Sorry about the odd format,but it was the only way that I could get the diagram off the data sheet.I've cursed that damn Acrobat Reader I don't know how many times! You have to unzip the folder,then open it using the clipboard viewer (in Windows,don't know if it will work with a Mac).I will be going to the parts place later this week,so I will get a couple of 1R resistors and try them.
I took Carlos's advice and tried a Zobel on the output of my opamp buffered IGC and it does make a slight difference in tightening up the bass. Actually bass also seemed to go a tad lower.
Of course, CFM is never satisfied
and is making me try another value resistor. I think that he is on commision from the resistor manufacturers!
Of course, CFM is never satisfied
and is making me try another value resistor. I think that he is on commision from the resistor manufacturers! Nuuk said:
That's it, that's really it.
Your amp is driving better your speakers than before.
The amp needs a zobel.
On some cables/speakers it's more noticeable, but it doesn't muck up the sound, at least in most situations.
So... I will always use it.
Nuuk said:Of course, CFM is never satisfied
Well... because I recommend 2.7R.
Because that gave me very good results, on my tests.
Curiously, after all the fuss, it's exactly what National recommends on the LM4780 datasheet.
One of the best datasheets BTW, very complete.
Sheldon said:
That's what a Zobel network is for, but the high frequencies can enter the circuit from many places, not just at the input. EM radiation from many sources (spikes on the mains, phones, CB radio) can be picked up either through the circuitry itself or the speaker cable to trigger oscillations.
The problems that a Zobel network addresses can be very subtle. Oscillations may only start under certain conditions, perhaps when someone nearby pulls a big current from the mains (induction motor perhaps), or when a passing taxi uses its radio. Often it only occurs with longer speaker cables, or only when the are layed in certain locations and orientations (there's a particularly bad place in my room where my speaker cables have to run alongside some RF coaxial cable which has caused headaches in the past). I've made an amp that's only unstable in the summer too!
Once the oscillation starts, if you're lucky it will be immediately obvious and the problem can be addressed. In the worst case the oscillation is of a very high frequency and low amplitude, not giving any immediate signs that anything is wrong apart from a slightly fuzzy sound.
Thus I always measure the output of my amps with an oscilloscope when connected with long cables, or maybe a capacitor. Sometimes the Zobel proves unnecessary, but I only go without if I'm sure the setup isn't going to change e.g. changing cables or speakers.
I've learnt the hard way (several times
) trying to keep component counts down to a minimum, skimping on a couple of pence.Placement is important. Putting them across the amp output terminals isn't as good as putting them on the PCB itself, with proper attention to layout although it can still help.
OK, I'm delurking...
No GC in the works here, but thoughts of a pair have been residing in my head for a while. Once a suitable enclosure and funding present themselves, I'll try a pair.
How would I go about estimating Zobel values for the chips of my choice? Is compensation more dependent on the IC itself, speaker/cable impedance, amplifier circuit, PCB layout, or what?
Links to where I should have looked before asking?
...Anyone for tennis?
Tim
No GC in the works here, but thoughts of a pair have been residing in my head for a while. Once a suitable enclosure and funding present themselves, I'll try a pair.
How would I go about estimating Zobel values for the chips of my choice? Is compensation more dependent on the IC itself, speaker/cable impedance, amplifier circuit, PCB layout, or what?
Links to where I should have looked before asking?
...Anyone for tennis?
Tim
I may be wrong, but I don't think the exact values are too critical.
So, if you use a 100n cap you could try resistor values of say 2R7 (which Carlos tells me is suggested in the NS application data) to say 6R8. I am currently using 4R7 on my 3875's and will try some other values when I can get hold of them.
2.7R + 0.1uf is on the LM4780 datasheet and works very well for me.
1R and 0.22uf is also recommended for other National chips. I tested and didn't like it. It affected the treble.
If you don't have 2.7R resistors, use something around this value, even if you have to parallel two resistors.
But the cap should be 0.1uf.
1R and 0.22uf is also recommended for other National chips. I tested and didn't like it. It affected the treble.
If you don't have 2.7R resistors, use something around this value, even if you have to parallel two resistors.
But the cap should be 0.1uf.
Rod Elliot (sound.au.com) did a piece on cable termination and zobel's. I can follow all the details, but it seems to suggest that the appropriate values for a zobel depend more on the speakers/load than on the amplifier. This was in connection to a discrete amplifier, but the internal topology of many chip amps isn't that different.
i've tried the zobel for a couple of days, desoldered them again and...
The amp seems to have a little more control in lower frequencies with the zobel (but only slightly). With my b&w dm220i's a basskick in eva cassidy's 'bridge over troubled water' on live@blues alley is still a bit boomy.
Bass is also somewhat deeper than without the zobel. haven't tried the output resistors yet. our local electronics shop doesn;t have non inductive 3w resistors.
It also seems that the amp is less prone to 'click and pop' due to power line pollution. It could also be that the neighbours are on vacation ...
i'm using a 2R7 0,5w el cheapo metal film resistos and a 0,1uf multilayer cap soldered between the output pins of brian's boards.
All in all, a slight change for the better! I'll keep 'em.
The amp seems to have a little more control in lower frequencies with the zobel (but only slightly). With my b&w dm220i's a basskick in eva cassidy's 'bridge over troubled water' on live@blues alley is still a bit boomy.
Bass is also somewhat deeper than without the zobel. haven't tried the output resistors yet. our local electronics shop doesn;t have non inductive 3w resistors.
It also seems that the amp is less prone to 'click and pop' due to power line pollution. It could also be that the neighbours are on vacation
...i'm using a 2R7 0,5w el cheapo metal film resistos and a 0,1uf multilayer cap soldered between the output pins of brian's boards.
All in all, a slight change for the better! I'll keep 'em.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.