advertisement


No sound output from a DAC - any simple checks I can do?

The Woodside PCB is very good quality, I just reflowed all joints first, then careful use of solder sucker (keeping tip away from PCB when activating), then finish up with braid to release all pins. No damage, easy job, just take your time. I never bother with anti-static for domestic electronics either…I never knowingly killed anything as a result.

I really would not cut pins on the S1…just in case! My last stash cost me chf45 each, don‘t think I’ll ever get that lucky again (I still keep meaning to list my failed one on eBay as faulty and to see how much it really goes for :))
 
few methods of removing chipset without damaging the board or chip

I liked the method of using lots of solder and heating the mass to release the legs

other method of 2 soldering irons looked interesting
 
Thanks again for all the replies. Just taken this photo for reference of the regulator nearest the large blue cap with the slight scorch mark on the side:

20210522_075504 by Chris To Go, on Flickr

Definitely a LM340T5/7805 fitted, putting the picture here as a useful reference for me for later, when it comes to seeing the + and - connections for when fitting the new tantalum capacitors.
 
The TDA1541A has three supplies 15V, 5V and -5V. It has a poor reliability track record.

The datasheet says nothing about the sequencing of these supplies, but it is common to have to manage supplies like these so that the 15 is never less than the 5, none can ever be reversed a little etc. The datasheet does state that no supplies can be below zero at all, a common problem at switch off or when something goes wrong and often triggers destructive latchup
 
Whatever it is, the same concerns about sequence apply. The data sheet says nothing about this. This has to be one of the most unreliable chips ever made.
My experience in far too many years is that if not actually frying, only ICs exposed to external transients from outside the equipment, fail often
 
I've fixed many players with dead chips but it's not a massive number. My CDI is on its original DAC and has been constantly powered for 30 years give or take a few weeks.
 
My CDI is on its original DAC and has been constantly powered for 30 years
Which would make sense if it is a power supply sequence bug. Yours has done 260000 hours, very long if he chip just dies of heat and lifetime
 
I've fixed many players with dead chips but it's not a massive number. My CDI is on its original DAC and has been constantly powered for 30 years give or take a few weeks.

Ever had one with the -5v rail open circuit inside the 1541 MJS? Normal fault I've seen is for them to go noisy when warm.

Thanks, Richard
 
Thanks again for all the replies. Just taken this photo for reference of the regulator nearest the large blue cap with the slight scorch mark on the side:

20210522_075504 by Chris To Go, on Flickr

Definitely a LM340T5/7805 fitted, putting the picture here as a useful reference for me for later, when it comes to seeing the + and - connections for when fitting the new tantalum capacitors.

@starbuck - as mentioned upthread, your photo is of one of the +5v regulators. You need the one circled (badly) in red below, along with the associated tants (and yes, they're 4.7uF 35v items). I'm about to change the tants in mine with 10uF, at least for the -5v rail. And measure smoothing caps...and swap back in an S1 :)

Edit to add: in photo below, the missing tant is under the PCB...I was lazy when I fixed it!

51196566705_a8ef3d6ce3_k.jpg


Richard
 
Whatever it is, the same concerns about sequence apply. The data sheet says nothing about this. This has to be one of the most unreliable chips ever made.
My experience in far too many years is that if not actually frying, only ICs exposed to external transients from outside the equipment, fail often


I think the problem is that the TDA1541, any variant, actually dissipates quite a lot of heat overall - c.0.7W constant, in a 28pin DIP - and an awful lot of implementations didn't help this; skinny traces of thin copper on single-sided PCBS etc - neither did Philips suggest any heat-sinking.

Naim's early implementations have it sat on a lot of thick, wide copper in the PCB (compared with competitors), and so perhaps like Mark's, my CD2 is at c.25yrs in with no issues. Perhaps rare.

Also, perhaps, Phillips didn't much care for longevity; this dac family was after all their first couple of steps, the first to market for 16 bit for domestic digital audio - and it bought time during which Philips invented& developed 'bitstream' as a route to a much cheaper process (no more precision diffusion required for current division by emitter-scaling) and offering a route to far higher accuracy, from single supplies, vastly lower power consumption via cmos, massive logic integration etc.

And funnily enough - bitstream, sigma-delta, or whatever you want to call the approach and its closely-allied versions like PWM - remains the only real route forward to this day.





ETA:

Supply sequencing isn't a thing in the 1541; the internal current bit sources are between the -5 and -15 v supplies; the +5 supplies the ECL that generates the output above the -5v supply (in fact AC current on the +5v supply is the complement of the output signal) Dropping any one - or more - of the supplies, does not damage the chip; and since everything is a current source, or a complementary form of, sequencing (even voltage tolerance on supplies..!) does not appear to matter at all here. Agreed,it can be fatal for many cmos/bitstream dacs.
 
Wow - it just took me 1.5 hours to get the 1541 out cleanly, son of a bitch! And I still broke 1 pin. PCB still OK (see photo below), it is really good quality, but the damned through board copper is just so hard to get free of solder. I also changed the input and output decoupling tants on the -5v reg - *note*: the input tant to the -5v regulator was around the wrong way!!!! Check yours @starbuck - it's a negative rail, so the pin with the + needs to connect to ground. I wonder if this is why the TDA failed?

Measured one of the smoothing caps, over spec and no ESR but 2.1% vLoss. I do not have replacements in stock that will fit, but am going to order some.

Anyhoo - an S1 back in the WS3 :) Listening now, stunning - need to do a proper back-back with the Naim CDI at some point, but its much better with the S1 than I remember.

51196820905_1ea9b20ea7_k.jpg


Richard
 
It is very unusual for a tantalum top survive long when fitted backwards, they are incredibly fragile when abused.
 
Ever had one with the -5v rail open circuit inside the 1541 MJS? Normal fault I've seen is for them to go noisy when warm.
Thanks, Richard
I think I've seen one entirely dead one but I can't remember checking the voltages on it. All the rest went noisy to some degree.
 
I've replaced maybe half a dozen faulty ones and they all exhibited the usual static type noise which got worse as the chip warmed up.
 
It is very unusual for a tantalum top survive long when fitted backwards, they are incredibly fragile when abused.

I just measured them both - capacitance OK, ESR of 2 ohms (maybe normal for tants?). I’d swear it was mounted wrong way round on input to -5v regulator, so it had a reverse voltage of around 10v but still survived?
 
@starbuck - as mentioned upthread, your photo is of one of the +5v regulators. You need the one circled (badly) in red below, along with the associated tants (and yes, they're 4.7uF 35v items). I'm about to change the tants in mine with 10uF, at least for the -5v rail. And measure smoothing caps...and swap back in an S1 :)

Edit to add: in photo below, the missing tant is under the PCB...I was lazy when I fixed it!

51196566705_a8ef3d6ce3_k.jpg


Richard

Thanks for your reply and picture, Richard. I had been looking at the one with the heatsink attached that was mentioned as a possible cause of problem further back in the thread, as that was the one with some slight scorching on the nearest large blue capacitor. I have looked again following your post and can see that the regulator in that position is different, as seen in the picture below:

P5230001 by Chris To Go, on Flickr

It is, I think, a Motorola 7906C which when looking on ebay/google, doesn't seem that common in the UK. Is an MC7906C an equivalent, these seem more readily available?
 
Before going any further regards ordering any bits, I have recovered the Arcam from the shed and measured pins 15, 26 and 28 on the TDA1541A in there. Results are:
pin 15 = -6.77v;
pin 26 = +3.009v;
pin 28 = -13.51v.

These obviously differ from the suggested voltages upthread, does the difference indicate the TDA1541A in the Arcam is also no good, or does it sound like it could be useable?

Again, thank you all for your patience and help.
 


advertisement


Back
Top