Another can of worms...

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am assembling some analog active crossovers and I would like some suggestions on capacitor and resistor types.

What does everyone like?

Now what do you like that I can afford? :D

Part of the BOM that was generated is below.

Part Value Device Package Description
B1 2KBP 2KBP RECTIFIER
C1 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
R1 POT-1/4P POT-1/4P
R2 1M TH5MM 0204/5
 
Build your AXO with JFET input opamps (TL074 works a treat and is dirt-cheap). This allows you to use the cheapest NP0 capacitors (values under 1nF, typically) in conjunction with high value resistors (100k and up).

I prefer to avoid lower capacitor value such as below 1nF. The effect with HF is not good. With lower C, the accuracy may suffer due to parasitic capacitance of the PCB. With lower C the R then need to be higher. This may increase noise, react with parasitic capacitance, introduce phase shift, lower the bandwidth, and so on.
 
I'd prefer replacing 'resistor' by 'capacitor', because resistors usually are much cheaper than capacitors.


Best regards!

True, but the filters use way more resistors than capacitors. Plus, the capacitors make a way bigger difference ;), although there are a lot that will work very well. MKP types are great.

I've used different kinds of film caps. I like polypropylene and MKP. I've used caps I took out of nice equipment from the 80s/90s; I don't know what they are but they work great and I have a lot of them. Differences are subtle to non existent. I make other people listen too for what it's worth, and solicit their opinion.

I prefer to build with new parts. I do buy some film caps x10 but I still go through a lot more resistors. A typical circuit will need 2x the same value cap two different values, and 8x the same value resistor two different values.
 
Design your filters so you use the same value resistor as much as possible. Then buy 10 lots and save a little money.


There is a thread regarding (What do you think about) Passive Crossover. Those who have the skill for passive design might have issues with rolling active crossovers (Me included).


So, for me, TL074 is a waste of time. True, it is cheap, but who's looking for cheap sound :D


And the resistors, the values are all over the place because the circuitry is designed to match the opamp being used. Many good opamps are actually unique, meaning that for acceptable result they cannot be used in conventional circuit (where people can do an op-amp rolling)
 
I second the poster who commented about reducing the number of capacitor values in the circuit. I'm my own designs where I wanted high accuracy I selected 1% Vishay MOO and played with resistor values to reduce the 8 stages of filter to only use 3 values of C.
All resistors are 0.1% welwyn arcol 0.6W types and all values are 1k to 30k max for low noise.

I have used both lf353 ne5532 and lme49720 with little difference between the latter pair. The I get input wasn't bad especially the lowly 353 but phase inversion is something to watch and input diodes protection is a good idea
 
I prefer to avoid lower capacitor value such as below 1nF. The effect with HF is not good. With lower C, the accuracy may suffer due to parasitic capacitance of the PCB.

Sure that's possible if you go too low. I tend to keep my cap values so that the parasitics don't account for more than roughly 2% of the cap's value. Board parasitics are less of an issue than non-linear input capacitance of JFET opamps though.

With lower C the R then need to be higher. This may increase noise, react with parasitic capacitance, introduce phase shift, lower the bandwidth, and so on.

Sure, the aim is to have higher Rs, this improves dynamics due to lower signal-correlated noise on the supply rails. The noise issue can be taken care of by running at higher signal levels, after all resistor noise is a function of the SQRT(R) so doesn't increase as fast as the dynamics go up.

btw TL074 whilst cheap doesn't produce anything like cheap sound when correctly treated.
 
Last edited:
btw TL074 whilst cheap doesn't produce anything like cheap sound when correctly treated.

Recently i read someone mentioned something like:

People rolled op-amps trying to improve the performance with supposed to be better op-amps and think that everything is improving until they put back the old NE5532... :D

The point was that all opamps have "op-amp sound". It will always be true if we are still ignoring the facts about (TL071/2/4) measured performance.
 
TL074 can sound pretty lousy if you load it for example, even with loads which are widely considered benign for opamps (like say 5k).

I only used such opamp for input buffer. But LF353 is not much difference in price so TL071/2 is history except for a quad like TL074 which has high performance to price ratio (last time I bought it less than 0.1 USD a piece), but still LF347 (the quad of LF353) is only 7x the price. It is much less than a glass of coffee.

Even expensive JFET input op-amp can be so noisy (BJT reference: NE5532 with 5 nV/SqrHz). And the gate capacitance issue as you mentioned, makes non-inverting configuration is to be avoided when possible.
 
There is a thread regarding (What do you think about) Passive Crossover. Those who have the skill for passive design might have issues with rolling active crossovers (Me included).


So, for me, TL074 is a waste of time. True, it is cheap, but who's looking for cheap sound :D


And the resistors, the values are all over the place because the circuitry is designed to match the opamp being used. Many good opamps are actually unique, meaning that for acceptable result they cannot be used in conventional circuit (where people can do an op-amp rolling)


The original circuit calls for TL074n not my choice but designers spec.

TI says the TL084 is a drop in replacement(pin compatible) and better so that is what I will be using.

I work with what I have. ;)
 
Below is the BOM generated by EAGLE but I have no clue how to convert this to an orderable BOM on Mouser or DigiKey.

I will be looking into it this week and hopefully be able to place an order this weekend.

Part Value Device Package Description
B1 2KBP 2KBP RECTIFIER
C1 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C2 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C3 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C4 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C5 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C6 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C7 CP-E-025X060 E-025X060 POLARIZED CAP
C8 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C9 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C10 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C11 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C12 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C13 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C14 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C15 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C16 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C17 .01uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C18 .005uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C19 .005uF TH-NPC050-075X040 C050-075X040 Through-Hole Capacitor
C20 TH-POL-ELECTE-075X160 E-075X160
C21 TH-POL-ELECTE-075X160 E-075X160
C22 47uF 63V TH-POL-ELECTE-050X100 E-050X100
C23 TH-POL-ELECTE-075X160 E-075X160
C24 TH-POL-ELECTE-075X160 E-075X160
C25 47uF 63V TH-POL-ELECTE-050X100 E-050X100
CRIGHT1 COPYRIGHT18 COPYRIGHT18 COPYRIGHT-18
IC1 TL084P TL084P DIL14 OP AMP 4 ch
IC2 TL084P TL084P DIL14 OP AMP 4 ch
IC3 7815 78XXS 78XXS VOLTAGE REGULATOR
IC4 7915 79XXS 79XXS VOLTAGE REGULATOR
INSP1 INSP: INSP: INSP: INSPECTION mark on board silk screen
JP1 PINHD-1X2 1X02 PIN HEADER
JP2 PINHD-1X2 1X02 PIN HEADER
JP3 PINHD-1X2 1X02 PIN HEADER
JP4 PINHD-1X2 1X02 PIN HEADER
KK1 FK218 FK218 FK218 HEATSINK manufacturer Fischer/distributor Buerklin
KK2 FK218 FK218 FK218 HEATSINK manufacturer Fischer/distributor Buerklin
LD1 T-HOLE 5mm T-HOLE5_SIDE LED5MM-SIDE
MT1 PAD_6-32_RECT PAD_6-32_RECT 6-32_SQR_LG 6-32 clearance square mounting pad
MT2 PAD_6-32_RECT PAD_6-32_RECT 6-32_SQR_LG 6-32 clearance square mounting pad
MT3 PAD_6-32_RECT PAD_6-32_RECT 6-32_SQR_LG 6-32 clearance square mounting pad
MT4 PAD_6-32_RECT PAD_6-32_RECT 6-32_SQR_LG 6-32 clearance square mounting pad
QC1 QC: QC: QC: QUALITY CHECK mark on board silk screen
R1 POT-1/4P POT-1/4P
R2 1M TH5MM 0204/5
R3 1M TH5MM 0204/5
R4 10K TH5MM 0204/5
R5 20K TH5MM 0204/5
R6 10K TH5MM 0204/5
R7 51K TH5MM 0204/5
R8 2M TH5MM 0204/5
R9 2M TH5MM 0204/5
R10 100K TH5MM 0204/5
R11 100K TH5MM 0204/5
R12 200K TH5MM 0204/5
R13 200K TH5MM 0204/5
R14 200K TH5MM 0204/5
R15 200K TH5MM 0204/5
R16 200K TH5MM 0204/5
R17 200K TH5MM 0204/5
R18 R-US_VTA55 VTA55 RESISTOR, American symbol
R19 2.2K TH5MM 0204/5
R20 R-US_VTA55 VTA55 RESISTOR, American symbol
R21 1K TH5MM 0204/5
R22 51 TH5MM 0204/5
REV1 REV: REV: REV: REVISION LEVEL mark on board silk screen
SN1 SN: SN: SN: SERIAL NUMBER mark on board silk screen
TEST1 TEST: TEST: TEST: TEST mark on board silk screen
X1 1751248 1751248 MKDS 1/ 2-3,5 Printklemme
X2 1751264 1751264 MKDS 1/ 4-3,5 Printklemme
X3 1751248 1751248 MKDS 1/ 2-3,5 Printklemme
X4 1751248 1751248 MKDS 1/ 2-3,5 Printklemme
X5 1751251 1751251 MKDS 1/ 3-3,5 Printklemme
X6 1751248 1751248 MKDS 1/ 2-3,5 Printklemme
 
And the schematic. . XO's.jpeg
 
The original circuit calls for TL074n not my choice but designers spec.

TI says the TL084 is a drop in replacement(pin compatible) and better so that is what I will be using.


It's a standard basic circuit (not for "audiophile"). Most quad will probably work but TL074/TL084 are the most popular quad out there. TL074 is JFET-input, quad of TL071 while TL084 is BJT-input, quad of TL81. Bipolar (BJT) will work better in this circuit.


I don't work with quad because I don't work with "simple" circuits. But I have plenty of LF347 and OP400.


Your crossover frequency is 1/(2*pi()*sqr(C*C*Ra*Rb)). C=10n, Ra=100k, Rb=200k. So I think you can be flexible with the values, in case you need to change the BOM for a reason.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.