grege said:
Geez Greg- are you trying to make me feel old???
🙂 "Twenty years" - that's harsh! My job was to put together the comms network for the NSW lotteries - and a good slurp of VB at the end of the day was most welcome. Better would be some Sauv-Blanc from NZ... Food was not optional - Morton Bay Bugs absolutely required....🙂
As for test equipment - I like the idea and think it could really help out the DIY community - let's make it happen...
Cheers,
Bill
Hear no zoble, see no zobel, speak.........
"I also had an amp in one system, that was badly distorting. Adding Zobel didn't fix the problem. Changing for different speaker cables did." 😕
Maybe less capacitive ones? A responsible person would have said "wait a minute....... what's going on here". Were the speaker cables defective? What if some unsuspecting guy puts something like a parallel ribbon cable or a home made Cat 5 speaker cable (which are very popular with DIYer's I have even built some to play with) ?
You are so busy defending an indefensible position and nursing your ego that you are doing other chip amp builders a disservice and flushing your credibility down the drain. You throw Mr. Rasmussen's design out there seemly without a clue that it has a low pass filter on the input (and incorrect value feedback resistors I believe unless it has a gain of over 50 and what looks like some pretty big DC offset with bipolar inputs) and a 0.22 ohm buildout resistor on the output. All three of which have a big effect on cable and speaker loading and the amount of RF that enters the amp to excite high oscillations. And still I see no questions about what the network actually does. You anecdotal evidence about no one having problems flies in the face off the reports have received from a very respected designer whom I trust, and who often is my guru on many audio design questions. I don't think I need to mention his name since I am pretty sure you know who is.
Rather than endure any more abuse for trying to be helpful, I will leave the chimp amp cult to monkey around to their hearts content and for certain members to hurl their feces at some other passer by.
Well, I'll not be a monkeys uncle,
Fred
"I also had an amp in one system, that was badly distorting. Adding Zobel didn't fix the problem. Changing for different speaker cables did." 😕
Maybe less capacitive ones? A responsible person would have said "wait a minute....... what's going on here". Were the speaker cables defective? What if some unsuspecting guy puts something like a parallel ribbon cable or a home made Cat 5 speaker cable (which are very popular with DIYer's I have even built some to play with) ?
You are so busy defending an indefensible position and nursing your ego that you are doing other chip amp builders a disservice and flushing your credibility down the drain. You throw Mr. Rasmussen's design out there seemly without a clue that it has a low pass filter on the input (and incorrect value feedback resistors I believe unless it has a gain of over 50 and what looks like some pretty big DC offset with bipolar inputs) and a 0.22 ohm buildout resistor on the output. All three of which have a big effect on cable and speaker loading and the amount of RF that enters the amp to excite high oscillations. And still I see no questions about what the network actually does. You anecdotal evidence about no one having problems flies in the face off the reports have received from a very respected designer whom I trust, and who often is my guru on many audio design questions. I don't think I need to mention his name since I am pretty sure you know who is.
Rather than endure any more abuse for trying to be helpful, I will leave the chimp amp cult to monkey around to their hearts content and for certain members to hurl their feces at some other passer by.
Well, I'll not be a monkeys uncle,
Fred
Attachments
This is rather indelicate but
ummmmmmmm... ahhhh...... well there is no real nice way to ask. What is he holding with his foot? Oh wait a minute, I am fairly certain it is his other foot and that's his big toe, never mind.
ummmmmmmm... ahhhh...... well there is no real nice way to ask. What is he holding with his foot? Oh wait a minute, I am fairly certain it is his other foot and that's his big toe, never mind.
It took me a moment to realize that what's in the lower half of the picture is actually his right toe 🙂
I see Fred posted a similar observation. Great minds think alike🙂
I see Fred posted a similar observation. Great minds think alike🙂
jamesjung21 said:
I am using small, compact 0.5w resistors by Phoenix Passive Components, available from Digikey. An example part number is: ppc221xct, which is for the 221ohm resistors used on the LM3875 boards. The same part numbers used to correspond to the BC components compact metal film resistors, but it looks like Digikey changed suppliers. I am content with the performance of these resistors, and work well in the design, for a fraction of the cost of the Caddock resistors.
The 4.7uF capacitors in the basic kits are Panasonic FC capacitors, instead of the much more expensive BlackGate 4.7uF N capacitors.
I haven't determined the resistors for the basic kits, but I want to use Mills MRA-5 resistors for the premium kits. I might end up using these for the basic kits, if I can get a large enough order together to bring the price to a reasonable level.
By the way, the final version of the boards will contain placement for zobels before the output resistors, so that hopefully everyone will be happy. I will probably install them for my own personal use, as I would just be using the 4780 for the lower freq part of my future active system. I don't feel the need for more power than the LM3875 provides for my regular Seas Thor TL speakers, and my other speakers as well.
--
Brian
Re: Hear no zoble, see no zobel, speak.........
Well Fred, maybe it's time to change your guru?
The cables in question were Goertz. The guy calls me and asks what's going on. There are high frequency distortions and nobody knows what is causing them.
I realize that the amp doesn't have Zobel at the output and I quickly prepare a jumper, that supposedly should fix the problem. I'm driving 50 km across the town and finally arrive at the guy's house. The distortions are real. I place my magic jumpers at binding posts, and guess what? The distortions don't disappear. We try few other things and eventually I propose to try different cables. We put some Monster cables between the amps and Hales speakers and suddenly everything is fine.
So what the **** your Zobel network is good for?
PS: Should I ask how much business your guru is loosing because of kits sold on this forum? Let's ask Brian how many oscillation reports he had received because his boards din't have Zobel.
At the prices Brian asks for his kits, I'm pretty much sure he already sold more amps than your guru friend.
My post here is only a direct response at Fred's comment, and has nothing to do with my personal attitude towards a person we choose to call "guru". I aplogize if I hurt anybody's feelings.
Fred Dieckmann said:
Well Fred, maybe it's time to change your guru?
The cables in question were Goertz. The guy calls me and asks what's going on. There are high frequency distortions and nobody knows what is causing them.
I realize that the amp doesn't have Zobel at the output and I quickly prepare a jumper, that supposedly should fix the problem. I'm driving 50 km across the town and finally arrive at the guy's house. The distortions are real. I place my magic jumpers at binding posts, and guess what? The distortions don't disappear. We try few other things and eventually I propose to try different cables. We put some Monster cables between the amps and Hales speakers and suddenly everything is fine.
So what the **** your Zobel network is good for?
PS: Should I ask how much business your guru is loosing because of kits sold on this forum? Let's ask Brian how many oscillation reports he had received because his boards din't have Zobel.
At the prices Brian asks for his kits, I'm pretty much sure he already sold more amps than your guru friend.
My post here is only a direct response at Fred's comment, and has nothing to do with my personal attitude towards a person we choose to call "guru". I aplogize if I hurt anybody's feelings.
Well Fred, maybe it's time to change your guru.
"There are high frequency distortions and nobody knows what is causing them. " PD
Maybe no one you know ........ but plenty of people on this forum know exactly what is causing them. Would half a dozen names be sufficient?
"So what the **** your Zobel network is good for?" - PD
That's funny, it sounds pretty personal to me..............
First off they aren't my Zobels. I didn't invent the thing ............ I just know what is and how it works, something you seem absolutely determined not to learn. Maybe you are trying to keep anybody else from learning either.
Secondly you made my point for me. The cables you couldn't get the amp to work with are the very type of situation that a Zobel is designed to deal with. The fact you couldn't get it to work might very well be due your implementation rather than the fault the Zobel. The directions in the data sheet are pretty specific and for a reason. I don't think hanging the thing outside the amp on banana plugs is what they had in mind. A cable that capacitive might lead someone with any curiosity about how the network works, to try a smaller resistor and or larger capacitor and to mount the network where it belongs on the PCB.
Speaking of PCBs......... I honestly wonder if point to point wiring is the way to go for the power supply and ground connections. Even with very short wire leads I wonder if wide PCB traces might offer a much lower inductance for filter caps to IC connections and in particular for the zobel network. I think my guru is safe from the competition, all things considered. I am sure he is not putting all his efforts into one type and configuration of product. When he stops making great sounding products and giving me excellent advice maybe I will think it over. Besides, Who knows ........... maybe the gain clone as we know it is a passing fad. There is after all only so much you can do with a design that is pretty much out the data sheet aps circuit and there are only so many parts you can swap out. Thanks for the monkey poo, but I brought my umbrella with me. Maybe you can throw rocks next time......
"There are high frequency distortions and nobody knows what is causing them. " PD
Maybe no one you know ........ but plenty of people on this forum know exactly what is causing them. Would half a dozen names be sufficient?
"So what the **** your Zobel network is good for?" - PD
That's funny, it sounds pretty personal to me..............
First off they aren't my Zobels. I didn't invent the thing ............ I just know what is and how it works, something you seem absolutely determined not to learn. Maybe you are trying to keep anybody else from learning either.
Secondly you made my point for me. The cables you couldn't get the amp to work with are the very type of situation that a Zobel is designed to deal with. The fact you couldn't get it to work might very well be due your implementation rather than the fault the Zobel. The directions in the data sheet are pretty specific and for a reason. I don't think hanging the thing outside the amp on banana plugs is what they had in mind. A cable that capacitive might lead someone with any curiosity about how the network works, to try a smaller resistor and or larger capacitor and to mount the network where it belongs on the PCB.
Speaking of PCBs......... I honestly wonder if point to point wiring is the way to go for the power supply and ground connections. Even with very short wire leads I wonder if wide PCB traces might offer a much lower inductance for filter caps to IC connections and in particular for the zobel network. I think my guru is safe from the competition, all things considered. I am sure he is not putting all his efforts into one type and configuration of product. When he stops making great sounding products and giving me excellent advice maybe I will think it over. Besides, Who knows ........... maybe the gain clone as we know it is a passing fad. There is after all only so much you can do with a design that is pretty much out the data sheet aps circuit and there are only so many parts you can swap out. Thanks for the monkey poo, but I brought my umbrella with me. Maybe you can throw rocks next time......
"First off they aren't my Zobels. I didn't invent the thing ............"
Yet you keep talking about them, like it's the only thing that matters in the amp. You can only get personal, with personal questions. Being disappointed with Zobels shouldn't sound like that.
"Secondly you made my point for me. The cables you couldn't get the amp to work with are the very type of situation that a Zobel is designed to deal with. The fact you couldn't get it to work might very well be due your implementation rather than the fault the Zobel. The directions in the data sheet are pretty specific and for a reason. I don't think hanging the thing outside the amp on banana plugs is what they had in mind. A cable that capacitive might lead someone with any curiosity about how the network works, to try a smaller resistor and or larger capacitor and to mount the network where it belongs on the PCB."
Well, Goertz cables are reported to cause problem with many other amps, and each of them probably have Zobel installed. So your attempt at blaiming my implementation will not hold.
Since this thread is made specifically to deal with Zobel, let's ask other people to share their experiences with GCs, Zobels, oscillations and what was helpful to eliminate those (especially the dangers and damages the lack of Zobel created).
"........... maybe the gain clone as we know it is a passing fad."
I've heard it already a year ago, and the GC is getting better than ever nowadays. It seems like everybody is building one.😉
Yet you keep talking about them, like it's the only thing that matters in the amp. You can only get personal, with personal questions. Being disappointed with Zobels shouldn't sound like that.
"Secondly you made my point for me. The cables you couldn't get the amp to work with are the very type of situation that a Zobel is designed to deal with. The fact you couldn't get it to work might very well be due your implementation rather than the fault the Zobel. The directions in the data sheet are pretty specific and for a reason. I don't think hanging the thing outside the amp on banana plugs is what they had in mind. A cable that capacitive might lead someone with any curiosity about how the network works, to try a smaller resistor and or larger capacitor and to mount the network where it belongs on the PCB."
Well, Goertz cables are reported to cause problem with many other amps, and each of them probably have Zobel installed. So your attempt at blaiming my implementation will not hold.
Since this thread is made specifically to deal with Zobel, let's ask other people to share their experiences with GCs, Zobels, oscillations and what was helpful to eliminate those (especially the dangers and damages the lack of Zobel created).
"........... maybe the gain clone as we know it is a passing fad."
I've heard it already a year ago, and the GC is getting better than ever nowadays. It seems like everybody is building one.😉
Re: Well Fred, maybe it's time to change your guru.
I thought you were a better reader. It sounds funny in your context. Obviously those who didn't know what is causing them, were uneducated customers who are not regulars on this forum.
Speaking of PCBs, I don't know, they may actually be better, but I don't think you can tell me for sure.
Whatever I tried so far (PCBs), brings very good results.
Fred Dieckmann said:
I thought you were a better reader. It sounds funny in your context. Obviously those who didn't know what is causing them, were uneducated customers who are not regulars on this forum.
Speaking of PCBs, I don't know, they may actually be better, but I don't think you can tell me for sure.
Whatever I tried so far (PCBs), brings very good results.
Networks and other notes
It seems there are three things in this thread for now.
First Zobel Networks
I have built quite a few amps using chip amps without the output networks without problems, however they were all in very controlled applications. In a self powered subwoofer you have control over everything connected to the amp. In a stand alone amp you do not. National needs to recomend a solution that will be unconditionally stable, and that leads to their network. It is derived from an understanding of the phase margin at unity gain crossover (which changes with load), the range of load impedances possible with the output network and possible capacitive loads. It may be possible to still get the amp to oscillate with the full network but then I would look at excessive inductance in the supply bypassing or a grounding problem.
I would like to see a simple derivation of an optimum network. I have always resorted to cut and try.
I have also built high performance amps with no output network that were fine as long as there was some series R or L isolating the output (at least a few inches of wire). But in later production we added an RL network to be safe. I could hear the difference but it was small.
Second item- I also have an extensive shop with lots of very specialized audio test equipment that I can't use fully, that I would allow others to use (with obvious limits). I can even help analyze the results.
Third- the class A amp is an interesting concept but few chip amps would tolerate the load for very long. And it needs some reconfiguring to work.
-Demian
It seems there are three things in this thread for now.
First Zobel Networks
I have built quite a few amps using chip amps without the output networks without problems, however they were all in very controlled applications. In a self powered subwoofer you have control over everything connected to the amp. In a stand alone amp you do not. National needs to recomend a solution that will be unconditionally stable, and that leads to their network. It is derived from an understanding of the phase margin at unity gain crossover (which changes with load), the range of load impedances possible with the output network and possible capacitive loads. It may be possible to still get the amp to oscillate with the full network but then I would look at excessive inductance in the supply bypassing or a grounding problem.
I would like to see a simple derivation of an optimum network. I have always resorted to cut and try.
I have also built high performance amps with no output network that were fine as long as there was some series R or L isolating the output (at least a few inches of wire). But in later production we added an RL network to be safe. I could hear the difference but it was small.
Second item- I also have an extensive shop with lots of very specialized audio test equipment that I can't use fully, that I would allow others to use (with obvious limits). I can even help analyze the results.
Third- the class A amp is an interesting concept but few chip amps would tolerate the load for very long. And it needs some reconfiguring to work.
-Demian
And I Quote...
Hey guys, I have read this whole thread and I definitely have my own opinions on the matter, but hey who am I, so... I will quote a man I respect very much, G. Randy Slone, from pages 214/215 of his book "HIGH-POWER AUDIO AMPLIFIER CONSTRUCTION MANUAL"
"Unfortunately, in some cases, the amplifier will quickly pass from the oscillator phase to the destruction phase to the smoking phase to the dead-champ phase. This adverse reaction is not because we happened to be unlucky! It is due to a lack of understanding regarding the fundamentals of amplifier stability and the negligence of the following through with a systematic approach to assuring it."
I have seen fires in electronics started from what may have been considered as an "unlikely to happen" oscillation. The last thing I want to worry about when I leave my house every day is if some designer somewhere left out a few components for one reason or another and put my safety and well being at risk. This may seem a little extreme but all my amps have a zoebel network, turn on delay, inrush current limiting and DC protection circuitry. Just my choice, and I have to date been very happy with the way my systems sound when compared to others.
Anyway... AGAIN I in no way wish to take sides as I respect both parties in this dispute, but I felt the need to add to the thread.
FWIW Peter, if your amplifiers sound half as good as they look then you must be extremely happy with them, your finished products are always beautiful, even the ones you profess to being ugly.
Hey guys, I have read this whole thread and I definitely have my own opinions on the matter, but hey who am I, so... I will quote a man I respect very much, G. Randy Slone, from pages 214/215 of his book "HIGH-POWER AUDIO AMPLIFIER CONSTRUCTION MANUAL"
"Unfortunately, in some cases, the amplifier will quickly pass from the oscillator phase to the destruction phase to the smoking phase to the dead-champ phase. This adverse reaction is not because we happened to be unlucky! It is due to a lack of understanding regarding the fundamentals of amplifier stability and the negligence of the following through with a systematic approach to assuring it."
I have seen fires in electronics started from what may have been considered as an "unlikely to happen" oscillation. The last thing I want to worry about when I leave my house every day is if some designer somewhere left out a few components for one reason or another and put my safety and well being at risk. This may seem a little extreme but all my amps have a zoebel network, turn on delay, inrush current limiting and DC protection circuitry. Just my choice, and I have to date been very happy with the way my systems sound when compared to others.
Anyway... AGAIN I in no way wish to take sides as I respect both parties in this dispute, but I felt the need to add to the thread.
FWIW Peter, if your amplifiers sound half as good as they look then you must be extremely happy with them, your finished products are always beautiful, even the ones you profess to being ugly.
Re: Re: Hear no zoble, see no zobel, speak.........
Thinking about it now, the guy was actually biwiring big Hales with two sets of cables attached to my GC monoblocks.
He was using Goertz for mids and highs and some other cable for woofers. The distortions were real, but not to the point that it really make you upset, and on some material you couldn't even hear them.
Considering that type of load connected to an amp, I'm actually quite satisfied with it's overall performance and that idicates that even without Zobel, those amps can take a lot and still behave quite satsfactory. I mean, they didn't blow, after all😉
And he had them setup like that for at least a week. I don't hear anynore complains, since the Goertz cables were changed, he still uses biwiring.
Peter Daniel said:
Thinking about it now, the guy was actually biwiring big Hales with two sets of cables attached to my GC monoblocks.
He was using Goertz for mids and highs and some other cable for woofers. The distortions were real, but not to the point that it really make you upset, and on some material you couldn't even hear them.
Considering that type of load connected to an amp, I'm actually quite satisfied with it's overall performance and that idicates that even without Zobel, those amps can take a lot and still behave quite satsfactory. I mean, they didn't blow, after all😉
And he had them setup like that for at least a week. I don't hear anynore complains, since the Goertz cables were changed, he still uses biwiring.
Can we put it to a vote?
It seems that this has really gone too far - and I apologise for having any part of that. I guess my parting comment is that I would like to see the Zobel - *as an option* - and *on the board* (since I don't think that an external implementation meets National's recommendations/technical requirements).
Can we just put it to a vote in order to see what the consensus is (and I realize that the folks doing the PCB can do whatever they want - they're not bound by ANY of this 🙂 )...?
I vote: Make the option for an on-board Zobel available.
I'm hoping we can come to some agreement on this Zobel issue soon - 'cause THEN I want to move on to the "input caps" - AND the solar-powered digital attenuator stage with telepathic input that I think should be considered....
Just kidding!!!! 🙂 🙂 🙂
Regards,
Bill
It seems that this has really gone too far - and I apologise for having any part of that. I guess my parting comment is that I would like to see the Zobel - *as an option* - and *on the board* (since I don't think that an external implementation meets National's recommendations/technical requirements).
Can we just put it to a vote in order to see what the consensus is (and I realize that the folks doing the PCB can do whatever they want - they're not bound by ANY of this 🙂 )...?
I vote: Make the option for an on-board Zobel available.
I'm hoping we can come to some agreement on this Zobel issue soon - 'cause THEN I want to move on to the "input caps" - AND the solar-powered digital attenuator stage with telepathic input that I think should be considered....
Just kidding!!!! 🙂 🙂 🙂
Regards,
Bill
Re: Can we put it to a vote?
Whether people want to use it or not, footprints for the zobel networks will be included on the pcb. I will post a new screenshot of the layout when I get it finished.
Anyone have component suggestions (type, not value) for the zobel? National recommends a 1/4 watt 2.7 ohm resistor, and a 0.1uF capacitor. I thought about SMD, as Fred suggested, but I am considering not using any SMD parts on the pcb, to make it easier for the average user of the pcbs.
--
Brian
netgeek said:
Whether people want to use it or not, footprints for the zobel networks will be included on the pcb. I will post a new screenshot of the layout when I get it finished.
Anyone have component suggestions (type, not value) for the zobel? National recommends a 1/4 watt 2.7 ohm resistor, and a 0.1uF capacitor. I thought about SMD, as Fred suggested, but I am considering not using any SMD parts on the pcb, to make it easier for the average user of the pcbs.
--
Brian
Re: Re: Can we put it to a vote?
Well, my vote would be for leaded parts and not SMD.
BrianGT said:
Well, my vote would be for leaded parts and not SMD.
Re: Re: Can we put it to a vote?
I'd think standard leaded components - not SMD - should suffice. If the difference between leaded vs. SMD causes problems, then there are other problems to worry about - and if that's true, then why build them at all? Beyond "DIY" at that point...
🙂
BrianGT said:
I'd think standard leaded components - not SMD - should suffice. If the difference between leaded vs. SMD causes problems, then there are other problems to worry about - and if that's true, then why build them at all? Beyond "DIY" at that point...
🙂
Zobel Components
Having fried more than a few Zobel resistors I would push for a 1 or 2 watt part (I would use 5W for my applications). (I like to confirm power bandwidth claims.) And it should be noninductive. The cap should also be non inductive since the networks function is to be a conjugate of an inductor.
Using it without the RL network still leaves the amp open to oscillation with the right output capacitance.
-Demian
Having fried more than a few Zobel resistors I would push for a 1 or 2 watt part (I would use 5W for my applications). (I like to confirm power bandwidth claims.) And it should be noninductive. The cap should also be non inductive since the networks function is to be a conjugate of an inductor.
Using it without the RL network still leaves the amp open to oscillation with the right output capacitance.
-Demian
It still seems to me like 2W Panasonic is the best overall choice. Regarding the cap, I'd go with anything that fits the space on the board.
- Status
- Not open for further replies.