rpocc

Joined: Jan 23, 2010 Posts: 17 Location: Russia
Audio files: 1
|
Posted: Tue Jun 26, 2012 11:18 am Post subject:
Reducing the Doepfer A-138a/b noise with OP470 Subject description: To make story shorter replace A-138's TL064 with OP470 and have fun. |
 |
|
Hi, everyone.
I found very simple way to dramatically improve signal/noise ratio of a Doepfer modular system. Here i assume use of the A-138 inevitable as it is necessary utility component for creating complex, fat patches and also it is very common module to pass sound to sound system, recorder, etc.
We all use mixers all time — to mix several oscillators, mix CV (with linear mixers), to mix gate signals and get simple sequencer. As also do VCAs or filters, mixers contain amplifier. In case of mixers we are dealing with an operational amplifier working at little gain or even unity gain.
Unfortunately, complex components such as OpAmps are a bit noisy things. Moreover, mixers often may work with wide dynamic range. All of this leads to problems with noisy output of the whole system. Did you experienced a situation where you've made some very pleasant sound utilizing your lovely custom VCOs, softened it with Lo-Pass filter, set up some modulations, but it all can be easily ruined by VCA-Mixer chain? You try to increase volume, but either mixer's input is overloading or thermal noise becomes clearly audible. That's why we need modules to be the same quality level and that's why some critical components must be precise and expensive.
I've done some investigations about OpAmps noise characteristics and done some experiments with my modular. The following is what i've found.
The A-138 mixer has extremely simple schematics: four passive attenuators built around quad OpAmp and one more attenuator for master level. That's all. But it can be simply overloaded and is really noisy component. On the same volume any audio source connected directly after VCA has less noise than identical setup passed thru one A-138 channel.
The OpAmps used in Doepfer modules most of time is:
LM324, TL062, TL064, TL084.
Look at the characteristics taken from ST and AD datasheets:
Code: |
TL062/064 (Low-power version)
Equivalent Input Noise Voltage: 42 nv/√Hz
Channel Separation: 120 dB
LM324 (Old design)
Equivalent Input Noise Voltage: 40 nv/√Hz
Channel Separation: 120 dB
TL072/074/082/084 (Low-noise)
Equivalent Input Noise Voltage: 15 nv/√Hz
Channel Separation: 120 dB
|
As we can see, LM324 is not the best existing amplifier for audio purpose however it's actually good model and there's much worse models usually used only in power applications.
Now let's take a look on one AD's modern component. OP470 is expensive operational amplifier dedicated especially for audio. It has the almost best electrical characteristics and it has supply and input conditions that perfectly fit for modular synthesizer purposes: ±18v supply, rail-to-rail input, large gain, hi-linearity, etc. Also it is pin-to pin compatible with LM324 (not any HQ OpAmp do!!!). So, it fits as direct replacement for OpAmps used in certain models.
Code: |
AD OP470
Equivalent Input Noise Voltage: 5 nv/√Hz
Channel Separation: 125—155 dB
|
I've got trial amount of 6pcs OP470AY from China at price of 5$ each and replaced standard OpAmp on the A-138 with it. The amplification chain noise just almost disappear and even dynamic range has increased (distortions are gone away)! So, i got nice possibility to increase both input and output level. I did not measured s/n ratio before and after replacement because it was clearly audible, as i said before. I believe, it's close to 40—60 dB. My test setup was like A110Sine->A-106-6LP2 (Closed, no reso.)->A130->A138.
The A-138 costs about 50 Euro, so you need to add just about $5 ($15 for military version) to get high-quality OpAmp and you'll get a much better mixer.
Also i got better input dynamic range of the A-106-6 replacing input OpAmp (the chip next to CEM3379). I tried other modules but got no significant result with Dark Energy, A-116, A-130, A-119, A-120, A-121. However I believe OP470 must fit to almost any simple mixer-like modules used for audio such as panners, x-faders, matrix mixer, maybe to some pre-amps or VCAs. Also many modules such as oscillators have output buffers or input pre-amps at there's non-zero chance to make it better. I'll return to my investigations later when i receive Hi-end VCO and VC Panner modules.
It would be highly appreciated if somebody who will try to improve his own mixer inspired by this post, he will take the opportunity to test other modules, even non-Doepfer. The replacement is simple done with screwdriver (or chip-removing tool) because every IC on Doepfer modules is placed thru DIP-panel. And if IC placed right (see key) there's no risk to damage anything.
You can also try to install other OpAmps, like TL-072 (but changes would not be dramatic) or maybe another AD, THAT, etc production. It's also interesting to hear about another successful chip replacements.
Best regards,
Dmitry Shtatnov _________________ Modularsynth.ru — BLog
SSSR Labs — Eurorack Modules
Quorum — My first band
Eternal Wanderers — my second band |
|
EdisonRex
Site Admin

Joined: Mar 07, 2007 Posts: 4570 Location: London, UK
Audio files: 171
|
Posted: Fri Jun 29, 2012 4:18 pm Post subject:
|
 |
|
I have not had the opportunity to welcome you here, so Dmitry!
This is an interesting post for a couple of reasons. You quite rightly call out the specs of the Doepfer modules (which are also used by many others). It's a valid point about the mix path too.
I'd also be surprised if this wasn't said somewhere in these forums before, probably many times. Still, it is good advice. _________________ Garret: It's so retro.
EGM: What does retro mean to you?
Parker: Like, old and outdated.
Home,My Studio,and another view |
|