Low-distortion Audio-range Oscillator

Jump to Latest
#3,642
Bob Cordell said:
Bear in mind that the phone system long ago used open-wire transmission on poles where the wires were feet apart. That's what I was referring to. Scaling 300 ohm twinlead probably gets you there.

Cheers,
Bob
Click to expand...

Bob the subscriber loops today are current loops. Were those old trunks current loops as well.
One telephone guys up here was telling me they have loops with as much a 300Vdc on them.
 
#3,643
Bob Cordell said:
Bear in mind that the phone system long ago used open-wire transmission on poles where the wires were feet apart. That's what I was referring to. Scaling 300 ohm twinlead probably gets you there.

Cheers,
Bob
Click to expand...

I thought those were telegraph wires, usually not in a loop but using mother earth for the return.

I know that a lot was done to make long distance work. That's where the loading coils that landed in the MIT cables originated. Here is a timeline with details I did not know: History of the Telephone Line and even more: Applications: Telecommunications - The Evolution of Telephone Cable

"By 1887 all of the newly manufactured cables were metallic circuit cables. Some of the early telephone cable applications were across the East River Bridge, under the North River between New York and New Jersey and across the Delaware River from Philadelphia to Camden. There were numerous cable manufacturers. The cables made by the different manufacturers were very similar but not identical. The cables contained up to 100 copper wires. They were insulated with cotton, cotton impregnated with paraffin, gutta percha or a rubber compound, and then in wrapped in lead."

"There were two major improvements made to telephone cable in the late 1880's. The first was the issuance of a specification for a standard type of telephone cable in 1888. The specification outlined a metallic circuit or twisted pair cable. The wires were 18 B & S gauge copper conducting wire (40 mils in diameter) covered with at least two layers of cotton and sheathed in a 97% lead, 3% tin alloy pipe. The spaces in the core and between the core and the pipe were filled with an insulating material."

If by the 1890's all wires were twisted pairs I don't think they were 600 Ohms.

I wonder how magical a gutta percha insulated lead wrapped cable would sound. The lead jacketing is also used for high power distribution lines.

The learning never stops on DIYaudio. . .
 
#3,644
You can buy RG58 at 75 ohms or at 50 ohms. I think it's the same cable and they hand you different chart for each. That 75 ohms is a character impedance so at what frequency is it 75 ohms and what is it at all other frequency? I have not been able to get an answer to this to date.
 
#3,646
davada said:
You can buy RG58 at 75 ohms or at 50 ohms. I think it's the same cable and they hand you different chart for each. That 75 ohms is a character impedance so at what frequency is it 75 ohms and what is it at all other frequency? I have not been able to get an answer to this to date.
Click to expand...

An RG58 that claims 75 Ohms is neither. Perhaps its mismarked RG59 which is 75 Ohms. Its real obvious what the impedance is with a network analyzer or TDR. Above audio it should be pretty stable until you get to microwave frequencies. I have cables that are within 2 Ohms of 75 to 3 GHz.
 
#3,647
1audio said:
An RG58 that claims 75 Ohms is neither. Perhaps its mismarked RG59 which is 75 Ohms. Its real obvious what the impedance is with a network analyzer or TDR. Above audio it should be pretty stable until you get to microwave frequencies. I have cables that are within 2 Ohms of 75 to 3 GHz.
Click to expand...

Probably mislabeled. It is an old role laying in a corner at work. But we wouldn't have anything other than 75 ohm. Interesting that it can maintain a uniform impedance over such a large bandwidth. The TDRs at work display attenuation in dB over distance. I haven't seen an option in the menu to display impedance.
 
Last edited:
#3,648
Here's a composite opamp which will be suitable as output driver (as long as modest output currents are fine): LT1115-LT1468_composite_opamp_r1.pdf

I'm in a hurry so just a few notes: This is a composite opamp topology which is easy to compensate (and relatively well behaved under overload conditions)--just increase the attenuation of R1-R2 until sufficient phase margin at the unity loop gain frequency is available. The shown values are derived by simulation and not verified by experimentation. For gains other than unity make sure to scale C2 for low HF gain peaking.

The basic distortion performance of this combo will be beyond any practical need (<< -140 dB THD at 1 kHz up to voltage/current clipping), the details are in the implementation (feedback network, layout and power supply effects).

Samuel
 
#3,649
<<-140dB thd buffer

Samuel Groner said:
Here's a composite opamp which will be suitable as output driver (as long as modest output currents are fine): LT1115-LT1468_composite_opamp_r1.pdf

The basic distortion performance of this combo will be beyond any practical need (<< -140 dB THD at 1 kHz up to voltage/current clipping), the details are in the implementation (feedback network, layout and power supply effects).

Samuel
Click to expand...

That should do it for when a buffer/driver is needed. Thank you.


THx-RNMarsh
 
#3,650
davada said:
Bob the subscriber loops today are current loops. Were those old trunks current loops as well.
One telephone guys up here was telling me they have loops with as much a 300Vdc on them.
Click to expand...

I was just referring to how the 600 ohm standard came about long ago.

Copper twisted pair loops have -48V across them when unloaded by a legacy handset. For safety, higher voltages are prohibited on conventional twisted pair copper.

Cheers,
Bob
 
#3,651
1audio said:
I thought those were telegraph wires, usually not in a loop but using mother earth for the return.

I know that a lot was done to make long distance work. That's where the loading coils that landed in the MIT cables originated. Here is a timeline with details I did not know: History of the Telephone Line and even more: Applications: Telecommunications - The Evolution of Telephone Cable

"By 1887 all of the newly manufactured cables were metallic circuit cables. Some of the early telephone cable applications were across the East River Bridge, under the North River between New York and New Jersey and across the Delaware River from Philadelphia to Camden. There were numerous cable manufacturers. The cables made by the different manufacturers were very similar but not identical. The cables contained up to 100 copper wires. They were insulated with cotton, cotton impregnated with paraffin, gutta percha or a rubber compound, and then in wrapped in lead."

"There were two major improvements made to telephone cable in the late 1880's. The first was the issuance of a specification for a standard type of telephone cable in 1888. The specification outlined a metallic circuit or twisted pair cable. The wires were 18 B & S gauge copper conducting wire (40 mils in diameter) covered with at least two layers of cotton and sheathed in a 97% lead, 3% tin alloy pipe. The spaces in the core and between the core and the pipe were filled with an insulating material."

If by the 1890's all wires were twisted pairs I don't think they were 600 Ohms.

I wonder how magical a gutta percha insulated lead wrapped cable would sound. The lead jacketing is also used for high power distribution lines.

The learning never stops on DIYaudio. . .
Click to expand...

I think the characteristic impedance of twisted pairs used in telephone cables is in the neighborhood of 100 ohms.

Cheers,
Bob
 
#3,652
davada said:
You can buy RG58 at 75 ohms or at 50 ohms. I think it's the same cable and they hand you different chart for each. That 75 ohms is a character impedance so at what frequency is it 75 ohms and what is it at all other frequency? I have not been able to get an answer to this to date.
Click to expand...

RG58 is 50 ohms, RG59 is 75 ohms, RG92 is I think 91 ohms.

The cable looks like its resistive characteristic impedance down to extremely low frequencies if it is extremely long.

Cheers,
Bob
 
#3,653
Samuel Groner said:
Here's a composite opamp which will be suitable as output driver (as long as modest output currents are fine): LT1115-LT1468_composite_opamp_r1.pdf

I'm in a hurry so just a few notes: This is a composite opamp topology which is easy to compensate (and relatively well behaved under overload conditions)--just increase the attenuation of R1-R2 until sufficient phase margin at the unity loop gain frequency is available. The shown values are derived by simulation and not verified by experimentation. For gains other than unity make sure to scale C2 for low HF gain peaking.

The basic distortion performance of this combo will be beyond any practical need (<< -140 dB THD at 1 kHz up to voltage/current clipping), the details are in the implementation (feedback network, layout and power supply effects).

Samuel
Click to expand...


Thanks Samuel.

I give it a go.
 
#3,654
Bob Cordell said:
I was just referring to how the 600 ohm standard came about long ago.

Copper twisted pair loops have -48V across them when unloaded by a legacy handset. For safety, higher voltages are prohibited on conventional twisted pair copper.

Cheers,
Bob
Click to expand...

That's what I thought. I was shocked when he told me that. I don't think this was for a subscriber loop. He said "well that's what the piece of equipment requires."
 
#3,656
CFA op amp output/buffer usually gives more current and can be faster over a range of local output gains too for higher internal loop gain composites

Walt shows composite audio line driver circuits in his Op Amp book http://waltjung.org/PDFs/ADI_2002_Seminar_Ch6_Audio_Drivers_I.pdf

of course Walt mostly talks to other engineers and assumes they can figure out the stability requirements
and today I doubt the AD811 would be his 1st choice - its what they were making then and its available in DIP - but its not magic

http://www.ti.com/lit/an/sboa015/sboa015.pdf shows a little of composite amp stability

I have a few composite amp sims in http://www.diyaudio.com/forums/soli...ain-composite-op-amp-circuits.html#post512806
 
Last edited:
#3,657
davada said:
50 or 600 ohms is a gross impedance mismatch to 10k or 50k ohms. There is strong possibility what you're observing is destructive signal reflection from the mismatch. I believe the old 600 standard was for this reason. It was a real pita to have to bring every input and output to 600 ohm. I think it was phased out to save the cost. It is very real concern in RF and one of the things I have to know about to do my job. I deal with it daily among other things in the 25 square kilometers of RF transmission line I'm responsible for maintaining. The effect is much more pronounced at the lower end of the RF spectrum 5mHz to 45Mhz band for us. Lower frequencies propagate more easily because they require less energy to propagate and the cable capacitance has a lesser effect on them. The cable causes a delay VOP, velocity of propagation, which is a percentage of the speed of light. Most coax cable we use is typically 0.87 VOP - 87 percent. In audio the small cable delay is probably of no consequence because the wavelengths are long but reflections exist from DC to daylight. The magnitude of reflection is proportional to the mismatch, To get a sense of this put a graphical TDR on 100' of coax and measure with the cable end shorted, then open and finally terminated with a matching resistor.

I'm not clear why this is largely ignored in the audio realm.


On the 50 ohm attenuator; I did some back of the envelope calculations on this last night.
For a 20Vpp signal we would require a 4W resistor at 50 ohms. For a constant 50 ohm output impedance the input resistance will vary increasingly and the power requirement lessens but we would still be looking at 1W to 2w parts. To do it right 0.1% parts would be required. I think I have to agree with Bob that the power requirements make it to costly.
If 50 ohms is required padding the output with T networks would be a better option but it would still require a headphone size power amp to do it which is not unreasonable.
Click to expand...
Hello,
I believe the 600ohm comes from the tube days and transformer coupling. All devices use 600ohms for their transformers to maintain exact impedance coupling. It was fairly easy to do since transformers were needed any way for tube circuits.
 
#3,660
JCX's link explains the diodes. They help recovery from overload.

Maybe Samuel can offer insight on using a composite amp inside the oscillator. With so much gain available the amplifier distortion can be reduced even more. However it may not be the limiting factor in distortion in the first place.
 
You must log in or register to reply here.
Top Bottom