Flat Cable Wire Stripper

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Lets say 1ft of cable between each of the ten devices. 100ft of cable at the output. Output impedance 50 ohm. Cable terminated with 10K ohm.

Roundabout -3db at 50K Hz cascaded ten times.

This isn’t the exact situation but it will help me understand what’s going on.
 
Last edited:
I am not smart enough to calculate the overall response of ten cascaded "filters" with a -3 dB rolloff starting at 50 kHz. My hunch sez the wiring isn't the overall issue.

Bri
 
I am not smart enough to calculate the overall response of ten cascaded "filters" with a -3 dB rolloff starting at 50 kHz. My hunch sez the wiring isn't the overall issue.

Bri
Although I didn’t go to The Institute for Audio Research, I did look at the class materials. Good old Al Grundy had you go through all these sorts of systems engineering problems. Sigh.
 
So you have 10 x 100ft in series? If each has a 3dB rolloff at approx 50KHz - that’s for 600Ω source impedance - for 50Ω source impedance the rolloff is at 558KHz - you should not be having a problem anywhere if your output is at 50Ω . A stack of 10 RC filters in series each dropping 3dB at rolloff will cascade the -3dB point downwards in frequency as the rolloff is a slope and not a sharp cutoff but at that impedance I imagine you would struggle to get rolloff below 500KHz.
Doesn’t sound like a cable issue but some device mismatch in the chain - I presume these runs are to and from gear spaced well apart? I’d try bypassing each piece of gear one by one with a shunt cable and see which one has a problem.
I’ve passed digital audio through standard mic cable for a reasonable distance without degenerating the square wave - studio to studio 25M (82ft) apart there and back no problems at 48KHz. Also we ran tape machine SMPTE time code from the digital studio to the analog tape studio and sometimes MIDI at 31.25KHz up the same lines - 2 x 24 way multicores. We forgot to run MIDI cables when we built the second studio, it seems to drop out at more than 15M with ordinary MIDI cable anyway, mic cable seems to work better.
 
In post #61, Paul was "brainstorming" 10 devices with 1' cable between each and THEN 100' going to a final destination. Each of the 10 devices has a -3 dB at 50 kHz. My post #62 suggests 10 cascaded devices, each with -3dB at 50 kHz response, might result in a "sag" at 20 kHz.

Bri
 
In post #61, Paul was "brainstorming" 10 devices with 1' cable between each and THEN 100' going to a final destination. Each of the 10 devices has a -3 dB at 50 kHz. My post #62 suggests 10 cascaded devices, each with -3dB at 50 kHz response, might result in a "sag" at 20 kHz.

Bri
At 1ft between each device the cutoff is at 55GHz
 
The 1ft between devices was supposed to negligible. I was interested in 100ft of cable with a source impedance of 50 ohm and terminated with 10K ohm.
 
If you are getting a dip at 20KHz then what I said in my previous post about bypassing each unit one by one might be useful to see if there’s one in the chain with a problem with the next or previous device. If you isolate the problem device then you put it back in, first taking the output from the second device back down the chain - see if problem stays or goes - if the problem goes then it’s the previous device interacting causing the problem - if not then take the output to the second device further up and see if the problem goes - if it does then it’s the next device interacting causing the problem. If not - ie. the problem doesn’t go in either scenario, then it’s the device itself that has a problem.
 
Maybe I'm getting senile, like both of the current USA prez candidates....LOL! uh oh....ahem.

Forget the cabling for a minute. I AM SAYING that an audio device with a single-pole (6 dB/octave) HF rolloff and a cutoff frequency (-3 dB point) at 50 kHz will result in approx. -0.6 dB rolloff at 20 kHz.

Someone less senile than me...please verify that math!

So cascade 10 of those single pole filters with a 50 kHz cutoff (the -3 dB point for each) .... seeing a response being down 1 dB or more at 20 kHz seems plausible.

Cabling will add to all this rolloff mischief.

Bri
 
Thanks Brian. Luckily I have preview channels that can be experimented with. I can try bypassing some but not all active stages. I’ll start there. If that looks promising I’ll change some compensation capacitors and see what happens.

BTW it’s actually better than I described. All the active stages are pretty flat to 50k and roll off more like 75K. This confirmed some of my suspicions though.
 
Maybe I'm getting senile, like both of the current USA prez candidates....LOL! uh oh....ahem.

Forget the cabling for a minute. I AM SAYING that an audio device with a single-pole (6 dB/octave) HF rolloff and a cutoff frequency (-3 dB point) at 50 kHz will result in approx. -0.6 dB rolloff at 20 kHz.

Someone less senile than me...please verify that math!

So cascade 10 of those single pole filters with a 50 kHz cutoff (the -3 dB point for each) .... seeing a response being down 1 dB or more at 20 kHz seems plausible.

Cabling will add to all this rolloff mischief.

Bri
The more devices in the chain the greater the downwards rolloff - it depends largely on the rolloff slope for each device as to how far down it all goes at 20KHz - but I think there’s a bandit device in the chain which could cause this. Old caps - the bane of my life! When I’ve been putting a pile of stuff in series for mastering including D/A outputs and A/D returns there is not usually a noticeable dip in HF response. But any gear that needs recapping becomes obvious when you start including it in a chain as it affects the following devices and colors the preceding.
 
Hiya Oz! We're on the same page! In post #74 Paul mentions "pretty flat" at 50 kHz and a rolloff at 75 kHz. Even after 10 cascaded circuits with those specs, IMHO the entire system should have a smaller (but "visible") droop at 20 kHz.

As for caps in the signal path.....we agree, but those usually affect the LF end of the band. But, I've seen all sorts of possibilities.

Bri
 
I haven’t investigated this yet in great detail. It could actually be a measurement error. The top end doesn’t sound like it goes on for days so I am susspicious. I will first try to eliminate a measurement error before getting too far in the weeds.

I have suspected cascading roll off of active stages. I’m talking about changing the value if the opamp compensation capacitor in some stages.

IIRC I had to push the recommended value the most with unity gain stages for wide bandwidth. Does that fit opamp theory?
 
Hiya Oz! We're on the same page! In post #74 Paul mentions "pretty flat" at 50 kHz and a rolloff at 75 kHz. Even after 10 cascaded circuits with those specs, IMHO the entire system should have a smaller (but "visible") droop at 20 kHz.

As for caps in the signal path.....we agree, but those usually affect the LF end of the band. But, I've seen all sorts of possibilities.

Bri
Agreed. If there’s a significant drop then there should be a crook device somewhere in the chain. A small but acceptable drop is what I would envisage correctable by EQ prior to or in the chain - it should have a very shallow curve at that frequency. For working off tape it does become important to have as much of your HF unaffected as possible. I started out when the only medium was tape, no digital - it was only on the stuff I worked on doing my apprenticeship at Philips - I worked in a division of Philips on the first laser disc video players - 12” laser discs back in 1971 & 1972. The tape world in studios was about getting as much top end onto tape as you could because then tapes were mainly running at 15ips - when 30ips became a standard the HF loss problems started to fade.
 
I haven’t investigated this yet in great detail. It could actually be a measurement error. The top end doesn’t sound like it goes on for days so I am susspicious. I will first try to eliminate a measurement error before getting too far in the weeds.

I have suspected cascading roll off of active stages. I’m talking about changing the value if the opamp compensation capacitor in some stages.

IIRC I had to push the recommended value the most with unity gain stages for wide bandwidth. Does that fit opamp theory?
The worst that can happen if you start mucking about with the compensation caps is oscillating circuits feeding capacitive loads. The value best suited is dependent on the capacitance of the load.
 
@RoadrunnerOZ

I think @Gold may have some testing to do to see what Rogue (if any?) in his system of 10 cascaded devices might cause his reported -1 dB at 20 kHz. I can see a possibility of just accumulated small droops per stage all adding together....ASSuming there isn't one or more rogues in that cascade.

When I was young/dumb I was hired to design/build/deliver a studio console with specs of "DC to blue light". I eventually agreed with the client that 1 dB down at 75 kHz was acceptable to ensure pretty flat response at 20 kHz and minimal phase shift.

I struggled to keep that goal with Then Available opamps....in my case the new 5534 chips. Each stage in the signal path (which was very short) required HF response well into several hundreds of kHz.

On the LF end....."oversized" coupling caps.

http://brianroth.com/projects/m77/m77.html
Not bragging, but trying to come up with something that is as flat as possible without becoming a radio transmitter in the "audible range" is a headache! lol

Bri
 
Back
Top