Question on creating a discrete device OpAmp of the CA3080

[mdeffects]

Hey guys,

Haven't posted on here in ages, but had a recent question come to my mind regarding Discrete Opamp circuitry and design. I have been experimenting with a CA3080 based design, which is quite difficult given there aren't many available, and they are very pricey haha. But, I had the thought, what if I made my own discrete version of the chip? I feel like the performance would improve a lot with it being discrete, and potentially have a much lower noise floor? Again, this is my first time looking into this, so I am just overall curious how I can create a Discrete Opamp from the attached base schematic of the CA3080 or if it isn't quite that simple.

Thanks for all of your help!

[idy]

Maybe you should know ca3080 is not an opamp, but OTA, Operational Transconductance Amp.

Opamp changes gain according to ratio of resistor in feedback loop to *another* resistor. OTA changes gain based on an input (current or voltage, that's where I get fuzzy. But I'm guessing current, because BJT.)

[Rob Strand]

I believe some people have already made an attempt at that.   You might even find posts on this group.

The hardest thing to get right is the matching of the two transistors in the differential pair.   If these aren't
matched you get feed-through from IABC to the output.   What that means is as control current
changes a DC shift appears at the output.  That produces ticks and other undesirable artifacts.
What you need to do is match Vbe at a few currents.  Ideally the gains also match.

The diodes are probably best implemented using the same transistors you are using  throughout.   To form a diode connect the base to the collector.

Beyond that you would need to design the circuit from scratch rather than just copying the internal schematic.   For the current mirrors you ideally want matched parts there as well.  Matching means it will work well over the entire range of IABC.   You can sometimes put matched emitter resistors in the mirror to help matching at high currents but they don't help at low currents.   I think this has been looked at on the group also.

Using monolithic transistors or matched pairs can help but it can blow-out the cost and it certainly restricts sourcing the parts.  You have to trade that off against manually matching parts.  Manually matching is helped by buying parts from the same batch.

Here's one of the old threads,
https://www.diystompboxes.com/smfforum/index.php?topic=87319.0

[idy]

One thread here.https://www.diystompboxes.com/smfforum/index.php?topic=87319.0
The wise guys say it is hard to get the precision matching that comes with ICs.

[merlinb]

Haven't posted on here in ages, but had a recent question come to my mind regarding Discrete Opamp circuitry and design. I have been experimenting with a CA3080 based design, which is quite difficult given there aren't many available, and they are very pricey haha. [/url]

Just use an LM13700.

[Rob Strand]

Let's not forget this thread,
https://www.diystompboxes.com/smfforum/index.php?topic=125268.0

[anotherjim]

You might find the work done to make a replacement BA662 OTA interesting...
http://wiki.openmusiclabs.com/wiki/BA662

[bowanderror]

Is it possible? Yes. Will it perform like a CA3080? Probably not. Will it be cheaper than an LM13700? No chance.

This topic came up on another forum, and I'll post the summary here:

"Build an OTA? It's a fun experiment, but a lot of effort for not a lot of benefit. All of the transistors should be matched, not just the pairs, and they should all be at the same temp. Easily achieved when they're all on the same die. As one of my early mentors told me: 'Why buy the cow when milk is so cheap?' He wasn't talking about electronics, but in this case, it's applicable."

You can get cheap matched NPN pairs like the DMMT3904 or BCM547, but by the time you factor in all of the costs & get the thermal management nailed down, you won't get anywhere near the ~$1.30 you'd be paying for an LM13700 - and it's a dual OTA.

[ElectricDruid]

<deleted>
I didn't see Rob had mentioned the earlier AS3080E thread further up
<deleted>

[noisette]

Not sure if it helps, but this old post from René Schmitz´s site features a discrete 2040 vcf clone, which is an OTA (you could use one of the structures as vca also...), just not the 13700, and very good sounding, while not clean at all
https://www.schmitzbits.de/rs2040.html