Ptownkid
Well-known member
When a scope is called a 100mhz scope, I realize that is measures up to 100 mhz, but what does that mean in terms of audio?
How much is neccessary?
How much is neccessary?
osiloscope query
- Thread starter Ptownkid
- Start date
Help Support GroupDIY Audio Forum:
Ptownkid
Well-known member
Including tube gear?
[quote author="Ptownkid"]How much is neccessary?[/quote]
Baby humans can only hear up to about 20kHz, and it gets worse as we get older. Add to this the fact that there's an inverse relationship between the bandwidth of a scope and its resolution (Volts/division), and it becomes apparent that a 10 or 20MHz scope is actually better than a 100MHz one, as far as audio work is concerned.
[quote author="Ptownkid"]Including tube gear?[/quote]
Transistors, tubes, ICs, or digital, you still won't be able to hear past 20KHz (unless you're one of those freaks of nature that claim to be able to go to 25KHz, or whatever :razz
. As a starting point, look for a 10-20MHz scope with a minimum resolution of 1-5mV/division.
Peace,
Al.
Baby humans can only hear up to about 20kHz, and it gets worse as we get older. Add to this the fact that there's an inverse relationship between the bandwidth of a scope and its resolution (Volts/division), and it becomes apparent that a 10 or 20MHz scope is actually better than a 100MHz one, as far as audio work is concerned.
[quote author="Ptownkid"]Including tube gear?[/quote]
Transistors, tubes, ICs, or digital, you still won't be able to hear past 20KHz (unless you're one of those freaks of nature that claim to be able to go to 25KHz, or whatever :razz
. As a starting point, look for a 10-20MHz scope with a minimum resolution of 1-5mV/division.Peace,
Al.
bcarso
Well-known member
You can generally infer when something is oscillating indirectly, with a little experience, although it may be comforting to actually see the waveforms on the 'scope. For some areas of DIY like class D amps you might well want even more than 100MHz, but most of the time on my 2236 I leave the bandwidth limit switch engaged which makes it a 20MHz 'scope, as Scenaria recommends. And as Al points out, the higher BW comes at the inevitable expense of noise.
I've had solid-state amplifier circuits screaming at 3GHz, out of the question for any but fancy sampling scopes to look at directly. But I determined the frequency by the pattern of peaks and nulls in d.c. currents vs. distance of my hand from the circuit, and any scope probing would have likely interefered with the oscillation anyway. The cure was little damping R's until it went away. As I remarked somewhere, I more or less gave up on high-freq discretes with leads at least for the internals of amp circuits, and SMD is not easy to breadboard.
100MHz with Thomson filter response typical of 'scopes corresponds to a step response 10% to 90% of abut 0.35/100MHz = 3.5ns. That's fast enough to see more or less what's going on in class D amps, and most of the time you will be looking at the ringing in your probe ground lead unless this is very short. But you would be amazed to see some of the garish ringing with proper probing in parts with seemingly short leads when you get a faster 'scope.
I've had solid-state amplifier circuits screaming at 3GHz, out of the question for any but fancy sampling scopes to look at directly. But I determined the frequency by the pattern of peaks and nulls in d.c. currents vs. distance of my hand from the circuit, and any scope probing would have likely interefered with the oscillation anyway. The cure was little damping R's until it went away. As I remarked somewhere, I more or less gave up on high-freq discretes with leads at least for the internals of amp circuits, and SMD is not easy to breadboard.
100MHz with Thomson filter response typical of 'scopes corresponds to a step response 10% to 90% of abut 0.35/100MHz = 3.5ns. That's fast enough to see more or less what's going on in class D amps, and most of the time you will be looking at the ringing in your probe ground lead unless this is very short. But you would be amazed to see some of the garish ringing with proper probing in parts with seemingly short leads when you get a faster 'scope.
> How much is neccessary?
For what?
I started with an Eico 427, which claimed 450KC (yes, 0.4MHZ), but even that was bogus. It would not display over 100KHz except as a small blur. I learned a LOT from that piece of crap (good value for price, but still a poor excuse for a scope).
For synthesizer class, I deliberately re-built a scope with 20KHz response. In that application, if you can't hear it, why should you see it? (It also let me use $0.33 transistors instead of $5 transistors, and a major savings in power supply filtering: less MHz can be very much cheaper.)
Transistor loudspeaker amps used to love to oscillate at 2-3MHz. I have a 5MHz scope for that. But the last oscillator I built (supposed to be a headphone amp) would not focus the waves so it was screaming higher than that. And at bcarso's level of skill, GHz screams are possible, and beyond the ability of any home-shop scope to capture. For small personal DIY, you can't possibly afford a scope to cover any possible crap that can come out of a hot amplifier. I think an awful lot of good commercial audio gear was developed with 1MC-10MHz scopes, and some with no scope at all (scopes didn't exist as standard gear before 1935, and were still uncommon when WWII started).
> Including tube gear?
Most tubes can't hardly amplify past 10MHz. Usually the problem is getting to 100KHz or 1MHz. 20mHz scopes are not that expensive new from China or with careful second-hand shopping.
Yes, I had tube FM radios: they did not amplify 0-100MC, they amplified a 1MC band around 100MC. You can, and have to, tune-out the tube capacitance to get past a few MHz, and you usually know when you are doing that. Wire-wound resistors can sometimes tank-up and oscillate; carbon is your friend. Two bottles and a few inches of wire (parallel tubes) can sometimes Barkhausen in the microwave band: tight conservative layout with damping resistors may be cheaper than a super-scope.
> there's an inverse relationship between the bandwidth of a scope and its resolution (Volts/division)
Only if Money Is Constant. If Money is allowed to increase, there are many ways to improve Gain-Bandwidth product. Ringing-up the top octave (at the expense of everything past cut-off) is too-too trivial. Next you take less gain (at higher bandwidth) per stage, then increase the number of stages ("throwing money at the problem"). When gain-per-stage approaches Unity, "distribute" the gain along several tubes feeding a delay line ("throwing big money").
And in the 20MHz-100MHz, a number of $3 chips will boost milliVolts up to Volts, adding gain to low-price low-gain wide-band scopes (though the cost is so low, most scopes have all the sensitivity you need for Signal, and Noise always needs dedicated preamps).
For what?
I started with an Eico 427, which claimed 450KC (yes, 0.4MHZ), but even that was bogus. It would not display over 100KHz except as a small blur. I learned a LOT from that piece of crap (good value for price, but still a poor excuse for a scope).
For synthesizer class, I deliberately re-built a scope with 20KHz response. In that application, if you can't hear it, why should you see it? (It also let me use $0.33 transistors instead of $5 transistors, and a major savings in power supply filtering: less MHz can be very much cheaper.)
Transistor loudspeaker amps used to love to oscillate at 2-3MHz. I have a 5MHz scope for that. But the last oscillator I built (supposed to be a headphone amp) would not focus the waves so it was screaming higher than that. And at bcarso's level of skill, GHz screams are possible, and beyond the ability of any home-shop scope to capture. For small personal DIY, you can't possibly afford a scope to cover any possible crap that can come out of a hot amplifier. I think an awful lot of good commercial audio gear was developed with 1MC-10MHz scopes, and some with no scope at all (scopes didn't exist as standard gear before 1935, and were still uncommon when WWII started).
> Including tube gear?
Most tubes can't hardly amplify past 10MHz. Usually the problem is getting to 100KHz or 1MHz. 20mHz scopes are not that expensive new from China or with careful second-hand shopping.
Yes, I had tube FM radios: they did not amplify 0-100MC, they amplified a 1MC band around 100MC. You can, and have to, tune-out the tube capacitance to get past a few MHz, and you usually know when you are doing that. Wire-wound resistors can sometimes tank-up and oscillate; carbon is your friend. Two bottles and a few inches of wire (parallel tubes) can sometimes Barkhausen in the microwave band: tight conservative layout with damping resistors may be cheaper than a super-scope.
> there's an inverse relationship between the bandwidth of a scope and its resolution (Volts/division)
Only if Money Is Constant. If Money is allowed to increase, there are many ways to improve Gain-Bandwidth product. Ringing-up the top octave (at the expense of everything past cut-off) is too-too trivial. Next you take less gain (at higher bandwidth) per stage, then increase the number of stages ("throwing money at the problem"). When gain-per-stage approaches Unity, "distribute" the gain along several tubes feeding a delay line ("throwing big money").
And in the 20MHz-100MHz, a number of $3 chips will boost milliVolts up to Volts, adding gain to low-price low-gain wide-band scopes (though the cost is so low, most scopes have all the sensitivity you need for Signal, and Noise always needs dedicated preamps).
Ptownkid
Well-known member
Wow, great reply PRR, many thanks.
