Cheap 24/96 DAC, Revision B.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Disabled Account
Joined 2003
Check the voltage monitor datasheet to see what it's supposed to output (I'd measure mine but dissasembling the enclosure means two days putting it back together). If you are getting lower than it should be, perhaps something is trying to draw too much current from it, which souldn't be the case. If so, disconnect stuff from it's output and see if the voltage goes to whatever it's supposed to be.
 
I checked, and the output pin should be Vdd. I disconnected the reset line to the CS43122 and left it connected to the CS8420 and the output measured Vdd. I disconnect the reset line to the CS8420, it's Vdd. I never measured it with the output pin connected directly to the CS43122 because I don't want to ruin the PCB. So it seems like the CS43122 is drawing too much current on that pin. I suppose I do have to buy another $14 CS43122, don't I?
 
Disabled Account
Joined 2003
To check that without ruining the PCB, just use a jumper wire to connect the monitor to the CS8420.

Try things in this order (when using the multimeter, touch probes on the PCB tracks rather than the connectors of the components as that makes sure that you don't miss any cold joints):

Check the 47K resistors. I think that the pins they connect to shouldn't draw much current, so there shouldn't be much voltage drop over them.

I still think that it is more likely that there's a soldering problem, such as cold joints or tiny bridge between two pins. Examine the board with a magnifying glass and you may find your problem.

One thing that may draw too much current is a shorted capacitor or resistor, but your current measurements seem fine. Using the multimeter in resistance measuring mode to check for shorts might be a problem since it puts voltage on the probes. I'd only do that after taking the big ICs off. If you get to that point without finding the problem, build the hot air pencil instead of using a braid to remove the CS43122. Then you can safely check the resistors and capacitors with the multimeter, and if you find the problem you can try putting the DAC chip back in.

If you end up getting a new CS43122, use the hot air pencil instead of risking frying the part again.

Hopefully at least the CS8420 is fine, but you haven't got to that point yet.

Or, you can do like I did and rebuild the whole thing on the spare PCB, this time putting in the ICs last after checking everything else is working... :cannotbe:

BTW, IMHO if you are only driving cheap computer speakers, using a standalone DAC instead of the sound card is not worth it. Take a look at the Aleph-X and some DIY speakers, or the cheaper route, a set of HD-580 and DIY headphone amp.
 
Everything else is fine. No bridges, no shorts. No excess voltage drops across the resistors or resistor arrays. I was planing on using an exacto knife or solder pick and lift each pin, one by one. I suppose I will have to buy a brand new CS43122. I have done everything I can with what I have. Maybe my best option would be to just buy a new one...
 
Disabled Account
Joined 2003
I hope you heeded my warning about using the digital out from your computer. You can't rely on a multimeter to show you the output voltage as the frequency is too high. If you did use the computer output, your CS8420 may be fried. As for the DVD player digital output, it's extremely unlikely not to be +/-0.5V, but you may want to check the manual, so you can eliminate the possibility that this is the problem.

Use the hot air pencil when you are soldering in the new CS43122 or you may have to replace it once again.

Bending the pins on the ICs multiple times can break the fine internal connection.

Good luck.
 
I have definately used the output from my dvd player. It's a brand from hong kong, called "Universal". It plays all regions of DVDs, outputs no macrovision, has 6 channel analog audio output, S/PDIF coax and optical, (which the audio output is set to S/PDIF, not analog) composite, s-video (which is dim, because the output amplifier for the luma and chroma is practically non existant), component video, and progressive scan component and VGA outputs. I also have a pioneer DVD player with S/PDIF coax output as well that I have tried and still nothing. So, the CS43122 is removed, I just await a new one.
 
Disabled Account
Joined 2003
Did you make the hot air pencil? I know I keep repeating that, but that's only because I think it may save you from frying something again.

About the DVD player: I don't see how the stuff you said has anything to do with the S/PDIF voltage. Check the manual to see what it is. Also note that the DAC will only take uncompressed stereo PCM. Playing a DVD with multichannel sound will not work with the DAC. If you want to test 96kHz/24bit uncompressed audio recordings, there are some 24/96 DVDs such as from Chesky. Although I'm not sure that all DVD players will pass through 24/96 PCM, maybe someone can say something on that.
 
ATM I'm using a audio CD. My DVD player has two S/PDIF settings, raw and PCM. Since a CD is 16bit 44.1kHz 2 channel PCM, I'm using raw output. I know multichannel audio will not work unless I set it to PCM. I think I know what caused it to die, in terms of heat. I have a small, fine tip for my weller, and now since it's removed, I'll allign the new one and just touch the pads to cause the solder to flow. This will minimize heat. I'm afraid I might do more harm than good with the air pencil, I'll need to practice first. Now I just wait for the CS43122 to come...
 
Also, I did some more modifications to the power supply when I built it. I really never saw a purpose for C5 and C6, along with their bleeder resistors. According to the datasheet of the LM337, 1uF caps would have done the job find without bleeder resistors. I made a mistake when ordering parts, I ordered 470uF instead of 4700uF. I then calculated the time constant and used caps and resistors I had on hand to get close to the time constant as 3.5 seconds as I could. I have no idea why these caps and resistors are there, but I put some with the same time constant as the ones in the schematic. Secondly, I used 120 ohm, 1/8 watt (all digikey had in bulk packaging) and a 2k trim pot for adjusting the regulator. I may or may not replace it with resistors when I'm done, though it's a good idea too. I may when it's ready to be put in the chassis. The size of the diodes seemed like overkill, but I kept as close to the schematic as possible. I used your CLC network you showed me for the digital suplies.

When I first assembled it, I plugged it in and I noticed L2 was smoking. So I went though everything and found nothing wrong. Pluged it in again and it still smoked. Some one on IRC told me I might have the filter cap backwards. I looked, and I did. I never knew an electrolytic cap would conduct when reverse voltage is applied. That was why L2 was smoking, because it was a direct short. Silly me... I should have realized that lesser potential should be placed on the negative side, like the negative voltage. Well, the cap got warm, but never blew because it wasn't plugged in long enough. It wasn't really hurt. Just thought I'd share my experience :p
 
Disabled Account
Joined 2003
The resistors at the end of the power supply are not for bleeding, but to make sure that there is a minimum amount of current flowing so the regulators are functioning correctly. Although you could instead change R1(4).
While 1uF "would have done the job" that doesn't mean they would have done the best job. Larger capacitance should result in smaller voltage variation under a changing load (the regulators do not have perfect instantaneous regulation).
I've had electrolytics backwards, but I never actually had a problem from that...
 
Ok, then I'll change them.

While I wait for the CS43122 to arrive, what circuit are you using to convert optical to coax? I have a toshiba optical receiver I bought, but what TTL to coax circuit are you using? I built this but I don't have a 74HC04. I would imagine using a transformer coupled circuit would be better? I used an MH7404 and it's TTL, instead of the CMOS 74HC04, and it never worked. The high level of the toshiba receiver is 4V, so I imagine I need a CMOS inverter for it to work.

I have built an S/PDIF source. It uses the DIT4096 which converts 24 bit i2s audio from the service layer decoder of an XM radio into S/PDIF. Either TTL or coax transmission can be used very easily, but the coax transmission is not working for me. This schematic works really well. It works with ALL XM receivers because they all contain the same decoder which has a 24 bit i2s output. Page 18 of the DIT4096 datasheet has the output schematic for optical and coaxial outputs. Optical output works, but coax does not, even though I'm using the same transformer, resistor, capacitor and RCA connector. Since this does not work (I have both connected, and optical works fine, coax is silent) I need to know the circuit you are using. To test, I'm using an Harmon Kardon AVR225 and tested both coax and optical inputs, but the coax output using the circuit on page 18 of the DIT4096 datasheet does not seem to work.

Also, I question the transformers. The core is loose and moves around. Is this normal? Is this why the coax transmission is not working? And because of this, will my DAC work?

Thanks.
 
Disabled Account
Joined 2003
Unless the core is broken or a coil is broken, it should not be a problem. To avoid microphonic effects from vibration I would add a drop of epoxy or glue to make it stay put.

I would suggest not to use TOSLINK. All optical recievers, including the high end AT&T glass fiber stuff, are inferior to a proper coaxial connection in terms of jitter. These are not my opinions, but of Jocko, who has worked with the optical fiber industry.

To convert TTL to S/PDIF levels, I used the following:
http://www.andrewkilpatrick.org/mind/spdif/spdif_circuits.gif
Note that the circuit, while simple (it only needs one HC04) only impedance matches the S/PDIF side; so the conversion should be right at the TTL source, not after a cable. I used this to get proper coax out of my TTL output sound card, and the HC04 5V and ground can be taken right from the sound card or computer power supply. This might be even better if you put a 1:1 transformer after the circuit; this way with transformers on both sides you can approach the AES/EBU balanced digital connection in performance.

The best cheap connection option IMHO is Jon Risch's suggestion of Belden 1506A coaxial cable. Bare solid copper conductor, foamed teflon insulator with very low dielectric constant, and foil-and-braid shield. Thinner than regular coax but still 75 ohm. That's what I used, and I added an external braid outside grounded at the source end only and with a 0.01uF ceramic on the other side (so the cable doesn't act like an antenna at RF frequencies). With that even if I put the cable along power cords I can't hear any interference.

Also note that if the input jack is not right on the PCB and you run wires from the chassis to the PCB, the internal digital wiring should also be 75 ohm coax to maintain the characteristic impedance, or you get reflection problems. Jacks and connectors should also be 75 ohm. That's easy with BNC, but if you also use RCA, then it's not possible. A compromise for RCA is to use the appropriate Canare connectors.
 
Prune,

Most of my devices have only optical jacks so I need that conversion circuit. The toslink transmitter with a high of 4V will register as a high with a 74HC04? According to the datasheet it does, so I'll trust that. One interesting difference is that your circuit has a double inversion. The TTL input is inverted, then inverted again (and why are 3 gates conected in parallel, probobly for more current draw?) then the impedance is matched to 72 ohms for the coax transmission. My circuit doesn't use a buffer (the double inverter) but uses two inverters in parallel.

Thanks.
 
I received the CS43122 today and as I suspected, it does not work. I'm getting annoyed. I wish I could tell that both ICs were operating as they are suppose to. I get silence. For testing purposes I'm using a BNC to RCA connector for connecting to my DVD player. Nothing works. The DVD player's output works fine. Is the CS8420 configured to receive consumer S/PDIF? I measured the voltage at the CS43122 reset pin at it was around 600mV or so, but the output at the output pin of the voltage monitor is 4.98V (all with respect to ground). I get noise when I touch the test clip of my meter to the reset pin. I'm trying one more thing. After that, I guess I'll start saving up for a new PCB and parts...
 
Still nothing. I'm really running out of ideas. Seems like I've wasted a little more than $300. I wish I had a scope to see if i was getting a clock signal. I get 2.5V or so on my meter (measured with respect to ground of course) from the epson clock output. I have no idea what the problem could be without a scope. Any ideas?

Thanks.
 
The "other" thing I tried was with that reset pin. The reason for the voltage level on that pin is because it was floating. I never reconnected the reset pin to the voltage monitor. Now I am sure the CS43122 and CS8420 is getting the reset pulse, but it still is not working. I get a pop when I apply power and disconnect it, as jwb says. Voltage is 5V at U1 and U5, and 4.9 at the digital supply of the CS8420 (I forget the name of the that regulator). I get 5.4V at the adjustable regulator. The CS43122 was fried, and it's ok now that I have the replacement, but I'm not sure what else can be wrong. This is really getting me bummed because of all the money I spent. The transformer has the correct resistance, I reflowed all the pads and joints just incase. All I get now is silence. I don't have a scope, but I'm almost sure the reset pulse is working ok. Anyone have any other ideas?

Thanks.
 
Disabled Account
Joined 2003
Is your connector working fine? In any case, you shouldn't use a BNC connector if you are going to be using an adapter to change to RCA; that defeats the purpose of using BNC in the first place. If your DVD player has an RCA jack, use an RCA to BNC cable (make your own).

The CS8420 will recieve S/PDIF and AES3. But not if you've fried it...
The regulator outputs are fine at least.
Don't resolder too many times as you may fry something.
If you used the multimeter to check some resistances, it might have fried the CS8420.
You should try unsoldering the CS8420 and soldering it again.
Post a high resolution scan of both sides of your board so I can look at your soldering.
I'd do the soldering for you if I lived closer...
 
My connector is working ok. One thing I wonder is if I really fried the CS8420 or not... After a while, it heats up, just like you said, and the CS43122 as well. This is what I'm curious about if I really fried something. The diode test on the meter might have killed it, but I really hope not. I think my last options here are to buy a new CS8420 and possibly another epson resonator, but I doubt I'll go that far. I could have fried it with too much heat as well. Is there any way to tell id both IC's are operating without a scope?

I'd give you a high res scan, but I have no working scanner or digital camera. Possibly if I still get no where, maybe you can take a look at it for me?

Thanks.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.