Re: Zobel Components
Good point- how about 1/2 watt minimum, 1 watt preferred - and choice of cap up to the builder? Maximum PCB footprint (depending on how much space is available) will accomodate everyone?
1audio said:
Good point- how about 1/2 watt minimum, 1 watt preferred - and choice of cap up to the builder? Maximum PCB footprint (depending on how much space is available) will accomodate everyone?
I wonder how those Panasonic ones are like, compared to Resista (By Roe, Germany) resistors...🙄
Re: Re: Can we put it to a vote?
BTW: Fred weren't you totally uninterested in everything labeled LM38xx? I suspect that Mr. D already has a Gainclone
I think that a Zobel network and output filter is a typical "can't hurt to add and you can always omit those"".BrianGT said:
BTW: Fred weren't you totally uninterested in everything labeled LM38xx? I suspect that Mr. D already has a Gainclone
Re: Zobel Components
The resistor should have good HF properties and not get too hot under the conditions you deside. I think that max 2 watts is sufficient unless you aren't going to test the amp with quarewave.1audio said:
Note also that the cap must be able to take the AC voltage (the music signal), more important for powerful amps.Peter Daniel said:
Do any of you use a calculator.........
The magnitude of the impedance of 0.1uf is around 80 ohms at 20 KHz.
The resistor in the Zobel network sees a small portion of the voltage at this frequency (voltage divider anyone?) and 1/2 that for every octave decrease ie. 1/4 of that at 5 kHz. Before someone runs the numbers and tells me that the resistors power dissipation at 20 kHz full scale continuos output, exceeds the resistors power rating, let me know how your tweeters hold out during this test. The power level at very high frequencies is much smaller than midrange levels ( and very much less than low frequency power levels) for music. If you are absolutely paranoid put 2 (or more) high quality low inductance resistors in parallel for greater power handling and even less parasitic inductance. No wire wound resistors by the way.
If your amp is oscillating long and hard enough to burn up a 1/4 watt resistor in a correctly built Zobel, you have other problems with the amp. The whole point of the Zobel is to keep the amp from oscillating to begin with.
I just can't believe the confusion over a circuit with one resistor and one capacitor in it. I would really suggest that people without even this level of basic electronics FOLLOW THE DIRECTIONS IN THE DATA SHEET EXPLICITLY. If you don't how something works to begin with improving the design is very unlikely. There is a good reason that they chose this resistor to be some small like a quarter watt. When all else fails, read the directions. I do think I will keep up with the Chip Amps threads for the sheer entertainment though.
"I suspect that Mr. D already has a Gainclone" Had, I stripped it for parts.........
The magnitude of the impedance of 0.1uf is around 80 ohms at 20 KHz.
The resistor in the Zobel network sees a small portion of the voltage at this frequency (voltage divider anyone?) and 1/2 that for every octave decrease ie. 1/4 of that at 5 kHz. Before someone runs the numbers and tells me that the resistors power dissipation at 20 kHz full scale continuos output, exceeds the resistors power rating, let me know how your tweeters hold out during this test. The power level at very high frequencies is much smaller than midrange levels ( and very much less than low frequency power levels) for music. If you are absolutely paranoid put 2 (or more) high quality low inductance resistors in parallel for greater power handling and even less parasitic inductance. No wire wound resistors by the way.
If your amp is oscillating long and hard enough to burn up a 1/4 watt resistor in a correctly built Zobel, you have other problems with the amp. The whole point of the Zobel is to keep the amp from oscillating to begin with.
I just can't believe the confusion over a circuit with one resistor and one capacitor in it. I would really suggest that people without even this level of basic electronics FOLLOW THE DIRECTIONS IN THE DATA SHEET EXPLICITLY. If you don't how something works to begin with improving the design is very unlikely. There is a good reason that they chose this resistor to be some small like a quarter watt. When all else fails, read the directions. I do think I will keep up with the Chip Amps threads for the sheer entertainment though.
"I suspect that Mr. D already has a Gainclone" Had, I stripped it for parts.........
Re: Do any of you use a calculator.........
I love to discuss those simple circuits with you, Fred. My recomendation was based on another circuit that I've built before (which had Zobel). It was A75 from Mr. Pass, and incidentally he recommends 1W for Zobel resistor. But this was way before chip amps craze. I wonder why he would use such big wattage resistor in that project? But what do I know? I'm not an amp designer.
In A75 article, it says:
"At the output node we see an RC load network of R31 and C8, which
helps stabilize the amplifier when it looks out at the real world. It damps
out some of the strange high-frequency resonances which develop in
driving real loads and cables. This network has been employed on every
(working) solid state amp ever made."
BTW, I didn't notice Zobel in any of Aleph amps, why is it?
Fred Dieckmann said:
I love to discuss those simple circuits with you, Fred. My recomendation was based on another circuit that I've built before (which had Zobel). It was A75 from Mr. Pass, and incidentally he recommends 1W for Zobel resistor. But this was way before chip amps craze. I wonder why he would use such big wattage resistor in that project? But what do I know? I'm not an amp designer.
In A75 article, it says:
"At the output node we see an RC load network of R31 and C8, which
helps stabilize the amplifier when it looks out at the real world. It damps
out some of the strange high-frequency resonances which develop in
driving real loads and cables. This network has been employed on every
(working) solid state amp ever made."
BTW, I didn't notice Zobel in any of Aleph amps, why is it?
Re: Do any of you use a calculator.........
When GC became popular on this forum, Mr. D e-mailed me saying that he built one and he puts it (the GC) on the very short list of the best sounding amps. He must have built something better with those parts since then. But don't ask him what, chances are he won't tell you😉
Fred Dieckmann said:
When GC became popular on this forum, Mr. D e-mailed me saying that he built one and he puts it (the GC) on the very short list of the best sounding amps. He must have built something better with those parts since then. But don't ask him what, chances are he won't tell you😉
Going back down this drill as follows:
1) power Bandwidth of an LM 3886 100KHz
2) Rated power 68W
3) Xc of .1 uF at 100 KHz approx 16 Ohms
4) Z equ. at 100 KHz w/ 2.7 Ohms = 18.7 Ohms
5) Pd of resistor at 100 KHz = 7.7W (to be safe I ran a spice analysis)
6) Pd of resistor at 20 KHz = .32W
This is why I proposed 5W. In practice with a 20KHz bandwidth (CD) and audio program you can see why National recommends .25W but that is pretty marginal I think.
-Demian
1) power Bandwidth of an LM 3886 100KHz
2) Rated power 68W
3) Xc of .1 uF at 100 KHz approx 16 Ohms
4) Z equ. at 100 KHz w/ 2.7 Ohms = 18.7 Ohms
5) Pd of resistor at 100 KHz = 7.7W (to be safe I ran a spice analysis)
6) Pd of resistor at 20 KHz = .32W
This is why I proposed 5W. In practice with a 20KHz bandwidth (CD) and audio program you can see why National recommends .25W but that is pretty marginal I think.
-Demian
Regarding the watt rating of zobel resistors: Here are two reasons why a 5W rating may make sense.
1- If for some reason you intend to test the amp with HF signals (50kHz?) a lesser rating may be too little.
2- In the event you encounter oscillation pro0blems the zobel can get very hot - possibly too hot for a 1-2W rating.
I normal operation the zobel resistor may well be OK at 1W or 2W, but particularly in case 2, above, it gets really annoying to run through a bunch of smoked resistors while trying to track down and fix an oscillation.
1- If for some reason you intend to test the amp with HF signals (50kHz?) a lesser rating may be too little.
2- In the event you encounter oscillation pro0blems the zobel can get very hot - possibly too hot for a 1-2W rating.
I normal operation the zobel resistor may well be OK at 1W or 2W, but particularly in case 2, above, it gets really annoying to run through a bunch of smoked resistors while trying to track down and fix an oscillation.
Re: Re: Do any of you use a calculator.........
Best sounding amps...life is full of surprises....coming from Mr. DPeter Daniel said:
This was probably one of the silliest threads I've read on this forum. The two things I learned from going through and reading the whole thing was:
1. Fred really seems to attack very specific things.
2. Peter can get defensive.
No one seemed to be down with attacking the lack of input coupling caps? Any reason? Fred? I might learn something!
Naz
1. Fred really seems to attack very specific things.
2. Peter can get defensive.
No one seemed to be down with attacking the lack of input coupling caps? Any reason? Fred? I might learn something!
Naz
naz said:
And the rest of us watching Gladiators with wine and fruits and toss flowers into the ring every now and then... ohs and ahs...
sorry that I could not refuse the opportunity being off topic.
Chris😀
A good recommendation!sam9 said:
I don't think this thread is silly at all because many people don't know why we must have output filters. The serve actually a purpose but sometimes you can also omit them. It's only to know when....
... and Mr. D is building gainclones
sam9 said:
Your comment makes sense, but so does Fred's comment that
we want to avoid wirewound resistors because of their high
inductance. Paralleling several smaller resistors seems the
way to go if shooting for a high wattage.
Hi,
Slone and Self say 10ohm/2-5W and 100nF.
Wirewound resistor has inductance about 0,5uH what isn't problem here, IMHO.
Regards
Slone and Self say 10ohm/2-5W and 100nF.
Wirewound resistor has inductance about 0,5uH what isn't problem here, IMHO.
Regards
Brian,
IYO, is the idea of a switch out?
If you want something along the same lines, sony provides a switch in some of their SACD players that applies a filter to counter harmful ocsilations in amps. There was a lot of discussions on this item, and most people prefered to leave the switch off to get the most of their high frenquency data. But the sony litterature clearly indicated that leaving the switch off may be harmful to the amp.
If people are interested, I can dig out the exact context.
IYO, is the idea of a switch out?
If you want something along the same lines, sony provides a switch in some of their SACD players that applies a filter to counter harmful ocsilations in amps. There was a lot of discussions on this item, and most people prefered to leave the switch off to get the most of their high frenquency data. But the sony litterature clearly indicated that leaving the switch off may be harmful to the amp.
If people are interested, I can dig out the exact context.
"Your comment makes sense, but so does Fred's comment that
we want to avoid wirewound resistors because of their high
inductance. Paralleling several smaller resistors seems the
way to go if shooting for a high wattage."
I have not given much though to to the inductance of the Zobel resistor. However, I have wondered about it with respect to the RE resistors typically found at the output. No doubt that's why I have seen photos of amps somewhere with the 5W sandcast replaced with 3 parrallel smaller (2W?) resistors. So long as we are on the subject, the finicy could analyse the effect of (claimed) 0.5uH inductance of the resistor typically (like that word!) used to dampen the output inductor.
As an aside, I worry about values I cant measure myself such as very small inductances. My DMM, for intance won't measure low enough to verify the inductance of output inductors, so I rely on Wheeler's formula. I have seen a project to build a an inductance meter, http://www4.tpgi.com.au/users/ldbutler/RFInducMeter.htm. A single range would satidfy my curiosity, it looks really simple but I bump into the problem of where to find a dual ganged 450pF variable cap. Nothing remotely close in digikey, mouser etc.- at least not that I can see.
we want to avoid wirewound resistors because of their high
inductance. Paralleling several smaller resistors seems the
way to go if shooting for a high wattage."
I have not given much though to to the inductance of the Zobel resistor. However, I have wondered about it with respect to the RE resistors typically found at the output. No doubt that's why I have seen photos of amps somewhere with the 5W sandcast replaced with 3 parrallel smaller (2W?) resistors. So long as we are on the subject, the finicy could analyse the effect of (claimed) 0.5uH inductance of the resistor typically (like that word!) used to dampen the output inductor.
As an aside, I worry about values I cant measure myself such as very small inductances. My DMM, for intance won't measure low enough to verify the inductance of output inductors, so I rely on Wheeler's formula. I have seen a project to build a an inductance meter, http://www4.tpgi.com.au/users/ldbutler/RFInducMeter.htm. A single range would satidfy my curiosity, it looks really simple but I bump into the problem of where to find a dual ganged 450pF variable cap. Nothing remotely close in digikey, mouser etc.- at least not that I can see.
- Status
- Not open for further replies.