sam9 said:
If you can, leave out the output inductor. An amp should be stable without it. It does nothing good, other than make the scope pictures pretty with cables attached.
Definitely move the Zobel net to the output terminals, and use a separate ground return to star ground. All it does on the card is inject noise into the ground system, which can cause big problems.
Is the Zobel network necessary, in general and specifically for the Slone amps? From what I read, it seems a lot of good-sounding amps have left it out. In Elektor amp articles, they always keep those components optional. "If you face problems, you can try adding these..." -- type of thing.slowhands said:
Both Rod Elliot sound.au.com and D. Self (in his book on amplifiers)have done some stuff on benefits of both of these (Zobel and output inductor). My interpretation is the the output inductor becomes less needed the shorter the speaker cables and visa versa. (I use it because I tend to have lonish speaker cables.) The zobel is generally beneficial although in some cases ther may already be one at /in the speakers.
One benefit of diy is you can build to your specific situation, whereas when you build for others to use, you have to assume the worst.
One benefit of diy is you can build to your specific situation, whereas when you build for others to use, you have to assume the worst.
By the way, there is a "rule of thumb" whose accuracy I can't vouch for that says the optimal value for zobel resistor is chosen the same value as the nominal impedance of the speakers. So if you have 4 ohm speakers use a 4 ohm zobel resistor. I would be interested to hear if there is any validity to this.
In Self's "Audio Power Amplifier Design Handbook" he reports expiriments conducted with and without a zobel. He clain HF instability in the output section if omitted even if the the speaker load is even "mildly inductive"
As for the output inductor, he claims that by experiment he found that if the inductor is omiited (or shorted - the same thing) a 100nF of capacatance parallel to the load can result in destructive ocillations. He distingushes this from the dampened ringing seen in the standard capacative load test. ***LET me add that I tried this very thing when building my 11.4 and got the same results - only quickily hitting the off switch saved a pair out output BJTs. At the time I had not read Self's book so I wasn't expecting the such a spectacular result - my scope just went totally bonkers!** Just a speculation-- a CFB output section may be more suseptable than an EF.
Self is as full of opinions as anyone (some contoversial), but one thing I like about him is that he actually checks most of them out at least in simulation and often with a real amp.
As for the output inductor, he claims that by experiment he found that if the inductor is omiited (or shorted - the same thing) a 100nF of capacatance parallel to the load can result in destructive ocillations. He distingushes this from the dampened ringing seen in the standard capacative load test. ***LET me add that I tried this very thing when building my 11.4 and got the same results - only quickily hitting the off switch saved a pair out output BJTs. At the time I had not read Self's book so I wasn't expecting the such a spectacular result - my scope just went totally bonkers!** Just a speculation-- a CFB output section may be more suseptable than an EF.
Self is as full of opinions as anyone (some contoversial), but one thing I like about him is that he actually checks most of them out at least in simulation and often with a real amp.
"Does this mean that practically any speaker with a coil and a moving diaphragm will need a Zobel?"
Unless the spaker itself already has one. ."
"I didn't know it can get this bad this easily"
There a lot of varibles that can have an influence. The parameters vary wildly from one speaker design to the other, some cables can have an influence, and two amps with the same schematic but different PCB layouts can react differently. Thus it is also within the range of possabilities that one combination of factors results in a stable situation while another oscillates readily. I doubt any are "beyond the laws of physics" but there are just a lot of permutations that make predictability hard. One way to deal with this is just to anticipate the worst.
You might want to look at this: Rod Elliots's web site (sound.au.com), the article on "Cable Impedance
Unless the spaker itself already has one. ."
"I didn't know it can get this bad this easily"
There a lot of varibles that can have an influence. The parameters vary wildly from one speaker design to the other, some cables can have an influence, and two amps with the same schematic but different PCB layouts can react differently. Thus it is also within the range of possabilities that one combination of factors results in a stable situation while another oscillates readily. I doubt any are "beyond the laws of physics" but there are just a lot of permutations that make predictability hard. One way to deal with this is just to anticipate the worst.
You might want to look at this: Rod Elliots's web site (sound.au.com), the article on "Cable Impedance
"I simulated Slone's circuit .It is not linearity at output high level ,about 17-18V,.perhaps that's non-linearity of input or VAS stage"
Are you talking about the "11.4" amp? I set one up in LTSpice and found nothing unusal. More important several here (including me) have built it. Your concern does not seem consistent with experience of actually building and listening to one even when the output is fairly close to the rails which are `30V in my case.
BTW, what do you mean by non-linearity? Frequency response? THD? All amps are non-linear in some respects, at least until someone produces one that truly has zero distortion.
Are you talking about the "11.4" amp? I set one up in LTSpice and found nothing unusal. More important several here (including me) have built it. Your concern does not seem consistent with experience of actually building and listening to one even when the output is fairly close to the rails which are `30V in my case.
BTW, what do you mean by non-linearity? Frequency response? THD? All amps are non-linear in some respects, at least until someone produces one that truly has zero distortion.
Update,
I received my power supply parts, and started wiring things up. Here is Randy Slone’s new low impedance power supply board, I’m using 8 10,000uf caps but the board can hold 10 caps. The transformer is an 800VA 50+50. Anyway back to work!! I will let you guys know how it sounds very soon!!!!!!
I received my power supply parts, and started wiring things up. Here is Randy Slone’s new low impedance power supply board, I’m using 8 10,000uf caps but the board can hold 10 caps. The transformer is an 800VA 50+50. Anyway back to work!! I will let you guys know how it sounds very soon!!!!!!
Attachments
tcpip,
The case I'm using for now used to be a Kenwood 5 disk CD player that never worked well from new, the aluminum plate is from a car amp, and I added the handles to it. This is a temporary case, as I plan to build a 4 or 5 channel amp in one enclosure. As for the heatsinks they came as a kit, which consisted of PCB all components and heatsinks. This is my first attempt at building an amplifier, so I decided to buy it as a kit form. He does sell things separately at his web site (seal electronics). I just finished setting the bias and it is ready for action! I will be connecting it up tomorrow morning, and see how she performs!!!! It’s too late now kids are sleeping.
The case I'm using for now used to be a Kenwood 5 disk CD player that never worked well from new, the aluminum plate is from a car amp, and I added the handles to it. This is a temporary case, as I plan to build a 4 or 5 channel amp in one enclosure. As for the heatsinks they came as a kit, which consisted of PCB all components and heatsinks. This is my first attempt at building an amplifier, so I decided to buy it as a kit form. He does sell things separately at his web site (seal electronics). I just finished setting the bias and it is ready for action! I will be connecting it up tomorrow morning, and see how she performs!!!! It’s too late now kids are sleeping.
Good job! Looks great and this just might be a reason for me to start building a case for my amps 
Make sure that you tighten those blue connectors properly , and let us know how it sounds. You may also try to up the bias current a tad bit (have my OPTI-MOS kits running at 80mA instead of 40mA) and everything is better sound quality wise 
Make sure that you tighten those blue connectors properly
, and let us know how it sounds. You may also try to up the bias current a tad bit (have my OPTI-MOS kits running at 80mA instead of 40mA) and everything is better sound quality wise
What's your power supply voltage? I have mine on +/-63V and the heatsinks ran warm to the touch at 100mA (around 40-45 deg celcius) so I backed it up to 80mA just to be 'safe'. It also depends on your speakers, mine are 4ohm, 86dB/1W sensitivity and they tend to 'suck up' more power than 8ohm higher sensitivity ones.
The OPTI-MOS highly benefits (sonicly) from the input tantalum caps removal (shunt with a piece of wire here) if your preamplifier permits (does not have any DC offset larger than 5-10mV). The preamp DC offset shouldn't be any trouble because the 470uf/35V in the feedback loop assures unity gain at DC.
Also solder the fuses in their slots (use a 100W soldering gun) to avoid excessive resistance causing rail dropping on heavy bass passages
The higher bias shouldn't cause any problems except for the 'small' additional heat; the amps are stable as a rock and do not run into thermal problems due to L-mosfets use (of course, if you get to mail Mr Slone about these he will tell you that is totally unnecessary and the fact that everything seem to sound better is just a subjective claim).
Just hook everything up, paying attention to the blue connectors that do not stand much fiddling around (I ended having the PS lines soldered because of this), run the amp a couple of days at 'defaults' and then up the bias/remove the input caps. If you feel there's any difference, leave it like this. If not back it up to 40mV between the two test points and remove the wire
Oh, I also tried changing the green mylar caps for metallized polypropylene ones and the sound got more congested in the dynamics department so leave those as they are, they do their job good.
The OPTI-MOS highly benefits (sonicly) from the input tantalum caps removal (shunt with a piece of wire here) if your preamplifier permits (does not have any DC offset larger than 5-10mV). The preamp DC offset shouldn't be any trouble because the 470uf/35V in the feedback loop assures unity gain at DC.
Also solder the fuses in their slots (use a 100W soldering gun) to avoid excessive resistance causing rail dropping on heavy bass passages
The higher bias shouldn't cause any problems except for the 'small' additional heat; the amps are stable as a rock and do not run into thermal problems due to L-mosfets use (of course, if you get to mail Mr Slone about these he will tell you that is totally unnecessary and the fact that everything seem to sound better is just a subjective claim).
Just hook everything up, paying attention to the blue connectors that do not stand much fiddling around (I ended having the PS lines soldered because of this), run the amp a couple of days at 'defaults' and then up the bias/remove the input caps. If you feel there's any difference, leave it like this. If not back it up to 40mV between the two test points and remove the wire
Oh, I also tried changing the green mylar caps for metallized polypropylene ones and the sound got more congested in the dynamics department so leave those as they are, they do their job good.
i do believe that the CF output is more susceptable to oscillations..now this is just an educated guess..but if you look at the CB junctions of the output sections, you will see that the miller caps and the resistor make a 2 pole circuit...very reminiscent of a phase shift oscillator...just my pennys worth..i could be wrong
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.