EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: Michaelaudio on February 08, 2021, 03:10:11 pm
-
So i build a BT speaker with a Selfmade preamp and 3 TPA3116 boards. 2x a stereo Board 1x Mono Board from eBay. When i Connect the Outputs of my Filter to the Inputs, the noisefloor is really loud and when i Turn Off power, i have a Pop. When there is nothing connected to the boards its all quiet (besides u turn the pot. in the middle then its noisy too, dont know why) and i have no turn off pop. So i guess the noise is loading some Caps that are causing the pop when turning Off, but why is it so noisy? I guess i missed something on my Preamp, but what?
Boards:
https://www.ebay.de/itm/173916201431 (https://www.ebay.de/itm/173916201431)
https://www.ebay.de/itm/284035213328 (https://www.ebay.de/itm/284035213328)
-
You can make a screenshot with hold alt + print screen button.
Good luck
-
Not sure what you have tried to achieve with those sallen key filters, but it is useless if what you wanted was an active crossover. Will not simply work as you've drawn it by multiple reasons!
Mainly because:
a) HF, MF and LF channels after combining back together will not result in the original signal.
b) three-way crossover needs compensation of phase delay, can't use separate filters, need to use 2 chained two-way splitters
c) can't feed a Sallen-key filter with a high impedance source, if the filter is not designed with that in mind - which I guess it is not (I mean your SUB LPF).
d) can't supply the bias to the Sallen-key filter by slapping a random voltage divider to one of the resistors
d) lacking any radio interference input protections.
Look & learn:
https://www.linkwitzlab.com/xo_eq.htm (https://www.linkwitzlab.com/xo_eq.htm)
https://sound-au.com/project125.htm (https://sound-au.com/project125.htm)
https://sound-au.com/project78.htm (https://sound-au.com/project78.htm)
https://sound-au.com/project09.htm (https://sound-au.com/project09.htm)
https://sound-au.com/project148.htm (https://sound-au.com/project148.htm)
https://sound-au.com/project81.htm (https://sound-au.com/project81.htm)
https://sound-au.com/project123.htm (https://sound-au.com/project123.htm)
Designing an active 3-way crossover is not just slapping some filters together. Disapointment guaranteed otherwise! 3-way analog crossover has zero "simple solutions".
But you could definitely make a simple 2-way one, just by using a highpass filter and difference signal for the LF path, such as that: https://sound-au.com/articles/derived-xovers.htm (https://sound-au.com/articles/derived-xovers.htm)
-
I quick read some articles, and i have some stuff to do i guess.
I still have some quesions:
- what do u mean with high impedance source in the Bass Filter? The 2 resistors that tie l and r together?
- guess U mean the resistor devider in the Bass filter? I guess its done that way If u have to use single supply? read it in a PDF from TI about Filters with single supply, and it works
I def. need a 3 way Crossover, since the speaker is complete and works fine, beside the noise, that i causing a Turn of Pop. It even seems that the Outputs sonst have much noise, which is weird, so im not sure If the Filter ist the Main Problem. But maybe it is, and i have to make a new one. I'll draw something and Show my schematic here then.
-
What Yansi said.
I'm at a loss at what your filter circuit is supposed to do. Just at a glance I see:
HP
HP
HP-LP ->probably BP?
HP-LP ->probably BP?
HP-LP ->probably BP?
???
Please explain your intentions, that will make helping you possible.
-
Right, its 2x a High Pass (L+R) and 2x a band Pass for the mids and one for the Bass.
-
Right, its 2x a High Pass (L+R) and 2x a band Pass for the mids and one for the Bass.
Understand.
But why do you have a BP for the bass? Do you have a subwoofer as well?
-
OK, you need to do two things first:
1:
Swap the lower LP and HP filters. This will eliminate the (c) problem that Yansi mentioned. The parallel connection of R29 and R30 will then replace R35 of the LP filter.
2:
Your single-supply approach with a distributed "virtual ground" through several voltage dividers is not the right solution. You need to make one low-noise bias source, preferably active, that is stable and low noise.
In general, using dual supplies and a real ground works much better.
-
Isnt that the way how u tie left and right Signal together? At least i have seen it in every Subwoofer Preamp so far. Also when i swap LP (170hz) and High Pass (27hz) then i would just get frequencies below 27HZ wouldt i?
Some kind of low noise regulator? But i would have to use many of Them, since i cant Just tie them all together.
I had the noise Problem before with another Crossover, but instead of deviders i used Z diodes for the biasing. I seapped all diodes with resistors and the noise was gone. Maybe the 7815 is too noisy
Id prefer using a real ground, but its battery powered speaker with single supplied amp boards.
-
Isnt that the way how u tie left and right Signal together? At least i have seen it in every Subwoofer Preamp so far. Also when i swap LP (170hz) and High Pass (27hz) then i would just get frequencies below 27HZ wouldt i?
Some kind of low noise regulator? But i would have to use many of Them, since i cant Just tie them all together.
I had the noise Problem before with another Crossover, but instead of deviders i used Z diodes for the biasing. I seapped all diodes with resistors and the noise was gone. Maybe the 7815 is too noisy
Id prefer using a real ground, but its battery powered speaker with single supplied amp boards.
Yeah, had a brain fart, sorry.
Just scrap the Bass HP filter completely and stay with the LP. Then you can sum L+R easily.
For the DC biasing, use an opamp (you have one spare now) as a voltage follower with a nicely decoupled input voltage divider.
This will supply the bias for the two input amps and R8, R15, R19 and R25. R33 is out anyway.
Scrap R2, R3, R7, R9, R14, R16, R18, R20, R24, R26 (and of course R32, R34)
-
When i tie R8,15,19, 25 together, wouldt that have an effect in the other Filters? The Signal in R8 is a different one than from R19, wouldt they influence mutually?
-
When i tie R8,15,19, 25 together, wouldt that have an effect in the other Filters? The Signal in R8 is a different one than from R19, wouldt they influence mutually?
In a normal circuit with bipolar supplies, those resistors would be grounded. What you need is to provide a "virtual ground" for them, and this is done through a voltage follower with low output impedance supplying a bias voltage to them.
Single-supply design is not easy.
-
I would add at least the blocking capacitors at the dividers, something like 10-100uF..
-
Yes, when i add decoupling Caps, the noise is reduced. I could live with that amount of noise, its pretty quiet. But the problem ist i have still s huge turn off pop, and dont know how to get rid of that.
@Benta, yes but when i tie all the Points together to one Voltage followern Output, that wouldt be so good bc of the Signals of the different stages, i guess?
-
Yes, when i add decoupling Caps, the noise is reduced. I could live with that amount of noise, its pretty quiet. But the problem ist i have still s huge turn off pop, and dont know how to get rid of that.
That's caused by the capacitors charging, when the power is applied. Using a bipolar power supply will help, but the only way to eliminate it is to use switches (relays or analogue switch ICs) and a timer to only connect the output, after the power has been applied for long enough for the capacitors to charge. Even then, there still may be a faint click, as capacitors never really charge to 100%.
-
@Benta, yes but when i tie all the Points together to one Voltage followern Output, that wouldt be so good bc of the Signals of the different stages, i guess?
If your voltage follower is low impedance output (as it should be), the different signals won't see each other. This way to do it is standard.
-
Suggest to original poster to research and read the many fine texts and papers on crossover and filter design
All of this is very well documented.
Finally most well designed audio preamps and crossovers use a split power +/- 12 or 15V and have an FET, relay or other turn on transient muting ciircuit.
Bon Chance
Jon
-
@Zero999, i dont have a Turn ON pop at all, its a turn OFF pop, probably couse some Caps are discharching at the Inputs, but it only Happens when i Connect the outputs of the Crossover.
@Benta, now i got it.
As i said, i cant use dual supply, and i have a Chinese amp board that works with single supply too, and it hast absolutely No Turn ON or Off pop
-
@Zero999, i dont have a Turn ON pop at all, its a turn OFF pop, probably couse some Caps are discharching at the Inputs, but it only Happens when i Connect the outputs of the Crossover.
Don't you think it may be due to the output caps (C4, C8, C13...)?
-
I think so, but i cant get rid of them ofc, or maybe use 1uf instead of 10?
Edit: Just shorted C4, still a turn off pop, doesnt seem to be the Output caps.
-
Here's my preamp circuit - HP and LP filters at around 80Hz for a subwoofer. Digital volume control - I specified too slow optos - should have been 6N137, and the main channels could benefit from an additional gain stage on the output. Note the half supply bias generator, and the shunt resistors on the outputs.
(https://www.eevblog.com/forum/projects/selfmade-preamp-noisey/?action=dlattach;attach=1171498)
-
That Looks more professional :-+ I have digital Volume Control too, but its a litle wondom (Sure) Board, which comes at verry first, so the Outputs of this Board are going into the filter Board. I'll try the shunt resistors, If that isnt going to Work, i'll have to make a new Board i guess, then with a low impedance source for the bias.
Edit: i Just found Out my amp boards already have a shunt, 1k in the stereo Boards and 20k on the mono board. I wanna know what is causing this Pop, its not the noise and cant bei the Caps with a 1K shunt across them. What is it?
Why isnt ur Input ground actually going to ground? I dont quite understand the First Input Stage.(yea its a voltage follower but whats with the Inputs ground and C9/7 and R22/23)
And how did u safe that Pic? when i use snipping tool the pic is too tiny and U cant read anything.
-
I thought about switching the outputs off before Power is removed, would this Work? Note that the Der in the Signal path is a N-ch ofc.
-
Maybe this will do it.
-
Why isnt ur Input ground actually going to ground? I dont quite understand the First Input Stage.(yea its a voltage follower but whats with the Inputs ground and C9/7 and R22/23)
And how did u safe that Pic? when i use snipping tool the pic is too tiny and U cant read anything.
The input stage is to avoid ground loops with a head unit - it's basically a differentail amplifier. I copied it from a commercial car amplifier circuit.
Kicad schematic has File - Export - Drawing to clipboard. I then paste into a new image in Irfanview, save as png
-
Ah okay.
I guess now this should Work. I delayed the Turn ON/Off time a little, so that the fets can Turn ON or Off the signal paths before Power is removed.
-
Here we go again. It doesnt work. I get a weird Waveform Out of the First Stage, and only ~ 2.5Khz and above. Why? I can only imagine that its the biasing, but what is wrong with it?
-
Looks like a mistake around your input buffer.
The input signal should connect to the + input. The 47K resistor should connect between the + input and 7.5v
Your input signal is being shorted to the output of the buffer but above 2.5KHz the output impedance of the opamp rises and some distorted signal gets through
-
Oh dear ofc! As a voltage follower should be..
Why doesnt it Work like that with AC voltage? BC there is no voltage drop in the Feedback Loop?
-
I don't understand your question. Can you rephrase it?
-
Why u cant use the - Input with the voltage follower. I die some Research and i got it now. Refreshing my knowledge, the Output would Go lower, when the - Input is a Higher voltage than on the + Input, therefore u wouldnt get much in the Output, Just how it is in my circuit rn.
-
Filter works now, but there 2 Problems left.
1. No Output in the LPF, after the 2 10K resistors, the Signal is pretty weak, and in the Output u have no Signal at all. But No short anywhere.
2. When i Connect R28 to the +Batt (i had to replace R28 with a Diode, ofc) then the BSS84 wont turn off when OUT (p6/U6) is high. Its still 20V at the Output. When i Connect the Diode (R28) to the source of Q2, then BSs84 turns Off, but then the Signal Mosfets are only switching Off, after Q2 switched Off, and i want it to be vise versa. But for that i have to Connect R28 to the +Batt. Verry Strange
-
You have a direct connection from C22 to 7.5v
That's shorting your signal out. Replace it with a large value resistor. That should get you sound.
The whole arrangement is a little strange though
-
Oh, i placed c22 bc i didnt want that current is flowing constantly theu r22, but shouldnt be a Matter i guess, i will try that.
Yea its Something different, but since i didnt know how to get rid of the Turn off pop with another way, i came up with this, and it should Work. But somewhere the BSS84 gets a Voltage on the Output, when there shouldnt be any Output .
-
By "strange" I was refering to the filters
why dont you build your 33Hz filter around U1C/U1D?
I think it will be more benificial to have that filter in the common signal path in terms of overall phase response
ethier way , R19/R20 should be connected directly to U1C/U1D and not after C3/C8 which is part of another filter - sort of as you had it before but increase the value to 20K
For active sallen and key filters you dont need to scale one set of values to be 10 times the oter like you'd do with passive filters
-
Tbh, that didnt csme to my mind, yes would have been Vetter to do it that way.
Oh my Bad, i corrected that in the current schematic, there ofc connected directly.
Now im Sure i'll get the LPF to Work, Just the weird Mosfets behavior left.
Edit: Nope, C22 isnt the fault. Somehow the 7,5V are shorting the Signal out. When i connect R19/20 directly to R22, the Signal is completely gone. I dont get it, thats the Same with the other 2 LPF of the midrange, and its working fine there.
It must be the OP amp, bc in this Case the 7,5 arent Busses with the audio Signal. Maybe i'll Change the verry First Input high Pass to the 33hz, then Connect R19/20 directly to R22 and removed the 7,5V from R22, then it should Work.
-
If you look at your schematic closely, you'll see it's not the same
I mentioned in an earlier post that your shorting the signal out with that connection to 7.5v
You dont need R22. You can kill two birds with one stone by using R19/R20 to mono the signal and act as the first resistor to the LPF
You do need a DC path from the + input of the opamp to 7.5 and the easiest way to do it is
remove C21, R21, C22, R22 (and that 7.5v connection) and connect the common of R19/R20 to C23/R23
-
Got it, thanks. Will Test it tomorrow, but i think it will work finally.
-
So the Filter works now, but BSS84 doesnt turn off when i Press the Push button. The Gate gets imidiatly a high Signal, but the drain stays high. I removed R29, but that didnt Changed anything. Seems +5V isnt enough to Switch it Off, guess i need the Same voltage thats in the source.
Edit: i had a brainfart, i need to do it Just Like the mosfet that switched the Power... That solches this Problem.
Only one Thing left: The preamp is really noisey! When i Put a 220uf in the Output of the Voltage follower (for biasing) it Just gets rid of the higher noise. With resistors deviders and a 10uf in each one, i hadnt noise at all. Why is this one so noisey?
-
I'll Post that Here too, See the updated schematic below.
When i push the Button in the On-state, Q8 doesnt turn off emidiatly, its drain stays at 7.5V till C38 is discharged. But that Happens ONLY when i connect any Output of a Filter to an amp, when there is so Connection to an amp, it Just works fine, and Q8 Turns off emidiatly. WHY? i dont get it, it makes No sense
-
I think you could simulate your "amp only" problem by connecting a resistor of say 10K across one of the output