Frequency counter

>I have two scopes (both Heathkit) but no multimeter!

Well if they have a calibrated time base you can measure frequency. Too lazy to push a few buttons on a calculator ?

Get it as close to one cycle per division as possible, that is a good tradeoff between accuracy and easy to read.

Most scopes have ten divisions horizontally, count as accurately as you can how many cycles are in those ten divisions. Then it is 10 math. I can do it but I am just not good at expressing how anymore. The period of one complete cycle gives you the period and frequency is the reciprocal of that, like 1/X. The sweep reads in per division so if it is around one cycle per division it may be 8, 10, 13 cycles across ten divisions. Use that number, either a tenth of it to get an answer in Hertz or the whole thing would give you an answer in tens of Hertz.

That method is usually close enough unless you get into RF or beyond the bandwidth of your scope. How high a frequency do you need to measure ? If you are even thinking of audio some don't go low enough almost, at least spectrum analyzers don't when you get the ones that go up to the GHz.

I never had a frequency counter until a few years ago and now that I have one, I haven't used it in at least a year.

Since you are here I am going to assume that you didn't mention a project from T Mobile or the government because you are not doing one. You do not need a 52GHz counter. Fact is your best bang for the buck is probably used. I got an HP 5314a for less than $100. (not to sell) If you can squeak by on 1GHz you do not have to spend $300 on it.

 

[default]

On Tue, 16 Jul 2019 22:30:00 -0700 (PDT), jurb6006@gmail.com wrote:

I have two scopes (both Heathkit) but no multimeter!

Well if they have a calibrated time base you can measure frequency. Too lazy to push a few buttons on a calculator ?

Get it as close to one cycle per division as possible, that is a good tradeoff between accuracy and easy to read.

Most scopes have ten divisions horizontally, count as accurately as you can how many cycles are in those ten divisions. Then it is 10 math. I can do it but I am just not good at expressing how anymore. The period of one complete cycle gives you the period and frequency is the reciprocal of that, like 1/X. The sweep reads in per division so if it is around one cycle per division it may be 8, 10, 13 cycles across ten divisions. Use that number, either a tenth of it to get an answer in Hertz or the whole thing would give you an answer in tens of Hertz.

That method is usually close enough unless you get into RF or beyond the bandwidth of your scope. How high a frequency do you need to measure ? If you are even thinking of audio some don't go low enough almost, at least spectrum analyzers don't when you get the ones that go up to the GHz.

I never had a frequency counter until a few years ago and now that I have one, I haven't used it in at least a year.

Since you are here I am going to assume that you didn't mention a project from T Mobile or the government because you are not doing one. You do not need a 52GHz counter. Fact is your best bang for the buck is probably used. I got an HP 5314a for less than $100. (not to sell) If you can squeak by on 1GHz you do not have to spend $300 on it.

Another thing about old test equipment is that it has knobs to play
with. There's nothing more frustrating to me than test gear that has
a few buttons to control all the functions. It seems to make minor
adjustments I have to take my focus off the equipment I'm working on
and step through some ill-conceived menu on a screen that's only
readable from a narrow angle.

 

[Ralph Mowery]

In article <7e2bc462-1013-44bf-b55b-0f5e6ff4dcc7@googlegroups.com>,
jurb6006@gmail.com says...

Well if they have a calibrated time base you can measure frequency. Too lazy to push a few buttons on a calculator ?

Get it as close to one cycle per division as possible, that is a good tradeoff between accuracy and easy to read.

Most scopes have ten divisions horizontally, count as accurately as you can how many cycles are in those ten divisions. Then it is 10 math. I can do it but I am just not good at expressing how anymore. The period of one complete cycle gives you the period and frequency is the reciprocal of that, like 1/X. The sweep reads in per
division so if it is around one cycle per division it may be 8, 10, 13 cycles across ten divisions. Use that number, either a tenth of it to get an answer in Hertz or the whole thing would give you an answer in tens of Hertz.

It all depends on how accurate you want your frequency to be. I doubt a
n analog scope will get to with in a few cycles at 150 MHz.

I do have a newer digital storage scope that cost about $ 300 and it
does have many functions such as a counter and voltmeter built in. I
have not checked it to see how accurate the counter or volt meter is
because I have not needed to get very close so far as I have other
instruments that I know how acuratge they are if I need to be on the
money.

 

[Cursitor Doom]

On Wed, 17 Jul 2019 07:14:28 -0400, default wrote:

Another thing about old test equipment is that it has knobs to play
with. There's nothing more frustrating to me than test gear that has a
few buttons to control all the functions. It seems to make minor
adjustments I have to take my focus off the equipment I'm working on and
step through some ill-conceived menu on a screen that's only readable
from a narrow angle.

Exactly! Trying to navigate around some convoluted menu tree that some
autistic code-writer assumed would be self explanatory to the user is
good enough reason alone to stick with vintage test gear IMO.




--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.

 

[Tom Gardner]

On 16/07/19 23:51, Phil Allison wrote:

Tom Gardner Asshole wrote:


My apologies to all.


** And you have plenty to apologise for - fuckhead.



I was careless and presumed I was replying to the OP, not
to the "famous" Phil Allison.


** Lot more famous than you are, and for the right reasons too.


...... Phil

Q.E.D.

 

[Phil Allison]

Some total asshole puked:

Tom Gardner Asshole wrote:


My apologies to all.


** And you have plenty to apologise for - fuckhead.



I was careless and presumed I was replying to the OP, not
to the "famous" Phil Allison.


** Lot more famous than you are, and for the right reasons too.


...... Phil


Q.E.D.

** Wot a fuckwit retard.

 

>It all depends on how accurate you want your frequency to be. I >doubt a n analog scope will get to with in a few cycles at 150 >MHz.

It won't, but you don't always need that much accuracy. I know ways around all kinds of things. Want to adjust the free run of the 38KHz oscillator in an FM tuner ? Who needs that accuracy ? Radio stations. So you just put your scope in free ruin and rock it in to the frequency when it is locked to a station. Then you off tune it and use the frequency control on the FM tuner to make it stationary. Now you can get to fractions of a Hertz. Same thing with the old 3.579545 MHz oscillators in color TVs. Before SAW filters, the IF frequency in a TV was not critical at all, you could align those coils to quite a range and have it come out right. Before the ceramic filters in FM IF strips, which are almost the same as a SAW filter, that 10.7MHz was not critical either. In an AM tuner, the IF can be anywhere from like 420KHz to 490KHz and it'll work just fine.

I lived for decades without a frequency counter. And it would not have come in handy with some of the frigged up things I have had to do.

Here's an example. Remember how VCRs work ? Video carrier is FM at 4.2MHz, color under is hetrodyned down to 629KHz. in the more extended play time speeds there is no guard band between the slashes on the helical scan tape and the chrome bleeds from track to track so they rotate the absolute phase (which includes the burst) 90Âş which makes the crosstalk cancel out by use of a COMB filter.

So I was working in Willowick, Ohio which is a bit of an unusual demographic. People there don't really have alot of money, but all those $150,000 houses out there they bought for $35,000 many years ago. There were incompetents out there and one of them was actually named Jeff, which was almost a concern for me... Had one old Italian guy whose bigscreen he screwed up. he says "Some years ago that guy would not have walked out of here". Nice to make your acquaintance, and I actually mean that. If you had a convergence problem this guy's solution was to defocus the affected tube. Now if that isn't bad enough we were getting in a bunch of Funai built VCRs that would not record.

Someone had screwed with some controls, carrier frequency, deviation, white clip, all that in the FM modulator for the recorder. That requires thousands of dollars in equipment to align properly.

Well it did.

I stuck a dummy tape in a VCR and hit play. I loosely coupled the signal going to the record heads on the defective VCR to the one in play with the dummy and it picked up the signal just like it would if it were playing a tape. Easy.

That is how to get shit done and why I got how I got so how I got. Other shops would just condemn them but we stacked them up, seventy bucks, seventy bucks, seventy bucks. I could do about ten of them in a day. Probably more.

And FM alignment, I can do it just fine with an air signal. The only thing I can't do is the stereo separation but I got that solved now. Those little FM transmitter for send like your CD or MP3 player to your car radio ? Those things are all digital, the separation out of them is perfect. There is really not much to adjust on a digital stereoplexer. Frequency ?

All a digital stereoplexer does is switch between the two channels at 38KHZ generating a DSB suppressed carrier signal, then puts that to a flip flop to make the pilot frequency and filters it to a sine wave. Oh, and I suppose that must be adjusted for phase... big deal. I can figure that out with any FM stereo tuner which doesn't even have to be aligned for separation, just adjust the phase for maximum L-R, separation that is.

That is how it is done. Up until a couple years ago I never owned a frequency counter. In fact now in the last couple of years I think I have used it about five times, and two of them were just playing around with it.

So I got me a nifty HP 5314A. I am thinking of taking it off the bench on onto the shelf. I am not in a hurry though, I have nothing to put there right now.

One thing irks me about this HP, I don't know if the input level controls go up or down when clockwise. (or whichever) It is not in the manual, I started looking for it on the print but it does not specify.

Well it's HP so what did I expect ? I will not buy their computers. I think I got my last HP printer. It is a great printer but it costs $300 to fill it.

You can throw a frequency counter together fairly easily. All you need is a digital (as in base ten) counter, and oscillator and something to gate it accurately. You know the operation of a DVM is not all that different. It has a calibrated gate but the signal is current controlled by the input ramping up a calibrated cap. There are also cap checkers that work in a similar fashion.

When I first started, and that involved fixing things under the bench...all I had was a VTVM. I remember my first dog. Color killer control works but no color lock, barber pole effect. I have no idea how many people tried to fix it but I am the one who did. The color killer control, like a 1.5 meg or whatever feeds the burst gate tube with the proper pulse to only allow the burst through which then locks the local oscillator. the wiper varies a pulse to the color killer circuit, which was only there to reduce noise on black and white shows, but in this case the pot was open. It had a bad spot. You rotate the control and the color would come in and go out as expected but no sync. The pulse was there at one side but not at the other. The point though is that the pulse off the flyback which is what runs all this was not getting to the burst gate where it was specifically needed.

I also invented the dim bulb tester. Sure other people also did, but I did. I ran down a short in a clothes dryer. We lived in a rental and the place had a fuse box. some things re out upstairs and this one fuse keeps blowing..

We tried yanking stuff, unplugging and changing the fuse but they kept on blowing. Finally I said "Gimme a light bulb". It was the same socket back them, and the bulb lit. I unplugged the dryer and it went out. Problem solved.. I also figured out how to do it with cars.

Frequency counter ? Just put the probe in my mouth and wait for a number. Then to determine if it is GHz or MHz note if my testicles are in my neck. Why spend all that money ?

 

[Dimitrij Klingbeil]

On 2019-07-15 16:33, Peter Percival wrote:

I'd like to buy a frequency counter but I was afraid they'd cost
thousands. Apparently, "only" hundreds -
https://uk.rs-online.com/web/p/products/6654919/?tpr=2. Any
recommendations?

Hi

Some can indeed cost thousands, but that would be the high end ones like
an Iwatsu SC-7217A. Nothing wrong with that, it's likely a fine
instrument. 12 digit resolution at a 1 second gate time means single
picosecond time interval measurement resolution internally (at this
resolution, it would have to measure the ratios of time intervals rather
than counting pulses that the simpler "normal" frequency counters do).

On the other hand, a frequency counter can be had for a lot less, the
most basic ones do not even cost "hundreds". A search on Aliexpress
turns multiple variants for around $60 - $80, some of them with RF
inputs to 1 or 2.4 GHz. A "bare board" frequency counter module costs
even less, but of course you're not getting a housing and mains power
supply with one. There are reviews of some of these things on EEVblog.

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

The choice of frequency counter comes down to your frequency range and
accuracy needs. There are versions for everyone and basically they all
come in one of 3 basic types (from simplest to most complex):

1. Basic "gated edge counter"
This type of device simply enables a count-up mechanism for a specified
time (called the "gate time") and counts how many edges of one polarity
occur during this time. The number of edges divided by the gate time
gives the value of the frequency measured. However, if the frequency is
low, there are not enough edges to get a decent resolution, so the low
frequency resolution (and low frequency accuracy) becomes very poor.

This is the cheapest type of counter that can be made, but I would NOT
recommend buying one because of the really useless performance in the
low frequency ranges. There are other more reasonable options not too
expensive. The time of the simple "direct counting" passed years ago.

2. Recipocal counter
This device may switch to direct counting mode and work like a basic
counter at highest input frequencies, but will switch to a different
method at moderate and low frequencies. Instead of counting how many
pulses of an input signal have occurred during a fixed time interval, it
measures how many pulses of an internal precise high frequency reference
signal have occurred during some (variable and typically automatically
selected) number of cycles of the input signal. The ratio between these
numbers of cycles multiplied by the internal reference frequency gives
the measured input frequency. This counter type maintains a high
measurement resolution down to low frequencies. However it is still
dependent on counting integer numbers of cycles, therefore it cannot
achieve high resolutions without using either very high internal
reference frequencies and (/or) very long gate times.

This is the type in most common use nowadays. Decent oscilloscopes (and
some signal generators that have a frequency measurement input) also
tend to use this method internally. I think, this is the most advisable
type to a beginner who wants a good but not too expensive instrument.

However, 9 digits resolution is about as far as this device will do. An
older Iwatsu SC-7207 will do that, as will other less expensive ones.

3. High resolution frequency / time interval counter
This device overcomes the integer number of cycles limitation by
measuring the time intervals that elapse for some numbers of cycles at
the input frequency. Because the time intervals can be measured with
sub-cycle accuracy (with respect to the reference frequency), higher
resolutions are possible than with the traditional pulse counting
approach. Time intervals can now be measured down to picoseconds, but
require some specific and rather expensive hardware to do so.

13 digits of resolution can be had (Iwatsu SC-7217A at 10 seconds), but
this type of device is way into the "time nuts" and metrology territory.
It's rather unlikely that you will need one in day-to-day use unless
that use involves laboratory work with time and frequency standards.

And you'd need an atomic frequency standard (a "cesium clock" or even a
hydrogen maser) as a reference to go with it, otherwise it'd be lots of
resolution without a commensurate accuracy

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

Additionally, you should keep in mind that other instruments (notably
oscilloscopes and some signal generators with an external input) can
also measure frequencies in addition to providing their main function.

A Siglent SDS1000X-E series oscilloscope is not too expensive, works
reasonably well in its main function as an oscilloscope, and it also
has some frequency measurement capabilities, even multiple ones:

1. Dedicated frequency measurement from the trigger circuitry
This works like a dedicated recipocal counter (type 2 above) that is
hard-wired into the trigger circuitry of the oscilloscope inputs. It has
6 digits of resolution and about 5.5 digits of accuracy (going down to
about 4.5 digits accuracy below 1 kHz) and can measure anything from 10
Hz up to the maximum input frequency of the oscilloscope model.

2. Automatic measurement from waveform parameters
This works like measuring the waveform period on the screen and than
calculating the frequency from that, but it does so automatically. The
algorithm seems able to pick up the periods as needed to maintain an
accuracy of approximately 0.2 % in the frequency measurement.

3. Manual cursor measurement
Just as the name says. You move the on-screen cursors with a knob and
the oscilloscope displays time interval and frequency that correspond to
the difference in cursor positions. This is limited by manually moving
the cursors as you cannot place them more precisely than one pixel on
the LCD screen. About 0.5 % accuracy approximately.

4. Integrated FFT function
This can pull weak signals out of noise and other stuff but the
resolution strongly depends on the FFT parameters. Fortunately with
a 1 million points FFT and a choice of window functions available, one
can pull a frequency at up to 5 digit resolution out of the spectrum
given some dexterity and a well chosen set of parameters.

5. Data capture and manual analysis (just for the sake of completeness)
Basically, you save the waveform as a text file, copy it to a PC and
process it with a mathematics program like Octave (Open-source program
for the Matlab language). This is laborious as you'd have to program
much of your own processing algorithms but it can pull weak signals out
of noise or overlapping stronger signals and has lots of flexibility.

A 4 channel 100 MHz model in the above oscilloscope series (SDS1104X-E)
is about $500 and a 200 MHz version of same (SDS1204X-E) is about $800.
It either case, that would give you a reasonably capable 4 channel
oscilloscope with a basic but acceptable day-to-day use frequency
measurement ability.

So, if your requirements are modest, you may not even need a dedicated
frequency measurement instrument if another multifunction instrument
can provide that capability as a secondary function.

Regards
Dimitrij

 

[Peter Percival]

Dimitrij Klingbeil wrote:

On 2019-07-15 16:33, Peter Percival wrote:
I'd like to buy a frequency counter but I was afraid they'd cost
thousands.  Apparently, "only" hundreds -
https://uk.rs-online.com/web/p/products/6654919/?tpr=2.  Any
recommendations?

Hi

Thank you for a very comprehensive reply.

Some can indeed cost thousands, but that would be the high end ones like
an Iwatsu SC-7217A. Nothing wrong with that, it's likely a fine
instrument. 12 digit resolution at a 1 second gate time means single
picosecond time interval measurement resolution internally (at this
resolution, it would have to measure the ratios of time intervals rather
than counting pulses that the simpler "normal" frequency counters do).

On the other hand, a frequency counter can be had for a lot less, the
most basic ones do not even cost "hundreds". A search on Aliexpress
turns multiple variants for around $60 - $80, some of them with RF
inputs to 1 or 2.4 GHz. A "bare board" frequency counter module costs
even less, but of course you're not getting a housing and mains power
supply with one. There are reviews of some of these things on EEVblog.

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

The choice of frequency counter comes down to your frequency range and
accuracy needs. There are versions for everyone and basically they all
come in one of 3 basic types (from simplest to most complex):

1. Basic "gated edge counter"
This type of device simply enables a count-up mechanism for a specified
time (called the "gate time") and counts how many edges of one polarity
occur during this time. The number of edges divided by the gate time
gives the value of the frequency measured. However, if the frequency is
low, there are not enough edges to get a decent resolution, so the low
frequency resolution (and low frequency accuracy) becomes very poor.

This is the cheapest type of counter that can be made, but I would NOT
recommend buying one because of the really useless performance in the
low frequency ranges. There are other more reasonable options not too
expensive. The time of the simple "direct counting" passed years ago.

2. Recipocal counter
This device may switch to direct counting mode and work like a basic
counter at highest input frequencies, but will switch to a different
method at moderate and low frequencies. Instead of counting how many
pulses of an input signal have occurred during a fixed time interval, it
measures how many pulses of an internal precise high frequency reference
signal have occurred during some (variable and typically automatically
selected) number of cycles of the input signal. The ratio between these
numbers of cycles multiplied by the internal reference frequency gives
the measured input frequency. This counter type maintains a high
measurement resolution down to low frequencies. However it is still
dependent on counting integer numbers of cycles, therefore it cannot
achieve high resolutions without using either very high internal
reference frequencies and (/or) very long gate times.

This is the type in most common use nowadays. Decent oscilloscopes (and
some signal generators that have a frequency measurement input) also
tend to use this method internally. I think, this is the most advisable
type to a beginner who wants a good but not too expensive instrument.

However, 9 digits resolution is about as far as this device will do. An
older Iwatsu SC-7207 will do that, as will other less expensive ones.

3. High resolution frequency / time interval counter
This device overcomes the integer number of cycles limitation by
measuring the time intervals that elapse for some numbers of cycles at
the input frequency. Because the time intervals can be measured with
sub-cycle accuracy (with respect to the reference frequency), higher
resolutions are possible than with the traditional pulse counting
approach. Time intervals can now be measured down to picoseconds, but
require some specific and rather expensive hardware to do so.

13 digits of resolution can be had (Iwatsu SC-7217A at 10 seconds), but
this type of device is way into the "time nuts" and metrology territory.
It's rather unlikely that you will need one in day-to-day use unless
that use involves laboratory work with time and frequency standards.

And you'd need an atomic frequency standard (a "cesium clock" or even a
hydrogen maser) as a reference to go with it, otherwise it'd be lots of
resolution without a commensurate accuracy

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

Additionally, you should keep in mind that other instruments (notably
oscilloscopes and some signal generators with an external input) can
also measure frequencies in addition to providing their main function.

A Siglent SDS1000X-E series oscilloscope is not too expensive, works
reasonably well in its main function as an oscilloscope, and it also
has some frequency measurement capabilities, even multiple ones:

1. Dedicated frequency measurement from the trigger circuitry
This works like a dedicated recipocal counter (type 2 above) that is
hard-wired into the trigger circuitry of the oscilloscope inputs. It has
6 digits of resolution and about 5.5 digits of accuracy (going down to
about 4.5 digits accuracy below 1 kHz) and can measure anything from 10
Hz up to the maximum input frequency of the oscilloscope model.

2. Automatic measurement from waveform parameters
This works like measuring the waveform period on the screen and than
calculating the frequency from that, but it does so automatically. The
algorithm seems able to pick up the periods as needed to maintain an
accuracy of approximately 0.2 % in the frequency measurement.

3. Manual cursor measurement
Just as the name says. You move the on-screen cursors with a knob and
the oscilloscope displays time interval and frequency that correspond to
the difference in cursor positions. This is limited by manually moving
the cursors as you cannot place them more precisely than one pixel on
the LCD screen. About 0.5 % accuracy approximately.

4. Integrated FFT function
This can pull weak signals out of noise and other stuff but the
resolution strongly depends on the FFT parameters. Fortunately with
a 1 million points FFT and a choice of window functions available, one
can pull a frequency at up to 5 digit resolution out of the spectrum
given some dexterity and a well chosen set of parameters.

5. Data capture and manual analysis (just for the sake of completeness)
Basically, you save the waveform as a text file, copy it to a PC and
process it with a mathematics program like Octave (Open-source program
for the Matlab language). This is laborious as you'd have to program
much of your own processing algorithms but it can pull weak signals out
of noise or overlapping stronger signals and has lots of flexibility.

A 4 channel 100 MHz model in the above oscilloscope series (SDS1104X-E)
is about $500 and a 200 MHz version of same (SDS1204X-E) is about $800.
It either case, that would give you a reasonably capable 4 channel
oscilloscope with a basic but acceptable day-to-day use frequency
measurement ability.

So, if your requirements are modest, you may not even need a dedicated
frequency measurement instrument if another multifunction instrument
can provide that capability as a secondary function.

Regards
Dimitrij

--
"He who will not reason is a bigot;
he who cannot is a fool;
he who dares not is a slave."
- Sir William Drummond

 

[whit3rd]

On Friday, July 26, 2019 at 11:10:58 AM UTC-7, Dimitrij Klingbeil wrote:

On 2019-07-15 16:33, Peter Percival wrote:
I'd like to buy a frequency counter...

The choice of frequency counter comes down to your frequency range and
accuracy needs. There are versions for everyone and basically they all
come in one of 3 basic types (from simplest to most complex):

1. Basic "gated edge counter"
This type of device simply enables a count-up mechanism for a specified
time (called the "gate time") and counts how many edges of one polarity
occur during this time. The number of edges divided by the gate time
gives the value of the frequency measured. However, if the frequency is
low, there are not enough edges to get a decent resolution, so the low
frequency resolution (and low frequency accuracy) becomes very poor.

This is the cheapest type of counter that can be made, but I would NOT
recommend buying one because of the really useless performance in the
low frequency ranges

That's perhaps true for RF work, but if one wishes to kit out a Geiger counter,
or other random-pulse source, it's the function you want.

Many multimeters come with a 'low frequency range' that does exactly this.

There are so many counter options on the high-end 'counter' boxes that
a tech and I puzzled for a quarter hour on how the timebase had gotten
set up on a fancy Fluke counter. We ended up just ignoring the absolute
values (which seemed to indicate an odd gate time) and looking at ratios.

My recommendation: try before you buy, the most amazing awkwardnesses
show up in these digital boxes, and a test instrument ought NOT ever be awkward.