¿ªÔÆÌåÓý

  1. Test-Equipment-Design-Construction
  2. Topics

Has any had any experience with the Renesas ICS511MILF ?

 

I had never heard of the ICS511, so looked it up. It could be very useful, but I think it's geared for much higher frequencies (like crystal oscillator multiplying), not audio like you describe. You probably would be better off using all CMOS like 4000 series or 74HC, which both have the classic 4046 PLL that you mentioned.

You can make an input amplifier with a self-biased CMOS inverter string like with 4049 or 74C(or HC)04 "UB" (un-buffered) style, to save power, and skip the JFETs. If the input signal will be big enough, a 74HC14 would be simpler yet. I never use TTL anymore, except for maintenance on old gear. If you go with a 4046 type, I think it has this input buffer function built in already (maybe not the old original 4000 series one, but I'm pretty sure the 4046A, 4046B, and 74HC4046(X) do), so all you'd need is a coupling capacitor and some protection from abuse. So, two CMOS ICs could mostly do it, say a 74HC4046 and a 74HC390.

The HP counter you recalled is likely the 5345A reciprocal counter, based on time difference to frequency readout conversion - about 9-10 digits resolution at almost any frequency, with one second gate time. No waiting around to count millions of actual cycles, it just needs to make an accurate time period measurement (with 500 MHz internal clock, plus interpolation), then calculates frequency. The more cycles available, the more it can average to improve the answer.

Ed
 

I built one of these IB1103 Heathkit counters in the mid to late '70s. Audio, up to 10KHz could be multiplied by 10X, 100X or 1000X.

Here is the schematic:

On Fri, Jul 7, 2023 at 8:58?PM Ed Breya via <edbreya=[email protected]> wrote:
I had never heard of the ICS511, so looked it up. It could be very useful, but I think it's geared for much higher frequencies (like crystal oscillator multiplying), not audio like you describe. You probably would be better off using all CMOS like 4000 series or 74HC, which both have the classic 4046 PLL that you mentioned.

You can make an input amplifier with a self-biased CMOS inverter string like with 4049 or 74C(or HC)04 "UB" (un-buffered) style, to save power, and skip the JFETs. If the input signal will be big enough, a 74HC14 would be simpler yet. I never use TTL anymore, except for maintenance on old gear. If you go with a 4046 type, I think it has this input buffer function built in already (maybe not the old original 4000 series one, but I'm pretty sure the 4046A, 4046B, and 74HC4046(X) do), so all you'd need is a coupling capacitor and some protection from abuse. So, two CMOS ICs could mostly do it, say a 74HC4046 and a 74HC390.

The HP counter you recalled is likely the 5345A reciprocal counter, based on time difference to frequency readout conversion - about 9-10 digits resolution at almost any frequency, with one second gate time. No waiting around to count millions of actual cycles, it just needs to make an accurate time period measurement (with 500 MHz internal clock, plus interpolation), then calculates frequency. The more cycles available, the more it can average to improve the answer.

Ed

 

The ICS511 is intended for crystals from 5 to 27 MHz, or clock inputs from 2 to 50 MHz.? See the datasheet here:



Some multimeters have a low frequency counter built in.? Does yours?? Do you have an oscilloscope?? You can also measure frequencies with an oscilloscope.
 

There weren't many ways to do it, back then. It was intended for two way radio work, where doing reciprocal calculations was too slow while setting control tones for squelch circuits.

On Sun, Jul 9, 2023 at 4:05?PM Jeff Green <Jeff.L.Green1970@...> wrote:

Wow Heath did it the hard way!

Thanks for the view how it had to be done.

?

I found these designs:

?

Page 125 of

?

Page: 41

?

And

?

All are variations on a theme.

?

?

I've breadboarded the last one with a 2N3819 as an input buffer which drives two 2n3819 in a Schmidt trigger. This drives the 4046 perfectly. [I was given a box filled with Fairchild 2n3819s so I plan on using them for every project I have.

I added a 3rd divide by 10 for a X10, X100 and, X1000. The last one has some jitter, the LSD varies a bit even with a very stable input. I'll try tweaking the 4046 timing values to see if that helps.

?

My unit will work down to 10mV at ~0.1Hz.

?

We had a battery operated analog clock (boo hiss, I'm dyslexic) that was dropped. I saved the drive electronics because I thought a 1.5V 1 pulse per second device might come in handy one day. I divided it by 10 and have a ~0.1Hz. It is interesting to put it in the freezer and see the frequency drop to 0.85Hz at 0F

0.1Hz X 1000 = 100Hz.

85Hz / 1000 = 0.86Hz

I thought about adding X10,000 but that'd be sort of ridiculous.

?

This is where one needs a frequency counter that does reciprocals. Measure the time between zero crossings, do the math and presto, you have the actual frequency in seconds.

?

I missed the lower frequency limit of the?Renesas ICS511.

?

I know it will work down to at least 1MHz because a classmate used one to multiply a 1MHz TXCO to 6MHz.

?

?

1 - 4 of 4