I have a DBX 163X compressor. I upgraded all the capacitors in terms of age and quality changes. I am in the OPAMP area and there are 4 in number. What upgrades can I make to these in terms of clarity? I don't mind adapting to fit.
OA1 LF353
OA2 5534N
OA3 4558DD
OA4 4558DD
I was thinking these Operational Amplifiers because of the ULTRA LOW NOISE
OA1 AD8599
OA2 AD8597
OA3 AD8599
OA4 AD8599
Is there an issue here? I already have browndog adapters ready and the OPAMPS in hand. I attached the diagram so it could be easier to assess.
OA1 LF353
OA2 5534N
OA3 4558DD
OA4 4558DD
I was thinking these Operational Amplifiers because of the ULTRA LOW NOISE
OA1 AD8599
OA2 AD8597
OA3 AD8599
OA4 AD8599
Is there an issue here? I already have browndog adapters ready and the OPAMPS in hand. I attached the diagram so it could be easier to assess.
Attachments
The schematic is still too small to read blown up. You may have to slice up bits and post to allow us to read it.
It doesn't take much to beat a 4558 in the hiss dimension, particularly at high gain. I found significant hiss improvement in a 50x gain circuit with ST33078 costing $.38 ea. The phosphor bronze socket was $1. I did have to put local (1" away) supply bypass cap in (for 2 op amp packages) and parallel the feedback resistor with a 33 pf disk cap to eliminate oscillation. 1 0.1 uf disk cap +rail to - rail was enough, I didn't bypass the analog ground. I drilled holes in the board and bent the leads over to the rail traces to solder them.
LF353 is so old I'm not going to look the datasheet up.
5534 is not bad in the noise dimension, even though a 1974? design. You might not hear improvement. Do the experiment. I wouldn't monkey with surface mount to start with. Besides 33078, NJM2043 and 2068 are inexpenxive low noise parts, and LM4562 was only about $2. But you've spent the money on "ultra low noise", do your experiment & report back. Warning, any amp faster than 1 v/microsec slew rate may need the bypass caps to avoid oscillation if your board doesn't have them now.
On hiss upgrade, don't forget to update high value (100k up) carbon comp or carbon film resistors with metal film. They do hiss less, for a cost of about $.20 each or less.
It doesn't take much to beat a 4558 in the hiss dimension, particularly at high gain. I found significant hiss improvement in a 50x gain circuit with ST33078 costing $.38 ea. The phosphor bronze socket was $1. I did have to put local (1" away) supply bypass cap in (for 2 op amp packages) and parallel the feedback resistor with a 33 pf disk cap to eliminate oscillation. 1 0.1 uf disk cap +rail to - rail was enough, I didn't bypass the analog ground. I drilled holes in the board and bent the leads over to the rail traces to solder them.
LF353 is so old I'm not going to look the datasheet up.
5534 is not bad in the noise dimension, even though a 1974? design. You might not hear improvement. Do the experiment. I wouldn't monkey with surface mount to start with. Besides 33078, NJM2043 and 2068 are inexpenxive low noise parts, and LM4562 was only about $2. But you've spent the money on "ultra low noise", do your experiment & report back. Warning, any amp faster than 1 v/microsec slew rate may need the bypass caps to avoid oscillation if your board doesn't have them now.
On hiss upgrade, don't forget to update high value (100k up) carbon comp or carbon film resistors with metal film. They do hiss less, for a cost of about $.20 each or less.
Last edited:
Hi Gophat,
Don't forget that your noise might be due to the voltage-controlled op amps and other components, so you might get a reduction in noise, but only if those op amps were responsible for it. The NE5532A is not a noisy op amp, so it would work quite well. If your dBx is still noisy, I don't think a quieter op amp is going to help much.
-Chris
Don't forget that your noise might be due to the voltage-controlled op amps and other components, so you might get a reduction in noise, but only if those op amps were responsible for it. The NE5532A is not a noisy op amp, so it would work quite well. If your dBx is still noisy, I don't think a quieter op amp is going to help much.
-Chris
If you are replacing older, slower op-amps with newer faster high speed op-amps, I hope that you have an O-scope good to 20 to 50 MHz. It's real easy to turn the circuit into an oscillator.
The LF353 should be replaced with an upgraded FET opamp---an OPA1642 if you can deal with a SOIC-to-DIP adaptor; an OPA2134 if you want a DIP. Both are significant improvements over the LF353. The 5534 is probably fine; but I would replace the 4558s with LM4562s. As indianajo says, 0.1uF supply bypass caps and a parallel 33pF feedback cap are VERY good ideas! 8599s have good voltage noise specs, but their current noise is rather high; so I would avoid them in higher-impedance circuits.
Last edited:
Note that the only opamps actually in the direct audio path are the '353 at the input and the 5534, everything else is sidechain or metering.
I would be careful about the 5534, the gain cell output is quite capacitive and they seem to have done some work around the 5534 to stabilize things.
The '353 can probably be safely upgraded, but is probably not the limiting factor for noise.
If your input levels are sane and you are not applying a stupid amount of compression, noise really should not be a problem.
Regards, Dan.
I would be careful about the 5534, the gain cell output is quite capacitive and they seem to have done some work around the 5534 to stabilize things.
The '353 can probably be safely upgraded, but is probably not the limiting factor for noise.
If your input levels are sane and you are not applying a stupid amount of compression, noise really should not be a problem.
Regards, Dan.
Hmmm.....perhaps not, but the LF353 is about 11 db noisier than an OPA1642 per datasheet info. That would seem to make a significant difference.
I was wondering who would spot that 🙂
I see Gophat hasn't bothered to acknowledge any of the replies in exactly the same way he did not with his only other post on the forum some 8 years ago.
But is it even 1dB noisier then the noise floor of the input signal, after taking into account the thermal noise from the feedback gain control pot when you turn the gain up?
A 10k resistor at room temperature is a noise source of ~1.8uV RMS, in a 20k bandwidth. I make the opamp noise only 2.2uV RMS (in a 20Khz bandwidth, at unity gain and ignoring current noise contributions, reasonable in a low Z circuit), and there is ~20k resistance as an V/I converter between the LF353 and the Blackmer cell, so a perfect, noiseless opamp would lower the noise floor by maybe 3dB (Turn the input gain up and the increased thermal noise from the feedback resistance completely swamps everything else).
My betting is that the opamps are not the limiting factor.
Regards, Dan.
A 10k resistor at room temperature is a noise source of ~1.8uV RMS, in a 20k bandwidth. I make the opamp noise only 2.2uV RMS (in a 20Khz bandwidth, at unity gain and ignoring current noise contributions, reasonable in a low Z circuit), and there is ~20k resistance as an V/I converter between the LF353 and the Blackmer cell, so a perfect, noiseless opamp would lower the noise floor by maybe 3dB (Turn the input gain up and the increased thermal noise from the feedback resistance completely swamps everything else).
My betting is that the opamps are not the limiting factor.
Regards, Dan.
I'm not sure what your 10K resistor has to do with anything here. The noise contribution of the LF353 opamp is 2.54 uV, based on the datasheet's 18nV/√Hz voltage noise spec; the OPA1642's noise(5.1nV/√Hz) would be 0.72uV---about 11 db less. The THD of the OPA1642 is also an order of magnitude less than the LF353. Turning the gain up won't affect the signal-to-noise ratio in any case. However, it probably is true that the limiting factor here is the UPC1252H2 voltage-controlled amplifier. I can't find a datasheet on it (in English that I can read, anyway); so I'm not sure what it's noise factor or distortion is, but it's probably worse than either of the opamps. I'm just saying that an OPA1642 input stage is better than an LF353 input stage.
Last edited:
What I was getting at was that the thermal noise was of the same order as the noise from a distinctly unimpressive opamp, distortion is a whole other question, but for that I would be looking at a more modern gain cell (THAT Corp make some that are pin compatible but that use more modern processes) in the first instance.
Regards, Dan.
Regards, Dan.
I looked up THAT Corp; they do offer a VCA which looks promising--a 2181 with pretty good noise and distortion specs (-98 dbV output noise and 0.0025% THD) . Unfortunately, minimum order is 25 pieces at $4.61; it is also not 100% pin-compatible with the device used in the dbx163x; though it probably could be modified to accept the THAT device.
The THAT part is widely available in small quantities, Mouser/Farnell, the usual places all have them.
Note that the control input needs to driven from a low impedance source to get the THD performance, and that noise here turns directly into noise sidebands in the audio.
Regards, Dan.
Note that the control input needs to driven from a low impedance source to get the THD performance, and that noise here turns directly into noise sidebands in the audio.
Regards, Dan.
Yes, the THAT 2180 application notes specifically state that the control port must be driven by a low impedance (<50Ω), low noise source, with a 100 Ω resistor in series with a 1.5 nF capacitor from the control port to ground. It also states that even a 5534 driver will increase the noise output slightly----perhaps a super low noise opamp (ala LT1028) would be ideal. It certainly implies that the dbx163x could benefit from a replacement for the 4458s that are currently in the control port path.
- Status
- Not open for further replies.