Do we need Zobel with chip amps?

[Peter Daniel]

I was never using Zobel network with chip amps, and never found the need for it. I've heard only of few cases when it was really required and some report that the amp actually sounds better without it.

CHG is for chassis ground connection. We put it there just in case somebody wanted to use different point for chassis ground hookup. Since recommended transformer is not CT but double secondaries, I don't know myself if it's better to connect chassis to that point or somwhere on rectifiers boards. I was always using negative output to connect with chassis (from both channel's binding posts to one point on chassis).

We use indeed separate diode bridges for each cap and we made return connections as close to the filter cap GND terminals as possible.

The suggested power connections seem to be indeed much better. It's always good when somebody comes with a fresh outlook. Any comments about creating loop with those connections and possible influence?

[coredump]

Peter,

Peter Daniel said:
I was never using Zobel network with chip amps, and never found the need for it. I've heard only of few cases when it was really required and some report that the amp actually sounds better without it.[...]

so I may be the first one to report an improvement by using a Zobel; I use a GC based on Brian's 3875 kit. Initially I let it run in for about a week (without a Zobel) and then began serious listening for about a week, and after that fitted the Zobel.
The improvement was quite noticable. This may have to do with my speaker cable (DNM/Reson), which is not shielded at all and quite long (low capacitance, high inductance, perhaps prone to catch up RF), but this is only an assumption.
Anyway, the treble got cleaner and smoother, and I decided to keep it.

Zobel

Best regards,

Oliver

[Jamh]

I found Zobel more useful in the inverted design. Speaker cables for this amp need to be fast and not too thick, IMO. Same with connectors, not too much metal. You're compensating for a problem with the Zobel Network that needs not being a problem.

[coredump]

Jamh,

Jamh wrote:
I found Zobel more useful in the inverted design. Speaker cables for this amp need to be fast and not too thick, IMO. Same with connectors, not too much metal. You're compensating for a problem with the Zobel Network that needs not being a problem.

if you look at the photo you will see that both your points don't apply for my implementation.
But I sure may have other problems.

Best regards,

Oliver

[Fred Dieckmann]

Originally posted by Peter Daniel
I was never using Zobel network with chip amps, and never found the need for it. I've heard only of few cases when it was really required and some report that the amp actually sounds better without it.

CHG is for chassis ground connection. We put it there just in case somebody wanted to use different point for chassis ground hookup. Since recommended transformer is not CT but double secondaries, I don't know myself if it's better to connect chassis to that point or somwhere on rectifiers boards. I was always using negative output to connect with chassis (from both channel's binding posts to one point on chassis).

We use indeed separate diode bridges for each cap and we made return connections as close to the filter cap GND terminals as possible.

The suggested power connections seem to be indeed much better. It's always good when somebody comes with a fresh outlook. Any comments about creating loop with those connections and possible influence?

There are no loop problems from the additional power traces.


Zobles:

I would think the guys from National know what they're doing. You cannot tell what speakers and cables that will be used with this. It is part of the high frequency compensation for the amp. Telling people that you don't need them and leaving them out of the design is irresponsible. Don't stuff the parts if you think you don't need them, but why sell amps or PCB boards that are not the original design, just on a personal whim without a clue for the engineering reason for their inclusion ....... oh I forgot, you're an experienced amplifier designer. I guess the inexperienced builders can tack the parts on top of the board if they are not in the PCB layout. :

[Fred Dieckmann]

Originally posted by Jamh
I found Zobel more useful in the inverted design. Speaker cables for this amp need to be fast and not too thick, IMO. Same with connectors, not too much metal. You're compensating for a problem with the Zobel Network that needs not being a problem.

No...... you're second guessing the guys at National and other competent amp designers know what a Zoble network does it put it in an amplifier for a reason.

What is a fast speaker cable? In terms of inductance, capacitance, and
dieletric constant of the insulation. What difference does being thick make to the speed?

[Peter Daniel]

I'm not an experienced designer, but so far I didn't experience any problems regarding that. The same for at least 100 people who already built the kit.

Just out of curiosity, I installed Zobel on one of my amps right now, and the difference in sound is quite substantial. I can't say it's for worse, but it's definitely different. It seems like highs got cleaned out and are smoother, yet I can also detect coloration from a cap and resistor. I would have to listen more to decide what's better. But definitely worth investigating.

But those size resistor and cap might not be easy to place on the current board layout. I don't see any problem with mounting them directly to binding posts.

BTW, AudioZone offers a series resistor/capacitor jumper to those who have problems with matching speaker cables to the amp. So far there was only one such problem and Zobel network didn't help. Changing speaker cables eliminated that problem though.

[BrianGT]

Originally posted by Fred Dieckmann

Zobles:

I would think the guys from National know what they're doing. You cannot tell what speakers and cables that will be used with this. It is part of the high frequency compensation for the amp. Telling people that you don't need them and leaving them out of the design is irresponsible. Don't stuff the parts if you think you don't need them, but why sell amps or PCB boards that are not the original design, just on a personal whim without a clue for the engineering reason for their inclusion ....... oh I forgot, you're an experienced amplifier designer. I guess the inexperienced builders can tack the parts on top of the board if they are not in the PCB layout. :

I was under the assumption that the zobel could just attached to the binding post if there are problems with stability. This is what I have seen done on other designs, including the first amplifier I built a couple of years ago. I am clearly not an experienced designer of any sort, and am willing to consider any advice on this matter.

Also, what do you think about not having a feedback capacitor. On the National app note for the LM3875, it was marked optional, but it wasn't marked optional for the LM4780. I tried the LM3875 with, and without this capacitor, and didn't notice any difference on the scope. It is necessary to have this capacitor, with it providing unity gain at DC, and and also create a highpass filter with the inverting input resistor?

--
Brian

[Fred Dieckmann]

"I don't see any problem with mounting them to binding posts."

They are on the wrong side of the 0.1 ohm resistors and you are adding a lot of lead inductance in series with something that needs to be resisitive at high frequencies where the capacitor has become a very low impedance.. I would use a 1 watt metal film and a good 160V to 250V metalized polypropylene. Those don't take up much PCB and you could stand the parts on end and soldered the airbourne leads together if space is that tight.

[Peter Daniel]

I forgot about output resistors with that amp. As long as it does not interfere with current layout, I don't mind those. Probably Panasinic resistors from Digi Key would be fine here and some of those brown Panasonic caps from DK as well.