Clock to drive MAX7401

[mth5044]

MAX7401 datsheet: http://datasheets.maximintegrated.com/en/ds/MAX7401-MAX7405.pdf

I wanted to play around with this chip and apply a clock to it for setting the corner frequency. Datasheet states corner frequency can go from 1Hz to 5kHz, so I'd like to play around with the full frequency then trim it down later.

For the clock, the datasheet states "designed for use with external clocks that have a 40% to 60% duty cycle" as well as "drive [clock] with a CMOS gate powered from 0 to VDD" which would be 5V in this case. Finally, the equation they give is corner frequency = clock frequency/100.

I thought a CMOS 555 timer (like the LMC555) would work, but little googling for cmos gate clock brings back this seemingly simple circuit:



A few questions,

First, is this better than trying to figure out the CMOS 555 timer, if that even works? They would be somewhat similar in size, with the 555 being a smaller chip but needing a few more components.

When determining frequency, I'm not sure if the equation F=RC/2.2 is calling for straight up farads, or if it's calibrated to use micro or nano farads. I'm going to guess the equation is calling for at least microfarads, because when solving with farads, the resistance needed would be something like 1 Billion ohms to get the clock to at least 50Hz. I'm also guessing that the frequency is in Hertz and not kHz.

To make the chips corner frequency vary from 1Hz to 5kHz, the clock needs to go from 0.01Hz to 50Hz 100Hz to 50kHz. To do that, using a 10n cap (0.01 used in the equation) 1u cap (1 used in the equation), the resistance would have to be ~220 to 1.1Meg. So maybe to begin, a 1M linear pot with a small series resistance would suffice, feeding it white noise at first to make sure it works.

Anybody see any tragic flaws in that logic, or perhaps a better way to get it done? Does it really have to be driven by a CMOS clock? I could do it with a PICAXE, but I don't know if that would count.

Thanks all.

EDIT: I realized I had been figuring clk vs corner incorrectly. Working on getting it right now. EDIT2: I think I got it right this time.

[idiot savant]

I use these chips quite frequently. I just use a little 8 pin AVR, which would be the lowest part count. You should also check out rring's website:

www.circuitsalad.com

He has an envelope filter and ringmod using those filter IC's with a PLL for clock duties.

A cool trick with these IC's: You can tun them into notch filters by lifting the OS pin and connecting it to the signal input.

[mth5044]

Thanks, I'll check out the site!

So can I use a simple microcontroller's PWM out to control the MAX7401? Does that count as a 'clock' signal?

intereting note about the notch filter... So just split the signal and send it to the OS pin? Have any examples? Thanks!

[idiot savant]

Thanks, I'll check out the site!

So can I use a simple microcontroller's PWM out to control the MAX7401? Does that count as a 'clock' signal?

intereting note about the notch filter... So just split the signal and send it to the OS pin? Have any examples? Thanks!

Here's the app note about using the MAX74XX series as notch filters.


http://www.maximintegrated.com/en/app-notes/index.mvp/id/431

They have limited usefulness this way because of the lack of Q control, but they do sound kinda phaser-y. I was thinking that if you ganged 2 or 3 and offset the frequencies, that you could widen the notch that way, and get a faux phaser sound.

And, yes. MCU will clock these filters just fine. the 50% ish duty cycle requirement makes it slightly tricky.

On an AVR, you would use Fast PWM mode, set the output pin to toggle on compare matches(this gives 50% duty cycle).

Then, setup the ADC to read a pot, and use the ADC reading (scaled for your desired frequency ranges) as the Compare Match for your PWM timer.

Example--filter freq is equal to:

in FAST PWM mode, set to toggle on compare match

F_CPU / TIMER_RESOLUTION(your compare match or TOP) / 100