Gainclone 2.1 setup

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I built a Gainclone using 2 x LM3875 chips a year ago which I am very happy with. I have just built 2 full range speakers using Tang Band 3 inch drivers in Ikea salad bowls (you know the sort).

Now, obviously the 3 inch speakers have not got much bass, so I was thinking of building a small sub woofer and running it from a LM3886.

Now, my question is, how I go about 'splitting' the input signal into my sub amp...... Obviously the sub amp will take the right and left channel from the input and combine them into one output.

At the moment, each input is connected to a 4.7uF cap, which is then fed through a 220ohm resister into + of the chip. There is also a 22K resister connected to audio ground after the 4.7uF cap.

So, basically I want to know where I get the signal to put into my sub amp, as surely if I just simply got the input signal from the left and right channel, fed them both through 4.7uF caps and then 220 resisters and then connected them together to go into the chip, would I effectively be making the left and the right channel the same, and therefore my original Gainclone would be outputting a mixed mono output?

Would it be better to use 2 x LM3886 in bridged mode and feed the left channel into +ve of one chip and the right channel into -ve of the other chip?
 
Buy or build a summing amplifier with Low Pass filtering built in.

An MFB active filter is an inverting amp with 2pole low pass filter, all off one opamp.
The input resistor is simply doubled up to create two inputs. These two inputs sum the Left and Right channels.
The output can be fed to an inverting chipamp then to the speaker, or can be fed to a non-inverting chipamp then to the inverted leads of the speaker.
 
Thanks for the reply AndrewT.

Can you explain a little more about the MFB active filter please.

With an MFB active filter, can I leave my original amp as it is and simply build a monoblock with an MFB at the front end and 'feed it' with the same input signal as my original amp?

Thanks again AndrewT, you are always there when I need an answer :)


Is this the kind of circuit you are talking about..Multiple Feedback Bandpass Filter.

If so, I assume I build this type of circuit and use values of components to give me 0 gain, and then feed the 2 outputs into the front of an LM3886 amp?
 
Last edited:
Is this the kind of circuit you are talking about..Multiple Feedback Bandpass Filter
yes MFB active filter.
Note it is inverting.
Note that you double up on R1 to give two summing inputs.
gain = 0. NO!!!! Gain = 1times, or 1.5times, or 0.5times, or any other value you require
Note that the sum function adds the two inputs together.
If the stereo signal (two channels) are fed to the sumer and both consist of the same (mono) bass signal then the output is gain * [channel1 + channel2] you might need gain = 0.5times, if you use drivers and amplifiers all with the same sensitivity.

If you want wideband signals to pass to your speakers you can remove the high pass part of the bandpass. However many builders want to remove the sub-bass signals to prevent damage to their gear. You decide how low you want the bandpass to go to sound right for your situation.
 
Sorry AndrewT, when I said 0 gain, I actually meant no amplification of the signal, so I should have said '1'. Anyway, I see your point with the summing of the two channels, so I would have to work out the actual gain I require from the MFB filter.

I also originally asked how I would connect this to my original amp, would the following work?

MFB.jpg
 
Just looking at a few schematics for parallel and bridged LM3886 amps and I notice that the input is fed into 2 or more LM3886 chips.

Could I not simply build a bridged LM3886 amp, feed left channel to + of one chip, right channel to -ve of the other chip? The output from the 2 LM3886s would then be fed through a low pass filter and to the speaker.

So in actual fact I would have a stereo amp, where one channel is normal, the other is inverted and there is only one output.

Doing this, I don't have to alter anything in my original Gainclone to compromise the quality I get from it already. I could simply split the input signal to both amps. Would this work?

Any thoughts.
 
Last edited:
An MFB is, as the name suggests, a band-pass filter. I don't see how that would help you.

You want a stereo high-pass filter for your 3-inchers and then you can choose from a stereo set of low-pass filters followed by a mixer or a mixer followed by a single low-pass filter for the subwoofer. Something like this. You will have to adapt the frequencies and Q-factors to your needs of course.
 
MFB is not only a bandpass.
ESP shows the application of his bandpass filter.

The MFB can be used as a high pass or as a low pass.

Have you downloaded Ti filter software? The newer versions are not as easy to use as the old versions, I found about a decade ago.
Butterworth Q = sqrt(2)
L-R Q = 0.5
Two cascaded single pole Butterworth gives a 2pole L-R
Two cascaded 2pole Butterworth gives a 4pole L-R.

The MFB low pass is a very convenient way to sum two (or more) channels and low pass filter with presettable gain, presettable Q and presettable Fr each varied independently. Ideal for a 0.1 low bass speaker and very easy to then adjust for variable Q and variable Fr.
 
Last edited:
Thanks for the advice guys. I've just downloaded TI Filter suggested by AndrewT and installed it on my virtual machine on my iMac. I have to say I am even more confused now than I was when I started.

Here is what I understand...

I have a source which comprises of a left channel and a right channel (2 channels). I want to amplify the signals from these 2 channels to 2 speakers and also want to combine these channels and amplify the mono channel to feed a subwoofer. So, I need to take the signals L and R and then feed them into some kind of active filter arrangement. After the filters, I can end up with a Left channel, R channel and a mixed low frequency channel that I can feed into 3 separate amplifiers.

I have taken a look at the TI Filter software, and to be honest, I don't really understand much of it. I'll keep pursuing...


(edit) I just stumbled across this site... http://www.beis.de/Elektronik/Filter/Filter.html

Seems quite interesting and fairly easy to use. Any thoughts.
 
Last edited:
I've been looking at the website that I found and want to know what are the advantages/disadvantages of using more or less in the 'filter order'. Obviously I can see that the higher the order, the more op amps and components are required, but what about sound benefits from this.

I've been looking again at pacicblue's link to the 2.1 filter drawing and I have a few questions.

1. What are the crossover frequencies for the left/right and sub woofer channels (which components help us work this out)
2. Is the 5K pot used to adjust the gain of the sub channel.
3. I assume the sub woofer output is inverted.
4. There are 4 op amps listed in the drawing, is there any major difference between these?

Would there be any advantage of using a MFB type filter over the link that pacificblue posted.
 
Last edited:
Just found the website where bluepacific's link came from Subwoofer Filter Circuit

It says here that the crossover frequency is 200hz, this seems like the kind of thing I am after.

Would I simply connect my GainClone amp circuit to the outputs of this filter with the same 4.2uF cap and 22K resistor, or would this have to be changed. And also, where do I connect my 100k Log pot used for attenuating the amp. On my stereo amp, I have a dual pot connected at the input of my GainClones, would this be put at the front end of the active filter instead?

Thanks again :)
 
The main volume control comes before the gain stage that starts to split the signal into separate frequency bands.

You may need a sensitivity control before or after the low pass filter for the low bass channel. This channel will need to to be level matched with the two channel output. Once level matched that control is generally left unaltered until you change your speakers.
 
2. Is the 5K pot used to adjust the gain of the sub channel.
Yes, that is the sensitivity control Andrew mentions in post #13.

3. I assume the sub woofer output is inverted.
No, because inverted x inverted = non-inverted. The mixer stage is inverted and the output gain stage, too.

4. There are 4 op amps listed in the drawing, is there any major difference between these?
They have different advantages. The LM833 and NE5532 were specifically designed for audio use, so may have a slight edge for the satellite section. But in the case of filters you usually prefer FET inputs over BJTs, so the TL072 or the LF3.. are good choices, too, maybe even better for the subwoofer section. What they have in common are wide availability and reasonable prices.

Would there be any advantage of using a MFB type filter over the link that pacificblue posted.
If you understand the circuit and know how to calculate it, it is possible to save one op amp channel, because you combine the mixer stage with the first filter stage. With dual op amps you have that channel anyhow and would then be left with one unused op amp channel. And you would only have one inverted stage, so the subwoofer output would be inverted.

It says here that the crossover frequency is 200hz, this seems like the kind of thing I am after.
Generally the subwoofer should cross at 100 Hz or lower. With only 3 inch satellites this will of course only work for low listening volume and you may indeed need such a high crossing for louder listening. It is also worth looking at the satellites' natural roll-off. If it is somewhere near the crossing frequency you may get better results with only a 12 dB high-pass or even none at all.

Would I simply connect my GainClone amp circuit to the outputs of this filter with the same 4.2uF cap and 22K resistor
Yes. You should only skip those components when you are 100 % sure that you will never use the amp without the cross-over anymore. A roll-off at 1,7 Hz is quite low for a three inch speaker, so you could use those caps for the subwoofer amp and replace them with smaller and less expensive ones for the satellites.
 
Thanks pacificblue. I've been trying to find out why you can use more of less op amps to achieve 'the same' result. What I mean is, on the website I linked to in an earlier post (here it is again Active High-Pass Filter Design and Dimensioning), there is a 'Filter Order' option which obviously increases the number of op amps the higher the order you select. If you can get a highpass filter with only 1 or 2 op amps, why would you go for a filter order of 10 and use 5 op amps?

Sorry if these questions are silly, but I haven't had the time to read any books on filter design yet, I usually learn as I go, but I'm a little confused here.

If it isn't too much trouble, could you explain each part of the schematic and what each op amp is actually doing, because I have tried a few websites with filter designing guides etc. but cannot work out what the value of each component does in each part of the circuit.

Thanks in advance :)
 
Most opamp filters are 2pole (12dB/octave slope)
Cascade the filters (at the same frequency) and you add the slopes. 2active filters gives 4pole (24dB/octave slope)

Most normal Audio Band filters are in the range from 1pole to 4pole.

Look at unity gain Butterworth active filter.
Compare it to a Equal Component Value Butterworth active filter. This uses two extra resistors but in return it uses equal value components (for improved accuracy) and separates the Q adjustment from the Fr adjustment.
Finally look at the Multiple FeedBack Butterworth active filter. This uses one extra resistor and one extra capacitor, but now it separates gain from Q adjustment and is inverting.

Because all the filters are set to Butterworth and they are all 2pole they all have exactly the same effect in filtering the signal. There is a limit, usually quite far into the stopband where the slope degrades from the theoretical 2pole or 4pole with most filter types. It becomes quite complicated to attenuate this stopband error. For the time being ignore it and do further research later.

Cascade two Butterworth filters and the result is Q=0.5 and this is what Linkwitz Reilly uses for the L-R crossover. The Q of cascaded filters is the mutiplication (not addition) of the individual Q values).

All filters whether passive or active require the input and output conditions to meet strict limits. The usual is zero Rs and infinite Rload. These are fairly easy and cheap to achieve using buffers.
 
Last edited:
Thanks again for the replies. I've just started reading about filters from this site Fundamentals of Active Filters

It seems a good starting point and isn't too hard to understand.

Andrew, you said previously that "The main volume control comes before the gain stage that starts to split the signal into separate frequency bands.", so are you saying that in the Subwoofer-filter-diagram that pacificblue pointed us too, that the pot should be after the first opamp on both left and right channels. What would the difference be if it were before the entire circuit?

I think I'll read up a bit more of active filters and try building a few. I found 10 x LM833 at work, so I have some op amps, must have some caps and resistors hanging around (probably not the right values though!)

Obviously the tolerance of the components is key in these circuits, what tolerance components do you suggest and what type of capacitors are best for filter circuits?
 
If one does not care about accuracy and takes what comes out then use 5% resistors and 10% capacitors.
If you do care, use 1% resistors and 5% capacitors.
If you are a stickler for slope and turn over and channel matching, you're probably in for 0.5% resistor matching and 1% capacitor matching.
 
Most of the time the value will limit your choice. E. g. mica caps usually come in values below 1 nF. Bigger values are hard to get and expensive. Tantalum, like all electrolytics come in µF values. The usual values in active crossovers will be covered by ceramics and film caps. Ceramics come in different qualities. The good NP0 or C0G only come in too small values and the others are no match for a film cap.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.