Hello to you all. Sorry I haven´t been very active on this forum since I joined in 2004.
Now I found some real problem that I want to share and discuss with members who have interest or the desire to discuss about this matter.
In past year I slowly designed high quality phono preamp. Started with circuit itself, various ideas, pros and cons, power supply concepts, pcb design and so on.
Al those things have been tested in proto phase, carefully and slowly, every detail was thought through several times, everything was tested in real life for stability, even in some extreme conditions.
Input section of MC input circuit is based on Ti OPA1611 op amp that has best conditions to work flawlessly (power supply decoupling, nice short traces, high quality parts).
When everything was put together I made more testing and fine tuning for lowest noise and stability, I even selected OP amps for lower DC offset in given circuit and so on.
After all systems were GO, I started listening sessions on real turntable with real phono cartridge in my own system. Cartridge was moving coil Denon DL103R, on fine arm and base, a good testing rig.
Sound was good, after some hours of spinning LPs became very good, everything was fine, and in one moment of playing music one channel has gone silent with little louder Boom or PUC! I must say that phono stage has been working on test bench for cuople of day with no poroblem.
I quickly turned off phono stage, lower the volume on line pre, disconnected turntable from phono stage and check for problem.
First of all, coil in one channel of cartridge was burned, other thing, there was some 7V of DC on moving coil input, power supply rails were lower than normal so 7V is what is left from 13V.
OPA1611 has gone wild in one moment with no good or real reason!!!
I desoldered OPA and measure resistance between power V+ and positive input +IN pins and there was almost short with about 2 ohm resistance.
I never encountered this type of fail in all my years of working with OP amps, or servicing hundreds of audio devices, except two times, when my client had same issue on commercial phono preamplifier made by Italian manufacturer Gold Note, their model PH-10, which uses same series of OP amps only J-fet input type, OPA1641.
Now I am in phase of exchanging information with Texas instruments people, I gave them partial schematic of MC section, PCB layout of that section. They didn´t find single error in my design, but refuse to talk about potential problem in OPA16xx design.
This is dangerous failure, and it´s not isolated one!!
Do you have any similar experience or clue what is going on here?
Now I fear to use this OP inside my design, I don´t want more burned cartridges. I never encountered case that OP fail that resulted in shorted power rail to input node.
Best regards
Now I found some real problem that I want to share and discuss with members who have interest or the desire to discuss about this matter.
In past year I slowly designed high quality phono preamp. Started with circuit itself, various ideas, pros and cons, power supply concepts, pcb design and so on.
Al those things have been tested in proto phase, carefully and slowly, every detail was thought through several times, everything was tested in real life for stability, even in some extreme conditions.
Input section of MC input circuit is based on Ti OPA1611 op amp that has best conditions to work flawlessly (power supply decoupling, nice short traces, high quality parts).
When everything was put together I made more testing and fine tuning for lowest noise and stability, I even selected OP amps for lower DC offset in given circuit and so on.
After all systems were GO, I started listening sessions on real turntable with real phono cartridge in my own system. Cartridge was moving coil Denon DL103R, on fine arm and base, a good testing rig.
Sound was good, after some hours of spinning LPs became very good, everything was fine, and in one moment of playing music one channel has gone silent with little louder Boom or PUC! I must say that phono stage has been working on test bench for cuople of day with no poroblem.
I quickly turned off phono stage, lower the volume on line pre, disconnected turntable from phono stage and check for problem.
First of all, coil in one channel of cartridge was burned, other thing, there was some 7V of DC on moving coil input, power supply rails were lower than normal so 7V is what is left from 13V.
OPA1611 has gone wild in one moment with no good or real reason!!!
I desoldered OPA and measure resistance between power V+ and positive input +IN pins and there was almost short with about 2 ohm resistance.
I never encountered this type of fail in all my years of working with OP amps, or servicing hundreds of audio devices, except two times, when my client had same issue on commercial phono preamplifier made by Italian manufacturer Gold Note, their model PH-10, which uses same series of OP amps only J-fet input type, OPA1641.
Now I am in phase of exchanging information with Texas instruments people, I gave them partial schematic of MC section, PCB layout of that section. They didn´t find single error in my design, but refuse to talk about potential problem in OPA16xx design.
This is dangerous failure, and it´s not isolated one!!
Do you have any similar experience or clue what is going on here?
Now I fear to use this OP inside my design, I don´t want more burned cartridges. I never encountered case that OP fail that resulted in shorted power rail to input node.
Best regards
Damn a blown expensive MC .... not too fun 😢
I know from myself that being the designer and maker, makes you look for faults other places first 😉 ..... been ther many times with my blown power amps 😭 .... but I also know it's the cost of DIY 🙂
I recently had a problem with a DSP DAC I have designed. It also lost sound from time to time. Was sure it was the DAC that was faulty .... after a lot of investigation I found that a single strand from a clip off, of some wire was making a short .... only found it using a microscope ... go figure .... and yes my eyes are not what they wher when I was 20 😉
If it was really a general problem with these series, for which there must have been sold in the millions, TI would of course know and would have to act.
.... but then again you might be right ... what do I know ... just saying ..... think this is Ocaam's razor type of think 😉
In you design (and I'm not saying it it the cause, because I do not know), but switches can of course always make havoc with bouncing, would terminate them on both sides and maybe you also have, if the shown is not the full schematics, but a resistor on both sides I would think is always good to ensure no open circuit at any time.
Check solderings, both for bad solderings, and for shorts.
And an electronic microscope for the PC is worth gold 😉
I know from myself that being the designer and maker, makes you look for faults other places first 😉 ..... been ther many times with my blown power amps 😭 .... but I also know it's the cost of DIY 🙂
I recently had a problem with a DSP DAC I have designed. It also lost sound from time to time. Was sure it was the DAC that was faulty .... after a lot of investigation I found that a single strand from a clip off, of some wire was making a short .... only found it using a microscope ... go figure .... and yes my eyes are not what they wher when I was 20 😉
If it was really a general problem with these series, for which there must have been sold in the millions, TI would of course know and would have to act.
.... but then again you might be right ... what do I know ... just saying ..... think this is Ocaam's razor type of think 😉
In you design (and I'm not saying it it the cause, because I do not know), but switches can of course always make havoc with bouncing, would terminate them on both sides and maybe you also have, if the shown is not the full schematics, but a resistor on both sides I would think is always good to ensure no open circuit at any time.
Check solderings, both for bad solderings, and for shorts.
And an electronic microscope for the PC is worth gold 😉
In my DAC there are three OPA1641 in Walt Jung shunts and all three have been working for a long time without any problems. The only problem I have noticed so far is that they are a little more prone to self-oscillation in that regulator, but that can be easily solved.
I'll be testing them soon as I/V with a +/-15V supply and I certainly wouldn't want eight PCM1702s to fry.
I'll be testing them soon as I/V with a +/-15V supply and I certainly wouldn't want eight PCM1702s to fry.
Various thoughts- look up "walking wounded"+"esd". The part might have been partially damaged prior or during installation. You don't show your esd protection, but hopefully nothing came in the input that damaged it. I've seen many phono preamps with AC coupled cartridges. This seems like a good idea, but I don't know if it would have protected the cartridge because the cap has to be reasonably large. Worth looking into with expensive cartridges. Thermoworks no longer makes their close focus IR tool, that was perfect for spot checking ICs, but it would be nice to get a temperature reading on the part to be sure it wasn't too warm. Some opamps work very nicely until you get near the maximum supply, then they really dissipate too much power for longevity. Be sure the PCB layout gets the heat out of the chip. Can't remember the number but I had that problem with an OPA part that was highly touted. Naturally you checked for oscillation.
Your description fits to a latch up failure I sometimes experienced with CMOS-ICs. An ESD-pulse on an (arbitrary) IC pin ignites some intrinsic, parasitic thyristor that now shorts the supply terminals. As long as your supply delivers at least the latching current of typical 0.2~0.5amps, this current continues - until the die burns.
What you can do: Study the data sheet, is there at some non prominent place a specification of latching current? If so, I recommend a series resistor in each supply line that limits short circuit current below that value. Once triggered, the thyristor immediately blocks and no harm happens. This measure reduced failures of CPLDs in our boundary scan test field dramatically!
Furthermore I would not expose any IC-pin directly to external wiring - i.e. the non inv inputs here - but always add some series resistor 100-1000 Ohms. Which admittedly may be detrimental to lowest noise figures.
What you can do: Study the data sheet, is there at some non prominent place a specification of latching current? If so, I recommend a series resistor in each supply line that limits short circuit current below that value. Once triggered, the thyristor immediately blocks and no harm happens. This measure reduced failures of CPLDs in our boundary scan test field dramatically!
Furthermore I would not expose any IC-pin directly to external wiring - i.e. the non inv inputs here - but always add some series resistor 100-1000 Ohms. Which admittedly may be detrimental to lowest noise figures.
I have microscope for microsoldering and top series Weller soldering station with various pencils.
And I think my soldering is pretty neat.
So no errors here. Flux is alwas cleaned even they say "no clean".
And I think my soldering is pretty neat.
So no errors here. Flux is alwas cleaned even they say "no clean".
The PCB layout is less than ideal, is there a ground plane? The OP-Amp bypass capacitors should be ceramic smd NP0's. In the top photo the +Rail bypass capacitor is not located next to Pin7.
From page 20 of the OPA161x datasheet (layout guidelines):
From page 20 of the OPA161x datasheet (layout guidelines):
- connect a low-ESR,0.1-μF ceramic bypass capacitors between each supply pin and ground, placed as close to the device as possible
- Place the external components as close to the device as possible, keeping RF and RG close to the inverting input minimizes parasitic capacitance.
- Keep the length of input traces as short as possible. Always remember that the input traces are the most sensitive part of the circuit
- Refer to Figure 36 for layout details.
Last edited:
This problem reminds me of a preamp design of mine from around 35 years ago and which used a few 4017 CMOS chips in the discrete logic input selector. One of them, and it was always the same one, kept failing for no obvious reason. I seem to remember adding a decoupling cap directly across the power pins and perhaps a Zener clamp. It never failed again.
What I'm saying is it wasn't the chips, it was something weird with the design implementation. I would definitely try adding a couple of inverse parallel diodes across the supplies.
Could just one rail be dropping out momentarily?
What I'm saying is it wasn't the chips, it was something weird with the design implementation. I would definitely try adding a couple of inverse parallel diodes across the supplies.
Could just one rail be dropping out momentarily?
Don't judge by photo. Ground plane i present, it's multilayer pcb. Power pins have decopling caps on under 2mm of traces.
People, design is working perfecly, no untability issues related to bad layout.
Read my initial text please.
I still think you should look at the power supply. The internal block diagram for the chip shows a low impedance route (the protection diodes) from the supply pins to the inputs. If a rail dropped out and was driven to an incorrect polarity you have possible reason for what you have experienced.
No power supply rail dropping,
It isn't case only in this design.
I encountered same problem on commercial components.
Other channel which is the same board, works with no issues.
I am not newbie in this stuff.
It isn't case only in this design.
I encountered same problem on commercial components.
Other channel which is the same board, works with no issues.
I am not newbie in this stuff.
If the oscillation is caused by parasitics due to layout then it is possible that the other channel works without issues e.g. due to component variances.
Unless you have separate power supplies for each channel power supply rail dropping would have impacted both channels.
Mooly has a very good point. When you power up or down, if the rails don’t come up smoothly(eg + and - momentarily reversed), could you be damaging the chips protection diode? Do you have reverse voltage diode clamps across your PSU?
I would look at the circuits around the chip before looking for a chip fault.
(you could use a DC blocking cap in your MC inputs: load resistors on the input side, film cap of 1uF with 47k bias on the other side).
😊
I would look at the circuits around the chip before looking for a chip fault.
(you could use a DC blocking cap in your MC inputs: load resistors on the input side, film cap of 1uF with 47k bias on the other side).
😊
Why not to add an input coupling cap? It also may be choosen to low frequency rolloff at f<20Hz.