Thanks a lot again and this answers completely the question.
I will not do anything. Clearly I was trivializing the issue.
I will do for sure. It was clearly a bad idea.
Thanks a lot again, gino
Ginetto, your prayers are answered !
http://www.diyaudio.com/forums/swap-meet/294696-fs-custom-data-only-usb-cable-new-post.html
http://www.diyaudio.com/forums/swap-meet/294696-fs-custom-data-only-usb-cable-new-post.html
A contiguous return path for the single ended control signals that form part of the USB data protocol is a requirement, so you really need that return wire in there.....
Hi ! thanks but I understand now that I forgot to mention that price is also an issue
I was clearly underestimating the task.
I think I will try to select a good one off the shelf.
I should start saving money for a 4th hand spectrum analyzer instead ... I fell in love with the object.
I am about to move now but one day I will have one.
Thanks again, gino
Regarding noise and digital signal transfer in this case USB...
Firstly has the noise been measured, preferably at the DAC end of the USB cable...
If so levels and frequency...
Slap a ferrite round the cable as a start, this will provide common mode attenuation of some noise.
Adum devices provide galvanic isolation but do let some noise through as there is some capacitive coupling between input and output, they also only work at 12MHz. These devices are designed to cut ground loops and provide voltage isolation so could help. Layout is critical, so is filtering and the position of all these bits.
Unless it is a real problem I would stick with a decent USB cable, unless like some you believe in USB cables causing subjective differences to the sound... but then making your own is likely to make things worse.
When we do all the isolation and filtering we also have to take into account signal integrity, so we do simulations, measurements etc. it can be very interesting and when done properly involves a lot of work, outside of what the average DIYer can do usually so unless you have a problem keep it simple. I am having a discussion (argument with one person in particular who is hell bent on ridiculing and putting me down) regarding all these add on USB devices that are appearing on the market... personally I don't think they are needed and again we seem to have myths appearing; ALL USB 5V is noisy, USB cables have an effect on the sound even though the data is digital, so they would have to perform DSP duties as well as data transmission etc.
Characteristic Impedance, firstly +/-10% for the whole USB chain would be well controlled, because unless the PCB is manufactured as a controlled impedance board you probably wont even get that, most basic boards are designed using figures derived from say Saturn PCB toolkit by putting in the layer stack up details, so I would not worry to much for basic USB signal, when you get to DDR memory and gig Ethernet thinks get more critical, so use a cable ready made to the USB standard.
Firstly has the noise been measured, preferably at the DAC end of the USB cable...
If so levels and frequency...
Slap a ferrite round the cable as a start, this will provide common mode attenuation of some noise.
Adum devices provide galvanic isolation but do let some noise through as there is some capacitive coupling between input and output, they also only work at 12MHz. These devices are designed to cut ground loops and provide voltage isolation so could help. Layout is critical, so is filtering and the position of all these bits.
Unless it is a real problem I would stick with a decent USB cable, unless like some you believe in USB cables causing subjective differences to the sound... but then making your own is likely to make things worse.
When we do all the isolation and filtering we also have to take into account signal integrity, so we do simulations, measurements etc. it can be very interesting and when done properly involves a lot of work, outside of what the average DIYer can do usually so unless you have a problem keep it simple. I am having a discussion (argument with one person in particular who is hell bent on ridiculing and putting me down) regarding all these add on USB devices that are appearing on the market... personally I don't think they are needed and again we seem to have myths appearing; ALL USB 5V is noisy, USB cables have an effect on the sound even though the data is digital, so they would have to perform DSP duties as well as data transmission etc.
Characteristic Impedance, firstly +/-10% for the whole USB chain would be well controlled, because unless the PCB is manufactured as a controlled impedance board you probably wont even get that, most basic boards are designed using figures derived from say Saturn PCB toolkit by putting in the layer stack up details, so I would not worry to much for basic USB signal, when you get to DDR memory and gig Ethernet thinks get more critical, so use a cable ready made to the USB standard.
I have made cables with some random small silver wire with unknown impedance and the DAC showed signs of having trouble locking on to the signal upon cold start. So impedance is important. I have also used a cable without return and power with success. It varies with each USB device. Some require a return path (earth) but this can be connected elsewhere and not to the CPU.
It's easy to put a cable together but I doubt its worth the time unless you have compliant wire. Data transfer is extremely fast and unforgiving, even adding a ferrite can degrade the signal.
The best idea is to purchase a cheap cable and spend your time money on a regen device. I think theres 3 or 4 companies all using the same chip to accomplish this with reasonable reviews.
It's easy to put a cable together but I doubt its worth the time unless you have compliant wire. Data transfer is extremely fast and unforgiving, even adding a ferrite can degrade the signal.
The best idea is to purchase a cheap cable and spend your time money on a regen device. I think theres 3 or 4 companies all using the same chip to accomplish this with reasonable reviews.
It depends how far you go from the spec I am talking -/+ 15is % here, even the linked aerospace cable earlier is specked at +/-13 Ohms. Here I am talking about all the manufacturing tolerance you get something I am quite au fait with as I spend a lot of my time playing with signals and getting them from a to b with the utmost fidelity so I see these problems every day and know that unless there is an absolute requirement for controlled impedance PCBs a standard PCB will be ordered if for no other reason than cost. True controlled impedance boards are like boards made to IPC class 3 expensive (Interesting though I specify full class 3 boards quite often for the ultimate quality, why are they not used in audio if every little thing matters, the cost of some of the gear could justify having a board made and assembled to IPC class 3 standard)
ALL USB signalling requires a return path for the single ended signals that are part of the the USB data transfer protocol... To connect it away from the actual signal wires is stupid, I would suggest having a look at how digital signals propagate, its is not the earth and to separate the signal and its return path will cause an increase in noise and be detrimental to signal integrity. In fact the locking on was probably down to you not having a return path, I don't think you understand this signal propagation stuff much
Regen devices what do they do... try a basic USB hub first because that is all they are or make your own, why add more complications...
http://ww1.microchip.com/downloads/en/DeviceDoc/evb2412.pdf
Or just buy the evaluation board...
ftp://ftp.smsc.com/pub/usb/evb2412user.pdf
As I have said I believe these audiophile USB hubs are just jumping on a cash cow bandwagon and the need has been created by the manufacturers not by necessity,
ALL USB signalling requires a return path for the single ended signals that are part of the the USB data transfer protocol... To connect it away from the actual signal wires is stupid, I would suggest having a look at how digital signals propagate, its is not the earth and to separate the signal and its return path will cause an increase in noise and be detrimental to signal integrity. In fact the locking on was probably down to you not having a return path, I don't think you understand this signal propagation stuff much
Regen devices what do they do... try a basic USB hub first because that is all they are or make your own, why add more complications...
http://ww1.microchip.com/downloads/en/DeviceDoc/evb2412.pdf
Or just buy the evaluation board...
ftp://ftp.smsc.com/pub/usb/evb2412user.pdf
As I have said I believe these audiophile USB hubs are just jumping on a cash cow bandwagon and the need has been created by the manufacturers not by necessity,
Hi ! thanks but I understand now that I forgot to mention that price is also an issue
I was clearly underestimating the task.
I think I will try to select a good one off the shelf.
I should start saving money for a 4th hand spectrum analyzer instead ... I fell in love with the object.
I am about to move now but one day I will have one.
Thanks again, gino
Stavanger looks nice, do a fair bit of work for a firm in Asker, but do the work back here at base, one day I may get out there.
This looks a decent supply, just come through on email, may be a bit of overkill....
Robust 650 W supply meets MIL-STD-810F and ANSI/ISA standards for demanding applications
Robust 650 W supply meets MIL-STD-810F and ANSI/ISA standards for demanding applications
This is what return path discontinuities do, these are just small PCB based ones so separating a digital signal from its return path with a signal on a cable is even worse...
www.edn.com/Pdf/ViewPdf?contentItemId=4439672
http://bethesignal.com/wp/wp-conten...240_Return_Path_Issues_-MGB_Interconnects.pdf
So if the signal goes to the CPU so should the return path.
www.edn.com/Pdf/ViewPdf?contentItemId=4439672
http://bethesignal.com/wp/wp-conten...240_Return_Path_Issues_-MGB_Interconnects.pdf
So if the signal goes to the CPU so should the return path.
It's easily available in that it is a "USB(2) cable". That's the spec, right there. It says everything any consumer should need to know. If you want to know more, then the requirements are spelled out in the USB spec, which is trivially easy to find. If teasing out and understanding the relevant bits is difficult, then that's probably an excellent sign that just knowing it is a USB2 cable should be enough.
If you are truly concerned that a given cable truly meets the spec, then buy a couple of cheap budget no-name cables from ebay, or one value priced name brand cable from a more established seller. In either case, you are likely to end up with an in-spec cable, and if you get two of the cheap ones, you'll probably end up with two in-spec cables.
If you are worried about OEM provided cables, I'm sure you are thinking about this the wrong way. It's all down to the dynamics of mass market goods and manufacturing. Low cost requires high volume and good quality control. Without good quality control, high volumes are difficult, because no one will want to buy or distribute your goods, and the cost of returns and customer service will destroy profit margins. So, a cheap cable is very likely to work perfectly, because if it doesn't it gets expensive to manufactures and middlemen quickly.
The bigger concern with a cheap cable isn't whether it will work perfectly, the bigger concern is for how long it will do so. I think that mechanical durability is, arguably, harder to get right, and easier to get away with not getting right.
The thing is, knowing the detailed specs of a cable aren't going to protect you against getting a bad one, because you are only likely to get a bad one if something unintentional goes wrong, or if someone intentionally cuts corners and lies.
The good news is that it is likely to be obvious when/if a USB cable has problems. Diagnosis is easy too, just try another one. If you get your cables from a few sources, and keep an extra on hand, you'll be set. If that's still too uncertain for you, I'd probably tell you to get a cable from Belkin if you were in the US. They are a little more expensive than they need to be, but...
For what it is worth, I've had very few bad USB cables, and I think all of them were due to wear and tear under somewhat demanding circumstances. Flakey USB sockets have been a bigger problem, generally due to broken solder joints, and then, only on really cheap equipment.
Last edited:
There is also this
https://www.amazon.com/Optical-Cables-Corning-Cable-Meter/dp/B00JOJRF6K
Would love it if someone could do THD and noise measurements with and without it
https://www.amazon.com/Optical-Cables-Corning-Cable-Meter/dp/B00JOJRF6K
Would love it if someone could do THD and noise measurements with and without it
Hi and thanks again for the valuable advice.
Unfortunately I do not have the right equipment
(and skills ) to carry out this very fundamental measurement. What I can do is to see the overall THD+Noise of the usb soundcard connected to the pc. In the end this should give me the overall picture.
I will compare different usb cable to see if they can add noise or distortion.
But I am now not so hot on the idea of making a usb cable.
I understand better the difficulties now.
I have some to try out. Should I place it at the pc or at the soundcard end ?
I have one of these isolators. But I would prefer to do without.
Even if an excellent usb to spdif converter I have seen use them with excellent results. There is a xmos usb receiver > Adum isolator > output circuit for spdif/AES. They say it sounds fantastic.
So I guess that if correctly used this chip is not a bottleneck.
Thanks for the advice. I am abandoning the idea. Too complex.
For sure I will do it without any other hesitations.
I understand now better that a lot of different issues are present.
I will select a decent certified cable and live with that.
Thanks a lot again, gino
Maybe you mean Aker ? there is a big company with this name.
If you happen to come send me an message and we can meet here
I think so ... I just need 12V/1A or so.
But now that I found a way to measure THD+Noise I am more relaxed.
I have been left in panic by all these subjective listening tests.
I much prefer to see a nice spectrum ... I love this test a lot.
I do not understand why people prefer the scope. I prefer the spectrum analyzer. Much easier to understand than a scope.
If you pass by let me know.
Thanks a lot again, gino
And why would they be different...
It converts the data to light only, so it can be transmitted further...
https://www.corning.com/microsites/coc/ocbc/Documents/CNT-075-AEN.pdf
Then they are not USB compliant...
Or the devices you are using either take to much current and should be powered externally or they take to much inrush current so they again are NOT designed to be powered by a standard USB interface, either way something is not to USB sepcs, or are you just pushing the cable thing...
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.