Top notch electronics guru's needed

[Waterlogged]

I'm half surprised you haven't hit up DIYAudio.com for this project. somma them guys n gals are geniuses. lots of old-hands and engineers over there from all walks of audio design and implementation.

defect9. . . . .where the hell was this post when I started this topic?

First time I've heard of that site. I poked around some and wow...those folks are a little scary the way they debate the finer points of electronics (like metallized vs foil foil caps). Kind of like the way some of the LC guys debate pump or fan noise.

I'm pretty happy with the way things have turned out though, I just have a major decision to make on if I go ahead with the original plan to prototype using the breadboard and then do the real one or dump that idea and go to one of three plans.

Plan one is to go ahead and finish getting the parts I need to finish the breadboard and use that as is.
Plan two is to build a custom board and use the parts that were to go on the breadboard.
Plan three is to drop all the through hole parts and replace everything possible with a custom SMD board design. I'll still need the power amp chips I got but mounting them on SMD pads isn't going to be a problem and should actually make them sturdier for mounting the water block.

I'll probably have to price out each option to make the decision though.


Got two questions for anybody that can answer it. 1. Is dropping from a 0.12uF aluminum electrolytic cap to .1uF aluminum electrolytic cap going to make a noticeable difference for the one channel? I'm thinking that I likely won't be able to tell the difference. 2. I've been lining up Nichicon KW series for the aluminum electrolytic caps, would these a good choice?

[pby5cat]

The change from .12uF to .1uF could change the center fequency of that channel slightly, but most likey unnoticable.

Use Nichicon FM series due to lower impedence, then fill in with FC or HE series if need be.

[Waterlogged]

The change from .12uF to .1uF could change the center fequency of that channel slightly, but most likey unnoticable.

Use Nichicon FM series due to lower impedence, then fill in with FC or HE series if need be.

I can't find Nichicon FM or FC series...did you mean Panasonic? Also, is lower impedence the only reason to choose a general purpose cap over the audio grade (KW's) caps? Not trying to be a smartass, I'm genuinely curious why a general purpose would be recommend over a cap that was made for the purpose I planned to use it for.

1 more Q about the 0.12uF, would changing from an Aluminum Electrolytic to a Polyester Film affect the way the band sounds any? I'm asking because I found them in Polyester film.

[pby5cat]

Oh im sorry, all the "F's" are panasonic. The lower the impedance, the more current can go though the cap and lest impact the cap has on the circuit. The impedence is called the ESR, Equivelent Series Resistance. So when mobo companies say they use Low ESR caps in the CPU power circuitry, they are increasing the effeciency and lowering heat dissipation of the caps and inducters. Say if a normal cap is .8ohms and u have the CPU using 80Amp split between 8 phases, thats 10A per cap. so P=IV each cap has to deal with 8 watts. If that ESR was .08 ohms, the cap dissipates 0.8 watts.

Electrolytic to film is about the bigest change possible in caps. The films mostly have a lower ESR. When it comes to sound the ESR changes with frequency. Im not shure if you would know the difference, try both?

[defect9]

while you're pricing out options, consider something like this for your amp stage, as it's primarily all SMDs except the things that would be far too expensive to SMD. You can run it off the computer PSU directly, or off a meanwell switching psu (which is basically what your computer psu is, only the meanwell is obviously smaller). sounds amazingly clear and clean, even with floorstanders. and even running the bugger for hours on end with punishing music through 3-way floorstanders, it gets barely hotter than room temperature.

I still haven't found anything right for your EQ though. considered mentioning some of the smaller car-based EQs you could gut, but probably no better than what you've already found. just incase, here's the ones i found. option a option b

Oh, and polyester film is generally considered to sound way better than electrolytic when comparing same capacitance. basically consider electrolytic to be the least desireable in the audio path, but an excellent choice when used in a PSU filter for its polarity properties (or something like that, i'm no guru).

[Waterlogged]

Oh im sorry, all the "F's" are panasonic. The lower the impedance, the more current can go though the cap and lest impact the cap has on the circuit. The impedence is called the ESR, Equivelent Series Resistance. So when mobo companies say they use Low ESR caps in the CPU power circuitry, they are increasing the effeciency and lowering heat dissipation of the caps and inducters. Say if a normal cap is .8ohms and u have the CPU using 80Amp split between 8 phases, thats 10A per cap. so P=IV each cap has to deal with 8 watts. If that ESR was .08 ohms, the cap dissipates 0.8 watts.

Electrolytic to film is about the bigest change possible in caps. The films mostly have a lower ESR. When it comes to sound the ESR changes with frequency. Im not shure if you would know the difference, try both?

OK, I somehow understood what you said ...it's just...I can't find impedance values for the KW's on the PDF (Link), they have an "impedance ratio" though but, I can't compare it to the Pana's. On top of that, none of the series you listed go as low as I need them to for 3 of the cap values...but the KW's do. That was one of the 2 reasons I was looking at them, the other being they were "audio grade" which from what I understand (maybe mistakenly), are supposed to have decent impedance.

The GEQ has the following AE caps:

• 0.12uF 16V
• 0.22uF 16V
• 0.47uF 16V
• 1uF 16V
• 2.2uF 16V
• 4.7uF 16V
• 47uF 10V
• 47uF 16V
• 100uF 10V
• 1000uF 16V

The original plan was to replace all these with 50V Nichicon KW's but now I'm not sure.

while you're pricing out options, consider something like this for your amp stage, as it's primarily all SMDs except the things that would be far too expensive to SMD. You can run it off the computer PSU directly, or off a meanwell switching psu (which is basically what your computer psu is, only the meanwell is obviously smaller). sounds amazingly clear and clean, even with floorstanders. and even running the bugger for hours on end with punishing music through 3-way floorstanders, it gets barely hotter than room temperature.

I still haven't found anything right for your EQ though. considered mentioning some of the smaller car-based EQs you could gut, but probably no better than what you've already found. just incase, here's the ones i found. option a option b

Oh, and polyester film is generally considered to sound way better than electrolytic when comparing same capacitance. basically consider electrolytic to be the least desireable in the audio path, but an excellent choice when used in a PSU filter for its polarity properties (or something like that, i'm no guru).

The topic started off as something of a moving target. This is because I wasn't sure which direction I really wanted to go in on this project due to my lack of knowledge on the subject. I finally settled on consolidation reconstruction (shrinkage of the PCB to be more precise) of the GEQ I have here with newer components (there will be some selective theft along the way unless I can find replacements for the transformer and slider pots). This guarantees that I get a working product in the size I need the first time without trial and error (well, we'll see about the error part later)

If that's true about the film's, why don't I replace everything I can with them instead? That'll take care of the 6 lower valued caps and leave me with 3(4) AE type capes to source.

[defect9]

If that's true about the film's, why don't I replace everything I can with them instead? That'll take care of the 6 lower valued caps and leave me with 3(4) AE type capes to source.

if the electrolytics you're replacing are polar caps (having a distinct positive and negative) which most electrolytics are, the caps you replace them with must also be polar.

TBH, unless you have a really, really good system (every part of the chain, basically, with super sensitive speakers) to hear the difference, even though its considered "a big difference" in the audio world, it's only noticable in close monitoring A/B testing. also where in the audio chain it is will also affect how much difference is noticed anyway. if its tying to ground instead of as a pass-through, there will be less importance on it. still want a good cap though, and bang for buck the panasonics mentioned are great, i love em.

most times it seem to me AEs are picked for ease of sourcing and cost, and when polarity matters.

[Waterlogged]

if the electrolytics you're replacing are polar caps (having a distinct positive and negative) which most electrolytics are, the caps you replace them with must also be polar.

TBH, unless you have a really, really good system (every part of the chain, basically, with super sensitive speakers) to hear the difference, even though its considered "a big difference" in the audio world, it's only noticable in close monitoring A/B testing. also where in the audio chain it is will also affect how much difference is noticed anyway. if its tying to ground instead of as a pass-through, there will be less importance on it. still want a good cap though, and bang for buck the panasonics mentioned are great, i love em.

most times it seem to me AEs are picked for ease of sourcing and cost, and when polarity matters.

Crap...forgot about the polarity. That kills that idea...hmm, or does it? Looking at the current PCB, I have 3 of 10 bands that use film/ceramic caps in the RC filter circuit, 1 that uses all film caps in the RC filter circuit and the other 6 bands use film/AE caps in the RC filter circuit (ref post #20 for PCB image). I wonder if I can get away with replacing just the AE caps that are part of the actual RC filter circuit with films (these would be the 6 I suggested replacing earlier) and use the Pan's as suggested for the rest...or I could use the KW's for the lower values and the Pan's for the rest...but then again, I see that the KW's are available in lower voltages for some of the caps I need...GAHHH! Too many options!

*Deep breaths*. . . . .




*sigh*

So let me see if I got this right...If I go with all KW's (like in my original plan), I likely won't be able to tell any difference in the output but I'll be using slightly more power than if I use a mix of other general purpose caps...this about right? Seeing that all the caps I'm looking at are much newer (both in actual usage and materials/design) than the ones on the GEQ I'm basing this design on, the numbers should all be a little bit better shouldn't they? I mean we are talking about a Pyramid GEQ from the mid/late 80's. I can pick up a brand new Pyamid for like $30 US so, they aren't really considered "good kit" to start with.

[defect9]

the difference between caps from the 80's and caps from now is mostly just age and ESR (yes this is a generalization). Caps tend to become more volatile with age. replacing them basically updates their expiration date. honestly while manufacturing techniques of caps and all other audio circuit parts have improved, circuit design is still generally the same. A filter is a filter, whether using ICs or discrete components.

What makes 'great kit' versus 'radio shack base model' is in the tolerance of the parts used, and any extras in the circuit design to combat potential problems (which is an entire book in and of itself). it costs more money (generally) to guarantee better tolerances. That's why really good kit with .1% (or better) tolerance resistors and 1% or better tolerance caps cost so damn much. It means either the technique has to be ultra-rigid and controlled, or the parts have to be more rigorously tested and binned.

if pyramid uses +/- 1% resistors and +/- 10% caps (common), by the end of the audio chain in the EQ if you got completely unlucky and the values were nearly as far off as possible, you could have differences of beyond 3db (which in the wrong area of the audio band would be noticable without direct A/B comparison). However, as even Pyramid has basic testing standards, that would only happen if their test equipment malfunctioned or their tester was drunk/high/getting fired. Yes, each one will differ based on the batches of parts. and law of averages sorta states that odds are as long as they tested it to be within standards, it won't be noticably retarded. But as long as it falls within the acceptable tolerances (which pyramid still cares about, even if it's a budget item), you're gonna get a good piece of kit. I just wouldn't expect it to be the same as something you'd pair with $4000 speakers.

As this is portable and you're not building a quiet room specifically for listening to it, you very likely wouldn't hear that subtle stuff anyway. The EQ itself will shape the sound to your liking way more than your choice in caps. additionally, if you'd be getting the new pyramid and putting everything on a new PCB for space, you're probably just as good in keeping what you already have, replacing caps, replacing pots, and shrinking the footprint. 25v caps are usually a good compromise for keeping voltage overhead in a 12v system and reducing size vs the 50v. for some reason i thought there were also 16v caps out there.

[pby5cat]

I agree with everything defect9 has been saying about the caps.

[Waterlogged]

Well, I know for a fact all the resistors are 5% because I purchased exact replacements. The film caps are all "J" so their tolerance is 5%. I haven't gotten in depth with the ceramic's yet but, a couple of them say "K" on them so those are 10% and it's probably a safe assumption the rest are 10% as well (feel free to slap me for assuming ) The AE caps are going to be a bit trickier. They are made by a company that has "Chief" in the name (perhaps full name?) as that is the only identifying mark on them outside of temp, voltage and uF values. I've done a lot of searching and can not find anything about this company or their caps online. I'm hoping you guys will tell me that the color of the cap can help determine the tolerance value as I have 2 different colored caps (of same voltage and uF, 4.7uF/16V) on the board. The colors are a light blue and a lilac(or perhaps violet?) purple. 4 of the caps are blue, 2 of those are part of the filter circuit between the input wires and the first LM324 (the 1 that's only half used [marked #1 in pic]) and the other 2 are between the output of that LM324 and the power amp chips. The 2 purple ones are part of the 30Hz band filter circuit. Here are some pics of the caps. Maybe you guys can help me figure out tolerance values for these stupid things. I also have 2 pairs of caps that I can't pinpoint the manufacturer name/logo (looks like "CSE" to me), again can't find anything on this brand either.

Warning: First image is very large, do not open if you have a slow connection!



Honestly, this constantly having to fight history to find the info I need is really starting to piss me off. I'm half tempted to throw in the towel on this GEQ and get a brand new one and backwards engineer that one. With any luck, it'll likely have a similar layout with caps in values I can actually get relatively easy. Possibly save my sanity for the paltry sum of $27 + shipping. What do you think...should I?

[defect9]

you could always search the net for a 5-7 band schematic as well. instead of having to backwards engineer someone elses work, you simply translate the schematic onto a piece of breadboard or into some PCB software.

why physical eq instead of software eq?

BTW, you don't have to match exact tolerance for a circuit like this. tolerance is basically manufacturing value spread. 5-10% tolerance is just fine. if you're anal about being as exact as possible on the caps values, buy extra and measure them by hand with a multimeter that will read in microfarads and pick the ones that are closest to the intended value (like chipset binning, but simpler and faster).

[pby5cat]

you dont have to copy the tolerance values of the caps, u can use any tolerance you want as long as it is the same or lower.

The colors are done by manufacturers to deferential the companies cap series from each other, and to deter counterfeiters.

[Waterlogged]

you could always search the net for a 5-7 band schematic as well. instead of having to backwards engineer someone elses work, you simply translate the schematic onto a piece of breadboard or into some PCB software.

why physical eq instead of software eq?

BTW, you don't have to match exact tolerance for a circuit like this. tolerance is basically manufacturing value spread. 5-10% tolerance is just fine. if you're anal about being as exact as possible on the caps values, buy extra and measure them by hand with a multimeter that will read in microfarads and pick the ones that are closest to the intended value (like chipset binning, but simpler and faster).

From post #6

This is for a portable MP3 server, so the emphasis is on being 100% complete, portable and compact. Normally, I wouldn't bother going this far for something like this because WMP has a pretty decent Eq already and there are a host of third party software Eq's as well but, there is my problem, this isn't going to be a windows system. It's going to be a Ubuntu system and from looking around, there is only 1 half way decent music player that has a Eq built in and it's a bit flaky from the little bit I've been playing with it. During my writing this post, I just discovered another Linux music player (Banshee) that seems a bit better than the one I had found (Audacious) but I still like the idea of doing this via hardware better yet.

Since I posted that, I've done a little bit more playing around with Banshee and while it's a impressive improvement over Audacious, it still has a couple things I consider "problems", like when the track being played ends and advances to the next track, it doesn't highlight the new track when the program isn't selected (something the stock player [Rhythmbox] does properly). Also, the EQ preamp is touchy as hell which easily distorts the signal and you need to make a new preset to be able to manually adjust the sliders. Also, while I originally dreaded the thought of taking the route I chose, I find myself refusing to settle for a software GEQ now...something about the "unique factor" of having a hardware GEQ in a computer is now very appealing, I think it raises the build to a new level.

As for grabbing something off the net, I have a known working hardware unit here and many of the schematics I've found online are incomplete (missing parts list). The ones I find that are complete use chips that I can't get from Mouser or Digikey or I need to buy 1000 of them. That's why I've stuck with this thing for so long but it's starting to get on my nerves having to fight so hard to get info on almost all of the stuff it was made with. Starting over with something newer is very tempting , and still on the table at this point.


About the caps, thanks guys. I am still something of a noob at this and thought everything needed to be precisely the same. Any guesses as to what tolerance of the AE's might be?

I know you guys are recommending Panasonic FM's, FC's and Nichicon HE's but they simply aren't available in some of the values I need...are there any other caps you can recommend that come in the values I need?.........Hm, will SMD caps/resistors work just as well as standard through hole parts for this? (I'm guessing they will)...If so, lets concentrate on them. While they will be a headache to solder, it should allow me to more efficiently utilize the blank double sided PCB I have and make this unit as small as possible.

[pby5cat]

I found a 5 chan GEQ IC chip BA3812L, that only needs external pots and caps. When three IC are combined together it can make a 7 point stereo GEQ. http://www.datasheetcatalog.org/data...hm/ba3812l.pdf

About the Q factor mentioned in the data sheet, it controlles the bandwidth in the case of tuned or resenant circuits in ur case. http://en.wikipedia.org/wiki/Q_facto...trical_systems

Follow the equations in the data sheet to set up ur frequency points. Note that R0 and R1 are fixed, so play with C0 and C. Also it looks like C0 > C.

This chip has an output buffer but no amp, so you will need to provide an amplifier chip.

Check google shopping for the chip, its on ebay and one other company.

Operation voltage range is 3.5 to 16 volts, so 12volt is ok.

[defect9]

you're so gonna fry WL's brain with that lol.

Great find pby5cat. it'd be easy as pie (or cake) to add a small footprint TA2020 amp to that eq. Plus the simplicity really does lend itself well to SMD, assuming one can solder smd. last I checked, there's a way to bake SMDs onto a board with a solder designed for such a method, but I haven't googled that recently. so, instead i found a guide to smd soldering

http://www.avrfreaks.net/modules/Fre...ng%20Guide.pdf

I'm gonna be honest, at the rate this project is progressing, I'm considering making one for myself just for the funk of it, though it'd be attatched to a non-computer bound yet still 12v PSU system.

[Waterlogged]

Fry?...it's been sizzling on the back burner for a little bit now. When I looked over that stuff it went





Sorry pby5cat, I wouldn't even know where to begin with the math to figure out the bands.


This topic will be on a short hiatus until probably sometime mid/late next week at the earliest. I just ordered the new GEQ and will wait for it to show up. When it does, I'll crack it open and look it over to see if it's going to make life any easier for me, if so... I'll go with that design. If not, I'm going to forge ahead with the work I've already gotten done and just grab parts in the largest package SMD's I can (hopefully most of them will be 1206, though I'm pretty sure I'll need some 0805 already), for easier handling. Soldering shouldn't be too big a problem as this will be a from scratch build so it should be place the component, tack it, then solder the other side and clean up the tack if need be...sounds easy doesn't it? I'm not expecting things to go that smoothly though.

[defect9]

dang, and here I was about to start doing the math. I just about got the entire circuit figured out for layout too (don't ask how long it took staring at it to get to that point).

Looking forward to the continued progress of this one. It definitely falls in the realm of extreme.

[Waterlogged]

OK, haven't forgotten about this. I'm being jerked around by the company I ordered the new EQ from. Ordered on 3-16 and still do not have it yet even though the CC has been charged. I've been patient and understanding but the gloves are now about to come off. This company has a 10 rating on RR over the last 6 months, well, that's about to take a serious hit for starters. . .then I get rough.

[defect9]

Damn buddy, sorry to hear that. Always sucks when a company that's been great for so long falls down on the job.

Maybe it's a sign. You want me to get the math together on a custom eq? can't be all that difficult as it's just a collection of duplicated filter stages with tweaked values.

[Waterlogged]

Damn buddy, sorry to hear that. Always sucks when a company that's been great for so long falls down on the job.

Maybe it's a sign. You want me to get the math together on a custom eq? can't be all that difficult as it's just a collection of duplicated filter stages with tweaked values.

Thanks for the offer but I final got them to respond and things have been resolved (I didn't even have to get nasty with them). I'm guessing they just simply forgot/overlooked the email I sent asking about several substitute models I was interested in. Turns out they only had one that I have some minor concerns about (height) which costs more than the one I wanted, they sent it without charging me extra for the hassle I've been forced to deal with. Still not 100% happy with resolution but, it'll get me back up to speed on the project. Should have it by weeks end and then I'll start pulling it apart to discover it's secrets.

[defect9]

sweet, good to hear it's moving forward again.

[Waterlogged]

*sigh* Preliminary testing on the unit I got has not gone well. It turns out that it's missing something very important (something which I over looked in my haste to get something shipped to me ). . .the amplifier. I'll play with it some more but I'm pretty sure it's not going to work the way I need it to.

Currently tracking down the original unit that I wanted that has the amp in it, I just need to make sure the store that lists it in stock actually has it before placing the order. It's a different store and it's a refurb but that doesn't really matter to me at this point.

[defect9]

well, uh, damn again. I hope it aint a sign.

[cyberspyder]

A bit late to the concert, but consider some Class D amps or the venerable Chipamps.