Yeah IIRC you need at least a 10 volt differential to get the 13700 to perform well which is kind of a downer.
A naive CMOS copy would probably work from lower voltage and have a higher bandwidth but be worse everywhere else...
I have a couple of dozen VTC 2713 dual 70-MHz OTAs from ~25 years ago. Nice parts, but too much of a niche even back then. :(
Cheers
Phil Hobbs
When I do a Google search for a datasheet in addition to the datasheet a post of yours about it from 2 years ago is like the 4th link that comes up...
The one big market I can think of that would be interested in a modernized version would primarily want a lower supply voltage as the big feature
Why? V_BE is still ~0.6V, so why is a low V_CCmax a benefit? I'm a fan of 36V processes, myself.
Cheers
Phil Hobbs
Sorry what I meant was it would be nice to have an OTA of some kind you could operate off a single positive supply rail of 5 or even 3.3 volts, plus ground.
What's preventing you from running an LM13700 off 5V and ground? All it is is a bunch of current mirrors in a box.
IIRC the inputs are NPNs, so the CMR won't go to V-, but its output is a current, so you can do all sorts of stuff with it.
Cheers
Phil Hobbs
True the datasheet doesn't offer an absolute hard-limit on how low you can take the supply, just a "minimum recommended" of 9.5V. I guess there's just enough headroom to pull it off at 5 volts. I see at least a ~1.2 volt drop from the output mirrors on the + rail to the diff pair collectors, and at least ~1.2 volts up from the - rail to the diff pair emitters. Plus a little more headroom to keep the collectors of the amp bias mirror and the diff pair out of saturation.
The transconductance is only approximately linear over a range of a few
10s of mV even with the linearizing diodes applied so it ideally won't be seeing large signals, anyway.
So, what's wrong with LM13700? It isn't tested at that voltage, but there's no circuitry inside that drops out. It should function normally.
The program current is usually a resistor from Vcc to the Vbe drop of an NPN transistor, to Vee, so that program current will only be accurate if there's a power supply (Vcc-Vee) voltage much greater than Vbe. That's why one prefers not to use a low voltage supply.
A Wilson mirror works fine down to about V_BE+0.2 V, so it's not as bad as you say.
The "diodes" (really the input transistors of current mirrors) turn the input pair into a real translinear stage, so the voltage swing isn't important.
Cheers
Phil Hobbs
Agreed. Thanks for the suggestions. I have ordered the parts.
Being an analog guy, I liked the idea of feeding the signal through a hamburger.
Robert Miller
Robert Miller wrote on 10/25/2017 8:45 PM:
Not sure why you want an analog solution. This is a perfect application for a DDS. Modulating phase (really you want frequency modulation so the phase changes slowly) is duck soup since the thing is controlled by a phase ramp generator. You didn't specify resolution or any other details in the original post (I haven't bothered to read every post in the thread) but I doubt there are any issues that a $3, 16-bit audio ADC and an MCU can't handle. If you can work with the 10 or 12 bit ADC available in many MCUs you can do this in a tiny module available for under $5 and won't even need to build anything.
-- Rick C Viewed the eclipse at Wintercrest Farms, on the centerline of totality since 1998
He's an Analog Guy
It's annoyingly slow--like 1 MHz at its max I_ABC of 0.5 mA--so it's mostly used for audio stuff such as VCAs.
I did a fun PLL with it once--it used an RC VCO with an exponential voltage vs. frequency characteristic and a two-stage loop filter made with an LM13700. The OTAs were biased by a frequency-to-current converter, so that the poles and zeros tracked the VCO frequency, resulting in a loop which had the same phase margin at operating frequencies from 100 Hz to 30 MHz. (That was about 30 years ago, when I was thinking about selling lock-in amplifiers--they only went up to 100 kHz in those days.)
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
It can also be used to implement small signal voltage-controlled resistances, voltage-controlled capacitances, voltage-controlled inductances. Even floating ones!
Thanks Phil, The lockin we sell only goes to 3 kHz. :^)
I made may way through the theory section of the spec sheet, but got a little lost in the application circuits. (Too lazy to try and puzzle them out.) (I also have to hack up my own current mirror one day. I've read all about them, but until I make one of my own....)
I did notice a voltage controlled resistor, one could bend that into a voltage controlled All pass filter.
George H.
Ah. Start with a MAT14. They're fun. My fave is the switcheroo Wilson, where the diode-connected sections are diagonally opposite. That one has three separate feedback loops with only 4 transistors, and to leading order has no base current error either.
I taught myself electronics in large part by going through the circuits in datasheets and app notes, assuming they were junk, and figuring out why. ;)
(Not necessarily the best-tempered approach, but it worked at the time.)
My gizmo used the current-controlled-inverter method, where you hang a cap on the output. The unity-gain cross goes as I_ABC. The zero was made by summing the input and output, so that it was always right near the unity gain cross of the VCO + integrator.
Cheers
Phil
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
On a sunny day (Fri, 27 Oct 2017 13:53:27 -0400) it happened Phil Hobbs wrote in :
My idea is a big LC oscillator with a metal core in the L and a magnet close to it to saturate the iron, and fix the magnet to something that randomly moves, like a road surface or a wind vane, or a membrane that listens to a voice app reading usenet posts or trump tweets or.. or just use an electromagnet and feed it with white noise from an FM radio or ..
or just use an electromagnet and feed it with >white noise from an FM radio or ..
You're having way too much fun winding up the earnest lefties, but I'm not in your target demographic. ;)
Cheers
Phil Hobbs
With the phase changing each cycle, continuous, its not a sine wave. Arbitrary wave gen with random trigger?
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.