Yet another Volume controlers and source selections

AX tech editor
Joined 2002
Paid Member
Your'e right, I wasn't precise enough. Just to add - I didn't have an intention to blame anyone, especially not TI's engineers - rather I wanted to highlight a possible problem. Honestly, I didn't think the subject will cause such amount of feedback, therefore I didn't struggle too much to share all the prerequisites of my conclusions.
Again, I would be happy if I'm wrong and something went wrong due to my specific mistake, but, based on "test, collect&conclude" I had to conclude, that everything's fine with my design, the reason is somewhere behind.

One option would be (if that is possible) to measure the current from the two analog and the digital supply when sut down, to see if there is a specific one that causes the heating. That could lead to a direction where to look. Also the DC output voltage when shut down (if not zero) could be a clue.

jd
 
One option would be (if that is possible) to measure the current from the two analog and the digital supply when sut down, to see if there is a specific one that causes the heating. That could lead to a direction where to look. Also the DC output voltage when shut down (if not zero) could be a clue.

jd

Unfortunately, that's not possible already, as I have thrown away the defective chip and installed a new one...But. The heating were caused definitely by both analog supplies. Didn't need to measure - the +5V digital supply was staying cold, while both +15V and -15V analog regulator chips (
Lm317/337) were very hot, the only reason they remained alive - that I installed them on excessive heatsinks, which went very hot (was not possible to hold the finger on them). Normally I would also suspect an oscillation, but again - I had a case, where this occured during normal operation. The root cause was most probably a power-line spike/noise which forced the IC to "hang up" - and in such condition it started to act as a heater :) drawing current from both analog PS lines.
 
Yesterday I had an issue with PGA2310, want to post it here to check if someone may had similar thing. So, my PGA suddenly died. In the evening I turned off the device just usually (staying in standby, so the PGA and controller were on), and in the morning when I turned it on - it didn't work anymore.
On the PCB around the PGA traces of overheating are visible - the fiberglass even turned slightly brown, which means that for some reason the PGA just burned... Also, when I touched it (wasn't working already) it was still very hot.

I'm glad I saw this post (I don't check this site as much as I should, I guess). I've built a few pga boards (pc, perf, even hand wiring) and some of them have done this 'heating' thing on me! I thought it was my hacked up PSU (using a rail splitter chip to take 10v and make 2 fives from it, etc). I would remove power if the chip 'got hot' and let it cool for a bit then reapply power. it would always work fine after that but I never found what put it in that state. my controller is an arduino (in all cases) and I can't really blame the controller; it always seems to be working fine.

this happens on the DIP version (5v version) more often but I've seen it on the smd (12v version) sometimes, as well.
 
AX tech editor
Joined 2002
Paid Member
I'm glad I saw this post (I don't check this site as much as I should, I guess). I've built a few pga boards (pc, perf, even hand wiring) and some of them have done this 'heating' thing on me! I thought it was my hacked up PSU (using a rail splitter chip to take 10v and make 2 fives from it, etc). I would remove power if the chip 'got hot' and let it cool for a bit then reapply power. it would always work fine after that but I never found what put it in that state. my controller is an arduino (in all cases) and I can't really blame the controller; it always seems to be working fine.

this happens on the DIP version (5v version) more often but I've seen it on the smd (12v version) sometimes, as well.

I looked through the data sheet to see if it said anything about the power up sequence. Soms chips are critical with regard to that, they for instance want the digital supply to be up and stable before you should bring up the analog supplies, or something like that.
Didn't find anything though; anybody has more info on that for the PGA's?

jd
 
in my case, it had nothing at all to do with init'ing the chip. this would happen hours or even days after clean steadystate operation. it would just come out of nowhere.

once it gets into that run-away state, only a power cycle would fix it (sending bytes to it over the ctl channel does nothing)

This is 1:1 similar to my experience. It's definitely not from nowhere. The fact itself wouldn't be that much a problem, if there would be any kind of protection circuit within the IC. But there's not, obviously.
 
This is 1:1 similar to my experience. It's definitely not from nowhere. The fact itself wouldn't be that much a problem, if there would be any kind of protection circuit within the IC. But there's not, obviously.

You know the old say :" if you hear once is a rumour , if you hear for twice is a fact !"

I am also suspect a glitch or an transient spike from power source :
my PGA is on power stage with BD139/140 and zenner diode of15V and a lot of filtering (like 8600uF per side) and at the entering in transformer 230V I have a double filter with C and L on ferrite .
Never have such a problems like yours for almost now 3 years of daily usage .
 
Viktor, picture would help. Maybe somehing wrong with analogue supply, LM3x7 have internal thermal protection. I thing you should measure 5v power supply. Just to make sure there is no over voltage in power supply.
Didiet

Didiet,
+5V is actually the digital supply, the analogue rails running at +/-15 volts.
And everything is fine with the analogue supply - it has a 1000 uF before the LM3x7 and 4700 uf after, also 2x470 uF in close proximity to the PGA 2310's leads.
 
Last edited:
You know the old say :" if you hear once is a rumour , if you hear for twice is a fact !"

I am also suspect a glitch or an transient spike from power source :
my PGA is on power stage with BD139/140 and zenner diode of15V and a lot of filtering (like 8600uF per side) and at the entering in transformer 230V I have a double filter with C and L on ferrite .
Never have such a problems like yours for almost now 3 years of daily usage .

Actually, I suspect the glitch coming from DIGITAL part's supply. A PGA is a software-driven digital pot. The analogue stage is powered from bipolar analogue supply, and when the mentioned run-away state occurs - the circuit draws the power from analogue rails, BUT the state itself caused by wrong behaviour of DIGITAL (control) part of the chip (I suspect). Did anyone see a PCM2702 (USB DAC) loosing sync when a fridge turns on in the kitchen?? So, I suspect something similar in our case - so it has nothing to do with the analogue supplies, even if this current burns the IC at the end.
I am going to try better filtering, maybe increase cap after the regulator, but not sure how to check if it helped. Not eager to wait until this happens again.

Linuxworks, don't you remember what supply did you use to power digital part of your projects?
 
at the entering in transformer 230V I have a double filter with C and L on ferrite .

Will implement this. Anyhow, I would go back to the beginning and point to a fact, which has to be considered by designers, IMHO: such a run-away state IS possible, it CAN destroy the chip - would make sense to think about. Or, at least it can serve as a warning for further PGA23xx implementers.
 
Victor, i think 4700uF after LM3x7 is too big, 100uf is enough.

Maybe it's too big, but 4700 would never be worse than 100 in terms of spike filtering... Again, see my post above - I don't think the problem's coming from analogue supply. I measured the voltage across the regulators when the chip was naturally burning - and it was rock solid (+/-14,2 V exatly, in my case), so the PS was up and okay. The digital voltage I measure is 4.98 V across the PGA's leads.
It is bypassed with a Sanyo OSCon close to the IC.
 
I am dead serious.
Huh, scary! :)
But apparently my point isn't picked up.
Apparently, I did not notice a point to pick up other than "I know the answer, but not telling you..." Sorry...aren't you, by an accident, work for TI? Or may it be you are so much worried about TI because it somehow affects the business of "LinearAudio.nl"? Because the only point that shines through your posts is that some stupid ones try to blame the mighty TI, while they are just not clever enough to make the thing properly (like others can :) ). I did ask you already - why you take this in such an offensive-resisting manner - but you didn't answer. At the other hand, you simply ignored some obvious observations coming not only from me, which alltogether exclude your "points" from the list of possible reasons.

Others have came up with ideas, suggestions, tried to participate in 2-way idea exchange. Just checked back your posts - and sorry to state, your "point" is a menthorship tone floating around lack of proper approach of poor implementers. And I have no intention to prove you that everything's fine with my (and others) design - and there IS a problem with a particular IC and it's stated NOWHERE in TI's documentation. Or maybe TI members have access to some secret erratas?
I'm enough self-aware to trust myself and argumented replies of community members. Hope others will do as well.

Good luck with your project.

Thanks a lot, this seems to be the only useful idea coming from you.
 
Last edited:
for my various PSUs, what I often did was to use a TLE rail splitter to take 24 or 10v and create a virtual ground. the TLE chip would also get hot IIRC, once runaway started to happen.

one of the diagrams I followed had a 10ohm R between digi gnd and analog gnd. I think I did install this and it worked for a while, using a 5v from my arduino and the analog voltage came from a 317 based supply that was virtual gnd split (nothing was true bipolar in my test setups).

another circuit did not share that 10ohm 'gnd breaker' idea. but all of my 3 pga builds have had this heatup stuff happen. things would work just fine for days and even weeks. I would exercise the pga chip quite a bit (lots of IR remote up down requests and even a lot of glide/ramps inside for-loops).

power was always on to those chips. I did swap inputs and that might have been significant. perhaps shorting some inputs (1/4" trs plug siding thru the shorting metal jack) would cause it go freak out somehow? I stopped using rca plugs and have gone to stereo TRS to carry my line levels. you can't use trs's without shorting as they are just that way when you insert or remove them. this is the only thing I can think of; and it still does not explain the freak-outs that happen all on their own (no jacks being changed).