The clock signal must have two wires in very close proximity, either twisted pair or co-ax, it will have a return somewhere otherwise it is not going to work, there has to be a loop.
One of the worse things for clock signal integrity is not having an intimate return path for the clock signal.
Just to double check, you only have one wire for the clock? If so the return current will have a torturous path to follow, this will negate any benefits of adding an external clock.
One of the worse things for clock signal integrity is not having an intimate return path for the clock signal.
Just to double check, you only have one wire for the clock? If so the return current will have a torturous path to follow, this will negate any benefits of adding an external clock.
The clock has to have an intimate return otherwise its integrity (jitter etc) will suffer, this is the basics of signal integrity and signal flow, and especially with clocks. if you are going to mod a clock for better performance you need to do it properly, otherwise it is a waste of time, as you will be adding more problems than you are solving.
You need two wires for a signal to flow, so the return will be finding a path back somehow, if this path is not close to the clock signal you add signal integrity problems, and it is effect is the same as having cuts and slits in groundplanes with digital signals.
http://www.x2y.com/filters/TechDay09... _JohnWu.pdf
The 10 Best Ways to Maximize Emission from Your Product
Two basic documents that give some indication of how signals flow, from an earlier thread with a clock modification.
http://www.diyaudio.com/forums/pc-based/197116-xonar-st-stx-mods-50.html#post3836551
You need two wires for a signal to flow, so the return will be finding a path back somehow, if this path is not close to the clock signal you add signal integrity problems, and it is effect is the same as having cuts and slits in groundplanes with digital signals.
http://www.x2y.com/filters/TechDay09... _JohnWu.pdf
The 10 Best Ways to Maximize Emission from Your Product
Two basic documents that give some indication of how signals flow, from an earlier thread with a clock modification.
http://www.diyaudio.com/forums/pc-based/197116-xonar-st-stx-mods-50.html#post3836551
Your clock signal will be compromised, if you are going to mod a clock you need to do it correctly or not bother, you will be adding jitter not solving it... The return will be finding a path back through the mains. Why not do the job properly, if you don't want to take my word for it have a look around at the recommendations for clock modifications. You are not improving things you are adding more problems to the clock signal.
How do you know it is working better, have you checked the jitter levels with a network analyser.
Basic electronics, a signal will only travel if it has a return path....And as a clock is a critical signal then it must be routed/or wired properly otherwise you are wasting your time.
How do you know it is working better, have you checked the jitter levels with a network analyser.
Basic electronics, a signal will only travel if it has a return path....And as a clock is a critical signal then it must be routed/or wired properly otherwise you are wasting your time.
As you wont follow my good advice have a look at the recommendations from Tent Labs who do some of the best add on clocks:
http://www.tentlabs.com/Products/cdupgrade/xo2xo3/assets/XO23mounting.pdf
http://www.tentlabs.com/Products/cdupgrade/xo2xo3/assets/XO23mounting.pdf
The return can't find a way back through the mains as the mains is isolated by the transformers.
I'm not in any way dismissing your advice however my player is working and sounds better therefore the clock is working as expected.
Now at some point soon I will experiment further by grounding the clock to the player to see what difference it makes and I will feedback here with results.
Best wishes
Tom
I'm not in any way dismissing your advice however my player is working and sounds better therefore the clock is working as expected.
Now at some point soon I will experiment further by grounding the clock to the player to see what difference it makes and I will feedback here with results.
Best wishes
Tom
Sorry if I sound rabid over this, I am, one of the biggest problems on PCB layout is clocks and clock distribution, having done some quite funky ADC and DAC designs (50 ADC's on a board) as well as normal day to day DDR memory, gigabit Ethernet etc. you realise that clock integrity is paramount to the rest of the design working. Think of the clock as a systems heartbeat, the hot is the artery and the return is the vein, any problems with either of them and your heartbeat is going to suffer.
How sensitive clocks can be was illustrated by only a 12MHz clock on some codecs with only about 7mm of trace, the clock module had a very strong drive characteristic (so it could drive longer lines) and was causing problems on an un-terminated line. We had to simulate the PCB layout in our SIV package (signal integrity verification software) to work out the optimum termination, in this case an 82ohn series resistor to reduce the drive strength (slow the rising edge) to remove the overshoot and reflections it was causing that were adding amongst other problems an increase in jitter, which for this application was critical.
Regards
marc e
How sensitive clocks can be was illustrated by only a 12MHz clock on some codecs with only about 7mm of trace, the clock module had a very strong drive characteristic (so it could drive longer lines) and was causing problems on an un-terminated line. We had to simulate the PCB layout in our SIV package (signal integrity verification software) to work out the optimum termination, in this case an 82ohn series resistor to reduce the drive strength (slow the rising edge) to remove the overshoot and reflections it was causing that were adding amongst other problems an increase in jitter, which for this application was critical.
Regards
marc e
Thanks for being so specific on this. For some reason I had it in my head that the coax shielding or extra grounded twisted wire on the clock signal wire was only to stop interference and only needed grounding at 1 end.
On my cd63 I have 3 clock feeds (dac, decoder and a divided 1 for the servo) , all using solid core cat5 twisted pairs, but only grounded at the clock end and left floating at the pcb end. Sounds like it wouldn't be a bad plan to ground those at the pcb right next to the clock wires
Right added ground from nearest point on clock to nearest ground point on player (where original xtal caps terminated to) and can't really tell any difference.
Listening to "Upstairs at Erics" by Yazoo pre and post ground wire seems to sound the same although I have backed bass off a little to neutral position on amp post ground wire but i couldn't say it's because of the link.
Regards
Tom
Listening to "Upstairs at Erics" by Yazoo pre and post ground wire seems to sound the same although I have backed bass off a little to neutral position on amp post ground wire but i couldn't say it's because of the link.
Regards
Tom
£40 worth of components (if that) rather a lot of mark-up, its an oscillator with a PSU.
As to ground plane, I would expect any clock circuit to have a ground plane as standard.
A basic design with a dedicated PSU is going to be as good, from the design it looks like the output is straight from the oscillator, no output buffer to drive the wire that will connect the clock to the PCB. having worked on various projects where clocks are far more of a concern than audio (though the figures some want for audio, you would wonder) personally I think its a rip off, nothing special and as previously discussed, with the figures quoted you are getting into serious signal integrity territory you achieve those figures.
You could get a design, have a multilayer board made and populated for less than £400+...
A incorrectly implemented clock mod is likely to add jitter, this will add distortion, making the sound seem fuller, though without proper measurement or simulation (pref both) of the overall system and interconnects, layout, proximity to ground plane etc. you can only guess.
As to ground plane, I would expect any clock circuit to have a ground plane as standard.
A basic design with a dedicated PSU is going to be as good, from the design it looks like the output is straight from the oscillator, no output buffer to drive the wire that will connect the clock to the PCB. having worked on various projects where clocks are far more of a concern than audio (though the figures some want for audio, you would wonder) personally I think its a rip off, nothing special and as previously discussed, with the figures quoted you are getting into serious signal integrity territory you achieve those figures.
You could get a design, have a multilayer board made and populated for less than £400+...
A incorrectly implemented clock mod is likely to add jitter, this will add distortion, making the sound seem fuller, though without proper measurement or simulation (pref both) of the overall system and interconnects, layout, proximity to ground plane etc. you can only guess.
I'll leave that to someone with more idea of what is best for audio, there was a thread a bit ago that discussed the various factors involved, such as cycle to cycle jitter etc. I just do layout and get the signals from a to b with the minimum of added problems due to layout, cabling etc.
£40 worth of components? From where? Let me know and I'll change my own component suppliers.
5 or 6 years worth of continual development for free??
You guys make me laugh!
Clearly a £400 clock is not really in the DIY domain. Fidelity audio is a business with employees, premises, business rates, accountancy fee's etc etc!
Why don't you take a month of work without wages, design some electronics then sell for just the price of components. The point I'm making is don't be so short sighted! There is more to it than your hobby costs. The kit alone to measure and test the kit probabaly costs more than your whole systems did new!!
You have a choice, but let's not be too hasty suggesting that the price is unfair!!!
5 or 6 years worth of continual development for free??
You guys make me laugh!
Clearly a £400 clock is not really in the DIY domain. Fidelity audio is a business with employees, premises, business rates, accountancy fee's etc etc!
Why don't you take a month of work without wages, design some electronics then sell for just the price of components. The point I'm making is don't be so short sighted! There is more to it than your hobby costs. The kit alone to measure and test the kit probabaly costs more than your whole systems did new!!
You have a choice, but let's not be too hasty suggesting that the price is unfair!!!