Top notch electronics guru's needed

[Holst]

Ahhh, the 1980's bit explains alot. The cap size, the old looking design and the ancient components from manufacturers that went bust long ago also the lack of PDF documentation

The modern caps should be ok even if they are smaller, but just because the caps on the board are old dont assume they are lesser quality. This was probably an expensive bit of kit when new.

Get it working on the breadboard and then transfer to an etched board if you want to neaten things up.

[Waterlogged]

OK, so here's an image showing exactly how the 0.12uF cap is being used. I lit the PCB from the bottom so you can see all the components and traces at the same time.

The way the circuit is setup, it starts @ the 120KΩ on the right, signal then splits and goes to the LM324 pin 12 & the 332 Mylar cap, signal from the Mylar cap goes straight to the 0.12 cap, the other part of the signal comes out pins 13/14 and splits up, 1 path goes through a 560Ω resistor to the 0.12uF cap and the other through a 5.6KΩ to the ground rail. The 0.12uF cap then goes to the 1KHz slider.

[oohms]

if you want to upgrade those lm324, might want to consider something like OPA4134

[cdolphin]

Most of those capacitors are what actually do the filtering and amplification. I hope I am not preaching to the choir, is that of only two circuits classes.

For reference, Impedance is the frequency dependent variant of resistance, basically the point of your EQ system.

The basic setup for most EQs I have seen (I have only made one as a test for very limited frequency bands)
Involves piping the sound through a number of RC circuits with calculated impedance to match your goal EQ frequencies, then running that through an op amp with a voltage follower loop. Afterwards, the influence of this RC filter is moderated by a potentiometer (the impedance of which is simple its resistance)

I can attach a PDF of a simple lab we did last semester, it is written in very simple language and may be helpful for you to actually do, just to get an idea for what is going on.

[Waterlogged]

if you want to upgrade those lm324, might want to consider something like OPA4134

Actually, I am looking to replace 1 of the LM324's. Only half of it is used and I don't see the sense of wasting prime real estate on half a chip that does nothing. I'm looking at getting a single LM358 to replace it, from the documentation I've found, it looks to be a dual op amp from the same family. Can anyone confirm this? If this is the case, I've already got a NTE cross ref number for the chip that I can likely get locally, it'll give me an excuse at the very least to go back to the store where I got all the goodies I posted earlier and find out how my "ordered items" are coming along. Thinking I'll go with Digikey for future purchases as overwhelming as the site can be. They've obviously got a huge selection and decent prices from the looks of things and I can order only what I need which will also lower costs a good bit.

Most of those capacitors are what actually do the filtering and amplification. I hope I am not preaching to the choir, is that of only two circuits classes.

For reference, Impedance is the frequency dependent variant of resistance, basically the point of your EQ system.

The basic setup for most EQs I have seen (I have only made one as a test for very limited frequency bands)
Involves piping the sound through a number of RC circuits with calculated impedance to match your goal EQ frequencies, then running that through an op amp with a voltage follower loop. Afterwards, the influence of this RC filter is moderated by a potentiometer (the impedance of which is simple its resistance)

I can attach a PDF of a simple lab we did last semester, it is written in very simple language and may be helpful for you to actually do, just to get an idea for what is going on.

Thank you. You've confirmed something I thought I recalled finding on the net when I started this project but have been unable to find that link again (forgot to bookmark it and don't remember the search phrase used).

[Waterlogged]

Actually, I am looking to replace 1 of the LM324's. Only half of it is used and I don't see the sense of wasting prime real estate on half a chip that does nothing. I'm looking at getting a single LM358 to replace it, from the documentation I've found, it looks to be a dual op amp from the same family. Can anyone confirm this?

Anyone?

LM358

LM324

Please note that the PDF's also list a LM324A and LM358A which are a bit different and are not the chips being used.

[Waterlogged]

OK, so I went back to the electronics store where I got most of the goodies for this project so far for the NTE928M (LM358) and I can definitely say...soakage! $4.44 (before the 8.75% sales tax) for a chip I can get for under $ .50 from Digikey. I also talked briefly with the guy that ordered my stuff and he thinks it "should" be in sometime next week...another reason to shy away from there, it'll be almost 3 weeks since I ordered it when I get it. Anyway, I got the chip home, pulled it out of the bag and noticed that those w**kers over at NTE didn't even actually make the chip. I can clearly make out the TI logo and LM358P underneath the NTE printing when I hold it at a certain angle to the light.

You can barely see "LM358P" hiding under the NTE printing here.



Some clever pic taking reveals the truth behind the manufacturer



Anyway, I'm seriously toying with the idea of doing a custom circuit board for this and have gone so far as to resurrect my Shuttle SS51G with Windows Millennium on it so I can use my trusty IMSI TurboCAD 7.0 that has tons of electronics parts drawings already drawn up that I can use to help lay it out. I'm also considering going SMD as much as possible for the custom board as it'll help miniaturize the board even further. It'll mean reordering almost everything from scratch from Digikey but, I think it could really be worth the added cost/headache(?). I'll just keep going through the motions with what I've currently got/have coming and use this one as a valuable learning curve "experiment" or "proof of concept" run.

[pby5cat]

Waterlogged, on post 31, it is ok to use the LM358.

I recomend mouser.com for parts, it has a very advanced search function, and all the parts have data sheets.
http://www.mouser.com/

[Waterlogged]

Waterlogged, on post 31, it is ok to use the LM358.

I recomend mouser.com for parts, it has a very advanced search function, and all the parts have data sheets.
http://www.mouser.com/

Thanks for the confirm pby5cat.

I actually plan to use them as backup for whatever Digikey doesn't have. I had a good chuckle when I was getting the NTE928M, I noticed a bag on one of the shelves that still had a Mouser sticker on it, I had to keep myself from laughing out loud.

[Waterlogged]

Geez, I'm a bit bored ATM and got to thinking that if NTE "repackages" the LM358 chips from TI, then maybe they do it from other Manu's as well...I soon found my answer.

This is the 20w Power amp chip. It clearly says "µPC1230H2" (just like my original chips ).


Looks like NTE may be nothing more than an industry parasite...it doesn't really matter to me as long as they keep making chips I need available.

[skuldarin]

just make sure to have adequate cooling for the amp chips. it's looking like you've ended up with alot of parts that have increased capabilities. you CAN overdrive amp chips to a certain point as long as you have the right cooling... kinda like going water cooled for computers.

infact, that being said... if the amp chips are just the right size, you might want to consider doing your board just right to allow for say... a small water block for a MOBO power regulator and use that for your chips if your tower is WC'd. you wouldn't have much worries about cooling flow in such a cramped space and you wont have degraded performance over time from heat soak.

just a couple ideas to throw at ya

[Waterlogged]

just make sure to have adequate cooling for the amp chips. it's looking like you've ended up with alot of parts that have increased capabilities. you CAN overdrive amp chips to a certain point as long as you have the right cooling... kinda like going water cooled for computers.

infact, that being said... if the amp chips are just the right size, you might want to consider doing your board just right to allow for say... a small water block for a MOBO power regulator and use that for your chips if your tower is WC'd. you wouldn't have much worries about cooling flow in such a cramped space and you wont have degraded performance over time from heat soak.

just a couple ideas to throw at ya

You know what, I'm one step ahead of you on that. I had the thought last week and have already found a block that'll work...was just a bit afraid to mention it for fear of ppl thinking I'd completely lost my senses...but then again, I'm putting a GEq in a computer sooooo...maybe I have lost my senses. I even PM'd a buddy to tell him about what I was thinking about doing and that "I may need help".

[pby5cat]

@Waterlogged: THIS IS XS!! watercooling things that make heat doesnt make you weird.

[Waterlogged]

@Waterlogged: THIS IS XS!! watercooling things that make heat doesnt make you weird.

Well. they are only 20W a piece but...point taken.





Got a new quest for this project. . .to incorporate a 12V DC PSU. I spent the better part of Sat poking and prodding around the transformer area to see what kind of requirements I was facing and I've got to say, I don't think they're too bad. I used one of these power bricks (Link) for the testing and all worked well but. . .it's too damn big to fit into the build. It is also far more powerful than what's needed which is the most important thing I learned from the testing. The test equipment was a Kill-A-Watt P4460 (Link) that was used to measure amp draw from the wall and a Klein MM1000 DMM for voltage. The peak numbers I saw were .61A draw from the wall, 8.6mV draw from where the power hits the PCB to the transformer in and 147.8mV through the transformer. These are absolute worst case numbers as I had the volume all the way up on the computer and on Youtube and I also had full gain on every band on the GEq (which will be unrealistic in everyday usage as I dislike a lot of 2K+ bands added to my music, too sharp). I'm thinking I can get away with something that will have 120V in (I can tie this into the normal power supply at the AC in terminal connectors) and have 12V 1A output and still have a fairly decent safety margin, would this be correct?

If so, this is where I need your help. I have an old USB HDD enclosure that bit the farm a while ago but the wall wart is still good. It has plenty of power per my guess work (12V 1.5A) but when I went to test it, well, it made the nastiest buzz in the speakers so that one is a no-go. It was while I was poking around looking for something similar that it occurred to me that I would probably be better off trying to put this on the the circuit board as well. Well, I've seen lots of diagrams on how to do this kind of thing but I have no clue if the power I'm going to get out of any of them is going to be clean enough for my needs. What I need is a design that is known to give super clean power at the ratings I actually need (i.e. no hum, buzz or static when I test it)...any takers? Typical diagram and parts list is all that's needed this time as I've gotten a bit more used to reading them since I started this topic.

[pby5cat]

So the input voltage of the GEQ is 12 volts?

To reduce the noise on that HDD power supply, put a 100uF+ cap across 12v to ground. make shure the cap is rated for 15+ volts, and if it is an electrolytic cap you match the palority.

What about your computer powersupply? ATX spec requires less then 120mVpp noise, and any good manufacturer runs at about 80mVpp

For low a noise PS you need a Linear power supply, which has a ripple of 4mVpp. A switching PS for your needs is about 120mVpp. Follow the link for a small linear PS, please check the dimensions in the data sheet, under case type B. http://www.mouser.com/ProductDetail/...V7BK3RezC40%3d

You can also use a DC to DC converter, which has a 75mVpp. the link is a 12volt in, 12 volt out. http://www.mouser.com/ProductDetail/...F6h9ilfp77c%3d

[Waterlogged]

So the input voltage of the GEQ is 12 volts?

To reduce the noise on that HDD power supply, put a 100uF+ cap across 12v to ground. make shure the cap is rated for 15+ volts, and if it is an electrolytic cap you match the palority.

What about your computer powersupply? ATX spec requires less then 120mVpp noise, and any good manufacturer runs at about 80mVpp

For low a noise PS you need a Linear power supply, which has a ripple of 4mVpp. A switching PS for your needs is about 120mVpp. Follow the link for a small linear PS, please check the dimensions in the data sheet, under case type B. http://www.mouser.com/ProductDetail/...V7BK3RezC40%3d

You can also use a DC to DC converter, which has a 75mVpp. the link is a 12volt in, 12 volt out. http://www.mouser.com/ProductDetail/...F6h9ilfp77c%3d

Hi pby, Thanks for taking the time to post but, it may be kind of moot now unless that same kind of setup will fix the new problem? I just knew someone was going to post as I was writing up the post below (damn that Murphy).


OK, so I found a wall wart that supplied clean power (old M$ Sidewinder joystick PSU) that is rated @ 12V 1.3Aand quickly found out...it's not enough juice. Volume was noticeably reduced and didn't sound quite as good as it has with the power brick. I then decided that I would try it with the NeoHE550 that's in my daily driver (the one I'm posting this from) and got this.

http://img703.imageshack.us/img703/3960/geuv.mp4

At video start to :10, computer is completely idle

At :10 to :20, I'm moving the cursor around.

At :24 mark, new noise is "reload every" doing a reload of the LC section page.

At :35 mark I open Ubuntu's system monitor and leave it open until the end of the vid.

Is this EMI or RFI? Can I filter these noises out and use the regular PSU to power the GEq? It would make this a lot easier if I can.

For reference, this (Link) is the PSU I will eventually be using in this build. It's the most powerful that will fit in the space I've been able to make for it. There is also an iStarUSA PSU (Link) that is a bit more with the same specs if that would be any better?

[pby5cat]

If you have an oscilloscope, put the probe on the 12volt and the clip on ground. post a picture of the screen showiing the noise. Make shure to include volts/div and time/div settings in your post.

Try a 100+uF capacitor like i talked about above, across the 12volt to ground, and see if the noise goes away with your computer PS.

I think the PS you linked are made by the same OEM. outside of the labels, everything is identical as far as i can see. The Istarusa should be better, due to the company making server stuff.

EDIT

A 7812 regulator would work. It takes 12 volts in and spits out 12 volts, but it is a regulator so it has ripple and noise suppressors and corrects for voltage fluctuations. you will need a .33uF filter input cap and 1uF output capacitor. Preferably tantium or polymer, somthing non electrolytic. The regulator and caps should be about $2. This chip will output 1Amp in free air, and up to 2.2Amps heatsinked or waterblocked. The mounting tab on the top is connected to ground, so use a grey or pink thermal pad and electricly isolate the screw head, so you dont ground out your water loop. http://www.mouser.com/ProductDetail/...jep01a5GgKU%3d\

Updated DC/DC conveter speced at 65mVpp noise http://www.mouser.com/ProductDetail/...2bzmSVcZJv4%3d

[Waterlogged]

No scope. . .anyway, I've got to take a step back from this for a couple reasons.

1. I find myself switching from "filter" to "separate PSU" on a almost hourly basis. I need to stop and figure out what will be the best decision for the project before I commit anymore time to this "power issue".

2. Yesterday, I went to replace the toilet and linoleum and found out I need to replace the sub floor as well. I already had a trip planned to FCPU today and as soon as I get back from that, I commence demo on the bathroom. It may be a day or so before I'm back.

[Holst]

I would be tempted to go the filter route.

I think it will be just as hard to build a separate PSU that will give the clean supply you need as it will be to clean up the pretty good 12v from the PC PSU.

A filter will be smaller and safer (as you wont be running mains power into the EQ)

[pby5cat]

Yes go with the filter idea.

First, try the 100+uF cap across the 12volt to ground right at the PCB connects. This should be an easy 10 min test. If it isnt better i would go with the regulator.

A 7812 regulator would work. It takes 12 volts in and spits out 12 volts, but it is a regulator so it has ripple and noise suppressors and corrects for voltage fluctuations. you will need a .33uF filter input cap and 1uF output capacitor. Preferably tantium or polymer, somthing non electrolytic. The regulator and caps should be about $2. This chip will output 1Amp in free air, and up to 2.2Amps heatsinked or waterblocked. The mounting tab on the top is connected to ground, so use a grey or pink thermal pad and electricly isolate the screw head, so you dont ground out your water loop. http://www.mouser.com/ProductDetail/...jep01a5GgKU%3d\

I am an electrical engineer, and i have electrical test equipement so id like to help if i can. PM me if you think of someway my equipement can help.

[Waterlogged]

Tried the cap (all the way to 2200uF), didn't seem to work.

Both you gentlemen have a PM.

[Waterlogged]

pby5cat, per my last PM to you, I finally got the chance to set the system up on the test bench (you have no idea how excited I was to be at this point of the project with all the sub systems ) with the 270W PSU and the highest wattage spike I saw was ~116W. This was while playing a CD and with everything but the actual HDD's I plan on using (had 1 3.5" SATA instead of 2 2.5"). Unless this is some sort of false sense of security it's trying to give me, I think the 270 is going to be fine. I also didn't seem to have any of the static, crackle or other noises with the 270W PSU that I got from my 550W NeoHE. I'm gonna play with it some more to make sure on that though.


Anybody know how much power a GPU fan header supplies? The card in question is a Zotac 8400GS (ZT84SEH2P-FSR). The fan it comes with is "rated" @ .11a according to the label but, I'm sure the header is capable of more than that, just don't know how much more.

[pby5cat]

Congrats on the test.

The physical fan header could probaly do 1amp safely, but the fan speed controller circuitry has an uknown current rating.

Atleast .11A is safe.

[Waterlogged]

Congrats on the test.

The physical fan header could probaly do 1amp safely, but the fan speed controller circuitry has an uknown current rating.

Atleast .11A is safe.

Thanks pby5cat. . .guess that rules that idea out. I was planning on tidying up the wiring a bit by putting that 10w item on the fan header but, that doesn't look like that's going to be a good idea after all. Probably for the best anyways, I've been playing with it some more and it seems my current custom cabling system is messing with the colors...I've found the appropriate connectors and pins @ Digi but I'm now thinking to get my order @ Mouser instead (they have some DMM fuses I need that Digi doesn't have). Just a matter of finding them @ Mouser now.

I've also had it with the local shop, I "ordered" some Mylar caps and some ¼w resistors from there on Jan 31 and haven't gotten a single phone call from them yet...not the way you should run a customer based business IMO. I'll be getting the stuff I didn't get from them likely from Mouser when I make my order.

[defect9]

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.