Follower after the final HPF section (maybe implementing a mild LPF while at it?) you can limit the cap's used again strictly to those in the filter.
Sayonara
what about that AC-coupled in and out jfet stage? it makes a high input impedance inveting amplifer.
also, what happens when the gain is dropped below 10 on the gainclones? for inverting you could just add a cap either in series with the input for highpass or in parallel with the feedback for lowpass, unless it would mess up with a gain < 10. national does this near the edge of the audio spectrum in the datasheets.
also, what happens when the gain is dropped below 10 on the gainclones? for inverting you could just add a cap either in series with the input for highpass or in parallel with the feedback for lowpass, unless it would mess up with a gain < 10. national does this near the edge of the audio spectrum in the datasheets.
Hi Flemming,
First, if you use FE206E (96dB/W(m)) and expose the tweeter directly to the amp’s out, then with buffer I’d strongly suggest usage of delayed power-up for the output opamps. (The only good buffer that I can think of at the moment and which still can make you pretty free of power up thump is the diamond transistor.)
I think the thing can be done two ways. First, you can use two whole buffer/filter/opamp circuits, one for each section. The second option assumes one buffer. I can only guess what will be better.
(those triangles at left are symbols for the buffers and not for the opamps
)
In fact, in both cases either section is band passed. As crossover elements, the coupling capacitor would be used for high-, and shunt capacitor for low-passing. Greg’s equations are ok with that addition, considering necessary capacitance for the LPF, R1 should be probably in the kOhm range, and hence 9k resistor should be calculated in parallel with Rout + R1. For the 2SK170 buffer biased between 5 and 10mA, Rout is between 30 and 25 Ohms, so if R1 is 1k or more, Rout can be neglected.
To make the designing of the filter more controlled: don’t tune the frequency by the resistors, first estimate what would be optimal values of the series resistors, then determine appropriate feedback resistor. Use the caps to tune the filters frequencies.
If you have some determined crossover frequency I could simulate the passive LLXO. Let me know. Also, what is the sensitivity difference between the drivers (if you intend to match their levels using the adequate gain in the amp sections)?
Second order filter might be a bit longer story.
Pedja
First, if you use FE206E (96dB/W(m)) and expose the tweeter directly to the amp’s out, then with buffer I’d strongly suggest usage of delayed power-up for the output opamps. (The only good buffer that I can think of at the moment and which still can make you pretty free of power up thump is the diamond transistor.)
I think the thing can be done two ways. First, you can use two whole buffer/filter/opamp circuits, one for each section. The second option assumes one buffer. I can only guess what will be better.
(those triangles at left are symbols for the buffers and not for the opamps
)An externally hosted image should be here but it was not working when we last tested it.
In fact, in both cases either section is band passed. As crossover elements, the coupling capacitor would be used for high-, and shunt capacitor for low-passing. Greg’s equations are ok with that addition, considering necessary capacitance for the LPF, R1 should be probably in the kOhm range, and hence 9k resistor should be calculated in parallel with Rout + R1. For the 2SK170 buffer biased between 5 and 10mA, Rout is between 30 and 25 Ohms, so if R1 is 1k or more, Rout can be neglected.
To make the designing of the filter more controlled: don’t tune the frequency by the resistors, first estimate what would be optimal values of the series resistors, then determine appropriate feedback resistor. Use the caps to tune the filters frequencies.
If you have some determined crossover frequency I could simulate the passive LLXO. Let me know. Also, what is the sensitivity difference between the drivers (if you intend to match their levels using the adequate gain in the amp sections)?
Second order filter might be a bit longer story.
Pedja
I have a question about using either passive or active XO that are located between power amp and preamp.The drivers are connected directly to the amps.
But what about the tweeter.Does it have to have some cup (say some 50uF) to protect it from DC.Maybe delay relay at the output of the amp would be enough.Any thoughts on that topic?
Bartek
But what about the tweeter.Does it have to have some cup (say some 50uF) to protect it from DC.Maybe delay relay at the output of the amp would be enough.Any thoughts on that topic?
Bartek
If I would use a cap on a tweeter, I would choose such value that it would create one pole of a filter, so the active crossover would need one pole less, usually in a range of 8u - 1uF and it could be a better quality than let's say suggested 50u.
But is the cup must,to protect tweeter or it can be omited?
It somehow makes the whole active XO less actractive!
A lot of benefits come from directly coupling the amps to the speakers (these are not my words).
Bartek
It somehow makes the whole active XO less actractive!
A lot of benefits come from directly coupling the amps to the speakers (these are not my words).
Bartek
If the offset of the amp is low, the cap shouldn't be necessary. However, I don't know how low is the low for a tweeter.
No,I have never build or have a tube amp.
I just would like to get the answer to the question:
Is the cup protecting the tweeter needed in case of using Active XO?
Let's say the amp is not exiding 100mV DC offset.
Bartek
I just would like to get the answer to the question:
Is the cup protecting the tweeter needed in case of using Active XO?
Let's say the amp is not exiding 100mV DC offset.
Bartek
it isn't needed, unless you foresee some case where the speaker is attached to DC. such as when the chip fails (if such occures), with a cap you save the tweeter but kill the amp. if you were making speakers for a buisness, then this would be important. if its for yourself, you just have to ask how much you care about a double-loss in case of failure.
Just be careful when building your amps, make sure there's no chance for either voltages to come in cantact and you'll be pretty safe. The LM series are pretty rugged if you don't massively abuse them and the most often failures I've had were when I connected the + and - voltages backwards.
It'd be better without the cap. If you're real worried about DC offset, match the impedances on the inputs with a resistor and pot on the + input, that should eliminate any meaningful amount (although the "usual" amount of DC offset on a "gainclone" is pretty low to begin with) of DC offset.
A cap before the tweeter is going to have to be large compared to the caps before the amp, and if you don't want this cap to degrade the sound you're going to have to pay a pretty penny (well, alot more than the other caps, we all have different ideas of expensive I know).
It'd be better without the cap. If you're real worried about DC offset, match the impedances on the inputs with a resistor and pot on the + input, that should eliminate any meaningful amount (although the "usual" amount of DC offset on a "gainclone" is pretty low to begin with) of DC offset.
A cap before the tweeter is going to have to be large compared to the caps before the amp, and if you don't want this cap to degrade the sound you're going to have to pay a pretty penny (well, alot more than the other caps, we all have different ideas of expensive I know).
I'm not worrying about DC offset and failure (that is up to me to do everything that it is low).I'm always very carefull.
I'm very interested to build active XO and I've heard some people put the cup on the tweeter.But than when you have two way speakers with only cup on the tweeter it doesn't seem like an upgrade from bi-amping (there are other advantages I know).
So now I know there is no need for this.Thanks.
I'm thinking about cloning Marchand active crossover.
They supply schematics on their website and some calculators for resistor values ,too.
Bartek
I'm very interested to build active XO and I've heard some people put the cup on the tweeter.But than when you have two way speakers with only cup on the tweeter it doesn't seem like an upgrade from bi-amping (there are other advantages I know).
So now I know there is no need for this.Thanks.
I'm thinking about cloning Marchand active crossover.
They supply schematics on their website and some calculators for resistor values ,too.
Bartek
- Status
- Not open for further replies.