Oscilloscope with capture and FFT?

[miguelito]

Hi all,

 

Looking for a scope that does both data capture and FFT in the audio range. Need it to have at least two channel inputs so I can compare signals.

 

Two use cases right now:

1- Checking line noise and effect of filtering

2- Adjusting the bias in my tube amp (for which I will need high voltage probes, I know...)

 

I am just a hobbist with some technical knowledge. Any recommendations short of a $15k Tektronix are welcome.

 

Thx!

[jtwrace]

I would read Tom's advice here.  He's also VERY responsive and helpful if you wanted to direct email him.  

[miguelito]

27 minutes ago, jtwrace said:

I would read Tom's advice here.  He's also VERY responsive and helpful if you wanted to direct email him.  

Thx. This looks like a pretty dated page. There are so many digital scopes on amazon... I am curious whether there’s something that I am missing here. 

[esldude]

https://quantasylum.com/products/qa401-audio-analyzer

$449

 

Not sure if this will do what you want.  It is more a measuring ADC for audio.  It also will do measures on speaker level signals off a power amp.  Say connecting across the speaker leads.  At least up to 20 volts RMS.  If you wish to monitor 120 volt power it wouldn't do. If you wish to see if power filtering cleans up noise in the amp output it would.   The software it comes with is useful. 

 

You can read the manual to it:

https://cdn.shopify.com/s/files/1/1631/5609/files/QA401_Users_Manual.pdf?17986742167759890824

 

As for setting bias in your tube amps, is there any benefit using an O-scope vs a plain volt meter?  What tube amp are you checking bias on?  Best thing I ever did for bias on tube amps was change the normal bias pot to a high quality 10 turn pot.  So much easier to get spot on bias and it seemed to stay put better too.  

[accwai]

Rigol MSO1074Z perhaps? Please read the datasheet and judge for yourself.

[mansr]

An oscilloscope isn't the best tool for this. The sample resolution (often 8 bits, up to 12 if you're lucky) isn't enough for measuring noise levels relevant to audio. The frequency resolution of the FFT function is usually also rather limited.

[opus101]

Definitely it would be worth looking at AverLab as the price is well within your budget (last time I checked I recall the starting point was $3k) : https://www.avermetrics.com/products/averlab/

 

Picoscope makes a high resolution USB scope for $1200 or so : https://www.amazon.com/Pico-PicoScope-16-bit-Oscilloscope-probes/dp/B0153KEQEW

[miguelito]

11 hours ago, esldude said:

As for setting bias in your tube amps, is there any benefit using an O-scope vs a plain volt meter?  What tube amp are you checking bias on?  Best thing I ever did for bias on tube amps was change the normal bias pot to a high quality 10 turn pot.  So much easier to get spot on bias and it seemed to stay put better too.  

My amp (Audio Note Kondo Ongaku) has two adjustments: bias and hum. The bias adjustment is done by checking the drop in a resistor inside with a volt meter as you said. There's also a hum adjustment (I don't know how this works circuit wise), and it is for this that I want the scope. In principle I could build a simple low-voltage diode bridge, add some caps, and measure the voltage on them (absolute level doesn't matter, just trying to minimize the hum). I have seen it done with a scope hence I would do it with such. In all honesty, I think the diode bridge+cap into a voltmeter would be more precise.

[miguelito]

I have seen these wackadoodle devices that people use to "listen" to line noise. How do I build such a thing?

[BrokeLinuxPhile]

The hum adjustment is probably a humdinger, or a variant of it.  A pot or two resistors, 1 leg on each phase, joined together at ground.  Often used when there is no center tap to fix the ground potential for each phase on a AC filament transformer secondary.

 

A scope recommendation is a tough call without knowing how often you would use it.  If you are a hobbyist and a DIY type, they are great to have and are used often.  Other than that, it will probably collect dust.  Have to consider also things like building a proper dummy load, buying a quality function/waveform generator, cost of GOOD probes, cabling, proper isolation of scope so results have proper context, probably a bunch of other stuff I'm forgetting.  Can be a black hole of money.

 

Barring purchase, look at rental or lease of a network analyzer or high precision scope.  There are lots of places out there that do this, and you can use lab grade equipment to analyze your system without dropping $10-20K.  The A/D converters are just way superior on high end units, it makes sub $1K scopes look like toys in comparison.  They are much better isolated from the line as well.

[miguelito]

2 hours ago, mansr said:

An oscilloscope isn't the best tool for this. The sample resolution (often 8 bits, up to 12 if you're lucky) isn't enough for measuring noise levels relevant to audio. The frequency resolution of the FFT function is usually also rather limited.

Good points. What would you recommend if I wanted to measure this?

[miguelito]

For hum adjustment, I am thinking something like the below on the speaker output might be good:

 

 

 

Additionally the resistor stepdown could make my Ongaku into a very expensive headphone amp!

[mansr]

9 minutes ago, miguelito said:

For hum adjustment, I am thinking something like the below on the speaker output might be good:

 

 

 

Additionally the resistor stepdown could make my Ongaku into a very expensive headphone amp!

Why not a (true RMS) AC volt meter?

[mansr]

18 minutes ago, miguelito said:

Good points. What would you recommend if I wanted to measure this?

What exactly do you want to measure?

[BrokeLinuxPhile]

A high quality rms voltmeter is a good idea, just read the hum measurement directly by putting it into AC mode.  Get a big heatsink, build a dummy load out of quality power resistors, hook up voltmeter leads across the load.  Should be that simple.  If you plan things out well, you probably won't even need to solder it up, and could use terminal strips and crimped connectors.  Put some thermal compound on bottom of the power resistors, then screw down to heatsink.  This way is better since hooking speakers up is a reactive load and can give odd measurements.

[BrokeLinuxPhile]

Something like this would work great:

 

https://www.amazon.com/Precision-2831E-Bench-Digital-Multimeter/dp/B072MKZ4DR/ref=pd_day0_hl_328_8?_encoding=UTF8&pd_rd_i=B004O8TUZ0&pd_rd_r=86372f3d-65da-11e9-b80c-09dbc8d3b34b&pd_rd_w=2ERyb&pd_rd_wg=8jkwM&pf_rd_p=ad07871c-e646-4161-82c7-5ed0d4c85b07&pf_rd_r=6G81YY96KP11SFGE06YA&refRID=6G81YY96KP11SFGE06YA&th=1

 

It's up to you if you want to pay extra for NIST calibration.  Considering what you are measuring and what you've probably spent on your system the extra $40-$50 is probably worth it for cert.

[miguelito]

42 minutes ago, mansr said:

What exactly do you want to measure?

Two things right now:

1- Line noise to assess what combination of

filters reduce it

2- Hum level out of the speaker terminal of my tube amp to adjust hum control to minimize hum. 

[miguelito]

31 minutes ago, BrokeLinuxPhile said:

A high quality rms voltmeter is a good idea, just read the hum measurement directly by putting it into AC mode.  Get a big heatsink, build a dummy load out of quality power resistors, hook up voltmeter leads across the load.  Should be that simple.  If you plan things out well, you probably won't even need to solder it up, and could use terminal strips and crimped connectors.  Put some thermal compound on bottom of the power resistors, then screw down to heatsink.  This way is better since hooking speakers up is a reactive load and can give odd measurements.

Agree though I don’t need precision, just minimize what I read. 

[accwai]

1 hour ago, miguelito said:

Two things right now:

1- Line noise to assess what combination of filters reduce it

 

You mean measure effect of AC line filters? Take a look at the following perhaps:

 

https://www.audioshark.org/power-conditioners-regulators-24/measuring-ac-power-line-noise-11713.html

 

Quote

2- Hum level out of the speaker terminal of my tube amp to adjust hum control to minimize hum.

 

A meter that can read true RMS for AC will do it. Something like this perhaps:

 

https://www.amazon.com/multimeter-autoranging-volmate-capacitance-connecting/dp/B07GBW7NCC

 

Just measure across the speaker, no need for a load resister. In fact, if you put your ear next to the appropriate driver, you'll be able to hear the hum drop down then come back up as you adjust the hum control. It's quite easy to find the precise lowest point by ear. But you might need a helper if your amp is stereo.

[miguelito]

13 minutes ago, accwai said:

A meter that can read true RMS for AC will do it.

Yes, I have a multimeter of this sort. I believe I do need to put a load on my amp though - a 16 Ohm / 10 W resistor would do I think.

 

Thx.

[Ralf11]

most speakers present complex loads to the amp, so performance with a pure resistance may not match what it does when you use it for music...

[miguelito]

1 minute ago, Ralf11 said:

most speakers present complex loads to the amp, so performance with a pure resistance may not match what it does when you use it for music...

I am not trying to assess that, just minimize the hum on the output. I could just connect a little speaker to it.

[esldude]

4 hours ago, mansr said:

Why not a (true RMS) AC volt meter?

There are a few variations of these hum balance adjustments.  I've only worked on push-pull amps that have them.  I seem to recall in some variations people suggest using a scope and getting symmetry to the hum bucking that isn't possible just doing it by ear or by voltmeter.  Of course it would be nice if someone knows how this amp has the hum adjustment done.  It'll be some variation on balancing AC filaments of course. 

[esldude]

This video shows the Oscope method with symmetrical clipping as a guide.  I've never done this method.  He gets to that portion around the 3 and a half minute mark.  Whether this is a good method for an SET I don't know.