A SMD BF862 as Option for FULL SMD Version
Hello Tea-Bag,
Maybe for considering for future proof of this work maybe a full SMD or Mixed version with cheap available BF862 J-Fets would be an viable option, ensuring prolonged usable life for all the effort taken within this project,
Just as an suggestion to consider . .
Best Regards,
Andreas
Hello Tea-Bag,
Maybe for considering for future proof of this work maybe a full SMD or Mixed version with cheap available BF862 J-Fets would be an viable option, ensuring prolonged usable life for all the effort taken within this project,
Just as an suggestion to consider . .
Best Regards,
Andreas
Toshiba JFets shortage/fakes is real. Myself. I have been basing some of the projects on the LSK170s. They are available - The sound does not appear to be markedly different.
Personally I have used the adapter boards SOT to TO-92 pin adapters with success. But me putting on SMD boards with my vision even under a big magnifier is troublesome.
Personally I have used the adapter boards SOT to TO-92 pin adapters with success. But me putting on SMD boards with my vision even under a big magnifier is troublesome.
Agree, no SMDs required.
Also agree, use off board PS - BiB, DCB1 or other.
Just my $0.02.
Magnifier a Standard Tool for SMD like the Micro Soldering Iron
Same here, I have to use a 3 x Magnifier A4 lenses for almost all my daily work with some other stuff,
not Audio, but with help of THIS magnifier lenses:
3X Full Page Magnifying Floor Lamp For Reading And Crafts
I can do ...
For the J-Fets, the BF862 SMD they are price and even SQ winners to the LSK ?
BRGDS,
Andreas
Same here, I have to use a 3 x Magnifier A4 lenses for almost all my daily work with some other stuff,
not Audio, but with help of THIS magnifier lenses:
3X Full Page Magnifying Floor Lamp For Reading And Crafts
I can do ...
For the J-Fets, the BF862 SMD they are price and even SQ winners to the LSK ?
BRGDS,
Andreas
Last edited:
I agree. Being able to play with the crossover frequency and slopes is a big plus, and will make the board more popular. Is 3-way functionality a necessary feature, or could xover boards simply be stacked for 3-way builds? Certainly a 2-way board would be more compact.
Using the BiB shunt boards seems like a good PS solution for the crossover board. As far as the front end, my vote would be to have a global volume control and buffer be placed on the crossover board. Though I guess if a DCB1 can serve double duty (PS and buffer), that still is only 2 boards for the build.
Variable LR2, LR4 XO
Here is a suggestion - clearly stolen from the B4 - to make our XO variable. Unlike the B4 which needs to meet everyone's needs, our diy XO can be simpler as I suspect most of us have at least a close idea of where we want the XO to be. To achieve some flexibility, I suggest placing a 4 segment DIP switch with 4 resistors 'below' each R in the schematic. By starting with a fixed R at the lowest expected XO freq, and putting the others in parallel a 4 position switch will give you 16 steps between F-lo and F-hi (say an octave or so). I have attached a simple spreadsheet illustrating how this could work. User input is F-lo (the lowest possible XO freq), F-hi (the highest), and C - chosen to be available and give reasonable values for R given the XO range. The spreadsheet calculates each step and the resulting freq. The user does have to enter the closest E48 resistor values to get the actual freq., but I included the list to make that easier. Just watch your powers of 10. Obviously E24 or even E96 values could be used. Someone smarter then me could probably make that happen automatically, but I'm a retired ER doc, not an electrical engineer.
Likewise, the same could be used for the 6dB slopes, but I am not sure if the equation is the same (I think so, but someone else needs to confirm.)
The spreadsheet is totally unlocked, so feel free to change/ improve and post.
OOPS! I tried to upload the .xls file,but keep getting the message it is an invalid file. It opens fine on my computer. Am I doing something wrong or is there a different procedure for uploading spreadsheets?
Here is a suggestion - clearly stolen from the B4 - to make our XO variable. Unlike the B4 which needs to meet everyone's needs, our diy XO can be simpler as I suspect most of us have at least a close idea of where we want the XO to be. To achieve some flexibility, I suggest placing a 4 segment DIP switch with 4 resistors 'below' each R in the schematic. By starting with a fixed R at the lowest expected XO freq, and putting the others in parallel a 4 position switch will give you 16 steps between F-lo and F-hi (say an octave or so). I have attached a simple spreadsheet illustrating how this could work. User input is F-lo (the lowest possible XO freq), F-hi (the highest), and C - chosen to be available and give reasonable values for R given the XO range. The spreadsheet calculates each step and the resulting freq. The user does have to enter the closest E48 resistor values to get the actual freq., but I included the list to make that easier. Just watch your powers of 10. Obviously E24 or even E96 values could be used. Someone smarter then me could probably make that happen automatically, but I'm a retired ER doc, not an electrical engineer.
Likewise, the same could be used for the 6dB slopes, but I am not sure if the equation is the same (I think so, but someone else needs to confirm.)
The spreadsheet is totally unlocked, so feel free to change/ improve and post.
OOPS! I tried to upload the .xls file,but keep getting the message it is an invalid file. It opens fine on my computer. Am I doing something wrong or is there a different procedure for uploading spreadsheets?
Dmt,
The Zip works just fine.
You can format the user input cells to "unprotect"
Then you "protect" the sheet, but without a password.
that allows the user to change any of the input cells, but prevents losing any of the formulae.
If the user wants/needs to change some part of the worksheet, they can "unprotect" sheet and make changes, then re-set the "protect" sheet.
The use of no password allows access for experienced users.
The Zip works just fine.
You can format the user input cells to "unprotect"
Then you "protect" the sheet, but without a password.
that allows the user to change any of the input cells, but prevents losing any of the formulae.
If the user wants/needs to change some part of the worksheet, they can "unprotect" sheet and make changes, then re-set the "protect" sheet.
The use of no password allows access for experienced users.
That's a very nice and useful spreadsheet. Thanks for sharing that.
I would think that 16 steps, of the builder's preferred range, would provide plenty of adjustability. I'm hoping to use the B1 active shed the lower frequencies from my speakers to a subwoofer or two. With your proposed 4 switch concept, that would give me 5hz increments between 30hz and 105hz for example. Plenty for my needs.
This seems like a nice compromise between adjustability and a more complicated build.
I would think that 16 steps, of the builder's preferred range, would provide plenty of adjustability. I'm hoping to use the B1 active shed the lower frequencies from my speakers to a subwoofer or two. With your proposed 4 switch concept, that would give me 5hz increments between 30hz and 105hz for example. Plenty for my needs.
This seems like a nice compromise between adjustability and a more complicated build.
Sallen-Key vs. Linkwitz-Riley
I always thought that two second order S-K filters in series created a 4th order L-R XO, but I note that on the B4, Nelson has included a switch which changes the 2C of the low pass circuit to 1C with no changes to the high pass. If I read the diagram correctly, the 2C version is labelled L-R and the 1C as S-K. Other diagrams I have seen of S-K filters had the 2C in the schematic. Can anyone explain?
I always thought that two second order S-K filters in series created a 4th order L-R XO, but I note that on the B4, Nelson has included a switch which changes the 2C of the low pass circuit to 1C with no changes to the high pass. If I read the diagram correctly, the 2C version is labelled L-R and the 1C as S-K. Other diagrams I have seen of S-K filters had the 2C in the schematic. Can anyone explain?
This is what I have always seen: the 12 dB version has all Rs equal and all Cs equal; the 24dB has 2R in the high pass and 2C in the low pass. I have not seen a 24dB alignment with equal Rs and Cs nor a 12dB with the 2R, 2C. Obviously this has implications in making a modular XO that can be 6,12,18, or 24 dB.
In looking at "B1 active cross.png" (which appears to be a copy of the B4 except for using all 2SK170s instead of complimentary pairs), it appears that the 'S' position will give all Rs equal, all Cs equal, but the 'L' position will give all Rs equal but 2C in the high pass. This asymmetry was my question. I know that the ratio of R1 : R2 (or C1:C2) sets the Q of the alignment, but what happens when the high pass and low pass do not have the same Q and why would you do that?
Perhaps "B1 active cross.png" is not an accurate copy of the B4 alignment, but the manual does say, "The L setting is provided to switch a capacitor on each filter so as to implement a “Linkwitz-Riley” value set when you are using 24 dB/octave setting."
Thanks for helping me understand all this.
the Linkwitz Riley is a special filter.
It is two cascaded Butterworth Filters to give Q=0.5 and F-6dB at the crossover, instead other filter types where we generally design for F-3dB at the crossover.
Sallen and Key (S&K) is a filter circuit topology where the active stage is unity gain. The S&K can like all other filter be set to any Q and any F-3dB frequency.
Changing the C of one leg changes the F-3dB frequency AND changes the Q value.
A Butterworth filter has a Q=1/sqrt(2) = ~0.707107
A cascade of two filters multiplies the two Q values. eg two B2 gives 1/sqrt(2)*1/sqrt(2) = 1/2 = 0.5
or Q=0.54 and Q=1.31 in cascade becomes Q=0.54*1.31 = ~0.707
Last edited:
Compare a unity gain S&K filter to a equal value active filter which has some gain.
The unity gain uses the RATIO of values to determine the Q.
The Equal value uses the active GAIN to determine the Q.
Equal Component Active Butterworth Low Pass Filter Calculator
This last filter tends to get used where the ease of setting Q (independant of F) is more important than cost, where the unity gain is the simpler with fewer components.
The B1 MUST be a unity gain type. S&K is the simplest of any unity gain filter.
The unity gain uses the RATIO of values to determine the Q.
The Equal value uses the active GAIN to determine the Q.
Equal Component Active Butterworth Low Pass Filter Calculator
This last filter tends to get used where the ease of setting Q (independant of F) is more important than cost, where the unity gain is the simpler with fewer components.
The B1 MUST be a unity gain type. S&K is the simplest of any unity gain filter.
Hi,
any news from here?
Not forgotten about. I am still uncertain we have come to a good conclusion of a design.
RodeoDave has done lots of work on this, and I like elements of Prooptiki's design as well. If a consensus is not drawn, and some other question remain unanswered, I will decide on something. We need to make sure we are not bothering anyone's intellectual property.
Regards,
Tea
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.