OPA1612AID's for the AMB Audio Behinger DCX2496 are on deep backorder from Mouser and Digikey. Is there another OPAMP that can be substituted and have the same performance.
NOTE: I realize I could stick just about any dual 8pin OPAMP in place but the goal here is high quality audio performace.
NOTE: I realize I could stick just about any dual 8pin OPAMP in place but the goal here is high quality audio performace.
2496 is digital speaker management
having analog in / analog out
that is where the opamps would be
this a repair? what is wrong with the opamps
Currently Mouser has about 7000 of the SOIC package in stock , and about 200 of the SON8 package in stock
not sure the issue
having analog in / analog out
that is where the opamps would be
this a repair? what is wrong with the opamps
Currently Mouser has about 7000 of the SOIC package in stock , and about 200 of the SON8 package in stock
not sure the issue
Last edited:
OEM analog board being replaced with a drop in upgrade board here:
https://www.amb.org/audio/kappadcx/
https://www.amb.org/audio/kappadcx/
No coupling caps. is that the cool thing.
nobody wants DC and offset in the chain.
gimmick products.
well regardless looks like SOIC-8 seems to be 7000 of them in stock
nobody wants DC and offset in the chain.
gimmick products.
well regardless looks like SOIC-8 seems to be 7000 of them in stock
The stock analog audio section of the DCX2496 is not that good. There are several tests online using a Sound Technology analyzer which demonstrates the deficiencies quite well. This is not audiophoolery. Now the mods online for the AES input - there I draw the line in the snake oil.
The output of my DCX2496 feeds a multichannel volume controller based on BB PGA2310 attentuator chips. Here there is a 22uf bipolar electrolytic cap on the input side. This will remove any DC coming out of the DCX2496. Yes, a pure DC coupled audio path is asking for trouble. But a strategically placed coupling cap is all you need to provide DC protection. The PGA2310 are very good attenuators, the best of the family which also includes the 2311, and 2320 that are of lower performance. I'm not worried about DC coming out of the 2310's. The amps are tube topology with DC coupling to the input stage tube grid. They are also PP Ultra Linear topology so there are two internal film caps on the plus and minus phase signals to the output tube grids. And of course there is an output transformer so the speaker drivers can never see any DC.
Yes TI has factory stock - minimum purchase 1500 pieces! Both Mouser and Digikey are on backorder. I just ordered all my parts.
The output of my DCX2496 feeds a multichannel volume controller based on BB PGA2310 attentuator chips. Here there is a 22uf bipolar electrolytic cap on the input side. This will remove any DC coming out of the DCX2496. Yes, a pure DC coupled audio path is asking for trouble. But a strategically placed coupling cap is all you need to provide DC protection. The PGA2310 are very good attenuators, the best of the family which also includes the 2311, and 2320 that are of lower performance. I'm not worried about DC coming out of the 2310's. The amps are tube topology with DC coupling to the input stage tube grid. They are also PP Ultra Linear topology so there are two internal film caps on the plus and minus phase signals to the output tube grids. And of course there is an output transformer so the speaker drivers can never see any DC.
Yes TI has factory stock - minimum purchase 1500 pieces! Both Mouser and Digikey are on backorder. I just ordered all my parts.
Last edited:
OK, that's a different suffix but may be perfectly acceptable. It costs more but is in stock!
Thanks for the heads up.
Thanks for the heads up.
Correct different suffix
1612 is a 1612
they all link to the same datasheet
its kinda comical to say a equivalent for a 1612 is a 1612
likely why the original suffix was chosen for the lower cost
you would likely need to ask TI the difference.
same opamp either way . More expensive package likely has better thermal range or G force ratings
this case it is more a savor since the more expensive unit
is less likely to get bought out quickly.
1612 is a 1612
they all link to the same datasheet
its kinda comical to say a equivalent for a 1612 is a 1612
likely why the original suffix was chosen for the lower cost
you would likely need to ask TI the difference.
same opamp either way . More expensive package likely has better thermal range or G force ratings
this case it is more a savor since the more expensive unit
is less likely to get bought out quickly.
Last edited:
The Q in the suffix means "Qualified for Automotive Applications", usually just a little better specs for a little higher price. Look also for p/n with a R in it, that just mean Tape & Reel....
Soundwise the same....
Soundwise the same....
Automotive qualification amongst other things involves very tight thermal cycling requirements - because vehicles experience this every day (or every drive) as well as a very wide temperature range, whereas even harsh industrial environments are usually fairly constant temperature. Packaging expansion/contraction is one route to failure in microchips, typically the bond wires get pulled off the die or fatique over time due to mechanical forces. No-one wants their car to stop working after a few years because one electronic component failed (and there may be 100's or 1000's of chips in a modern car).
I ended up going with the Ambs Lab board which is a rather more expensive option I think than modifying the original I/O PCB. Have you looked at the OPA1602? I was unwilling to pay the freight for almost a dozen OPA1612 for the Ambs Lab I/O and these provide excellent performance, they are about 7dB noisier than the 1612, but it is doubtful you would notice in practice? They are a few bucks each.
Your comment about the AES input mods is interesting. I did not know anyone was writing about mods for the AES input, I had a problem and solved it empirically.
Have you ever used the AES input? I am not sure if installing the AMBs lab I/O board or one of the other modifications I made to mine contributed to my problems, but I had terrible common mode noise issues with mine, all resolved interestingly enough by replacing the original AES input transformer and all of the wiring with a better input transformer + CM choke, and completely isolating the XLR input to AES/EBU input from the analog I/O board.
I use an SHD Studio with the DCX and when I selected USB input for calibration got a nice low level whine on all output channels of the DCX in addition to a bunch of odd back ground noises in other modes (LAN and AES input) which were mitigated in several steps ending with the wholesale replacement of everything ahead of the CS8420 input.. The only noise out of the DCX now is clean sounding white noise, no audible artifacts. (110dB mid and HF horns with 15dB of amplification DCX configured for 4Vrms at 0dBFS) [Of note is that nothing else connected to the AES outputs on the Studio made any noises at all.]
FWIW I think the quality of the electrical design in both the analog and digital portions of the DCX is pretty abysmal. There are plenty of flaws in that DSP board if you dig for them. I have the post 2016 version and the best thing about it is the chips used. After spending six months modifying it I am relatively happy with the performance.
Last edited:
Kevinkr,
As for the DCX2496, and many other AES audio products, the input wiring is not to strict RF standards. That said, the minor AES signal degradation caused by this is insignificant. If you are getting error free audio from the AES input, that's as good as it gets. I can't see where any modification to the DCX digital input would improve audio quality. As for the jitter quandary, wiring does not directly induce jitter but rather the reflections caused with severe impedance mismatch will cause errors.
I did purchase the AMB board but will be modifying the AES input feature wise. I am adding another XLR connector next to the power connector and wiring that directly to the AES side of the CHAN1/AES relay. Then jumping the relay to the analog input mode. This allows the AES signal to always be connected along with an analog signal and just use the menu control to switch. No need to swap cables around. I just used a piece of 24ga shielded pair cable. Just audio stuff, not even true 110ohm AES cable as it doesn't matter for that short length.
As for the DCX2496, and many other AES audio products, the input wiring is not to strict RF standards. That said, the minor AES signal degradation caused by this is insignificant. If you are getting error free audio from the AES input, that's as good as it gets. I can't see where any modification to the DCX digital input would improve audio quality. As for the jitter quandary, wiring does not directly induce jitter but rather the reflections caused with severe impedance mismatch will cause errors.
I did purchase the AMB board but will be modifying the AES input feature wise. I am adding another XLR connector next to the power connector and wiring that directly to the AES side of the CHAN1/AES relay. Then jumping the relay to the analog input mode. This allows the AES signal to always be connected along with an analog signal and just use the menu control to switch. No need to swap cables around. I just used a piece of 24ga shielded pair cable. Just audio stuff, not even true 110ohm AES cable as it doesn't matter for that short length.
Last edited:
I have the AMB Labs board and ended up removing all of the AES wiring from that board and the audio ribbon because EMI from the AES port resulted in significant audible noise on all channels. (low level tones and an audibly variable noise floor)
Removing all that wiring resulted in a very substantial improvement in noise floor, and replacing the AES input transformer with a better one finally solved my problem. Just very low level white noise now.
Hopefully you do not have the same problem, as I hinted mine is very heavily modified. It took me a while to figure it out.
I am relatively sure this problem exists in the original I/O PCBA design as well, but did not investigate to root cause because at that time I did not understand the implications of what I was hearing down in the noise floor.
I design consumer electronics for a living and generally don't run into the sort of EMC problems I have encountered in the DCX. My other AES components don't have these problems...
Removing all that wiring resulted in a very substantial improvement in noise floor, and replacing the AES input transformer with a better one finally solved my problem. Just very low level white noise now.
Hopefully you do not have the same problem, as I hinted mine is very heavily modified. It took me a while to figure it out.
I am relatively sure this problem exists in the original I/O PCBA design as well, but did not investigate to root cause because at that time I did not understand the implications of what I was hearing down in the noise floor.
I design consumer electronics for a living and generally don't run into the sort of EMC problems I have encountered in the DCX. My other AES components don't have these problems...
Last edited:
Hmmm, I might change my idea! Since I will have a dedicated AES input jack, and the AES interface circuit is on the digital board, why not just get rid of the AES ribbon pair and run a shielded pair from the new jack directly to the digital board? Can't hurt.
I also use those "pin them yourself" femail header connectors. I might also make a new cable harness with all the audio signals shielded pairs?
I also use those "pin them yourself" femail header connectors. I might also make a new cable harness with all the audio signals shielded pairs?
Last edited:
Have a look at the onboard transformer as well - I don't think the one Behringer used meets the minimum 30dB common mode rejection specification for AES, both Murata and Pulse AES transformers meet or exceed it. I settled on the Pulse transformer and circuit below CM choke and transformer on a separate PCBA, these are available at DigiKey:
I removed the original transformer from the DSP board.
Maybe overkill.. 😀
I removed the original transformer from the DSP board.
Maybe overkill.. 😀
That last part is one I have not done, but yeah that is a great idea and really cannot hurt. Twinax would seem to be the answer, although it might be a challenge to get six pairs to fit. (3 more if you do the inputs)
Perhaps a foil shield over all of them with kapton tape for insulation on top of the shield? I don't seem to need it, but there is no SMPS present in mine either, so...