DMM question?

How would we know? Presumably it has to read the changed frequency for long enough to decide that it has changed, then clear all the registers and accumulate enough new data to be able to work out what the new frequency is.

It may hold up the process for long enough to do it twice so that it is confident that the new frequency is stable enough to be worth displaying. It's not nice to write down four digits then find that the fourth digit is different when you look at at the meter.

Bill Sloman has NO Clue At All.:

------------------------------------------------

The Royal Plural baked up with utter bullshit. Wot a narcissist and what a old fool.

....... Phil

We make one frequency counter that actually measures period. It will count the time of N periods in some time interval, and compute frequency. Maybe it does something like that, measure the time of as many periods that fit into two seconds.

There's no way to get high resolution on a low frequency with a classic gated cycle counter. Observing periods for two seconds keeps the jitter down. Maybe.

..... Phil

We designed a circuit from scratch once to do the job. It took a lot of components to protect the input from a high voltage signal, and a few more to boost a low voltage swing signal up to a level where it could reliably drive a fast comparator - we got up to about 500MHz on the 50R terminated input and about 150MHz on the nominally 100k high-impedance input. There were a few ECL dividers behind the comparator to get the number of transition per second below the 50MHz that TTL (and the integrated counter chips we could buy at the time) could handle.

Quite busy. Happily we had triple extended Eurocard to put it on (with quite a lot of other circuitry).

On a sunny day (Sun, 7 Mar 2021 17:55:28 -0800 (PST)) it happened Phil Allison snipped-for-privacy@gmail.com wrote in snipped-for-privacy@googlegroups.com:

Maybe it just uses a prescaler and not a shorter gate time for higher frequencies?

Jan Panteltje wrote: =================

Upscaling by 100 ? Period to frequency conversion ? Frequency to voltage conversion ??

FYI I am currently monitor the AC supply frequency. Only the last digit ever changes.

.... Phil

On a sunny day (Sun, 7 Mar 2021 23:51:41 -0800 (PST)) it happened Phil Allison snipped-for-privacy@gmail.com wrote in snipped-for-privacy@googlegroups.com:

That is easy, you make a square wave from the mains that changes polarity every zero crossing. So you have ------------ ----------- | | |

---- ---------

then you make a gate like this by triggering only on the positive edge --------------------- | |

---- ---------

and feed it with high frequency pulses of say 1000 x fmains ||||||||||||||||||||||||||||||||||||||||||

now you count pulses in the gate time ||||||||||||||||||||||

--- -------

So say the mains is 50 Hz, you have a 50 Hz gate, and 1 MHz pulses in 20 ms you count (20/1000) * 1000000 = 20000 pulses

the time resolution is than 1 us ....

Good thing too, most power stations run in sync and phase :-) As to the last digit changing, here is from my past: I build 5 digit frequency meter in the sixties with 5 7490 decade counters, latches, and nixie drivers. However there was always that 1 count error thing. I solved it by not using a gate but a latch as input gate. That prevents the input changing when the count gate changes when input is high. Was my frequency counter for years until I donated it when I went abroad.

Jan Panteltje wrote: ================

..... Phil

On a sunny day (Mon, 8 Mar 2021 00:36:37 -0800 (PST)) it happened Phil Allison snipped-for-privacy@gmail.com wrote in snipped-for-privacy@googlegroups.com:

Sure, mine used a 100 kHz crystal back then, no micro-professors? existed yet.

Jan Panteltje wrote: ===============

Thanks.

( " micro-professors " LOL ! )

,.... Phil

On a sunny day (Mon, 8 Mar 2021 00:53:22 -0800 (PST)) it happened Phil Allison snipped-for-privacy@gmail.com wrote in snipped-for-privacy@googlegroups.com:

You may want to low-pass the mains a bit, sometimes there are big spikes, at least over here.

Most common would be some variant of period to frequency conversion by counting a fast precision internal clock between two rising zero crossings (or some other cleaner part of the waveform) of the signal. How much jitter and noise there is will determine how well it works.

These days you can't rule out some sort of flash DAC and software approach - that is how some of the signal generators do it.

Jan Panteltje wrote: ===================

IME the worst offender is right in my workshop. My 15W, desk fluoro lamp with choke ballast.

At switch off it kicks a 500 -1kV spike onto the local supply Upsets my triac, drill speed controller no end.

..... Phil

Here you go:

-------------------------

It's a hand held, low cost DMM drawing SFA from the battery.

...... Phil

You didn't provide the motivation for your question, so I just answered it, and since you have a low regard for the opinions of others, I thought that a simple "yes" would be unconvincing, therefore I provided you with a link to a complete working design.

How much is this SFA?

I expect a lot of micros could do a good enough job. For low power consumption, you'd be best off clocking just a hardware counter from the high frequency crystal, and clocking the processor at a lower frequency. Should be feasible with some off-the-shelf microcontrollers, though they would also need to use another counter as a prescaler to handle high input frequencies, and it might be tricky to search for the right prescaler ratio to use on the input signal and get that done fast enough to update quickly on large changes in the input frequency.

A small, low-power FPGA would be nicer, and some of those are now quite cheap. You could do better (maximum frequency and/or supply current) with custom silicon, if the sales volume would support the development cost.

WTF is SFA?

Sweet F- All, or not very much.

Lightning is an electrical discharge through ionised air. Getting the air hot enough to be sufficiently ionised is the crucial point.

Arc lamps could be seen as localised lightning, and it takes about 20kV for about a microsecond to start them - you have to wait until the electrodes are hot enough to boil off electrons.

Real lightning works on field emission - St Elmo's fire is the preliminary stage. Most of what electronic engineers organise is a little less capricious and a whole lot easier to organise.