pucho812
Well-known member
Is there any reason why most opamps have a max operational voltage of +/-18vdc. I know there are some exceptions but they are fewer and fewer these days.
Who decided on +/-18v
- Thread starter pucho812
- Start date
Help Support GroupDIY Audio Forum:
Physics... the typical maximum breakdown voltage for opamp silicon substrates is around 36V. Some modern op amps can handle more voltage.
For better reliability I never ran op amps much above +/- 15V
JR
For better reliability I never ran op amps much above +/- 15V
JR
pucho812
Well-known member
Learn something new daily. Yes very few go beyond that. Thanks jr.
Bo Deadly
Well-known member
Is there something about 36V to pn junctions in general or is it really decided by the fabrication tech / tool chain?
IDK.... There certainly are discrete semiconductors with breakdown voltages many times that. There are even tricks that some IC makers use to allow parts of the IC to operate at higher voltages than the rest. Some early power amp chips used that trick.
JR
k brown
Well-known member
Was that one of the reasons for the Exar 5532 ceramic package? It's max supply was +/-20v.
That was the best-sounding 5532 I ever used; I think I only have one left.
Newmarket
Well-known member
not really. this voltage spec' is common in all applications. not just audio.
It's related to the fabrication / electronic characteristics of ic manufacture.
Newmarket
Well-known member
fwiw I use opamps with high voltage rails eg +/-40V and higher on a daily basis.
See TI OPA445 and similar. These in SO8. Obvs heat dissipation can be an issue and the higher voltage types eg +/-90V have a thermal interface pad on the bottom that should be soldered to the negative rail. So awkward for diy and modding (I know this first hand).
But you won't get super low voltage noise Bipolar types in this range - they are FET and CMOS input types so low input bias current.
Going higher you can look at devices from APEX but they get very spendy
Yes for premium prices you can get higher rail voltages. I stopped following new op amp development closely after they routinely became faster than audio (back in the mid 1970s). Over the decades I have used truckloads of TL07x and 553x op amps as being more than good enough for audio. For lower input noise or increased output current it is easy enough to hang a few discrete devices onto and around the GP op amps.fwiw I use opamps with high voltage rails eg +/-40V and higher on a daily basis.
See TI OPA445 and similar. These in SO8. Obvs heat dissipation can be an issue and the higher voltage types eg +/-90V have a thermal interface pad on the bottom that should be soldered to the negative rail. So awkward for diy and modding (I know this first hand).
But you won't get super low voltage noise Bipolar types in this range - they are FET and CMOS input types so low input bias current.
Going higher you can look at devices from APEX but they get very spendy
Now we can buy off the shelf mic preamp chips that don't suck, and I/O input receivers/output drivers that don't suck. The work of audio design engineers has become much easier. Even modern VCAs have benefitted dramatically (thanks to improved IC processes)
I have been impressed with how much better CMOS has gotten. Early MOS op amps (like RCA BiMOS) were a yawn, but modern CMOS op amps are respectable.
Progress and technology advancement is good.
JR
Newmarket
Well-known member
You can also still get the TI OPA604, the single channel sibling of the discontinued OPA2604. Max supply rails of +/- 24V. Years ago the OPA2604 was kind of trendy in the types of equipment that liked to point out what devices were used in the marketing copy. The OPA2134 has mostly taken that spot in the market now. "Only" +/- 18V because most gear doesn't actually need headroom to over 30dBu (with symmetrical drive) any longer.
To answer the original question affordable junction-isolated semiconductor processes determined the 30 to 36V supplies we use today. Many of the earlier HV op amps (HA-2645) required expensive dielectric isolation to break the 40V barrier.
The 5532 (still) has an absolute maximum supply voltage of +/-22V and can be run on +/-20V in PDIP.
The 5534 single (and many others including duals if used properly) can be bootstrapped to much higher supply voltages with external transistors as was done by MCI's "Swinging Op Amp."
For a voltage-boosted NE5534 to have much value it needs to be current-boosted too.
MCI Swinging Op Amp Ad dB Magazine January 1979.
References:
Bootstrapped Op Amp Articles - Pro Audio Design Forum
Voltage-Bootstrapped Current-Boosted NE5534 Swinging Op Amp Audio Line Driver - Pro Audio Design Forum
Exar definitely had it going on with their 5532/5534.
The 5532 (still) has an absolute maximum supply voltage of +/-22V and can be run on +/-20V in PDIP.
The 5534 single (and many others including duals if used properly) can be bootstrapped to much higher supply voltages with external transistors as was done by MCI's "Swinging Op Amp."
For a voltage-boosted NE5534 to have much value it needs to be current-boosted too.
MCI Swinging Op Amp Ad dB Magazine January 1979.
References:
Bootstrapped Op Amp Articles - Pro Audio Design Forum
Voltage-Bootstrapped Current-Boosted NE5534 Swinging Op Amp Audio Line Driver - Pro Audio Design Forum
Last edited:
user 37518
Well-known member
Years ago I used to have an MCI forum with Larry Lamoray, one of the designers of the MCI JH-500. One time, we were discussing exactly this topic. The first two revisions of the JH-500 used to have MCI-2002 op-amps, which were an Intersil rebranded high-voltage op-amp, as you mention. The consoles used to run at +/- 36V, but they had to lower the rails down to +/- 32V in subsequent production runs, because the 2002's were getting fried very often.To answer the original question affordable junction-isolated semiconductor processes determined the 30 to 36V supplies we use today. Many of the earlier HV op amps (HA-2645) required expensive dielectric isolation to break the 40V barrier.
Exar definitely had it going on with their 5532/5534.
The 5532 (still) has an absolute maximum supply voltage of +/-22V and can be run on +/-20V in PDIP.
The 5534 single (and many others including duals if used properly) can be bootstrapped to much higher supply voltages with external transistors as was done by MCI's "Swinging Op Amp."
For a voltage-boosted NE5534 to have much value it needs to be current-boosted too.
MCI Swinging Op Amp Ad dB Magazine January 1979.
References:
Bootstrapped Op Amp Articles - Pro Audio Design Forum
Voltage-Bootstrapped Current-Boosted NE5534 Swinging Op Amp Audio Line Driver - Pro Audio Design Forum
About that time the 5534 came in and it was very superior to the 2002, particularly in terms of noise and distortion; they rebranded the metal can version with part number MCI-2003. They added transistors between the voltage rails and the output of the op-amp and this allowed using the 5534s with +/-32V rails. The design or idea—as Larry told me—came from an op-amp applications handbook, so it wasn't like they were inventing a revolutionary circuit. The result was a massive headroom, since the console was able to achieve a +/-32V voltage swing whilst keeping the maximum differential voltage between the supply pins of the the 5534 in a range between +/-15 and +/- 18 V. But every now and then you had to replace an op-amp, because they didn't always like this arrangement. Apparently this panacea wasn't as good as they thought, because IIRC late productions of the JH-500 (in revision D) ditched the +/-32V supplies along with the bootstrap transistors, and substituted them with more 'civilized' +/-18V rails; never to be seen again in their newer (and last console to display the MCI name) JH-600 series, albeit a much more crappy and dull console.
Something not mentioned in that ad you posted is that these high voltages produced a massive amount of heat in the console. I remember sitting in front of these consoles and feeling the radiated heat hitting my body, particularly my face; a very similar sensation to being a few feet away from a barbeque grill. Also, the bootstrap transistors used to fail from time to time, and the excessive heat in the console used to eat up the orange colored (and infamous) 85°C rated Siemens electrolytic capacitors in a few years, which resulted in the spilling of electrolyte and a layer of gunk covering the PCB. For the time being, I will not discuss the infamous Molex connectors or the "Red" IC sockets, which were an even bigger problem in their tape recorders, and are also responsible for the "I" in the community rebranded nickname for the MCI acronym: Munchy Crunchy Intermittent.
Even with all these issues, these consoles were (and are) awesome. Not long ago, a studio owned by a famous rock/indie band upgraded from a JH-528C to an API console. In about a one year span they sold the API and went back to the JH-528C.
The MCI JH-500 is like a 70s American muscle car: brute force and not very efficient; but loud, powerful and sexy. IIRC the JH-500 was the inspiration (or at least part of) behind the SSL 4000 series console.
To the OP: The trend right now is for voltages to reduce even more and more. "Regular" modern electronic designers gasp at the idea of using +/-18V rails, when the tendency is to use the lowest (single rail) voltage posible. In an age were the quest is to strive for a single sub 1V rail device, +/-18V seems like something barbaric.
Also, as have been mentioned before, the rails' maximum voltage has a lot to do with the semiconductor technology (and the fabrication process). For instance: GaAs which is routinely used in high-frequency applications has a low breakdown voltage. More recently, GaN has become the modern semiconductor wonder since it is capable of very high voltage breakdown and high electron mobility, which makes it very appealing for high-power and high-frequency applications. It has been used in almost any power application I can think of: from audio Class D power amps which require no heatsinks at all, RF/Microwave power amplifiers, and power electronics devices.
Last edited:
I first saw the swinging op amp in 1974. I built one as soon as I saw the Idea for Design in '74. I wrote:
I imagine they saw it in the Burr Brown book. Dogra is the first reference I've ever found for it. (Cited in earlier post.)
The PN5679/5681 plastic bootstrap transistors MCI later used to replace the metal 2N5679/2N5681 were incredibly failure-prone.
The JH-500C EQ section of the channel strip was an idea location to bake spit out of condensor mics. They did run hot and baked caps.
SSL used the HA-2645 HV op amp (originally Harris Semiconductor) as the 5 phase HV cathode driver for their plasma meters. I wonder what people use to replace them.
user 37518
Well-known member
Regarding the MCI-2002: about 15? years ago I found a replacement in Mouser for the 2002s, I don't remember how good it was in terms of performance, but it could handle the high-voltage rails. The original Intersil part could still be found NOS a few years ago (maybe still now?) on eBay and places like that. I bought several of them years ago (from Larry Lamoray himself) later to be sold faster than hot bread for about $30-40 USD a piece to the MCI community. The Signetics plastic 5534s are also fiercely sought after.
The MCI-2002s are really the crux of the JH-500 for owners today, almost any other part in the console is an off-the-shelf part, but the few 2002s that remain in the C rev (which is the most common) are becoming unobtanium. The B rev and the exceptionally rare rev A owners are basically screwed, since the consoles only use MCI-2002 op-amps. It is possible to substitute the MCI-2003 metal can op-amps with DIP8 NE5534s using adapters, some surgery is needed but it can be done. Some people have replaced the 2002s with 5534s and modified each module by adding LM317/337 regulators to lower the rails.
John Hardy made (and maybe still sells) very expensive daughter boards to replace the stock mic preamps in the JH-500 with his own designed based on the Jensen 990c op-amp, since the JH-500 included (but ditched later in the D rev) a Jensen mic trafo in every module.
Regarding the EQ, I loved the EQ. It was awkward to use since all the knobs could only turn right, and you had to hit a cut/boost switch depending on what you wanted. The biggest operational issue with the JH-500 up to revision C was that, IIRC, the insert points were located after the fader; which made using dynamic processors a drag, since compressors/expanders would compress/expand more or less whenever you adjusted the level of a channel in a mix. As always, people these days have found a solution to it—which involves some major hacking in the old sense of the word—in order to place the insert point where it belongs.
Last edited:
We explored this at length on Wayne's forum.
Back in the 70s I used a variant on that topology to refresh a hifi amp for my younger sister to play records through.
JR
user 37518
Well-known member
He provided a link to that thread at Wayne's forum a few posts ago.
user 37518
Well-known member
P.S. Isn't KA-electronics also Wayne? You sneaky you....
Newmarket
Well-known member
fwiw the "HV" opamps I mentioned are used with such a supply arrangement as the input signal is switching +/- several hundreds of Volts (sitting on a bias voltage that may be in the magnitude of big kVs)
