Will this work? Much-simplified OTA idea.

[earthtonesaudio]

I haven't tried it yet, but it's just a CMOS inverter interpretation of the old op-amp "voltage controlled current source" circuit you can find in many appnote collections or circuit cookbooks.

How well do y'all suppose this would work as an alternative to OTA-type applications?

[FrankHerbert]

If you go to http://users.ece.gatech.edu/~lanterma/ece4803/ there is a video on OTAs.  At about 24:00, he explains the guts of a CA3080 and eventually simplifies it into a 4-transistor discrete OTA.  Have fun!

[R.G.]

If you go to http://users.ece.gatech.edu/~lanterma/ece4803/ there is a video on OTAs.  At about 24:00, he explains the guts of a CA3080 and eventually simplifies it into a 4-transistor discrete OTA.  Have fun!

I thought "Oh, cool! A real lecture on OTAs." I was disappointed. It was nominally correct as far as it went, but I was put off by the non-description of what the other current mirrors in the 3080 do, and what they give you. It's not that complicated.

Maybe he was talking to a less technical audience that I thought.

[mdh]

Maybe he was talking to a less technical audience that I thought.

I take it you've never taught undergrads

[brett]

Hi
pardon my ignorance, but aren't three inverters equal to one inverter?  Why do you need three?
thanks

[earthtonesaudio]

Hi
pardon my ignorance, but aren't three inverters equal to one inverter?  Why do you need three?
thanks

Gain, gain, and more gain.  Greater open loop gain makes for more accurate closed-loop gain, which leads to a more predictable circuit, assuming the inverters are well-matched (on the same chip).  Alternatively, you could replace the three unbuffered inverters with one buffered inverter.

[R.G.]

How well do y'all suppose this would work as an alternative to OTA-type applications?

I meant to type this in earlier, but got distracted. As a signal OTA, it has some problems.

It's more of a voltage controlled current source (which you said) than an OTA. The current does vary with the input voltage, but it varies with the difference between the input and some ill-defined thresholds inside the CMOS chips. The "operational" part is kind of missing, as this usually implies a differential section where the difference between two inputs causes  some output.

The key feature of the OTA, however, is the variable transconductance. In this circuit, the transconductance is fixed by resistors to be some function of the input, feedback and current sense resistors. It doesn't vary as the transconductance of an OTA should.

Important aspects of an OTA are
1. An output best modeled as a controlled source/sink of current, with a largish voltage compliance compared to the power supply.
2. Output current proportional to the difference of two voltage inputs
3. The proportionality constant from 2 is a variable based on some third input.

[Cliff Schecht]

How well do y'all suppose this would work as an alternative to OTA-type applications?

I meant to type this in earlier, but got distracted. As a signal OTA, it has some problems.

It's more of a voltage controlled current source (which you said) than an OTA. The current does vary with the input voltage, but it varies with the difference between the input and some ill-defined thresholds inside the CMOS chips. The "operational" part is kind of missing, as this usually implies a differential section where the difference between two inputs causes  some output.

The key feature of the OTA, however, is the variable transconductance. In this circuit, the transconductance is fixed by resistors to be some function of the input, feedback and current sense resistors. It doesn't vary as the transconductance of an OTA should.

Important aspects of an OTA are
1. An output best modeled as a controlled source/sink of current, with a largish voltage compliance compared to the power supply.
2. Output current proportional to the difference of two voltage inputs
3. The proportionality constant from 2 is a variable based on some third input.

Good reply R.G., I was about to ask about the differential inputs and lack of Iabc myself. If you're looking for a transconductance amplifier then you may have something, but such a circuit would be much more inaccurate than a simple op amp/transistor based voltage to current converter. Not to shoot down your idea, but it may be somewhat ill conceived for a linear VCCS purposes. Now, it may sound damn cool and do something that nobody is expecting and so I'd still build it just out of curiosity.

This brings up an important difference between music electronics and just about any other style of electronics. With most EE's, you're going to find circuits that follow the rules of good design and people who use parts exactly as specified. Music electronics throws a lot of this out the window, you're real goal is to make something that sounds cool. If you "break" a few rules, more power to ya!

[earthtonesaudio]

Thanks for the thoughtful replies, guys.

I had some trouble wording the idea, but I suppose the main thing I'm looking for is not a substitute for an OTA so much as a voltage-controlled resistance.  I was thinking of the OTA-based filter and tremolo type circuits, so that's where the comparison came from, in my head at least.

Well... that, and an excuse to wring more circuits out of a logic chip that it was never intended to do.


I have an idea for a 1-chip-plus-a-few-transistors effect that would have variable compression/expansion, a flexible envelope filter, plus a clean blend, and this circuit idea is at the core, so I hope it works "well enough."

[tommy.genes]

The half-baked idea of the morning (I mean the idea is half-baked, not me. I'm at work):

Since folks on this forum and elsewhere have looked at just about every type of diode as a voltage-limiting device to cause clipping in voltage signals, are there any current-limiting devices that could cause clipping in a current signal?

-- T. G. --

[aziltz]

I have an idea for a 1-chip-plus-a-few-transistors effect that would have variable compression/expansion, a flexible envelope filter, plus a clean blend, and this circuit idea is at the core, so I hope it works "well enough."

that sounds promising!

[R.G.]

Since folks on this forum and elsewhere have looked at just about every type of diode as a voltage-limiting device to cause clipping in voltage signals, are there any current-limiting devices that could cause clipping in a current signal?

All depletion modes JFETs act as current limiting diodes. You short gate to source, and put current through it. It looks like a small resistor of rds value until the current rises near Idss, at which point it rounds over and limits at Idss.

There are commercial CLDs which are just JFETs sorted into current buckets with a tolerance.

You can cause a depletion JFET to limit at lower than Idss by putting a resistor in series with the source and connecting the gate to the outside of that. It adds a back bias from the current in the device and makes it limit below Idss.

[tommy.genes]

Thanks R.G. I'll have to pull out Horowitz and Hill to get my head around it, but it sounds like something worth exploring; to gain knowledge if nothing else.

-- T. G. --

[puretube]

The half-baked idea of the morning (I mean the idea is half-baked, not me. I'm at work):

Since folks on this forum and elsewhere have looked at just about every type of diode as a voltage-limiting device to cause clipping in voltage signals, are there any current-limiting devices that could cause clipping in a current signal?

-- T. G. --

E.g.: transistors?

preliminary...   

[tommy.genes]

So your schem is dated two days before my post, but you don't actually post it till about eight hours after. I think that's enough time to throw something on the breadboard and grab some scope traces. Just sayin'. 

Thanks for sharing. More to study...

-- T. G. --

[puretube]

So your schem is dated two days before my post, but you don't actually post it till about eight hours after. I think that's enough time to throw something on the breadboard and grab some scope traces. Just sayin'.  

Thanks for sharing. More to study...

-- T. G. --

"Dated" ?
look at the timestamp of the original pic...
(lower left corner: "Date/Time"...)  



BTW: "Discreted" OTAs for OTAs` sakes have been discussed in this forum earlier,  elsewhere...
(above example was meant for distortion purposes only).