Confused?
The MAX931 is a 12µs propagation delay comparator plus reference. Probably Bernard meant the MAX913.
Did you use a 50 Ohm coax cable terminated at the scope with a 50 Ohm resistor as in the instructions here:?
http://www.diyaudio.com/forums/showthread.php?postid=199928#post199928
Without that trick Guido's squarewave wouldn't look so nice either. And don't use a probe set to 1: 10. This will make matters worse. Besides that I have found no correlation between shape of the squarewave and sound trying many different comparators, Lcaudio's clock, Netaudio's clock and Guido's clock.......
We now have 40 posts dealing with a probably non-existent power supply problem. You did not like the sound of the Kwak-clock? Or did I miss that?
The difference in sqaurewave between your and Andrea's is indeed attributable to the different comparator. Sonically the sound between the AD8561 and the MAX913 is vanishing small.
Hi Bricolo,Bricolo said:
The MAX931 is a 12µs propagation delay comparator plus reference. Probably Bernard meant the MAX913.
Did you use a 50 Ohm coax cable terminated at the scope with a 50 Ohm resistor as in the instructions here:?
http://www.diyaudio.com/forums/showthread.php?postid=199928#post199928
Without that trick Guido's squarewave wouldn't look so nice either. And don't use a probe set to 1: 10. This will make matters worse. Besides that I have found no correlation between shape of the squarewave and sound trying many different comparators, Lcaudio's clock, Netaudio's clock and Guido's clock.......
We now have 40 posts dealing with a probably non-existent power supply problem. You did not like the sound of the Kwak-clock? Or did I miss that?
The difference in sqaurewave between your and Andrea's is indeed attributable to the different comparator. Sonically the sound between the AD8561 and the MAX913 is vanishing small.
Re: Confused?
Hi Elso,
Yes, I used a probe like this to measure the output signal.
Now I modified it to a 1kR probe like this one
http://www.sigcon.com/Pubs/straight/probes.htm
The signal looks the same, but it's an easier load for the comparator.
Why do you say this PS problem may be non existent
The KC isn't in the player now, It's 100% finished now but only since a few days (I recieved the 1uF wimas from Andrea this week)
I didn't test the MAX931, but it looks like on the paper. It has some kind of class A output stage, that prevents pulses on the PS rails when the output state is changing. The drawback is the high power consumption (40mA)
Please, Elso, can you explain me why my ferrite beads aren't ok between the regulator and the comparator?
Elso Kwak said:
Hi Bricolo,
The MAX931 is a 12µs propagation delay comparator plus reference. Probably Bernard meant the MAX913.
Did you use a 50 Ohm coax cable terminated at the scope with a 50 Ohm resistor as in the instructions here:?
http://www.diyaudio.com/forums/showthread.php?postid=199928#post199928
Without that trick Guido's squarewave wouldn't look so nice either. And don't use a probe set to 1: 10. This will make matters worse. Besides that I have found no correlation between shape of the squarewave and sound trying many different comparators, Lcaudio's clock, Netaudio's clock and Guido's clock.......
We now have 40 posts dealing with a probably non-existent power supply problem. You did not like the sound of the Kwak-clock? Or did I miss that?
The difference in sqaurewave between your and Andrea's is indeed attributable to the different comparator. Sonically the sound between the AD8561 and the MAX913 is vanishing small.
Hi Elso,
Yes, I used a probe like this to measure the output signal.
Now I modified it to a 1kR probe like this one
http://www.sigcon.com/Pubs/straight/probes.htm
The signal looks the same, but it's an easier load for the comparator.
Why do you say this PS problem may be non existent
The KC isn't in the player now, It's 100% finished now but only since a few days (I recieved the 1uF wimas from Andrea this week)
I didn't test the MAX931, but it looks like on the paper. It has some kind of class A output stage, that prevents pulses on the PS rails when the output state is changing. The drawback is the high power consumption (40mA)
Please, Elso, can you explain me why my ferrite beads aren't ok between the regulator and the comparator?
Re: Re: Confused?
Hi Bricolo,
The MAX931 is not suitable, much too slow. You need a 7ns delay comparator!
Please download the datasheet of the beads from Farnell recommended by me and compare the impedance with the beads you are using. www.farnell.com
As I said in that 18 pages long confusing thread about decoupling of digital IC's, the high impedance of the bead will raise the impedance of the powersupply. Anyway I found that "better" beads were not better in my clock. Version 8 of the Kwak-Clock has no beads between supply and FET and comparator. You can try a 47µF Oscon directly at the drain of the FET. I am not certain about the comparator not giving spikes back to the powersupply but the FET does for sure!
You are studying electronics? May I suggest a career-switch?
Edit:
The resistive input probe in your reference is interesting but you don't need the 1k resistor. The AD8561 has enough drive.
Bricolo said:
Hi Bricolo,
The MAX931 is not suitable, much too slow. You need a 7ns delay comparator!
Please download the datasheet of the beads from Farnell recommended by me and compare the impedance with the beads you are using. www.farnell.com
As I said in that 18 pages long confusing thread about decoupling of digital IC's, the high impedance of the bead will raise the impedance of the powersupply. Anyway I found that "better" beads were not better in my clock. Version 8 of the Kwak-Clock has no beads between supply and FET and comparator. You can try a 47µF Oscon directly at the drain of the FET. I am not certain about the comparator not giving spikes back to the powersupply but the FET does for sure!
You are studying electronics? May I suggest a career-switch?
Edit:
The resistive input probe in your reference is interesting but you don't need the 1k resistor. The AD8561 has enough drive.
Re: Confused?
Hello Elso,
Good to see you pointing towards the importance of 50 ohm measurement probes, I wouldn't call it a trick, but call it understanding of measurement principles.
While I more or less agree with your statement about square wave quality and sonic performance of clocks, too much overshoot leads to triggering protection diodes, which dramatically increases jitter, so the waveshape should look "decent".
And yes, it is very likely not a power supply problem, but a layout / EMC / RF signal integrity issue.
cheers
Elso Kwak said:
Hi Bricolo,
The MAX931 is a 12µs propagation delay comparator plus reference. Probably Bernard meant the MAX913.
Did you use a 50 Ohm coax cable terminated at the scope with a 50 Ohm resistor as in the instructions here:?
http://www.diyaudio.com/forums/showthread.php?postid=199928#post199928
Without that trick Guido's squarewave wouldn't look so nice either. And don't use a probe set to 1: 10. This will make matters worse. Besides that I have found no correlation between shape of the squarewave and sound trying many different comparators, Lcaudio's clock, Netaudio's clock and Guido's clock.......
We now have 40 posts dealing with a probably non-existent power supply problem. You did not like the sound of the Kwak-clock? Or did I miss that?
The difference in sqaurewave between your and Andrea's is indeed attributable to the different comparator. Sonically the sound between the AD8561 and the MAX913 is vanishing small.
Hello Elso,
Good to see you pointing towards the importance of 50 ohm measurement probes, I wouldn't call it a trick, but call it understanding of measurement principles.
While I more or less agree with your statement about square wave quality and sonic performance of clocks, too much overshoot leads to triggering protection diodes, which dramatically increases jitter, so the waveshape should look "decent".
And yes, it is very likely not a power supply problem, but a layout / EMC / RF signal integrity issue.
cheers
Re: Re: Re: Confused?
Hi Elso,
I didn't look at the Maxim comparator speed, but the output stage tppology got my attention.
The beads you recommended have:
25R@10MHz
65R@100MHz
Mine are
80R@10MHz
130R@100MHz
Maybe I didn't understant the role of a bead. I thought it was to add a high impedance to HF "garbage". Why has it to be "not too big" ?
I looked at the PS on the FET's drain, and it's much cleaner than the comparator's!
"You are studying electronics? May I suggest a career-switch?
"
No you may not. Except if you prefer EEs that answear "OK I'll do that ,DRs Kwak. Thank you for your help", without trying to understand anything of what they do...
The probe won't be used only on the KC, that's the use for the 1K resistor.
Alex
Elso Kwak said:
Hi Bricolo,
The MAX931 is not suitable, much too slow. You need a 7ns delay comparator!
Please download the datasheet of the beads from Farnell recommended by me and compare the impedance with the beads you are using. www.farnell.com
As I said in that 18 pages long confusing thread about decoupling of digital IC's, the high impedance of the bead will raise the impedance of the powersupply. Anyway I found that "better" beads were not better in my clock. Version 8 of the Kwak-Clock has no beads between supply and FET and comparator. You can try a 47µF Oscon directly at the drain of the FET. I am not certain about the comparator not giving spikes back to the powersupply but the FET does for sure!
You are studying electronics? May I suggest a career-switch?
Edit:
The resistive input probe in your reference is interesting but you don't need the 1k resistor. The AD8561 has enough drive.
Hi Elso,
I didn't look at the Maxim comparator speed, but the output stage tppology got my attention.
The beads you recommended have:
25R@10MHz
65R@100MHz
Mine are
80R@10MHz
130R@100MHz
Maybe I didn't understant the role of a bead. I thought it was to add a high impedance to HF "garbage". Why has it to be "not too big" ?
I looked at the PS on the FET's drain, and it's much cleaner than the comparator's!
"You are studying electronics? May I suggest a career-switch?
"No you may not. Except if you prefer EEs that answear "OK I'll do that ,DRs Kwak. Thank you for your help", without trying to understand anything of what they do...
The probe won't be used only on the KC, that's the use for the 1K resistor.
Alex
Re: Re: Re: Re: Confused?
The 1k resistor at the probe should be 50 Ohm ideally!. Only the amplitude of the signal is halved. Some scope already have a 50 Ohm input impedance. In that case you can omit the 50 Ohm resistor at the terminating end.
The idea behind the beads was preventing the spike coming from the FET entering the comparator and vice versa. So I grabbed the small beads in my toolbox and used these.
Later I wanted to do a "better job" and ordered some higher impedance beads. Much to my own surprise the sound was slightly less good. So I shortcircuited these beads.
Keep studying electronics!
Hi Alex,Bricolo said:
Hi Elso,
I didn't look at the Maxim comparator speed, but the output stage tppology got my attention.
The beads you recommended have:
25R@10MHz
65R@100MHz
Mine are
80R@10MHz
130R@100MHz
Maybe I didn't understant the role of a bead. I thought it was to add a high impedance to HF "garbage". Why has it to be "not too big" ?
I looked at the PS on the FET's drain, and it's much cleaner than the comparator's!
"You are studying electronics? May I suggest a career-switch?
No you may not. Except if you prefer EEs that answear "OK I'll do that ,DRs Kwak. Thank you for your help", without trying to understand anything of what they do...
The probe won't be used only on the KC, that's the use for the 1K resistor.
Alex
The 1k resistor at the probe should be 50 Ohm ideally!. Only the amplitude of the signal is halved. Some scope already have a 50 Ohm input impedance. In that case you can omit the 50 Ohm resistor at the terminating end.
The idea behind the beads was preventing the spike coming from the FET entering the comparator and vice versa. So I grabbed the small beads in my toolbox and used these.
Later I wanted to do a "better job" and ordered some higher impedance beads. Much to my own surprise the sound was slightly less good. So I shortcircuited these beads.
Keep studying electronics!
Re: Re: Confused?
This is possible.
When I bypass the bead with a small wire, the spikes are still here. I don't understand this.
when the bead is used, and that I measure the voltage across it, I see the spikes. So it's "caused by the bead"
When I bypass it, of course I don't see any voltage across the "'bypass wire", but between the comparatof supply pin and groud, the spikes are present
Guido Tent said:
This is possible.
When I bypass the bead with a small wire, the spikes are still here. I don't understand this.
when the bead is used, and that I measure the voltage across it, I see the spikes. So it's "caused by the bead"
When I bypass it, of course I don't see any voltage across the "'bypass wire", but between the comparatof supply pin and groud, the spikes are present
Re: Re: Re: Confused?
Hi
The voltage across the bead ofcourse is " spiky", but not at all relevant to the output signal.
What you should conisder is
- the voltage at the comperator supplies
- how sensitive the output is for that voltage
If the supply voltage is spiky, despite supply decoupling, I suspect you have a groundbounce problem, probably cuased by some RF currents running thruogh the grounplane.
succes
Bricolo said:
This is possible.
When I bypass the bead with a small wire, the spikes are still here. I don't understand this.
when the bead is used, and that I measure the voltage across it, I see the spikes. So it's "caused by the bead"
When I bypass it, of course I don't see any voltage across the "'bypass wire", but between the comparatof supply pin and groud, the spikes are present
Hi
The voltage across the bead ofcourse is " spiky", but not at all relevant to the output signal.
What you should conisder is
- the voltage at the comperator supplies
- how sensitive the output is for that voltage
If the supply voltage is spiky, despite supply decoupling, I suspect you have a groundbounce problem, probably cuased by some RF currents running thruogh the grounplane.
succes
Re: Re: Re: Re: Re: Re: Confused?
Hi,
I have no 470R here, but that's how it looks with the comparator connected directly to a coax with a 50R at the scope (no resistor between coax and comparator)
http://www.diyaudio.com/forums/showthread.php?postid=378023#post378023
Guido Tent said:
Hi,
I have no 470R here, but that's how it looks with the comparator connected directly to a coax with a 50R at the scope (no resistor between coax and comparator)
http://www.diyaudio.com/forums/showthread.php?postid=378023#post378023
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.