Op amps help! 3x3 12v matrix mixer

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
So many op-amps, which one to choose?

Hi! I'm working on a custom line level mixer for our freestyle soundbike setup.

What I am looking for at this point I don't think is particularly
complex. It is essentially a 3x3 array mixer, that is, three channels
in, three channels out, and the in->out bus mix is fixed. (This will allow me to mix the output of our 4-mic-to-one-line-level mixer with the mp3 player output and feed it out to a few things, mostly the 2x100w class-D amp.)

It needs to run at 12v, which if I am not mistaken means 12v single side
supply op-amps. I spent an evening reading about op-amps, and there
was a huge discussion on which "sounded better", and I kinda got lost
there, I now have a list of like 30 op-amps which are talked about on the forums. Some differences are obvious, like dual or quad, 12v or 15v, etc. But there's tons of audiophile discussion about one opamp being "brighter" or "more upfront" than the other, which I suspect is gearhead nerdiness that doesn't apply to my project. Even if I read 30 datasheets and narrow it down to 10 or so that fit my specs, how to choose one? or does it even matter?

My requirements are 9-12v single side supply operation, low power, decent sound, and breadboardable (not smp). By "decent sound" I mean I'd like fair
frequency response, but it's being used outdoors, not in a studio, so
this isn't for audiophiles.

In addition to this post on diyaudio
http://www.diyaudio.com/forums/head...oject-7x6-headphone-monitor-matrix-mixer.html
I found a great website with modular mixer circuits that I think can be
used with very slight modifications for this project.
Universal Preamp/ Mixer (Part 2)

Maybe a big op-amp matrix in the wiki would help? min voltage, max voltage, power use, frequency response, and "audiophile comments". I started such a list in the wiki under "op amps".

The real question I have though: what op amps do you recommend for this project?
 
The NE5532 is a low-cost dual audio opamp used in many professional and semi-professional mixers.

The OPA2134 is a dual opamp costing about twice that of the 5532, it is regarded as a better performer.

The LME49710 is an even better performing single opamp which costs about 4x the OPA because you have to buy 2.

Singles have better crosstalk performance, but the NE5532 is a perfectly adequate part in your application and you don't have to cry if you blow up a few experimenting.

w
 
You can make a fake signal ground for dual supply amps like wakibaki mentioned by taking two 5.1V zener diodes line to not-line, with suitable resistors top and bottom for the wattage. I'm using 22 ohm resistors on 5 w zeners. This is working great on 4 op amps, see the .pdf file at the end of http://www.diyaudio.com/forums/anal...performance.html?highlight=disco+mixer+mid-fi
Most commercial mixers use 4558's, which are slow enough to not need local power supply decoupling. 5532's have a big following on here, but most suppliers I looked at had the 14 pin kind, and I was replacing 4558's DIP 8's for hiss, so I used 33078's. I've also used 4560's which seem to be the same as 33078 performancewise. People on here say 4562 is even better (low noise, fast). All the fast op amps (not 4558's) need disc caps .01 to .1 uf within an inch of the IC on the power supplies, and need about 20 pf across the feedback resistor, to avoid oscillating. After I killed the oscillation with dime caps, my 33078's or 4560's sound great. The 4558's hissed a bit, is why I replaced them, but most commercial PA and disco mixers have them in them. I paid $.60 for the 33078's from farnell.com (newark in the US).
 
Last edited:
www.hifisonix.com
Joined 2003
Paid Member
There is no doubt that the OPA2134 is absolutely inferior to the SA5532/34 at loads greater than 600 Ohms. Please take a look at Douglas Self's comments on this device. The OPA2134 is another op amp masquerading as an audio grade device that absolutelt does not meet the grade. Currently (February 2011), there are only 3 op-amps to be recommended for audio: SA5532/32, LM4562 and direct derivatives, AD797. Period.
 
Currently (February 2011), there are only 3 op-amps to be recommended for audio: SA5532/32, LM4562 and direct derivatives, AD797. Period.
I'm pretty happy with predecessor NJM4560, and interchangeable ST33078. Might try 4562 someday, but I'm not breaking my neck to pay $6 shipping to find out. Listening to the 33078's right now, in fact, the PAS2 is too expensive to leave on all the time (100 watts). Little vague why everybody keeps calling it LM4562, when all the datasheets I have originate from JRC & TI, and the 1998 Peavey CS800s amp had actual NJM4560's in it. Who was first, NationalSemi or JRC? My paper catelogs stopped coming in 1980 when the Army called me back.
Besides the resistor loads of the two transistor amps, I tried the 33078 direct coupled to the grid of a 7199 tube last week (ST70). Sounds great at no load but the internal resistor in the mixer and some capacitance.
I test sound with piano records of a Steinway versus the real Steinway console located between the speakers. What to you use?
 
Last edited:
Currently (February 2011), there are only 3 op-amps to be recommended for audio: SA5532/32, LM4562 and direct derivatives, AD797. Period.

In this particular application, it's about low-voltage performance. Which means the NE5532 and AD797 are out. As are the Linear Tech amps (I noticed there is no minimum supply voltage spec in the LT1028/1128 datasheet, I'm just inferring that this family will fail below 8V).

Also, the AD8599 datasheet has no spec, but some of the curves suggest operation down to 4.5V.

The LM4562 family is spec'ed at 5V, which is nice.

The TI amps (OPA2211/1612/2141 or its quad-brother the 4141) are spec'ed even more aggressively, at 4.5V.

All this surprises me a bit. I would have thought 4.5V (5V +/- 10%) audio op-amps to be a big market.

BTW, the ADI and TI parts are all SMD, but can be made "breadboardable" by using Browndog adapters (or, like a crazy friend of mine does it, DIL adapter sockets and some spare metal leads).

--
Greetz,
MatchASM
 
Last edited:
I had the 33078's operating at +-3.3 VDC for a while before I figured out there were 330 ohm resistors in the disco mixer power supply, and 33078's use more power than 4558's. They sounded fine on mag phone in and line in from CD player, nominal 1v PP(?) However, the FM radio hisses more when the output is turned down that low, so to cope with 4 V PP signals I boosted the power supply on the disco mixer to 1n5344's, or +-7.9VDC.
NJM4562 has no minimums listed on gain bandwidth product or slew rate, which 33078 does have. RC4560 (TI) only has typical numbers for those. NJM4562 has good noise numbers, plus a lot of typical performance charts.
Datasheetcatalog has no listing for a datasheet on 4562 from national semiconductor, only NJM, so if they want to sell any they need to communicate better. Of course, who cares about diy'ers.
 
Last edited:
Just to throw a monkey wrench into the works, I think it's reasonable to consider DC-DC converters to take ~+12V and convert it into ±16-17V. That way, you can run nice amplifiers and handle full, professional levels with no need to worry about headroom and low voltage amplifier performance.

Right now, I'm working on a box that will take +12V in and make all of the needed voltages out of that, and to do this, I'm using a combination of switching and linear regulators. The TI LMR16030 seems to be a nice, easy to use switcher that can be wired up to provide a positive output, and the same part can also be wired to produce a negative output. So, given a random +12V battery-like input, you could have a nice, somewhat precision regulator to produce proper ±15 to ±17V rails.

You'll have to do a reasonably nice layout using some large pieces of PCB foil to get low impedances and low noise, but honestly, given the ~2MHz switching that you can do with these modern switchers, it's not hard to isolate their switching noise to one part of the board, and have extremely high performance circuitry not that far away.

Having said all of that, my vote for an op amp is the LME49710, and if you absolutely cannot afford them, the New Japan Radio NJM5534. Both are really very nice, especially when run as inverters. The LME49710 is absolutely amazing IMHO - low noise, low offset, easy to stabilize, huge gain bandwidth, high output drive, basically a perfect audio amp, and they're extremely inexpensive. The AD797 is a tiny bit nicer in some ways, but it's 5x the cost, much more difficult to stabilize, and has a huge input bias current and noise. Not so nice for general line level and line impedance circuitry, but in the right place, unbeatable, once all of its requirements are met properly.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.