Expensive Digital Audio Test Equipment

[Julian]

Just following up a previous thread where Scott recommended the Prism
DSA-1 and Sencore DA795 test sets. I have both units on hand now and
I have to say I prefer the Prism.

It has a more sturdy construction and a display that is easier to read
IMO. I find the menus keys easier to navigate on the Prism and the
scroll wheel awkward on the Sencore. There is a whole bunch of LED's
on the Prism front panel that show you instantly when there is a
problem. It locks up much faster than the Sencore and reads out on
several measurements to greater resolution. It also has a pass
through mode which means if you need to look at something that is live
on air, you can pop it into circuit loosing only a few seconds of
audio. The Sencore won't do a pass through mode and takes twice as
long to lock up so it cannot be used this way which is a real minus
for my possible applications. The Prism does not do greater than 48
kHz however and it is a "old" unit as opposed to the Sencore for which
they haven't even finished writing the manual! Other than that and
the $1,000 less on the price tag, I cannot find anything else I like
better about the Sencore.

The multiple LED's on the Prism are very handy indeed. I had a couple
of router outputs I suspected were bad and as soon as I plugged it in
I could see a carrier error light which instantly confirmed a fault.
I plugged it into other router outs and everything's all green. I
plugged it into an ISDN codec and instantly knew the sample rate was
marginally > 200 ppm off by a yellow light. After seeing the lights
you can quickly go to that section of the instrument and make detailed
test but you know literally in 2 seconds if you have a problem or not
and where to start looking.

A fine piece of engineering.

Julian

[Abemeister]

What did the Prism cost?

Also what instruments and what kind of router was that? Thanks!

On Jan 24, 3:25 am, Julian > wrote:
> Just following up a previous thread where Scott recommended the Prism
> DSA-1 and Sencore DA795 test sets. I have both units on hand now and
> I have to say I prefer the Prism.
>
> It has a more sturdy construction and a display that is easier to read
> IMO. I find the menus keys easier to navigate on the Prism and the
> scroll wheel awkward on the Sencore. There is a whole bunch of LED's
> on the Prism front panel that show you instantly when there is a
> problem. It locks up much faster than the Sencore and reads out on
> several measurements to greater resolution. It also has a pass
> through mode which means if you need to look at something that is live
> on air, you can pop it into circuit loosing only a few seconds of
> audio. The Sencore won't do a pass through mode and takes twice as
> long to lock up so it cannot be used this way which is a real minus
> for my possible applications. The Prism does not do greater than 48
> kHz however and it is a "old" unit as opposed to the Sencore for which
> they haven't even finished writing the manual! Other than that and
> the $1,000 less on the price tag, I cannot find anything else I like
> better about the Sencore.
>
> The multiple LED's on the Prism are very handy indeed. I had a couple
> of router outputs I suspected were bad and as soon as I plugged it in
> I could see a carrier error light which instantly confirmed a fault.
> I plugged it into other router outs and everything's all green. I
> plugged it into an ISDN codec and instantly knew the sample rate was
> marginally > 200 ppm off by a yellow light. After seeing the lights
> you can quickly go to that section of the instrument and make detailed
> test but you know literally in 2 seconds if you have a problem or not
> and where to start looking.
>
> A fine piece of engineering.
>
> Julian

[Arny Krueger]

"Abemeister" > wrote in message
ups.com

> What did the Prism cost?

I don't think that Prism knows how to sell anything for much less than about
3 $large. ;-)

[Julian]

On 24 Jan 2007 07:20:48 -0800, "Abemeister" > wrote:

>What did the Prism cost?

Both units are on evaluation. We haven't paid for either. The Prism
is over $6,000 and the Sencore just under $5,000.

>Also what instruments and what kind of router was that? Thanks!

The SAS 64000 router is part of a radio station. We have hundreds of
analog and digital cross points which interconnect all the rooms and
devices in the station's 3 production rooms, on air control room, main
equipment room and patch bays. For example all workstations are set
up so their record inputs are controlled from the router. That way
you can record from any console or playback machine (analog or
digital) in the entire building by simply assigning it so. I could
plug anything into a digital or an analog patch bay back in the rack
room and assign it to any record device in the building.

The questionable outputs tested bad with or without audio routed to
those channels. There is a carrier problem with those outputs
according to the Prism. Previously I wasn't sure if the problem was
with the audio card inputs or the router outputs. Now I know the
signal feeding the cards is questionable. I still haven't solved the
issue, but at least I'm looking in the right places now.

Julian

[Scott Dorsey]

Julian > wrote:
>
>The questionable outputs tested bad with or without audio routed to
>those channels. There is a carrier problem with those outputs
>according to the Prism. Previously I wasn't sure if the problem was
>with the audio card inputs or the router outputs. Now I know the
>signal feeding the cards is questionable. I still haven't solved the
>issue, but at least I'm looking in the right places now.

NOW it's time for the oscilloscope.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Julian]

On 24 Jan 2007 14:22:02 -0500, (Scott Dorsey) wrote:

>Julian > wrote:
>>
>>The questionable outputs tested bad with or without audio routed to
>>those channels. There is a carrier problem with those outputs
>>according to the Prism. Previously I wasn't sure if the problem was
>>with the audio card inputs or the router outputs. Now I know the
>>signal feeding the cards is questionable. I still haven't solved the
>>issue, but at least I'm looking in the right places now.
>
>NOW it's time for the oscilloscope.
>--scott

What's the scope going to tell me?

Julian

[Scott Dorsey]

In article >,
Julian > wrote:
>On 24 Jan 2007 14:22:02 -0500, (Scott Dorsey) wrote:
>
>>Julian > wrote:
>>>
>>>The questionable outputs tested bad with or without audio routed to
>>>those channels. There is a carrier problem with those outputs
>>>according to the Prism. Previously I wasn't sure if the problem was
>>>with the audio card inputs or the router outputs. Now I know the
>>>signal feeding the cards is questionable. I still haven't solved the
>>>issue, but at least I'm looking in the right places now.
>>
>>NOW it's time for the oscilloscope.
>
>What's the scope going to tell me?

First of all it'll tell you if you have any signal on the line at all.
Secondly it will tell you if the voltages are correct.
Thirdly it'll tell you if the waveform is nice and square, or if you
have lots of reflections from cable or termination problems.
Fourthly it'll let you know if the clock is any good or if the eye
pattern is jumping around a lot from jitter.

Ultimately, you can actually read the subcode and data out on the scope,
but it's a major pain and it takes a scope with good delay triggering.
It's not really the right tool for that, but it can work when it's the
only tool you have.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Julian]

On 25 Jan 2007 09:14:46 -0500, (Scott Dorsey) wrote:


>>What's the scope going to tell me?
>
>First of all it'll tell you if you have any signal on the line at all.
>Secondly it will tell you if the voltages are correct.

The Prism tells me that.

>Thirdly it'll tell you if the waveform is nice and square, or if you
>have lots of reflections from cable or termination problems.

Good point.

>Fourthly it'll let you know if the clock is any good or if the eye
>pattern is jumping around a lot from jitter.

I already know it is from the read out on the Prism.

>Ultimately, you can actually read the subcode and data out on the scope,
>but it's a major pain and it takes a scope with good delay triggering.
>It's not really the right tool for that, but it can work when it's the
>only tool you have.
>--scott

I will put the scope on the scope output of the Prism as the next step
but first I need to make some probes to get to the output of the
router.

Julian

[Scott Dorsey]

Julian > wrote:
>On 25 Jan 2007 09:14:46 -0500, (Scott Dorsey) wrote:
>
>>Fourthly it'll let you know if the clock is any good or if the eye
>>pattern is jumping around a lot from jitter.
>
>I already know it is from the read out on the Prism.

The Prism tells you only that it can't lock up... once it DOES lock up
it gives you a rough scalar number about jitter. I don't recall if you
can get a real jitter spectrum from it, but it's a nice thing to have.

I'm not sure the Prism is all that useful to tell you why it won't lock
up, when it won't lock up. But, it's true that just knowing that it won't
lock up is very useful information.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Julian]

On 26 Jan 2007 08:37:39 -0500, (Scott Dorsey) wrote:

>Julian > wrote:
>>On 25 Jan 2007 09:14:46 -0500, (Scott Dorsey) wrote:
>>
>>>Fourthly it'll let you know if the clock is any good or if the eye
>>>pattern is jumping around a lot from jitter.
>>
>>I already know it is from the read out on the Prism.
>
>The Prism tells you only that it can't lock up... once it DOES lock up
>it gives you a rough scalar number about jitter. I don't recall if you
>can get a real jitter spectrum from it, but it's a nice thing to have.

In this particular case, it shows a jitter figure and a eye closure
figure and it keeps changing. so I can see it is jumping around.

>I'm not sure the Prism is all that useful to tell you why it won't lock
>up, when it won't lock up.

How will a picture give me any more info about the why?

I did fix one problem already with the Prism I might not ever have
found otherwise. I had a digital DA coming up to 3 points on a patch
bay and as I measured the one output had twice the jitter and eye
closure specs as the other. then I checked the voltage and it was
about 60% of the voltage on the others too, so I suspected wiring.
The outputs of the DA are on Phoenix connectors and I had hot and
ground swapped on one connector. The interesting thing is that the
signal measured "no errors" even with the unbalanced signal, low
voltage and jitter. AES voltage spec can be as low as 2 V and it was
3.2 V so it technically passed although it was 6 V after I wired it up
the right way.

As you say, a scope will be a useful tool, if I can just learn what to
look for on it and how to interpret the results. the hand held Prism
unit is a very useful tool too when you're crawling around on the
floor or getting in behind equipment looking for information. Plus it
has that pass through mode and a watch dog function that logs results
over time.

Julian

[Scott Dorsey]

Julian > wrote:
>On 26 Jan 2007 08:37:39 -0500, (Scott Dorsey) wrote:
>
>>Julian > wrote:
>>>On 25 Jan 2007 09:14:46 -0500, (Scott Dorsey) wrote:
>>>
>>>>Fourthly it'll let you know if the clock is any good or if the eye
>>>>pattern is jumping around a lot from jitter.
>>>
>>>I already know it is from the read out on the Prism.
>>
>>The Prism tells you only that it can't lock up... once it DOES lock up
>>it gives you a rough scalar number about jitter. I don't recall if you
>>can get a real jitter spectrum from it, but it's a nice thing to have.
>
>In this particular case, it shows a jitter figure and a eye closure
>figure and it keeps changing. so I can see it is jumping around.

Ahh! Then you HAVE a scope, right there! I didn't know you could get
that.

>>I'm not sure the Prism is all that useful to tell you why it won't lock
>>up, when it won't lock up.
>
>How will a picture give me any more info about the why?

If the waveform is rounded at the corners, you have a high frequency rolloff
problem. If the waveform is rising instead of flat on top, you have a low
frequency rolloff problem. If the waveform has a chunk out of the corner,
it's a reflection problem. If it's jumping back and forth you have a clock
stability problem.

Looking at the eye pattern tells you a lot... how big the hole in the middle
of the eye is tells you how stable the clock is. If you see a nice big hole
you can be happy, but if you see only a tiny hole you invariably have a jitter
issue even if the system is locking up properly.

>I did fix one problem already with the Prism I might not ever have
>found otherwise. I had a digital DA coming up to 3 points on a patch
>bay and as I measured the one output had twice the jitter and eye
>closure specs as the other. then I checked the voltage and it was
>about 60% of the voltage on the others too, so I suspected wiring.
>The outputs of the DA are on Phoenix connectors and I had hot and
>ground swapped on one connector. The interesting thing is that the
>signal measured "no errors" even with the unbalanced signal, low
>voltage and jitter. AES voltage spec can be as low as 2 V and it was
>3.2 V so it technically passed although it was 6 V after I wired it up
>the right way.

That is slick! I like that!

>As you say, a scope will be a useful tool, if I can just learn what to
>look for on it and how to interpret the results. the hand held Prism
>unit is a very useful tool too when you're crawling around on the
>floor or getting in behind equipment looking for information. Plus it
>has that pass through mode and a watch dog function that logs results
>over time.

You don't need super resolution on the the eye pattern display, so the
little LCD might be fine enough.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Julian]

On 26 Jan 2007 17:05:21 -0500, (Scott Dorsey) wrote:


>You don't need super resolution on the the eye pattern display, so the
>little LCD might be fine enough.
>--scott

No display, just a number :-(

Julian

[Scott Dorsey]

Julian > wrote:
>On 26 Jan 2007 17:05:21 -0500, (Scott Dorsey) wrote:
>>You don't need super resolution on the the eye pattern display, so the
>>little LCD might be fine enough.
>
>No display, just a number :-(

That's not useful. You gotta have an actual display so you can see the
eye pattern. The eye pattern is useful only for qualitative measurements
for the most part, but good ones. As I said, the scope is handy.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Julian]

On 27 Jan 2007 08:09:25 -0500, (Scott Dorsey) wrote:

>Julian > wrote:
>>On 26 Jan 2007 17:05:21 -0500, (Scott Dorsey) wrote:
>>>You don't need super resolution on the the eye pattern display, so the
>>>little LCD might be fine enough.
>>
>>No display, just a number :-(
>
>That's not useful. You gotta have an actual display so you can see the
>eye pattern. The eye pattern is useful only for qualitative measurements
>for the most part, but good ones. As I said, the scope is handy.
>--scott

The Sencore has a scope function, but it displays the decoded analog
signal, not the digital signal. My thought on that one was "Cute, but
not very helpful." The Prism has only 2 lines of text, but I find the
LCD display hard to read on the Sencore.

Thanks for the suggestions about how to interpret the scope pattern.
I'll let you know what I find out next week.

Julian

[[email protected]]

Hello All, I have a one time use for either of the above tools and
found this board through a yahoo search. Is there anyone here that
would be willing to rent me either the sencore or the prism for 2 days
or so. I'm willing to fedex it back overnight if need be. I have a
dCS DAC that needs the world clock signal calibrated after a software
upgrade, and I'm not having much luck with my HP frequency counter.
It only offers one decimal place or 10ppm resoloution and the spec on
the DAC states that it needs to be calibrated to withing +- 1ppm. dCS
has provided me with the diagnostic tool for the calibration, and now
all I need is a tool to give me the world clock frequency. dCS
recomends the Prism unit, but it would appear that the Sencore
actually allows a greater degree of precision in getting the clock
signal spot on. I have been able to rent other test equipment in the
pas but this is too specialized of a piece of equipment to find at a
typical rental house. Please Please help. BTW if you are in the
Chicago, or Milwaukee area, I can come to you. Thanks Alot!!!

[Scott Dorsey]

> wrote:
>Hello All, I have a one time use for either of the above tools and
>found this board through a yahoo search. Is there anyone here that
>would be willing to rent me either the sencore or the prism for 2 days
>or so. I'm willing to fedex it back overnight if need be. I have a
>dCS DAC that needs the world clock signal calibrated after a software
>upgrade, and I'm not having much luck with my HP frequency counter.
>It only offers one decimal place or 10ppm resoloution and the spec on
>the DAC states that it needs to be calibrated to withing +- 1ppm. dCS
>has provided me with the diagnostic tool for the calibration, and now
>all I need is a tool to give me the world clock frequency. dCS
>recomends the Prism unit, but it would appear that the Sencore
>actually allows a greater degree of precision in getting the clock
>signal spot on. I have been able to rent other test equipment in the
>pas but this is too specialized of a piece of equipment to find at a
>typical rental house. Please Please help. BTW if you are in the
>Chicago, or Milwaukee area, I can come to you. Thanks Alot!!!


Sencore has an actual loan procedure... call them up, explain it, and
they will probably let you borrow one for a week or so to try out. They
are very good about this, and they do it in the hopes that you will fall
in love and buy one after you try it.

Prism may be willing do this as well but they don't stock much in the US
which makes it harder.

You could use a higher-grade HP, though. Tucker will rent you one of those.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Richard Crowley]

> wrote ...
> Hello All, I have a one time use for either of the above tools

To what does "above tools" refer? I don't see anything "above"?

> would be willing to rent me either the sencore or the prism for 2 days

Model numbers?

> signal spot on. I have been able to rent other test equipment in the
> pas but this is too specialized of a piece of equipment to find at a
> typical rental house. Please Please help. BTW if you are in the
> Chicago, or Milwaukee area, I can come to you. Thanks Alot!!!

People may know of sources/rental if you provided Make/Model.

[[email protected]]

The models are

Prism DSA-1 and Sencore DA795

I have contaced both manufacturers. Sencore is the only one that can
make one available for "demo" but the sales rep has to prequalify the
opportunity, and it's pretty obvious that this is the only thing I
would ever use this for. I'm a speaker designer. I only have 3
pieces of equipment that even use AES/EBU connections or have world
clock connections. If anyone knows a place that rents either or would
be willing to take some cash in exchange for the use of theirs, I
would apreciate it. Thanks!


What I have found with frequency counters is that even though mine has
a 9 digit readout, it only has resoloution to .1Hz at the most precise
setting. I haven't found too many others that do a whole lot more and
becasue the world clock signal isn't allways at a very high level, it
doesn't allways get picked up on the counter.

Thanks!

[Mike Rivers]

On Feb 6, 4:03 pm, wrote:
> What I have found with frequency counters is that even though mine has
> a 9 digit readout, it only has resoloution to .1Hz at the most precise
> setting.

I have a Fluke 1910A counter on my bench that I picked up at a hamfest
for about $15 that reads 44.0982 kHz coming out of my Mackie HDR24/96.
Is that close enough for your purposes?

How about measuring period rather than frequency? I can read 22.6767
uS, which converts to 44098.13 Hz.

[Scott Dorsey]

Mike Rivers > wrote:
>On Feb 6, 4:03 pm, wrote:
>> What I have found with frequency counters is that even though mine has
>> a 9 digit readout, it only has resoloution to .1Hz at the most precise
>> setting.
>
>I have a Fluke 1910A counter on my bench that I picked up at a hamfest
>for about $15 that reads 44.0982 kHz coming out of my Mackie HDR24/96.
>Is that close enough for your purposes?

Not good enough to set a reference clock, I think. The problem is that
in order to check your clock reference, you need a counter with a reference
that is an order of magnitude better than the clock you're testing. And
that means a rubidium standard or something comparable if you're trying
to adjust something like a network master clock.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Mike Rivers]

On Feb 6, 8:01 pm, (Scott Dorsey) wrote:

> in order to check your clock reference, you need a counter with a reference
> that is an order of magnitude better than the clock you're testing. And
> that means a rubidium standard or something comparable if you're trying
> to adjust something like a network master clock.

I understand the measurement theory, but just how accurate does an
audio sdata ample rate clock have to be?

[Scott Dorsey]

Mike Rivers > wrote:
>On Feb 6, 8:01 pm, (Scott Dorsey) wrote:
>
>> in order to check your clock reference, you need a counter with a reference
>> that is an order of magnitude better than the clock you're testing. And
>> that means a rubidium standard or something comparable if you're trying
>> to adjust something like a network master clock.
>
>I understand the measurement theory, but just how accurate does an
>audio sdata ample rate clock have to be?

Depends on what you want it for. If it's a CD player, it doesn't have to
be very accurate but it DOES have to have very low phase noise. If it's
a studio master clock, it should be a lot more accurate. If it's a network
master clock for broadcast, it ought to be a lot more accurate than that
if only because other stuff is being derived from the digital clock.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Julian]

On 6 Feb 2007 13:03:49 -0800, wrote:


>I have contaced both manufacturers. Sencore is the only one that can
>make one available for "demo" but the sales rep has to prequalify the
>opportunity, and it's pretty obvious that this is the only thing I
>would ever use this for. I'm a speaker designer. I only have 3
>pieces of equipment that even use AES/EBU connections or have world
>clock connections. If anyone knows a place that rents either or would
>be willing to take some cash in exchange for the use of theirs, I
>would apreciate it. Thanks!

Yikes. Time to improvise. Embellish a story that makes it seem like
you might buy one :-). Sencore and Prism both sent us one but we are
a state University department and it turns out we decided to actually
buy the Prism.

Julian

[Arny Krueger]

> wrote in message
ups.com

> What I have found with frequency counters is that even
> though mine has a 9 digit readout, it only has
> resoloution to .1Hz at the most precise setting. I
> haven't found too many others that do a whole lot more
> and becasue the world clock signal isn't allways at a
> very high level, it doesn't allways get picked up on the
> counter.


Begs the question why this application *needs* so much accuracy.

[Arny Krueger]

"Scott Dorsey" > wrote in message

> Mike Rivers > wrote:
>> On Feb 6, 8:01 pm, (Scott Dorsey) wrote:
>>
>>> in order to check your clock reference, you need a
>>> counter with a reference that is an order of magnitude
>>> better than the clock you're testing. And that means a
>>> rubidium standard or something comparable if you're
>>> trying
>>> to adjust something like a network master clock.
>>
>> I understand the measurement theory, but just how
>> accurate does an audio sdata ample rate clock have to be?
>
> Depends on what you want it for. If it's a CD player, it
> doesn't have to be very accurate but it DOES have to have
> very low phase noise. If it's
> a studio master clock, it should be a lot more accurate.
> If it's a network master clock for broadcast, it ought to
> be a lot more accurate than that if only because other
> stuff is being derived from the digital clock. --scott

Reading between the lines, the application might have something to do with
loudspeaker testing.

[Scott Dorsey]

Arny Krueger > wrote:
>
>Reading between the lines, the application might have something to do with
>loudspeaker testing.

For loudspeaker testing, three or four digits of precision is just fine.

After all, if the absolute pitch is off by a hundredth of a percent, it
will still be much less than the width of any of the device resonances that
you're trying to measure.

You probably care about clock phase noise, though, but this adjustment won't
affect that. (And you could argue that the sidebands caused by jitter will
be dwarfed by the sidebands created by typical speaker drivers, too, but
tweeters are getting better every day...)
--scott


--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[[email protected]]

To answer the questions of the above posts:

I would hope to have 3 decimal places of precision i.e. 44,100.001 Hz
This would allow adjustment to slightly beyond the spec of plus or
minus 1ppm. That is the ideal world. I would be happy with being to
measure to one less decimal place or 44,100.01 for example. Both the
Prism and the Sencore seem to allow that depth of measurement. Right
now the best I can get is 1 decimal place using my HP frequency
counter.

The application actually has little to do with testing and more as a
master reference to rule out the DAC as a less then perfect piece of
equipment. The dCS Delius is an aproximately $9,000.00 DAC and it's
considered by many to be the best in the world, barring it's bigger
brother. Basically the theory is that if there arn't any flaws in the
hardware before the loudspeakers, then any flaws heard are in the
design.

Expensive test equipment and pseudo-anechoic testing can only tell you
so much. In the end you still have to listen and tweak the final
product by ear.

I have designed technically correct speakers that sounded like crap.
So you can't rule out the human element.

After software updates the DAC needs to have it's clock recalibrated.
It can be one at the Mfg. , but the Mfg, is in the UK and charges
something like $350 / Hr. to work on these plus shipping a 45lb box
overnight to the UK. In the end this calibration comes to be
something like $1000.00 It doesn't make finnancial sense when I can
rent a super accurate counter and do the same thing in my shop. The
catch is that DAC's arn't allways coutner friendly becasue counters
have a varying sensativity to the incoming signal at different levels
of resoloution. That's why the recomended piece of equipment to use
is the Prism DSA-1. The sencore pretty much does the same thing, so
that's an option too.


I apreciate all the advice on here. In the end though, I really need
to rent one of the above pieces of equipment.


As a backup option I guess I could try a rubidium counter, or a very
high end tektronix occiliscope with the jitter program, but I would
much prefer the recomended pieces of equipment. Thanks!

[Arny Krueger]

"Scott Dorsey" > wrote in message

> Arny Krueger > wrote:
>>
>> Reading between the lines, the application might have
>> something to do with loudspeaker testing.
>
> For loudspeaker testing, three or four digits of
> precision is just fine.

Agreed. One of the more sophisticated loudspeaker design houses that I know
of (JAES papers and all that) does all their work with 12 bit converters.

> After all, if the absolute pitch is off by a hundredth of
> a percent, it will still be much less than the width of
> any of the device resonances that you're trying to
> measure.

Oh, yes. I figure that the piece of equipment being calibrated is probably
sorta high end, and therefore has high end specs for its performance
parameters. Those require high end test equipment to verify. Thing is,
speakers is sort of meatball surgery compared to many things that one could
think of.

> You probably care about clock phase noise, though, but
> this adjustment won't affect that.

I'd hope not!

> (And you could argue
> that the sidebands caused by jitter will be dwarfed by
> the sidebands created by typical speaker drivers, too,

I'm thinking -60 to -80 dB down.

> but tweeters are getting better every day...)

Maybe 70 dB down, one exception being compression drivers running at high
SPLs.

No good piece of digital audio gear should have jitter sidebands just 80 dB
down - they should be at least 100-110 dB down, maybe 120. You don't get
that info with a counter, or by setting a clock frequency.

[[email protected]]

I am aware that the clock frequency doesn't directly control Jitter.
That's not the objective. The DAC allready has one of the best
internal clocks in the industry. Unfortunately when the software is
upgraded to enable a few other functions, the clock needs to be set up
for the first time. Think of it as renstalling windows and needing to
reload drivers. Therefore I'm not trying to increase accuracy but
rather to simply set up a piece of equipment I have. Once again, this
has nothing to do with testing. My test setup is a completely
seperate piece of equipemnt and it's computer based. The clock
frequency on the DAC is the only variable that needs to be setup.
Everything else is permanantly calibrated and falls into place with
the clock frequency being set.


All I need is to know If there's someone who will let me pay them rent
for a day or so's use of this equipment, or if there's someone in the
Chicago / Milwaukee area that has one I can use at their shop. Thanks!

[Ben Bradley]

On 6 Feb 2007 20:01:01 -0500, (Scott Dorsey) wrote:

>Mike Rivers > wrote:
>>On Feb 6, 4:03 pm, wrote:
>>> What I have found with frequency counters is that even though mine has
>>> a 9 digit readout, it only has resoloution to .1Hz at the most precise
>>> setting.
>>
>>I have a Fluke 1910A counter on my bench that I picked up at a hamfest
>>for about $15 that reads 44.0982 kHz coming out of my Mackie HDR24/96.
>>Is that close enough for your purposes?
>
>Not good enough to set a reference clock, I think. The problem is that
>in order to check your clock reference, you need a counter with a reference
>that is an order of magnitude better than the clock you're testing. And
>that means a rubidium standard or something comparable if you're trying
>to adjust something like a network master clock.

Many years ago I got a Motorola-brand frequency counter at a
hamfest for about $60. It goes up to the usual 50MHz range, but also
has special "tone" modes for frequencies in the roughly 50Hz to 4kHz
range that gives 0.01 Hz resolution with one-second update time (it
apparently measures period and does a reciprocal to display
frequency). This was apparently designed for testing communications
equipment that generates audio tones for signalling.
What makes this a really nice piece of gear (that I didn't have a
clue about when I bought it) is the oven-controlled crystal oscillator
that I can beat with WWV at 20MHz (I run a short wire out the 1MHz TTL
output, and tune to 20MHz WWV with a shortwave receiver). I can (after
it's warmed up) zero-beat the adjustment and know it's accurate to
somewhere around 1 part in 10^8, whereas most plain "crystal
oscillators" I've seen can be as much as 100 Hz off at 1MHz (about 1
part in 10^4).
I can look up the model of this thing, but by the looks of it I
would guess that only a few were made for use by Motorola repair
shops. But there's surely lots of older equipment like this that have
more than sufficient precision and accuracy. If you want even more
accuracy, check out this site:
http://leapsecond.com/
and especially the the time-nuts list mentioned there.

>--scott

[Julian]

On 7 Feb 2007 08:03:33 -0800, wrote:



>I apreciate all the advice on here. In the end though, I really need
>to rent one of the above pieces of equipment.

I seem to be the only one here who has one. You'd be welcome to use
ours if you were in Seattle.

Julian