Re: Re: Re: Industrial SMD Gainclone - group buy
Konnichiwa
You clearly misunderstand what I propose, namely:
1) Input Buffer
2) 3rd order inverting highpass around the Power Op-Amp (Chip Amp) to drive the tweeter (good reason for 3rd order)
3) Inverting summing circuit that combines the Tweeter signal and the fullrange input signal to derive the lowpass for the LF Driver, LF Eq if required in the feedback loop of the LF Chip Amp
It means the addition of minimum of parts, the provision of the only possible "squarewave perfect" crossover and the needed LF eq with a minimum of added parts. No Op-Amp's other than the Power Amp's, buffer to be open loop.
Given that we'd replace simple build in X-Overs with a usually very low parts count - nope, using the right "perfect summing" crossover and equalising the woofer only works fine. Try it. The LF Eq can be tuned usually by ear and/or guesswork very close.
Sayonara
Konnichiwa
jackinnj said:
You clearly misunderstand what I propose, namely:
1) Input Buffer
2) 3rd order inverting highpass around the Power Op-Amp (Chip Amp) to drive the tweeter (good reason for 3rd order)
3) Inverting summing circuit that combines the Tweeter signal and the fullrange input signal to derive the lowpass for the LF Driver, LF Eq if required in the feedback loop of the LF Chip Amp
It means the addition of minimum of parts, the provision of the only possible "squarewave perfect" crossover and the needed LF eq with a minimum of added parts. No Op-Amp's other than the Power Amp's, buffer to be open loop.
jackinnj said:
Given that we'd replace simple build in X-Overs with a usually very low parts count - nope, using the right "perfect summing" crossover and equalising the woofer only works fine. Try it. The LF Eq can be tuned usually by ear and/or guesswork very close.
Sayonara
Re: Re: Re: Re: Re: Industrial SMD Gainclone - group buy
Konnichiwa,
I designed and build something along these lines in the mid 1980's. Trust me, it works quite well.
Subtractive X-Overs and wrapping the X-Over around the Power Amplifier are ancient concepts and have been repeatedly documented.
Sayonara
Konnichiwa,
jackinnj said:
I designed and build something along these lines in the mid 1980's. Trust me, it works quite well.
Subtractive X-Overs and wrapping the X-Over around the Power Amplifier are ancient concepts and have been repeatedly documented.
Sayonara
Re: Re: Industrial SMD Gainclone - group buy
The other reason I want to do this is because it's rather small as a project. My time is limited so I can only work on small projects.
Active cross-over is a good idea Thorsten, but not for me at the moment.
The challange for me here is to make a perfect solution with the given specifikations. The last QRP01 project went very well. A 100% correct pcb without any prototyping, perfect function out. Everybody are pleased. This was really cool to achieve such result. Ok, I worked on the design for over 6 months on and off.
The concept here is a simple buffered inverting Gainclone, nothing more. I want also to keep the number of parts down so it won't get to mighty for the skilled builder. Note that by skilled I mean people with a decent soldering iron, a pair of good tweezers... and steady hands. As long as you see well and don't tremble too much this is not going to be hard. My other SMD projects have been very easy to get going. Noone have had ANY problems. Can you believe that?
Unusual, how many SMD Gainclones are there out there, I mean complete products? This is the first so I may not call it common.Kuei Yang Wang said:
The other reason I want to do this is because it's rather small as a project. My time is limited so I can only work on small projects.
Active cross-over is a good idea Thorsten, but not for me at the moment.
The challange for me here is to make a perfect solution with the given specifikations. The last QRP01 project went very well. A 100% correct pcb without any prototyping, perfect function out. Everybody are pleased. This was really cool to achieve such result. Ok, I worked on the design for over 6 months on and off.
The concept here is a simple buffered inverting Gainclone, nothing more. I want also to keep the number of parts down so it won't get to mighty for the skilled builder. Note that by skilled I mean people with a decent soldering iron, a pair of good tweezers... and steady hands. As long as you see well and don't tremble too much this is not going to be hard. My other SMD projects have been very easy to get going. Noone have had ANY problems. Can you believe that?
Re: Re: Re: Re: Re: Re: Industrial SMD Gainclone - group buy
I expressed my interest in this project already on previous occasions, and if you are still game for that, we might do some prototyping. This of course requires your input with filter calculation and overall support. I believe Brian wouldn't mind coming up with few test boards and the whole project could be made on purely non-profit basis, as an excercise in proving ancient concepts
Kuei Yang Wang said:
I expressed my interest in this project already on previous occasions, and if you are still game for that, we might do some prototyping. This of course requires your input with filter calculation and overall support. I believe Brian wouldn't mind coming up with few test boards and the whole project could be made on purely non-profit basis, as an excercise in proving ancient concepts
To Kuei and Peter Daniel,
As a quiet lurker, I would greatly appreciate any progress made on the Subtractive-Active-Current-Clone (SACC) project--and I'm sure I'm not the only one!
As a start, a schematic would be very helpful.
As for the Seas coax, a few have the clear cone (TPX) that is reported to sound quite decent:
T17REXCOAX 17cm
T17RECOAXTV 17cm shielded
T18RECOAXTV 18cm shielded
These drivers are under ~$100 US, easily available in Europe and North America and I think they would make a great experiment base to try different enclosures (sealed, ported, open baffle) and different crossovers (simple passive--cap on tweet, complex passive, active, active with current drive). And the shielded ones could be used for home theater to boot!
The low cost and wide setup possibilities could give us a reference (small "r") from which we could speak a common language. (Another reference could be the Linkwitz headphones (Sony EX71SL ~$50 US) with notch filter--see Linkwitz site)
And finally, a Seas Coax could be used with a dipole woofer (http://www.mfk-projects.com/dipole_pa_system.htm scaled down for 15" with Eminence Beta 15s ~$50 US each) to produce a half-decent system (wide range, reasonable SPL) at low cost.
So please, any progress would be greatly appreciated!
Thanks,
Jeff
As a quiet lurker, I would greatly appreciate any progress made on the Subtractive-Active-Current-Clone (SACC) project--and I'm sure I'm not the only one!
As a start, a schematic would be very helpful.
As for the Seas coax, a few have the clear cone (TPX) that is reported to sound quite decent:
T17REXCOAX 17cm
T17RECOAXTV 17cm shielded
T18RECOAXTV 18cm shielded
These drivers are under ~$100 US, easily available in Europe and North America and I think they would make a great experiment base to try different enclosures (sealed, ported, open baffle) and different crossovers (simple passive--cap on tweet, complex passive, active, active with current drive). And the shielded ones could be used for home theater to boot!
The low cost and wide setup possibilities could give us a reference (small "r") from which we could speak a common language. (Another reference could be the Linkwitz headphones (Sony EX71SL ~$50 US) with notch filter--see Linkwitz site)
And finally, a Seas Coax could be used with a dipole woofer (http://www.mfk-projects.com/dipole_pa_system.htm scaled down for 15" with Eminence Beta 15s ~$50 US each) to produce a half-decent system (wide range, reasonable SPL) at low cost.
So please, any progress would be greatly appreciated!
Thanks,
Jeff
Now it comes to mind -- Nelson wrote an article for Speaker Builder in 1982 ("Phase Coherent Crossover Networks" SB 2/82) where he discussed the subtractive XO -- the magnitude response has a bump which, he stated, has to be dealt with via manipulation of the Q of the network.
FWIW, I modeled the results of Nelson's paper (He uses Sallen-Key -- without Q manipulation first. Here's a screenshot of the schematic and bode plots. I used MFB and the results didn't look right. I guess this is why Nelson is in the XO biz:
It looks worth a try:
FWIW, I modeled the results of Nelson's paper (He uses Sallen-Key -- without Q manipulation first. Here's a screenshot of the schematic and bode plots. I used MFB and the results didn't look right. I guess this is why Nelson is in the XO biz:
It looks worth a try:
Attachments
Konnichiwa,
Not quiet. The SUMMED RESPONSE is flat (and inherently so) at least it would be for co-incidental perfect drivers.
Now the Derived (subtracted) Channel output shows a "bump" and it MUST show this bump. This bump is neccesary and essential to make this thing do what all the kings men couldnt, namely put humpty dumpty back together.
The output from the individual driver matters diddly squat, it can have bumps and whatever, AS LONG as the summed response is FLAT.
Looks as expected and as it MUST to work. Where's the problem?
Cognitive dissonance as you expect two nice curves that can be overlaid to make a nice CROSSover on the paper?
If it did that you would various humpty pieces and various dumpty pieces with the discombulatory and disjointed sound.
Didn't use it right then.
BTW, if you mainpulate the derived path group delay you CAN make the slopes look nice and form the usuall Cross, you may then also take said cross and place it as headstone for the music having passed.
It is. In my own active speaker project I ended up using a 4th order LR type X-Over, all discrete with EQ and X-Over plugins on db-15 sockets. I used that because at the time I was building a dozend or so of these, never quite got back to subtractive but I do remember that it had something....
Sayonara
jackinnj said:
Not quiet. The SUMMED RESPONSE is flat (and inherently so) at least it would be for co-incidental perfect drivers.
Now the Derived (subtracted) Channel output shows a "bump" and it MUST show this bump. This bump is neccesary and essential to make this thing do what all the kings men couldnt, namely put humpty dumpty back together.
The output from the individual driver matters diddly squat, it can have bumps and whatever, AS LONG as the summed response is FLAT.
jackinnj said:
Looks as expected and as it MUST to work. Where's the problem?
Cognitive dissonance as you expect two nice curves that can be overlaid to make a nice CROSSover on the paper?
If it did that you would various humpty pieces and various dumpty pieces with the discombulatory and disjointed sound.
jackinnj said:
Didn't use it right then.
BTW, if you mainpulate the derived path group delay you CAN make the slopes look nice and form the usuall Cross, you may then also take said cross and place it as headstone for the music having passed.
jackinnj said:
It is. In my own active speaker project I ended up using a 4th order LR type X-Over, all discrete with EQ and X-Over plugins on db-15 sockets. I used that because at the time I was building a dozend or so of these, never quite got back to subtractive but I do remember that it had something....
Sayonara
I too am interested in this subtractive crossover. On first sight it looks as though the two combined signals cannot sum flat, but maybe I need to investigate it more.
Kuei Yang Wang, I am particularly interested in why you recommend 3rd-order as the optimum.
Digressing slightly, can 3-way crossovers be made with this topology?
Kuei Yang Wang, I am particularly interested in why you recommend 3rd-order as the optimum.
Digressing slightly, can 3-way crossovers be made with this topology?
Try to stick to the topic, please.
I have started to work on the design and I wonder how common balance signals are. Is it 1 of 100 or less? Are there anybody who are interested in balanced inputs and I'm putting this question to those who interested in this psrticular design?
I'm fighting now with the question balance input or not? It's quite possible to create some options but I want also to reduce extra alternative parts because the take up space. My goal is to keep it compact.
I have started to work on the design and I wonder how common balance signals are. Is it 1 of 100 or less? Are there anybody who are interested in balanced inputs and I'm putting this question to those who interested in this psrticular design?
I'm fighting now with the question balance input or not? It's quite possible to create some options but I want also to reduce extra alternative parts because the take up space. My goal is to keep it compact.
Hey guys,
I'm really inrterested in these guys:
http://www.pad2pad.com
I read about them in Aug PC magazine
They are now listed in our Bazaar
You use their software to lay out the circuit board and they will make it and populate it with any components from Mouser. You can also specify pads for componennts you want to add later. Wouldn't this be a good way to prototype? First we could order about 5 to test. These would be expensive each, but would be included in the overall cost of the subsequent larger order. In quantity the price drops a LOT (of course!
I'm really inrterested in these guys:
http://www.pad2pad.com
I read about them in Aug PC magazine
They are now listed in our Bazaar
You use their software to lay out the circuit board and they will make it and populate it with any components from Mouser. You can also specify pads for componennts you want to add later. Wouldn't this be a good way to prototype? First we could order about 5 to test. These would be expensive each, but would be included in the overall cost of the subsequent larger order. In quantity the price drops a LOT (of course!
peranders said:
does the problem relate to the issue of SOIC16W ? you can get the SSM2019BRN which is SOIC8 from Digikey for $4.61 -- for some reason their search engine doesn't readily come up with this part -- I am using the W devices -- I do the same with the Linear LT1014 in the W package.
i think that Analog sent some AD8620's in the pack of samples which went out 2 years ago -- these devices are $13 or so.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.