PCM1794A tube-output DAC.... 4 chips detroyed - diyAudio
Go Back   Home > Forums > Source & Line > Digital Line Level

Digital Line Level DACs, Digital Crossovers, Equalizers, etc.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 15th April 2012, 07:19 PM   #1
wmincy is offline wmincy  Malaysia
diyAudio Member
 
Join Date: Sep 2003
Location: Italy
Default PCM1794A tube-output DAC.... 4 chips detroyed

Recently I'm DIYing a PCM1794 based DAC, for me and for some friends of mine.

Well, the reason is very simple, just want to make an A/B test with the current DAC I'm listening to, TDA1541 NOS, and to share some fun with friends.

This project was started since last December and this couple of weeks I got some free time to focus on it. The schematic is nothing special just followed TI proposed designed confronting with Andrea Ciuffoli and Pavouk circuits.

I took some photos showing the main DIR PCB and DAC PCB.

Just like I wrote in the title, unfortunately I burnt 4 of PCM1794 up till now. I can't believe that these chips are so fragile...

I checked the layout many times but can't spot the problem.

Power supplies are shunt regulators which worked flawless and considering the current limiter that prevent short-circuit at the output.

The situation is this, once the power is on the 3.3VDD of DAC starts to decrease and the DAC increases in temperature which seems like a shortage somewhere. First with the 3.3VDD power rail, and a few seconds later the 5V VCC starts decreasing too.

This problem is so weird... Is there anyone having the same problem with this DAC? Any suggestion will be welcome!! Thanks.
Attached Images
File Type: jpg 1334509672940 copy.jpg (618.8 KB, 701 views)
File Type: jpg 1334509741642 copy.jpg (294.1 KB, 647 views)
File Type: jpg 1334509772989 copy.jpg (368.9 KB, 632 views)
File Type: jpg 1334509871252 copy.jpg (427.1 KB, 600 views)
__________________
the fever of DIY HI-FI will NEVER be cured !!!!
  Reply With Quote
Old 15th April 2012, 11:59 PM   #2
diyAudio Member
 
Join Date: Apr 2003
Location: Eastern Pennsylvania
Quote:
Originally Posted by wmincy View Post
Recently I'm DIYing a PCM1794 based DAC, for me and for some friends of mine.

Well, the reason is very simple, just want to make an A/B test with the current DAC I'm listening to, TDA1541 NOS, and to share some fun with friends.

This project was started since last December and this couple of weeks I got some free time to focus on it. The schematic is nothing special just followed TI proposed designed confronting with Andrea Ciuffoli and Pavouk circuits.

I took some photos showing the main DIR PCB and DAC PCB.

Just like I wrote in the title, unfortunately I burnt 4 of PCM1794 up till now. I can't believe that these chips are so fragile...

I checked the layout many times but can't spot the problem.

Power supplies are shunt regulators which worked flawless and considering the current limiter that prevent short-circuit at the output.

The situation is this, once the power is on the 3.3VDD of DAC starts to decrease and the DAC increases in temperature which seems like a shortage somewhere. First with the 3.3VDD power rail, and a few seconds later the 5V VCC starts decreasing too.

This problem is so weird... Is there anyone having the same problem with this DAC? Any suggestion will be welcome!! Thanks.
These sorts of problems are very difficult to diagnose from afar. It definitely sounds like you have blown those DAC chips if they get hot while the power supply voltage also sags. So long as you are not subjecting the DAC outputs to significant voltage, it should not be possible to damage the chip through it's outputs. They are inherently current limited by design.

My suggestion is to re-check a few basic things. First, check for any shorts between adjacent DIP socket pins, with the DAC removed, of course. Also, check for unexpected shorts from each socket pin to ground. Should you find no errors, then make sure that your point-to-point hand wiring is absolutely correct (confirm all hand wired paths between the terminating nodes of that path with a continuity checker), including that you have not misidentified pin-1. I can say from my own experience that we often see what we expect to see, rather than what is actually there.

Beyond those things, I can only suspect that you may be damaging the DAC chips simply through handling them should you not be observing proper anti-static procedures. Let us know what you find.
__________________
Ken

Last edited by Ken Newton; 16th April 2012 at 12:10 AM.
  Reply With Quote
Old 16th April 2012, 12:38 AM   #3
kevinkr is offline kevinkr  United States
diyAudio Moderator
 
kevinkr's Avatar
 
Join Date: Sep 2004
Location: Boston, Massachusetts
Blog Entries: 6
I'd almost suspect that the supply and grounds somewhere are swapped or you are delivering way more than 3.3V on the digital side or 5V on the analog side. This is a unipolar dac so make sure that the supply polarities are correct on all supply pins.

I've used the PCM1794A and PCM1798 successfully in a number of designs but do not use shunt regulators in any of them. Make absolutely certain that if the shunt regulators overshoot at power up that they do not exceed the absolute maximum supply voltage ratings of the PCM1794A - this is another thing I would suspect. (Strongly!)

Make sure that the supplies come up together, I'm not sure that the PCM1794A is/is not sensitive to power supply sequencing. (Mine come up in lock step well within a couple of ms.)

Handle according to good ESD procedure as these are static sensitive and will self destruct if exposed to excessive static discharges.

http://www.ti.com/general/docs/lit/g...a&fileType=pdf

FWIW I do not like those like those not so little dac boards as there should be decoupling of the reference and supply pins within mm of the device package for best performance. I'm wondering what sort of decoupling you have..

What are you doing for I/V conversion?

Can you post a schematic of what you are attempting to do? And unfortunately I have a lot of concern about your build technique which is not appropriate for high speed, high performance digital/mixed signal devices. I don't see any evidence of good low impedance digital grounds, power or decoupling and rather a lot of very small wire tack soldered here and there. I also see what appears to be a highly capacitive implementation of I2S..

FWIW Both the 1798 and more particularly the 1794A are excellent performers with careful implementation. Worth the effort to sort out.
__________________
"Extraordinary claims require extraordinary evidence." - Carl Sagan
  Reply With Quote
Old 16th April 2012, 02:42 AM   #4
diyAudio Member
 
Join Date: Jan 2008
Location: Virginia
At the first look, Vcc and Vdd looks ok. Sice you have the Iout pins bended out and missing on the board, I assume that you are inserting it right.

My first suspect would be the power supply. I never trust a shunt-type power supply. Not without crowbar protection, but sometimes even that is slow to react to voltage spikes (depends of the capacitor on output of regulator).
Click the image to open in full size.
  Reply With Quote
Old 16th April 2012, 03:51 AM   #5
diyAudio Member
 
5th element's Avatar
 
Join Date: Oct 2002
Location: England
I've used the PCM1792/94 for a long time now and find them to be quite resilient but they do not take well to certain things.

They will survive slightly higher power rails for a sort period of time. I once shorted the 3.3 and 5 volt rails together by mistake and one 1792 got very hot, smoked in fact, I quickly turned everything off, then solved the problem and the chip still worked perfectly and measured just the same.

One thing the chips do not like is having the ground pins unconnected by mistake, as in perhaps the pins aren't soldered down correctly. This will cause them to heat up and they will fail as a result. If you manage to turn them off before they totally cook themselves then they do survive, so they are quite robust as things go.

The PCM1792/4s will operate fine without anything connected to their outputs. If your I/V stage is causing problems then disconnecting it could be a way of finding this out. I had a discussion with an engineer at TI who basically said that the chip could possibly become unstable under capacitive loading, so if your I/V stage is somewhat different then this could be causing the issue.

As Ken mentioned the outputs are somewhat current limited, but this is more a case of the output current is usually controlled directly by the DAC chip as a result of how it functions. However under inappropriate circumstances it would probably be possible to force them into outputting too much current.

As you've mentioned that the supply rails are sagging this implies that something is forcing the chip into oscillation or that something is being forced to work harder then it should.

The outputs of the DAC chip are biased around the 10k resistor that connects to pin 20. The lower the resistors value the higher the output current. If you were to connect this pin to ground by mistake or have it shorted to something else then this would easily force the outputs into pushing more current then they are capable of. The 1792/4 always has a standing current flowing into the outputs, so even without any music playing the DAC is pumping current. If you disconnect the I/V stage then this current ceases to flow, if you measure the drawn current on the 5volt supply line it changes with the sampling frequency and rate of over sampling chosen, but typically it should not exceed around 50mA, if you disconnect the I/V stage this will fall down to only a mA or so. This might be helpful in figuring out where the problem is.
__________________
What the hell are you screamin' for? Every five minutes there's a bomb or somethin'! I'm leavin! bzzzz!
  Reply With Quote
Old 16th April 2012, 05:52 PM   #6
wmincy is offline wmincy  Malaysia
diyAudio Member
 
Join Date: Sep 2003
Location: Italy
Ken,
I revised the layout once again and did try point-to-point check: SSOP adaptor pin to IC pin -> socket pin to IC pin. There was no open circuit.

The same 3.3VDD provide both DIR9001 and PCM1794, however DIR9001 was still alive.

May be I didn't care too much about the ESD, but DIR9001 was soldered the same way PCM1794 was, so I'm quite confusing about this...


Kevinkr,
there is no way that I could swap the powers, since the output connector for 3.3V and 5V are totally different: 4-pin connector and 2-pin connector. I purposely choose two different connectors to avoid this problem happen.
The supply voltage was adjusted up to the value before connecting to the board.

Sorry for my unprofessional building technique, the resources are quite limited however the signals (LRCK, DATA, BCK, SCK) from DIR9001 were quite pleasant and my oscilloscope showed decent output wave.


SoNic_real_one,
I'm not that hostile against the shunt regulator, after all it provides a very stable voltage that satisfied my need. Please check out the attachment.

The Iout pins were bent so I could easily connect to tube grid.


5th element,
today I took another try with others PCM1794s handling with care, soldered them with solder iron, and again, 3 more were dead...

Just being curious, how did you solder them? I tried different soldering method (except reflow), would you mind share your method?

After the chip were burnt (burnt but without smoke), I measure the resistance between the 3.3VDD pin - DGND and 5VCC - AGND; I read a ~130ohm which differs from the initial resistance I measure before applying power, above some kohm.

I did try using normal voltage regulator instead of the shunt regulator, 7805 and TPS7A4533. But both of them got hot, too hot.


I think I should remove the socket and soldered the adapter directly on the PCB...
Attached Images
File Type: jpg 1334592392135.jpg (872.9 KB, 174 views)
File Type: jpg Sala-shunt-for-DAC.jpg (622.3 KB, 253 views)
File Type: png pcm1794a_sch.png (33.6 KB, 261 views)
File Type: png dir9001_spdif12_sch.png (31.4 KB, 241 views)
__________________
the fever of DIY HI-FI will NEVER be cured !!!!
  Reply With Quote
Old 16th April 2012, 06:57 PM   #7
diyAudio Member
 
5th element's Avatar
 
Join Date: Oct 2002
Location: England
Quote:
Originally Posted by wmincy View Post

Just being curious, how did you solder them? I tried different soldering method (except reflow), would you mind share your method?
Over the years I have used various methods for soldering SMD chips but all have worked. The first is simply soldering the individual legs with a very fine soldering iron and a magnifying lens. The second is using a large soldering iron tip and flooding the legs and pads with solder, then using solder braid/wick, to remove the excess solder. The third and by far the best imo is using solder paste. I solder the chip in place by soldering down 1 or 2 pins, usually two, one at each end of the chip and on opposite sides. I then wipe solder paste into and around the pins/pads, then heat with a heat gun/paint stripper hot air gun. This melts the solder and everything solders nicely into place by surface tension. Sometimes I need to use the solder braid if I've used too much paste, but this technique works the best.

Quote:
Originally Posted by wmincy View Post
After the chip were burnt (burnt but without smoke), I measure the resistance between the 3.3VDD pin - DGND and 5VCC - AGND; I read a ~130ohm which differs from the initial resistance I measure before applying power, above some kohm.
If the resistance changes then that means you've most likely cooked the chip real good. I have never damaged a PCM1792/4 due to ESD and I don't wear a static distarge bracelet either. I do make sure to connect myself to earth by touching something appropriate to discharge myself before handling chips though.

Quote:
Originally Posted by wmincy View Post
I did try using normal voltage regulator instead of the shunt regulator, 7805 and TPS7A4533. But both of them got hot, too hot.
Linear regs tend to get quite hot, even with low current, if the voltage they are stepping down is quite large. They obviously get a lot hotter if the current drawn is high, which it sounds like is happening here.


Quote:
Originally Posted by wmincy View Post
I think I should remove the socket and soldered the adapter directly on the PCB...
This is unlikely to solve your problem. Looking at the schematic for you I/V converters I can't help but notice that your circuit is quite different from what TI recommend you do on the outputs.

First of all the outputs of the DAC chip are designed to work into a virtual ground, ie zero ohm source resistance. From your circuit diagram is looks like your I/V stage isn't doing this. The second thing that jumps out at me is the 2n2 cap that you've got soldered across the outputs of the DAC. This is the one thing the TI engineer advised against me doing as the chip would be unlikely to enjoy the capacitive load.

As before I recommended disconnecting the DAC chip from the I/V stage and see what happens. From my experience chip wont burn if the I/V stage is disconnected, but the chip can burn if something goes wrong in the I/V stage.
__________________
What the hell are you screamin' for? Every five minutes there's a bomb or somethin'! I'm leavin! bzzzz!
  Reply With Quote
Old 16th April 2012, 08:40 PM   #8
diyAudio Member
 
Join Date: Apr 2003
Location: Eastern Pennsylvania
Okay, I may have spotted an anomaly. Your photo of the PCM1794A mounted on the green DIP-28 adapter board appears to show pin-19 (AGND1) wire shorted directly to pin-20 (IREF). Pin-20 should be connected to analog ground via a 10k resistor.
__________________
Ken

Last edited by Ken Newton; 16th April 2012 at 08:43 PM.
  Reply With Quote
Old 16th April 2012, 09:36 PM   #9
diyAudio Member
 
5th element's Avatar
 
Join Date: Oct 2002
Location: England
I believe that's a 10k surface mount resistor that wmincy has between pins 19 and 20.
__________________
What the hell are you screamin' for? Every five minutes there's a bomb or somethin'! I'm leavin! bzzzz!
  Reply With Quote
Old 17th April 2012, 01:20 AM   #10
diyAudio Member
 
Join Date: Apr 2003
Location: Eastern Pennsylvania
Quote:
Originally Posted by 5th element View Post
I believe that's a 10k surface mount resistor that wmincy has between pins 19 and 20.
Yes, I can see that now. My eyesight isn't what it once was. Looks like an SMD resistor, as seen edge on. I thought maybe I was seeing a piece of wire. It's very unfortunate, that he is having such catastrophic failures of multiple DAC chips without ready determination of the cause.
__________________
Ken
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Comments on PCM1794A DAC design novec Digital Source 54 4th June 2012 08:58 PM
My new tube DAC output Algar_emi Tubes / Valves 46 19th January 2011 06:57 AM
Micro DIR9001 + PCM1794A DAC theAnonymous1 Digital Source 88 14th June 2010 10:01 AM
FS: Twistedpear PCM1794A DAC Board pftrvlr Swap Meet 2 18th April 2010 04:52 PM
Anyone used a PCM1794A with balanced output? hifimaker Digital Source 11 18th March 2006 06:36 PM


New To Site? Need Help?

All times are GMT. The time now is 09:03 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2