Rewiring 7-band EQ

[whiterabbit]

I'll see if I can find a good 12v converter tomorrow and hopefully get it in the mail by the weekend.

Most of my time will be spent looking for a more relevant schematic. I'm pretty much doing various google image searches, following similar pictures to their original pages, looking for model numbers and following related source links to forums. If you have any suggestions for a more targeted approach....

I put in the order for a double set of the components you linked, in case I accidentally fry anything during the production process. Should have that in the next week or so.

I can take more pictures of the interior of the tuner panel if that might possibly be helpful. I could definitely send you some pieces to investigate if need be! I'm not sure who else I might send them to... You SERIOUSLY know your stuff!

I will get back to you tomorrow with any updates that I work out!

 

[whiterabbit]

Hey! So I got the components in the mail on Monday, and it's taken me a few days to find the time to get this all laid out according to your schematic.

I plan on reorganizing this and making it more compact and elegantly laid out, but I wanted to check with you first that I understand how everything is fit together.

First image is a regular pic, second has a colored overlay so you can see how I imagine the connections working.

Also I just want to say that in the final product obviously there will be no shorts; I know I can't actually have exposed wires running across each other like they are now- I just get a rough idea of the layout so I could more easily plan out how to make it more compact.

Waiting on my paycheck to get put into my account to buy a converter. And still roaming the interwebz in search of a better schematic when I get a few minutes here or there.

 

[KrisBlueNZ]

That looks OK. It's quite spread out though. You can save a lot of space by mounting the resistors upright instead of lying down.

You might want to try a program called diylc (DIY layout creator). It's a free Java program so it's multi-platform, and it allows you to create a nice-looking layout for stripboard. It might be just what you want.

One more thing to consider with your layout. The "summing nodes" for the two channels, which are pins 2 and 6 of the IC, should be electrically well isolated from other parts of the circuit. That means that if possible, the two resistors should connect close to the IC pin, and the track should be cut just beyond the resistor connections, so there is no unnecessary copper connection to those two pins.

 

[whiterabbit]

Oh, absolutely. I just wanted to make sure I had all the right connections in the proper places before I started bunching everything all together.

Can you describe in a little more detail how the summing nodes should be isolated? The board I'm using doesn't have any metal in/on it, so I figure I don't have to worry about that too much. Is the concern to avoid a short? or interference of some kind? I plan on posting a few pictures of how I've planned the final layout for the board before I start soldering things together

 

[KrisBlueNZ]

Just minimise the total amount of copper on the stripboard that is connected to it. That's all.

 

[whiterabbit]

FINALLY

I got this all soldered together this past Sunday but didn't have the time or energy to work the overlay up until tonight.... Have been spending a good bit of time working on building the new body that this whole project is going to go in. It's looking amazing so far!

Hopefully I did this right. I wasn't exactly sure where there was wiggle room for restructuring, but I tried not to slide stuff around an extreme amount.

Is there a way I can go about testing this without possibly starting a small fire? I *did* buy enough parts to make a full second circuit in case I seriously start melting stuff.

I got a 12v ac > dc wall wart. I think I read somewhere (I tend to do a lot of research right before I go to sleep so things get jumbled sometimes) that the voltage might vary slightly up or down from 12v- so I was planning on testing that with my multimeter at some point to see if it's true or not.

Unrelatedly, I've also started doing some research into building FX boxes (pedals, I guess). I spend a good bit of my time on the weekends doing electronic music production and I quite enjoy the idea of having a small audio rack on my desk with a bunch of simple, custom FX boxes I can route sample audio through and then record. I know (at least at first) that any effects I end up making will pale in comparison to the basic plugins in Ableton, butttttt making my own sounds WAY more fun and satisfying!!

 

[KrisBlueNZ]

That looks good at first glance. Nice work!

It's very unlikely you'll start a fire with it; it's not even likely you'll do any damage to anything even if there's an error. But you can start with the IC unplugged and make some measurements on the IC socket.

With the power conected and your voltmeter negative lead connected to the 0V rail. Check these pins on the socket for these expected voltages:

Pin 3: 6V ±0.5V
Pin 4: 0.00V
Pin 5: same as pin 3
Pin 8: 12V ±0.5V

If that's OK, power down, insert the IC and power up. The remaining pins (1, 2, 6 and 7) should now be 6V ±0.5V.

You may have a problem with noise from the wall wart. If it's the old fashioned kind with the heavy transformer, you get mains hum at twice the mains frequency. This can be fixed by adding a low-value series resistor and a large-value electrolytic capacitor as a filter, just like R9 and C6 in the buffer circuit (although you should use a larger electrolytic - say 1000 µF, 16V). Put this filter straight on the output of the wall wart, before the power wires split off to the graphic equaliser and the buffer board.

If it's the modern style of wall wart, which is very light and compact, the power supply is a type known as a switching supply, and the noise is switching noise, which sounds like whistles and/or hiss, often constantly changing in quality. This is best filtered using an inductor in series with each wire, then two capacitors in parallel across the wires after the inductors. This type of filter can be cascaded if one stage isn't enough.

For the inductors you could use something like http://www.digikey.com/product-detail/en/AIAP-02-221-K/AIAP-02-221-K-ND/3059881 and for the capacitors, http://www.digikey.com/product-detail/en/C420C104K5R5TA7200/399-4491-1-ND/818348 in parallel with http://www.digikey.com/product-detail/en/EEU-FM1V331/P12413-ND/613774

 

[whiterabbit]

w00w00

Great news!

Finally got my lazy ass around to testing the audio buffer circuit. It tested exactly as you said it would, so I didn't screw it up!! I was a little unsure with how I was rearranging it because of my total lack of experience reading circuit diagrams.

I've been reading more about op amps because I don't think I quite understand how this circuit works. I know what it DOES, just... a little fuzzy on the how. The amount of amplification is based on the resistor on the input (1M) divided by the resistor that connects to the output of the op amp (100k) ... so the output is 10x louder (?) than the input?

I guess I'm still not exactly certain how an op amp works in this circuit. It works as a comparator and tries to become balanced, and thereby draws from the power supply which then gets channeled to the output..? I'll get it eventually... the more I work with these kinds of things.

On to thinking about how this circuit will be worked in with the existing components. When I tested the EQ I put it between the panel with the volume pot and the amplifier panel. Initially it sounded like you were suggesting attaching the buffer to the volume pot. Why wouldn't it go between the EQ and the amplifier instead? Or even immediately before the EQ, after the L/R signals are sent out of the control panel?

 

[whiterabbit]

Ah, almost forgot... This is the adapter I picked up:

http://www.radioshack.com/product/index.jsp?productId=13360646

Probably could have gotten one for less online but I was impatient. Is this a switching supply, and should I pick up those parts from digikey you recommended in your last post?

 

[KrisBlueNZ]

That's great! That's an impressive result for a first project.

The buffer circuit consists of two inverting stages. Each stage has a gain of -1, which means the output amplitude is the same as the input amplitude but the signal is inverted.

I'll describe the left channel; the right channel is identical. There is an input resistor, R1, from the circuit input to the inverting input of the op-amp, and a feedback resistor, R2, from the op-amp's output back to the inverting input. The non-inverting input (pin 3) is held at half the supply voltage by the voltage divider R5/R6.

An op-amp with negative feedback (feedback through R2 from the output to the inverting "-" input) tries to adjust its output voltage so its inputs are at the same voltage. In this case, that voltage is half the supply voltage.

A rising voltage at the cirucit input, from the rising part of a signal waveform, tries to pull the inverting input node voltage upwards, via R1. To balance this change and keep the inverting input at the same voltage, the op-amp generates an equal but opposite swing at its output. The same happens in the other direction.

The result is that the op-amp's output is an inverted copy of the circuit's input signal. Because R2 and R1 are equal, the amplitudes are the same. You can vary R2 to change the gain; the gain can be calculated as:

G = - (R2 / R1)

The circuit's input resistance is equal to R1. That's because the inverting input of the op-amp is always held at the same potential (because of the op-amp's action) so the input signal just sees a resistance of R1 to ground. The output resistance is low, because the op-amp has a low-impedance output with significant drive capability. This is needed for the graphic equaliser, which has a relatively low-resistance input. This is why the circuit is called a buffer.

The reason for all the capacitors is that the signals are referenced to the 0V rail but the op-amp must have its inputs and output operating comfortably between its positive and negative supply pins. Op-amps cannot drive their outputs beyond their supply rails; most op-amps can only drive them to within a few volts of each supply rail. So both inputs and the output are biased at half the supply voltage (set by R5 and R6), and capacitors are used to pass the AC signal into and out of the op-amp while allowing a DC voltage difference.

Regarding positioning of the buffer. The buffer needs to go at the input of the graphic equaliser. Where exactly in the signal chain you put them is up to you. Probably the sensible way is to disconnect the signals that feed to the clockwise ends of the volume pot and feed them into the buffer, which feeds to the graphic equaliser, and the return from the graphic equaliser would go to the clockwise ends of the volume pot.

It's hard to be totally sure about this because I don't know the exact way the volume pot is connected in the existing unit. That's why a schematic would be very helpful, even a schematic for a non-identical but similar model.

It would also be interesting to know the resistance of the volume pot. This should be marked on it somewhere. I'm assuming it will be something around 1 megohm, maybe a bit less, which is why I set R1~4 in the buffer to that value.

My guess is that you will have a problem with noise from that switching supply, but try it first and see, unless you're already ordering other parts from Digikey, in which case you might as well get the filter components at the same time.

 

[whiterabbit]

Thank you for being willing to explain things multiple times until I understand! The more information you give me the more I am able to research into the right terms and processes, and put it into the context of this project. It is extremely helpful!

I was initially confused about the purpose of this buffer circuit; I thought it's purpose was to boost the level of the input signal being put through the EQ- which would then decrease the level.

IF I understand this correctly now, the buffer I've built will not change the level of the signal, but push out the input signal with an increased CURRENT so that the added impedance caused by the circuitry of the EQ does not drop the level of the output signal too much. Is that closer to what's happening?

I will look more closely at the volume pot tonight and see if I can trace the route of the ins/outs to make sure I know which I need to reconnect to which end of the EQ and buffer circuit.

I have not yet ordered the components from digikey, although I suppose I may as well do so now so when I get around to hooking all this up and testing it- I won't have to wait on delivery to start working on decreasing any switching noise.

I've been meaning to ask- is this sort of thing (audio circuitry and the like) your specialty, or something you've worked with much? Or has this all been general electrical knowledge? I have a lot of ideas on upcoming audio based projects that I'd kill to have a sounding board on, once I have time for them...

 

[KrisBlueNZ]

I was initially confused about the purpose of this buffer circuit; I thought it's purpose was to boost the level of the input signal being put through the EQ- which would then decrease the level.

IF I understand this correctly now, the buffer I've built will not change the level of the signal, but push out the input signal with an increased CURRENT so that the added impedance caused by the circuitry of the EQ does not drop the level of the output signal too much. Is that closer to what's happening?

Yes, that's right. It's called a buffer because it doesn't provide any voltage gain; if it had voltage gain it would be an amplifier, or an attenuator if the gain was less than 1.

It provides current gain, or impedance reduction. It has a high-impedance input that can be driven properly by a signal from a high-impedance circuit (most toob circuitry is high-impedance) and provides a relatively low-impedance output to suit the solid state circuitry in the graphic equaliser.

You may want to adjust the gain of the buffer circuit so that its output level is better suited to the voltage levels used by the graphic equaliser. If you do that, you'll have to provide an equal and opposite gain on the returns from the equaliser. If you add gain to the buffer, you can attenuate the return from the equaliser with just resistors, wired as a voltage divider.

Do you have, or have access to, an oscilloscope? It's the best way to measure the signal levels.

I have not yet ordered the components from digikey, although I suppose I may as well do so now so when I get around to hooking all this up and testing it- I won't have to wait on delivery to start working on decreasing any switching noise.

You may not need them. It depends on the quality of the adapter. The adapter will have SOME filtering in it already; it may or may not be enough.

I've been meaning to ask- is this sort of thing (audio circuitry and the like) your specialty, or something you've worked with much? Or has this all been general electrical knowledge? I have a lot of ideas on upcoming audio based projects that I'd kill to have a sounding board on, once I have time for them...

I have worked with audio, but not as a specialist. I'm happy to try to help with your ideas. Start new threads for each project and send me a PM so I don't miss them.

 

[whiterabbit]

I've ordered the parts you've suggested- just in case; I figure if I don't end up needing them for this project I can always find another use for them in a later project when the need arises!

I do not have access to an oscilloscope, although I've already started looking into them. Do you have any suggestions? I don't know what I might even possibly need in the future, so it's hard for me to pick between the available features.

In the future, I hope to eventually start building basic electronic synthesizers... I don't know how much that would impact the requirements of an oscilloscope. I imagine that ANY oscilloscope would be able to show me what type of wave form a given circuit is generating, but I honestly don't know.

With regard to the record player rebuild-

I looked at the volume pot and didn't immediately see any clear markings. Tomorrow I'm going to inspect all of them with a magnifying glass and a very bright light to see if anything turns up. If all else fails, I should be able to measure the resistance with my multimeter, right? (BTW, I think I'm getting the hang of using that thing... derpppp)


I've draw up a few preliminary ideas for my other projects and I will go post them now so I can start getting some feedback on the concept and design before I start buying parts.

 

[KrisBlueNZ]

If you're working at audio frequencies, you don't need a flash oscilloscope. I'd go for a second hand dual trace analogue scope with 20 MHz bandwidth. I wouldn't pay more than about USD 150 for it. It's a bonus if probes are included, as these are not cheap. Avoid Chinese brands. I like Hitachi scopes, and other reputable brands are Tektronix, Trio/Kenwood, B&K, and HP/Agilent.

You might rather buy a digital scope or an acquisition unit that plugs into a computer. There's a variety of them around and I haven't used any of them. I've heard that the cheap acquisition boxes are not much more than toys, but they might be usable for audio-only work. Dave Jones of eevblog (http://youtube.com/user/eevblog) has done some videos on them, and you can probably find reviews and comparisons.

You can measure the pot from end to end, assuming there are no other resistances with DC paths to it. If in doubt, disconnect anything that connects to the clockwise end, turn the pot fully anticlockwise, and measure from end to end.

Cool! I'll have a look for your new threads.

 

[whiterabbit]

Okay... so tonight I disconnected as many leads as I safely could from one of the sections of the volume pot and tried to measure the resistance. It's kinda hard for me to tell but it looks like there may still be one wire connected to the back/bottom of the pot which I couldn't really access well...

So looking at the pot from the top so there are 3 leads coming out with the shaft sticking out/away.

With my multimeter on 20M ohms I measured between the left and center leads: 0 on counterclockwise, 4.55 when rotated all the way clockwise.

right and center leads: 4.6 both clockwise/counterclockwise (no difference)

I also tried to move the big capacitor off the bottom of the pot but didn't want to disconnect any more leads and risking ruining things. I hope this was a helpful test...

 

[KrisBlueNZ]

OK, that sounds like it's a 5 megohm pot. That's higher than I expected. The buffer board has a 1 megohm input resistance. This may cause some loss of bass frequencies. You could change the four 1 megohm resistors on the buffer board to 4.7 megohms if you notice a problem with loss of bass.

 

[whiterabbit]

Hm, dunno why that posted twice...

Anyways- assuming I measured this right... you're saying increasing the 1Mohm resistors to 4.7M ohm resistors will increase the bass level?

Is there a possibility that this would harm the speakers in any way since the cones are 50 years old (again assuming that they haven't been replaced or serviced). Maybe this is something that I will have to just experiment with.

On the topic of the db level of the output- I did notice that when I had the EQ running in series with the rest of the components, even with all the bands pushed to the max on the EQ there was a huge drop in volume when the EQ was not being bypassed.... Initially I thought the buffer circuit you posted would handle that but it sounds like I need to change a few of the resistors on there to increase the gain as well as the current- is that right?

 

[KrisBlueNZ]

Yes, if you notice a lack of bass you should increase the four 1M resistors to 4M7. It's not ideal to have such high resistances at the input of an op-amp but it will be OK.

I doubt anything will harm the speakers. The amplifier will soft-clip at a level far below that level.

Maybe you do need to increase the gain. I would expect the EQ to have 0 dB gain (no gain, no loss) with all the controls at flat. When you bypass it, are you bypassing just the EQ, or the buffer AND the EQ?

If you need more gain, we should probably change the buffer design by adding a second dual op-amp to provide the gain, with a high-impedance non-inverting buffer stage. Are you prepared to do that?

 

[whiterabbit]

I spent last night temporarily connecting the buffer board and EQ to the wall wart. Instead of hooking the board to the volume pot I decided to try putting it inside the EQ itself. It just fits inside the empty space in the case. This seemed more secure and compact. I disconnected the leads from the L/R RCA in and put the buffer circuit in series that way.

I tested the sound quality this morning. The buffer works great! I don't think increasing the gain any more will be necessary.

However, I think it may be necessary to switch to the 4.7 megohm resistors. I hooked up my laptop to the input reserved for the record player with a 1/8" phono > RCA cable. I tested a few songs that I'm quite familiar with which I know to be quite bass-heavy, and with the help of the EQ I was able to get them sounding much closer to how they should sound by boosting the bass a few db and dropping the high end very slightly.

I also noticed a rather prominent ambient hum even with no music playing. It didn't sound like clicks or pops like what I expected based on your explanation of switching noise from the wall wart. Playing around with the EQ indicated that most of it was at both the high and low ends of the frequency range. So, <150Hz and 6KHz< or thereabouts. I wanted to get your opinion on resolving that.

Should I add the inductors and capacitors to the power supply and see if that helps any? Or is there a more targeted approach?

 

[KrisBlueNZ]

It can be hard to know how the hum is getting into the signal path. It's most likely coming from the power supply; if you can find a 12V battery to test with, you can find out for sure. You'll need a battery made from C-cells or D-cells. Don't use AA cells; they can't supply enough current. Don't use a car battery; they're dangerous because they can supply a huge amount of current. Something in between would be suitable. If you don't have any suitable battery, don't worry; it's most likely coming from the power supply.

Assuming it's the power supply, yes, add the inductors and capacitors. Try just one stage - that is, an inductor in series with each wire, then two capacitors (one of each type) in parallel connected across the supply after the inductors. This may fix only the high-frequency noise, or it may fix both.

How many inductors and capacitors did you get?