|
|
Post by Blackdawg on Jun 8, 2018 11:48:33 GMT -6
So Im working on a BOM and sourcing parts to build the DIY Racked MEQ hopefully to start my mastering setup with a SilverBullet.
the BOM calls for 11 OPA2604 but I can't find them anywhere. Are they out of production? I know the NE5532 would work possibly. But is a bit more sterile sound from what i've read as well as it will take more power.
I haven't A/B the two op amps so I don't really have personal preference yet. But would like to try the OPA2604s as directed in the BOM first..
|
|
|
|
Post by svart on Jun 8, 2018 12:06:43 GMT -6
I find the older OPA stuff to be strange. They absolutely impart a "sound" on the signal and are well known to do that, aka the "(Burr) Brown sound". I personally think the 5532 is a more neutral sound comparatively and most devices I swapped in 2604 and 2134 I ended up removing and putting in something else.
The OPA2604 is FET input IIRC, so using a BJT part like the 5532 would necessitate checking impedances to make sure the resistor values around the opamps are suitable for using something other than a FET opamp. Sometimes using such high values intended for FET opamps can cause issues with BJT parts.
I've started using the TLE2072 in some places that called for TL072 (another fet input opamp) and they seem OK, but such high bandwidth that you need to make sure you're using suitable feedback filtering to squash oscillations.
I'm also a fan of the LM4562, which is another BJT part but has fairly high input impedance so I'd wager it wouldn't cause an issue.
|
|
|
|
Post by svart on Jun 8, 2018 12:08:20 GMT -6
And as an aside, I might actually have 11 of the DIP OPA2604 somewhere. Might have to dig them out of a box. Some new, some pulls. I guess I could be persuaded to sell them.
|
|
|
|
Post by Blackdawg on Jun 8, 2018 12:31:46 GMT -6
Well the kit is a mastering parametric EQ based off the Sontec. So maybe i'll contact them to see if the NE5532 would a simple swap or not. I don't really need the EQ to be "vibey" i'd rather it be clean and open.
If its not a simple swap..I might hit you up to try and persuade you! haha
|
|
|
|
Post by jimwilliams on Jun 8, 2018 17:25:39 GMT -6
The OPA604/2604 can be troublesome in some layouts. Best to check with a scope for stray oscillations. It's also a rather dirty opamp with 11 orders of harmonics about -110 db down on an FFT. The sweep looks like a picket fence of harmonics. When it was released it got the "audio opamp" moniker. After the customers showed BB it was too dirty the catalog dropped the audio references as the OPA134/2134 had been released to replace it.
The 134 or 2134 is the better FET DIP replacement as long as the power rails are limited to +-18 volts DC. Better yet is the OPA1641/2 but that is SOIC only and would need an adaptor. The LME49720NA is another DIP that would probably work well. With those adaptors you could use the OPA1611/2.
Don't let the smoke out.
|
|
|
|
Post by Blackdawg on Jun 8, 2018 22:30:36 GMT -6
The OPA604/2604 can be troublesome in some layouts. Best to check with a scope for stray oscillations. It's also a rather dirty opamp with 11 orders of harmonics about -110 db down on an FFT. The sweep looks like a picket fence of harmonics. When it was released it got the "audio opamp" moniker. After the customers showed BB it was too dirty the catalog dropped the audio references as the OPA134/2134 had been released to replace it. The 134 or 2134 is the better FET DIP replacement as long as the power rails are limited to +-18 volts DC. Better yet is the OPA1641/2 but that is SOIC only and would need an adaptor. The LME49720NA is another DIP that would probably work well. With those adaptors you could use the OPA1611/2. Don't let the smoke out. I want to keep as much of the magic blue smoke in as possible! So an OPA1641 would work as long as the power rails are 18 volts? hmmm..according to the PCB, it has +-18v and 24v rails...damn. but I don't have a schematic to know if the opamps are getting just the 18 or what is using the 24. also..curious why the kit specs those out if they are so noisey. Its supposed to be a mastering EQ. Perhaps that noise is pleasant?? Its based on a Sontec.. |
|
|
|
Post by svart on Jun 9, 2018 7:56:27 GMT -6
The lme49720 and the lm4562 are the exact same part, just different numbers due to being marketed to different customers.
In fact the lme49860 is also the same part that's been tested and passed higher voltage testing and binned as a different part, not all that different to how Intel bins the same silicon as various speeds and models of CPU like the core, Pentium vs Celeron, etc.
|
|
|
|
Post by jimwilliams on Jun 9, 2018 11:17:30 GMT -6
And TI unfortunately discontinued the LME49860 dual and the single version. They are impossible to find now, even the resellers have a very hard time getting them. The lower voltage LME parts are limited to a 34 volt swing, not 36. +- 18 is too high. I'd just adjust the regulator voltage, stuff lives longer at +-17 volts. The hybrid design OPA1602 is another good choice. That is a collaboration between the National and BurrBrown design teams. It's a LME49720 opamp with the BurrBrown rail to rail output stage for greater headroom. That does +-18 volts at 2.5nv noise. The OPA2604 is 10 nv noise.
|
|
|
|
Post by Blackdawg on Sept 19, 2018 23:07:59 GMT -6
So bringing this back up.
Im in contact with DIY Racked and hoping to learn the voltage of the opamp rails soon.
If its +-18 what would you all recommend? the 1642 or or the 1602?
If its at +- 24v...idk?
|
|
|
|
Post by Blackdawg on Sept 20, 2018 9:32:07 GMT -6
So its 24v for the opamps.
Damn it.
|
|
|
|
Post by ulriggribbons on Sept 20, 2018 10:32:53 GMT -6
www.findchips.com/search/opa2604looks like a couple overseas companies are showing some stock? I checked my stash, I have 11 I could spare.... there also seems to be a bunch on ebay....
|
|
|
|
Post by svart on Sept 20, 2018 11:21:43 GMT -6
www.findchips.com/search/opa2604looks like a couple overseas companies are showing some stock? I checked my stash, I have 11 I could spare.... there also seems to be a bunch on ebay.... I've not had good results with overseas grey-market sources, or ones from ebay if it's a chip that commands a price premium. |
|
|
|
Post by ulriggribbons on Sept 20, 2018 11:29:27 GMT -6
I've done OK over the years, mostly for things like resurrecting old Lex reverbs (replacing RAM and the like), to replacing LCD's, and chasing down odd bits for consoles (old logic, the occasional funky opamp)
I agree it's a crapshoot, but apparently I'm decent at shooting crap =P
|
|
|
|
Post by Blackdawg on Sept 20, 2018 11:49:01 GMT -6
Im not sure I have the gear to test the opamps to make sure they are in good working order. As I want this for mastering Im a bit worried about going that route.
Looking more at the BOM the 24v power rail has trim pots and is the same components as the 18v one. I wonder if I can just turn it down to 18volts and run the 1642s
|
|
|
|
Post by svart on Sept 20, 2018 11:58:45 GMT -6
Something you can do is place some diodes in series with the + and - rails to drop the voltage down to the potential chips. Three series diodes will add up to around 2-ish volts of forward drop, etc.. Just decouple at the power pins of the opamps after the diodes.
There's a reason that not many opamps can do crazy high voltage rails.. It's not really needed.
Voltage headroom doesn't necessarily mean huge increases in clean gain or lower distortion. Going from 48V differential(+24/-24) to 36V(+18/-18) only reduces the potential headroom by about 2dB.
Also, you have to consider how close an opamp can get to the voltage rails, aka "rail-to-rail" operation, or not. Having an opamp without rail-to-rail operation but higher voltage rail capable might be the exact same outcome as a rail-to-rail opamp at a lower voltage. Operational gain also matters a lot. Loading also plays a huge role in how close some opamps can get to the rails in operation.
|
|
|
|
Post by Blackdawg on Sept 20, 2018 12:58:28 GMT -6
Something you can do is place some diodes in series with the + and - rails to drop the voltage down to the potential chips. Three series diodes will add up to around 2-ish volts of forward drop, etc.. Just decouple at the power pins of the opamps after the diodes. There's a reason that not many opamps can do crazy high voltage rails.. It's not really needed. Voltage headroom doesn't necessarily mean huge increases in clean gain or lower distortion. Going from 48V differential(+24/-24) to 36V(+18/-18) only reduces the potential headroom by about 2dB. Also, you have to consider how close an opamp can get to the voltage rails, aka "rail-to-rail" operation, or not. Having an opamp without rail-to-rail operation but higher voltage rail capable might be the exact same outcome as a rail-to-rail opamp at a lower voltage. Operational gain also matters a lot. Loading also plays a huge role in how close some opamps can get to the rails in operation. Interesting. Stepping a bit above my level of comfort now. Plus i have been told by the designer that going down to 18v at the trims would not work, it would mess up other parts of the circuit. Though Im not sure which ones or why that would would matter. Unless I need to swap the other opamps too. |
|
|
|
Post by Blackdawg on Sept 21, 2018 0:58:18 GMT -6
So there is an OPA604 in the circut too. Id assume if I swapped that out too I would be fine to knock the 24v rails down to 18. Not sure what the 604 is being used for its right by the L/Mid gain IN relay. THAT1240 and THAT 1246s are used everywhere else in the section of the board. Seems odd to make that one different?
between the 1612 and 1642 which is better? If im reading things right the 1612 is quieter yes?
|
|
|
|
Post by Blackdawg on Sept 21, 2018 1:08:21 GMT -6
Actually looks like the 1642 might be a better choice. Am i looking at spec right? jimwilliams svartOr does one sound better than the other? |
|
|
|
Post by svart on Sept 21, 2018 7:37:27 GMT -6
Something you can do is place some diodes in series with the + and - rails to drop the voltage down to the potential chips. Three series diodes will add up to around 2-ish volts of forward drop, etc.. Just decouple at the power pins of the opamps after the diodes. There's a reason that not many opamps can do crazy high voltage rails.. It's not really needed. Voltage headroom doesn't necessarily mean huge increases in clean gain or lower distortion. Going from 48V differential(+24/-24) to 36V(+18/-18) only reduces the potential headroom by about 2dB. Also, you have to consider how close an opamp can get to the voltage rails, aka "rail-to-rail" operation, or not. Having an opamp without rail-to-rail operation but higher voltage rail capable might be the exact same outcome as a rail-to-rail opamp at a lower voltage. Operational gain also matters a lot. Loading also plays a huge role in how close some opamps can get to the rails in operation. Interesting. Stepping a bit above my level of comfort now. Plus i have been told by the designer that going down to 18v at the trims would not work, it would mess up other parts of the circuit. Though Im not sure which ones or why that would would matter. Unless I need to swap the other opamps too. I meant placing the diodes at the opamps themselves and leaving the power supply alone, like this:  you'd essentially use as many series diodes as it took to get the voltage down to whatever you need. It's inefficient, but it'll work! |
|
|
|
Post by svart on Sept 21, 2018 8:06:51 GMT -6
So there is an OPA604 in the circut too. Id assume if I swapped that out too I would be fine to knock the 24v rails down to 18. Not sure what the 604 is being used for its right by the L/Mid gain IN relay. THAT1240 and THAT 1246s are used everywhere else in the section of the board. Seems odd to make that one different? between the 1612 and 1642 which is better? If im reading things right the 1612 is quieter yes? There is no reason to swap opamps if you make the voltage rails lower, if that is what you're inferring. There's a few things you need to consider when changing these things.. While some opamps are being run at higher voltages, it might be because they are unable to swing as far to the extremes as other opamps in the chain. You're essentially using higher rails to get the same amount of headroom before clipping. Some stages will reduce headroom due the amount of gain they have dialed in. An opamp running at +/-12v but used as a buffer (zero voltage gain) might have the same or better headroom as an opamp run at +/-24v with a lot of gain, and so forth. You also need to consider that the resistor networks, and filter responses are all designed to be used as a system, and changing from one type of opamp to another might change how they work. For instance, if the filter/feedback network around an EQ section is designed for the high impedance of a JFET input opamp, if you were to substitute another opamp with a much lower impedance, it might skew the filter response to something else.. or it might not. Hard to say without an in-depth look at the circuitry. I guess the above is saying that you need a FET input opamp to be sure of correct operation, and changing to something like the NE5532 might not work. However, I might have a solution that could save you.. The OPA604 is still in production and available in SOIC format. You can use dual-SOIC to single-DIP adapters for these. You'll retain the voltage rails, and the correct impedances, but the tradeoff is a little complexity and some extra work and cost. www.ebay.com/itm/10pcs-Dual-SOIC8-SOP8-to-DIP8-Adapter-PCB-Board-PIN-Mono-Opamp-OPA627-AD797/251302346672?hash=item3a82c97fb0 |
|
|
|
Post by Blackdawg on Sept 21, 2018 19:54:38 GMT -6
So there is an OPA604 in the circut too. Id assume if I swapped that out too I would be fine to knock the 24v rails down to 18. Not sure what the 604 is being used for its right by the L/Mid gain IN relay. THAT1240 and THAT 1246s are used everywhere else in the section of the board. Seems odd to make that one different? between the 1612 and 1642 which is better? If im reading things right the 1612 is quieter yes? There is no reason to swap opamps if you make the voltage rails lower, if that is what you're inferring. There's a few things you need to consider when changing these things.. While some opamps are being run at higher voltages, it might be because they are unable to swing as far to the extremes as other opamps in the chain. You're essentially using higher rails to get the same amount of headroom before clipping. Some stages will reduce headroom due the amount of gain they have dialed in. An opamp running at +/-12v but used as a buffer (zero voltage gain) might have the same or better headroom as an opamp run at +/-24v with a lot of gain, and so forth. You also need to consider that the resistor networks, and filter responses are all designed to be used as a system, and changing from one type of opamp to another might change how they work. For instance, if the filter/feedback network around an EQ section is designed for the high impedance of a JFET input opamp, if you were to substitute another opamp with a much lower impedance, it might skew the filter response to something else.. or it might not. Hard to say without an in-depth look at the circuitry. I guess the above is saying that you need a FET input opamp to be sure of correct operation, and changing to something like the NE5532 might not work. However, I might have a solution that could save you.. The OPA604 is still in production and available in SOIC format. You can use dual-SOIC to single-DIP adapters for these. You'll retain the voltage rails, and the correct impedances, but the tradeoff is a little complexity and some extra work and cost. www.ebay.com/itm/10pcs-Dual-SOIC8-SOP8-to-DIP8-Adapter-PCB-Board-PIN-Mono-Opamp-OPA627-AD797/251302346672?hash=item3a82c97fb0Yeah they told me that as well. Im a bit nervous to buy some of the ebay(OPA2604). If I did how could I tell they were good? My dad has scopes and things I could use. Could just use the 604 like you said though. Was going to use those adapter for the 1642 anyways but changing the voltage changes a lot of stuff like you said. |
|
|
|
Post by Blackdawg on May 17, 2020 13:33:00 GMT -6
So there is an OPA604 in the circut too. Id assume if I swapped that out too I would be fine to knock the 24v rails down to 18. Not sure what the 604 is being used for its right by the L/Mid gain IN relay. THAT1240 and THAT 1246s are used everywhere else in the section of the board. Seems odd to make that one different? between the 1612 and 1642 which is better? If im reading things right the 1612 is quieter yes? There is no reason to swap opamps if you make the voltage rails lower, if that is what you're inferring. There's a few things you need to consider when changing these things.. While some opamps are being run at higher voltages, it might be because they are unable to swing as far to the extremes as other opamps in the chain. You're essentially using higher rails to get the same amount of headroom before clipping. Some stages will reduce headroom due the amount of gain they have dialed in. An opamp running at +/-12v but used as a buffer (zero voltage gain) might have the same or better headroom as an opamp run at +/-24v with a lot of gain, and so forth. You also need to consider that the resistor networks, and filter responses are all designed to be used as a system, and changing from one type of opamp to another might change how they work. For instance, if the filter/feedback network around an EQ section is designed for the high impedance of a JFET input opamp, if you were to substitute another opamp with a much lower impedance, it might skew the filter response to something else.. or it might not. Hard to say without an in-depth look at the circuitry. I guess the above is saying that you need a FET input opamp to be sure of correct operation, and changing to something like the NE5532 might not work. However, I might have a solution that could save you.. The OPA604 is still in production and available in SOIC format. You can use dual-SOIC to single-DIP adapters for these. You'll retain the voltage rails, and the correct impedances, but the tradeoff is a little complexity and some extra work and cost. www.ebay.com/itm/10pcs-Dual-SOIC8-SOP8-to-DIP8-Adapter-PCB-Board-PIN-Mono-Opamp-OPA627-AD797/251302346672?hash=item3a82c97fb0So Im revisiting this whole thing. I think I will be going the route of the Dual SOIC8 adapters to the DIP8 adapters. Or Dual DIP8 to 2x Single Dip8. Not sure if one is better than the other as you can get the OPA604 in SOIC or DIP. Biggest advantage to me is that the SOIC adapter is the same overall foot print as a normal DIP8 opamp. |
|
|
|
Post by svart on May 19, 2020 11:06:03 GMT -6
There is no reason to swap opamps if you make the voltage rails lower, if that is what you're inferring. There's a few things you need to consider when changing these things.. While some opamps are being run at higher voltages, it might be because they are unable to swing as far to the extremes as other opamps in the chain. You're essentially using higher rails to get the same amount of headroom before clipping. Some stages will reduce headroom due the amount of gain they have dialed in. An opamp running at +/-12v but used as a buffer (zero voltage gain) might have the same or better headroom as an opamp run at +/-24v with a lot of gain, and so forth. You also need to consider that the resistor networks, and filter responses are all designed to be used as a system, and changing from one type of opamp to another might change how they work. For instance, if the filter/feedback network around an EQ section is designed for the high impedance of a JFET input opamp, if you were to substitute another opamp with a much lower impedance, it might skew the filter response to something else.. or it might not. Hard to say without an in-depth look at the circuitry. I guess the above is saying that you need a FET input opamp to be sure of correct operation, and changing to something like the NE5532 might not work. However, I might have a solution that could save you.. The OPA604 is still in production and available in SOIC format. You can use dual-SOIC to single-DIP adapters for these. You'll retain the voltage rails, and the correct impedances, but the tradeoff is a little complexity and some extra work and cost. www.ebay.com/itm/10pcs-Dual-SOIC8-SOP8-to-DIP8-Adapter-PCB-Board-PIN-Mono-Opamp-OPA627-AD797/251302346672?hash=item3a82c97fb0So Im revisiting this whole thing. I think I will be going the route of the Dual SOIC8 adapters to the DIP8 adapters. Or Dual DIP8 to 2x Single Dip8. Not sure if one is better than the other as you can get the OPA604 in SOIC or DIP. Biggest advantage to me is that the SOIC adapter is the same overall foot print as a normal DIP8 opamp. From a physical space perspective, the dual SOIC adapters will be better as the ones I've seen are vertical mount. The dual DIP ones I've seen are horizontal mount and will hit just about anything that sticks up around the PCB footprint. |
|
|
|
Post by Blackdawg on May 19, 2020 12:25:03 GMT -6
So Im revisiting this whole thing. I think I will be going the route of the Dual SOIC8 adapters to the DIP8 adapters. Or Dual DIP8 to 2x Single Dip8. Not sure if one is better than the other as you can get the OPA604 in SOIC or DIP. Biggest advantage to me is that the SOIC adapter is the same overall foot print as a normal DIP8 opamp. From a physical space perspective, the dual SOIC adapters will be better as the ones I've seen are vertical mount. The dual DIP ones I've seen are horizontal mount and will hit just about anything that sticks up around the PCB footprint. Yep that's what I found. Be fine on my MB2 build but not okay on the Mq5. So going with the SOIC ones. Never done surface mount soldering but not too worried. I also found a store in NYC that has a decent stock of OPA2604AP's as well so going to buy some of those for the hell of it. I've read that part of the reason the 2604's are dying is that TI actually has a ton of them they just aren't releasing them. This is because they actually don't spec out to run correctly at 24v. I found an interesting thread about how a guy was able to get them directly from TI but had to sign a document that said TI doesn't spec them to operate at 24v any more. Was interesting.. So figure I'll try both. If the 2604's die due to heat, I'll use the dual 604's which are still good. Plus maybe they sound different..doubtful but who knows. |
|
