or Connect
AVS › AVS Forum › Audio › Audio theory, Setup and Chat › External DAC Questions
New Posts  All Forums:Forum Nav:

External DAC Questions - Page 7

post #181 of 267
Quote:
Originally Posted by Glimmie View Post


The QSC amps are all switch mode power supplies. The amp circuits are standard class AB but the power supply is regulated. It is well known that regulated power supplies on power amps are very difficult to get right. the problem is you have one feedback loop inside another and the power supply feedback system is usually slower than the amp. There goes your slew rate. Douglass Self and other experts on the subject have also cited this.

I have both recent Self books before me and don't see support for the conclusions mentioned above.
post #182 of 267
Quote:
Originally Posted by arnyk View PostTrue for either tubed or SS amps.

With all due respect, do you realize that you challenged the benchmark in amp design? May I ask ?

 

What qualifies your marginalization of Jim's patent (1 of 9). Please forward your 'spice model' showcasing pseudo science? Thanks in advance Arny.

post #183 of 267
A few random thoughts. Take all this as opinion since I am too lazy to dig out all the references and provide links for everything I have learned through the years.

Class D amplifiers are pretty sensitive to the power rails and generally benefit more from regulated supplies than typical A/AB/etc. amplifiers. The output stage switches between the rails, after all, so power supply modulation gets straight to the output.

A lot of tube designs (preamps especially, and many amps) use a single-ended front end (and some, e.g. SETs, are single-ended throughout) and thus the largest distortion term is the second, which sounds "smoother" than the third to most of us. Differential circuits cancel even-order products to first order so odd harmonics dominate (as Arny said, there will always be a little of each).

Aside: The fundamental distortion series for a tube is factorial rather than exponential, so at the same op point a tube with the same gain as a transistor actually has lower distortion. Feedback makes all of that moot.

Aside 2: I once designed and built a differential tube preamp that had excellent specs. Everyone said it sounded "too solid state". I killed the differential input stage, just strapped it to single-ended so distortion went up and CMRR went down, and they all agreed it was a much better tube preamp. Go figure.

The vacuum tube gain characteristic is different and usually lower than transistors, and they have higher output impedance, and often less feedback, so tend to clip more "softly" than transistor circuits. They don't sound as bad, as fast, when over driven, at least initially.

Clipping makes sine waves more like square waves, and an ideal square wave has only odd-order products. Triangle waves have even-order components and sound "smoother" to people. Sine waves are smoothest, of course...
post #184 of 267
Quote:
Originally Posted by amirm View Post

Since the man is not here, your request is illogical. But you are here. Can you run the spice simulation and show what you want to show? It is a small circuit so shouldn't take you long. Is later today enough time to do it?


Carl Sagan suggest, bold claims require similar empirical evidence. I'm certainly able to analyze,  where is Arny's modeling data?

post #185 of 267
Quote:
Originally Posted by arnyk View Post

I have both recent Self books before me and don't see support for the conclusions mentioned above.

I have edition 3. Its clearly in there in the section on power supplies. He upholds the unregulated brute force power supply as the best bet. He does acknowledge the weight and international line voltage attributes of SMPS. So does Randy Sloan.

As for linear regulated power supplies, it is a known fact that most have a rising impedance with frequency. That's not to say you couldn't build an ideal linear regulator for power amplifiers but it's simply not worth the effort as the brute force approach works best audio performance wise.

In my 70 watt tube amps I regulated the driver circuitry but not the finals. I found this to be a good compromise for tube power amps.

EDIT: I should clarify I am referring to regulated power supplies that use a feedback system. Most do including a lowly 7805/12/15 regulator IC. But there are open loop topologies as well and while they have poor to no load regulation, they would be better for a power amp in theory. A capacitance multiplier is an example.
Edited by Glimmie - 1/29/14 at 3:26pm
post #186 of 267
Quote:
Originally Posted by Glimmie View Post

Quote:
Originally Posted by arnyk View Post

I have both recent Self books before me and don't see support for the conclusions mentioned above.

I have edition 3. Its clearly in there in the section on power supplies. He upholds the unregulated brute force power supply as the best bet. He does acknowledge the weight and international line voltage attributes of SMPS. So does Randy Sloan.

As for linear regulated power supplies, it is a known fact that most have a rising impedance with frequency. That's not to say you couldn't build an ideal linear regulator for power amplifiers but it's simply not worth the effort as the brute force approach works best audio performance wise.

In my 70 watt tube amps I regulated the driver circuitry but not the finals. I found this to be a good compromise for tube power amps.

I've experimented with powering power amps with regulated switchmode power supplies and had what I consider to be reasonable success. This kind of amplifier has been widely sold to consumers for decades in the aftermarket high performance automotive audio business. None of the problems described above seem to have afflicted the commercial products or my own experiments.

The solution seems to be pretty obvious. In aftermarket car audio the solution is both built into the equipment and sold separately as "Stiffening" capacitors. The basic idea is to sacrifice the ability to filter a high frequency switchmode power supply with relatively small capacitors and put in energy storage capacitors that would do proud to any power amp operating off of 50/60 Hz. The transients are supported by the power main supply capacitors just as they are with unregulated power line based power supplies.
Quote:
EDIT: I should clarify I am referring to regulated power supplies that use a feedback system. Most do including a lowly 7805/12/15 regulator IC. But there are open loop topologies as well and while they have poor to no load regulation, they would be better for a power amp in theory. A capacitance multiplier is an example.

Many of the zilliions of commercial variations on switchmode power supply technology seem to work best if they incorporate some kind of regulation. Must of this is revealed in modern pro audio and consumer power amps that already incorporate switchmode power supplies.

Crown's highly sucessful XTi series of power amps come to mind as a successful line of very powerful amplifiers that are based on switchmode power supplies and linear Class AB power amplifier circuits. They feature sufficient internal regulation that they are rated for use on power lines ranging from about 90 to 250 volts if memory serves. I've used them on 130 and 240 volt power no switches or adjustment required. I've had them on the bench and used them to amplify audio and they seem glitch-free. A 2KW XTi series amp weighs about 12 pounds, if memory serves.
post #187 of 267
Quote:
Originally Posted by Glimmie View Post

Note, I am not an audiophool. I don't believe in cables or other wacky products. But I do hear differences in these amps. And as an EE who has built amps from the component level I know all about measurements.

 

Should be an epiphany in amplifier design when Arny demos his 'spice modeling' of the detailed / patented circuit I've provided here:

 

http://www.hafler.com/techsupport/pdf/MAN1482B_P1000_man.pdf

post #188 of 267
Quote:
Originally Posted by bralas View Post

Quote:
Originally Posted by Glimmie View Post

Note, I am not an audiophool. I don't believe in cables or other wacky products. But I do hear differences in these amps. And as an EE who has built amps from the component level I know all about measurements.


 
Should be an epiphany in amplifier design when Arny demos his 'spice modeling' of the detailed / patented circuit I've provided here:

http://www.hafler.com/techsupport/pdf/MAN1482B_P1000_man.pdf

I notice that the manual departs from usual conventions by apparently not containing any specifications. I am therefore mystified why I even spent as little time with the manual as I have p there seems to nothing remarkable about it at all. Its only virtue, according to the manual, is reliability.

The amplifier itself appears to not be in current production, probably because it contributed to the well known financial failure of Hafler. I see that one is for sale here:

http://www.ebay.com/itm/HAFLER-P1000-Trans-Ana-2-Channel-Professional-MOSFET-Power-Amplifier-/201028415629

If someone showed that it had any sterling technical performance features that warranted even just this discussion... ;-)
post #189 of 267

Jim Strickland
Vice President of Engineering

A published AES member since 1970, Jim holds nine patents in the audio field and has been leading the engineering team at Rockford Corporation since 1985.

 

Arny how many patents do you hold? Also. don't belabor the said apparent "lack of performance" we are waiting for your spice model of this very circuit. As Amir mentioned it appears simple and elegant for the experienced. Please point out it's design (weakness) envelope. Thanks in advance!

post #190 of 267
Quote:
Originally Posted by arnyk View Post

I've experimented with powering power amps with regulated switchmode power supplies and had what I consider to be reasonable success. This kind of amplifier has been widely sold to consumers for decades in the aftermarket high performance automotive audio business. None of the problems described above seem to have afflicted the commercial products or my own experiments.

The solution seems to be pretty obvious. In aftermarket car audio the solution is both built into the equipment and sold separately as "Stiffening" capacitors. The basic idea is to sacrifice the ability to filter a high frequency switchmode power supply with relatively small capacitors and put in energy storage capacitors that would do proud to any power amp operating off of 50/60 Hz. The transients are supported by the power main supply capacitors just as they are with unregulated power line based power supplies.

That's quite true but consider the automotive problem. How do you generate more than three honest watts without some form of DC/DC converter. You don't have a choice outside of an MG set or auxillary alternator. The stiffening capacitor is mainly for voltage drop from the engine to the trunk, it's on the amp's power input terminals after all. Remember that you are pulling 30-40 amps at only 12 volts over an 8awg, possibly #6 wire during transiets. The car battery is surely as stiff as it gets but getting that energy reserve to the trunk is another matter. But you are correct in that adding large rail bypass capacitors in an SMPS application is good provided they don't prevent the power supply from startup. But many SMPS's don't like a cold 50,000uf capacitor across their terminals. They see it as a dead short and abort startup.
Quote:
Many of the zilliions of commercial variations on switchmode power supply technology seem to work best if they incorporate some kind of regulation. Must of this is revealed in modern pro audio and consumer power amps that already incorporate switchmode power supplies.

Crown's highly sucessful XTi series of power amps come to mind as a successful line of very powerful amplifiers that are based on switchmode power supplies and linear Class AB power amplifier circuits. They feature sufficient internal regulation that they are rated for use on power lines ranging from about 90 to 250 volts if memory serves. I've used them on 130 and 240 volt power no switches or adjustment required. I've had them on the bench and used them to amplify audio and they seem glitch-free. A 2KW XTi series amp weighs about 12 pounds, if memory serves.

Crown, QSC, are pro PA amps. That does not mean they sound any worse than good HiFi amps, in fact they often spec out even better. But the choice of a SMPS is for the exact reasons you, Self, and Sloan have stated. Weight, heat, and international usage are the primary reasons. And even on the HiFi front, if you have a 1kw sub woofer amp, you really don't have much of a practical choice. You can go class D or use a SMPS with class B. But once you get above 500-700 watts, that's the practical limit of a 50/60hz transformer in this application. And thanks to the electronics industry as a whole SMPS's are cheap and fairly reliable these days.

Still bottom line, the QSC amps sound different in my system than the Haflers (and the tube amps as well but that's expected). They both are below 0.04% THD in the 10 watt range and that's the limit I can measure with my Heathkit copy of the classic HP330. As to which is better - that's subjective. I prefer the Haflers. I really don't like the QSC's even though many consider them "high end"

And consider I too am using SMPS's with my new class A amps. Here again I really don't have a choice. There is very poor PSRR in a class A amp so do I build a +/-24 volt 60hz transformer based bipolar regulated supply that can deliver 12-15 amps per rail?

P.S. The Hafler amps I am using here are the DH220's. There are before the buyout to Rockford / Fosgate. Wasn't it some of the Dynaco 400 team that designed these original Hafler amps?
Edited by Glimmie - 1/29/14 at 4:45pm
post #191 of 267
Quote:
Originally Posted by Glimmie View Post

Still bottom line, the QSC amps sound different in my system than the Haflers (and the tube amps as well but that's expected). They both are below 0.04% in the 10 watt range and that's the limit I can measure with my Heathkit copy of the HP330. As to which is better - that's subjective. I prefer the Haflers. I really don't like the QSC's even though many consider them "high end"

 

Arny is going to model the Hafler design, report here why they preform so well...

post #192 of 267
Quote:
Originally Posted by bralas View Post

Arny is going to model the Hafler design, report here why they preform so well...

See my edit above. Different Haflers from different era in the company. Does it make any difference?
post #193 of 267
Quote:
Originally Posted by Glimmie View Post

Quote:
Originally Posted by arnyk View Post

I've experimented with powering power amps with regulated switchmode power supplies and had what I consider to be reasonable success. This kind of amplifier has been widely sold to consumers for decades in the aftermarket high performance automotive audio business. None of the problems described above seem to have afflicted the commercial products or my own experiments.

The solution seems to be pretty obvious. In aftermarket car audio the solution is both built into the equipment and sold separately as "Stiffening" capacitors. The basic idea is to sacrifice the ability to filter a high frequency switchmode power supply with relatively small capacitors and put in energy storage capacitors that would do proud to any power amp operating off of 50/60 Hz. The transients are supported by the power main supply capacitors just as they are with unregulated power line based power supplies.

That's quite true but consider the automotive problem. How do you generate more than three honest watts without some form of DC/DC converter. You don't have a choice outside of an MG set or auxillary alternator. The stiffening capacitor is mainly for voltage drop from the engine to the trunk, it's on the amp's power input terminals after all. Remember that you are pulling 30-40 amps at only 12 volts over an 8awg, possibly #6 wire during transiets. The car battery is surely as stiff as it gets but getting that energy reserve to the trunk is another matter. But you are correct in that adding large rail bypass capacitors in an SMPS application is good provided they don't prevent the power supply from startup. But many SMPS's don't like a cold 50,000uf capacitor across their terminals. They see it as a dead short and abort startup.

One common solution is to use some kind of "soft start" circuit.
Quote:
Quote:
Many of the zilliions of commercial variations on switchmode power supply technology seem to work best if they incorporate some kind of regulation. Must of this is revealed in modern pro audio and consumer power amps that already incorporate switchmode power supplies.

Crown's highly sucessful XTi series of power amps come to mind as a successful line of very powerful amplifiers that are based on switchmode power supplies and linear Class AB power amplifier circuits. They feature sufficient internal regulation that they are rated for use on power lines ranging from about 90 to 250 volts if memory serves. I've used them on 130 and 240 volt power no switches or adjustment required. I've had them on the bench and used them to amplify audio and they seem glitch-free. A 2KW XTi series amp weighs about 12 pounds, if memory serves.

Crown, QSC, are pro PA amps. That does not mean they sound any worse than good HiFi amps, in fact they often spec out even better. But the choice of a SMPS is for the exact reasons you, Self, and Sloan have stated. Weight, heat, and international usage are the primary reasons. and even on the HiFi front, if you have a 1kw sub woofer amp, you really don't have much of a practical choice. You can go class D or use an SMPS with class B. But once you get above 500-700 watts, that the limit of a 50/60hz transformer. And thanks to the electronics industry as a whole SMPS's are cheap and fairly reliable these days.

I know of no such limit on 50/60 Hz transformers. For example the Behringer EP-2500/4000 (basically same amp, different ratings) does indeed deliver 1200 wpc x 2 to 2 ohm loads if you provide it with enough AC power. It uses a classic brute-force 50/60 Hz transformer-bawed power supply. It is one reason why my test bench has 30 amp 230 volt service. ;-)

Quote:
Still bottom line, the QSC amps sound different in my system than the Haflers (and the tube amps as well but that's expected). They both are below 0.04% in the 10 watt range and that's the limit I can measure with my Heathkit copy of the HP330. As to which is better - that's subjective. I prefer the Haflers. I really don't like the QSC's even though many consider them "high end"

I believe you are talking about Heath's 5258 THD analyzer. With my proprietary mods mine has ca. 0.005% or less residual distortion. However I abandoned it to use pro audio sound cards for measurements, which have residuals in the 0.001 to 0.0003% range.
post #194 of 267
Quote:
Originally Posted by bralas View Post

Jim Strickland

Vice President of Engineering


A published AES member since 1970, Jim holds nine patents in the audio field and has been leading the engineering team at Rockford Corporation since 1985.

Arny how many patents do you hold? Also. don't belabor the said apparent "lack of performance" we are waiting for your spice model of this very circuit. As Amir mentioned it appears simple and elegant for the experienced. Please point out it's design (weakness) envelope. Thanks in advance!

My dear man it is your job to prove the positive attributes of this equipment. You made the claim(s), now its your job to provide the evidence. If I attempt this work and provide negative results people could always claim that was due to my negative attitude.

People who are incapable of this level of work might do well and try a different strategy than to trick/bully me into doing their work for them, don't you think?
post #195 of 267
Quote:
Originally Posted by arnyk View Post

One common solution is to use some kind of "soft start" circuit.
I know of no such limit on 50/60 Hz transformers. For example the Behringer EP-2500/4000 (basically same amp, different ratings) does indeed deliver 1200 wpc x 2 to 2 ohm loads if you provide it with enough AC power. It uses a classic brute-force 50/60 Hz transformer-bawed power supply. It is one reason why my test bench has 30 amp 230 volt service. ;-)

I said "practical limit". Surely there is no limit to the size you can make a power transformer - just look at a substation. But these days it just doesn't make sense in a 1kw+ audio amplifier. And 1kw+ is in the pro PA class.
Quote:
I believe you are talking about Heath's 5258 THD analyzer. With my proprietary mods mine has ca. 0.005% or less residual distortion. However I abandoned it to use pro audio sound cards for measurements, which have residuals in the 0.001 to 0.0003% range.

Yup, that's the one. I have seen other articles for improvements but like you, when that time comes for me to upgrade I'll go with a PC based solution.
Edited by Glimmie - 1/29/14 at 4:57pm
post #196 of 267
Quote:
Originally Posted by bralas View Post

Quote:
Originally Posted by Glimmie View Post

Still bottom line, the QSC amps sound different in my system than the Haflers (and the tube amps as well but that's expected). They both are below 0.04% in the 10 watt range and that's the limit I can measure with my Heathkit copy of the HP330. As to which is better - that's subjective. I prefer the Haflers. I really don't like the QSC's even though many consider them "high end"


 
Arny is going to model the Hafler design, report here why they preform so well...

Surely you jest. Professionals know that it is the responsibility of the person making the affirmative claim to provide the affirmative evidence.
post #197 of 267

Noted your edit Glimmie.

 

Arny has a very elegant (simple) circuit to model. My request is rhetorical in that I'm asking Arny to re-engineer inspired genius.

post #198 of 267
Re: Spice Simulation:

So I looked at that figure from the patent, and thought that it was simple enough to simulate in Spice. However, the components were not specified, and I wasn't about to select random MosFets and op-amps. The patent figure compares an amp with bipolar transistors with an amp with power MosFets, so it appears to put forth a topology for the feedback on a MosFet amp.

I then looked at the schematic in the manual, and it doesn't appear to be the same topology that is shown in the patent. Maybe I can't see it, so if someone could point out the feedback loop, I'd appreciate it.

However, nobody is going to perform a Spice simulation on that entire amp. It would take me days just to enter a schematic that large, and even then it would take forever (or even longer) to run a simulation that gets you meaningful data. Normally, a schematic is broken down into manageable pieces for simulation.

Then I looked at the amp itself, as posted on eBay, and thought "What's the point?". For a patent that claims no distortion, THD of .2% seems like an awful lot. My old Sansui has an order of magnitude less, and has twice the power.
post #199 of 267
When I was in high school, there was a biology teacher who would bet gullible students 25 cents that they couldn't do seventy-five pushups in front of the class.

Guys who thought they were up to the task would get up in front of the class and struggle and strain like crazy trying to get to 75. They'd make complete fools of themselves, all for the chance at getting a quarter.

The teacher would just laugh the entire time. And if somebody ever actually completed the seventy five pushups, the teacher would gladly hand the red-faced, trembling, gasping-for-breath student a quarter.

So, yeah, why would somebody go to all the trouble of simulating that amp just to prove a point to a stranger online - a stranger who just changed his username, no less.
post #200 of 267

"..the components were not specified"

 

http://www.hafler.com/techsupport/pdf/MAN1482B_P1000_man.pdf

 

All components are labeled. There is also detailed circuit description with block diagram.

post #201 of 267
Quote:
Originally Posted by MarkHotchkiss View Post

the schematic in the manual, and it doesn't appear to be the same topology that is shown in the patent. Maybe I can't see it, so if someone could point out the feedback loop, I'd appreciate it.

However, nobody is going to perform a Spice simulation on that entire amp.

http://www.fosgateaudionics.com/products/FAP-V1.asp

 

Dolby put forth much effort to model Jim Fosgate (Rockford/Hafler) Pro Logic II all tube processor.

post #202 of 267
Hi Bralas,
Quote:
Originally Posted by bralas View Post

"..the components were not specified"

http://www.hafler.com/techsupport/pdf/MAN1482B_P1000_man.pdf

All components are labeled. There is also detailed circuit description with block diagram.
You misunderstood - Figure 1 of the patent has no components labled. As I said in my post, that schematic would be Spice-able, if specific components were called out.

The schematic in the manual cannot be simulated at all, without breaking it down into manageable pieces.
post #203 of 267
Hi Bralas,
Quote:
Originally Posted by bralas View Post

http://www.fosgateaudionics.com/products/FAP-V1.asp

Dolby put forth much effort to model Jim Fosgate (Rockford/Hafler) Pro Logic II all tube processor.
Do you have a link?

I sure they didn't run spice on the entire amp as a monolithic unit. If they did, then they spent more human resources and more processing power then I have access to.
post #204 of 267
Quote:
Originally Posted by MarkHotchkiss View Post

Hi Bralas,
Do you have a link?

I sure they didn't run spice on the entire amp as a monolithic unit. If they did, then they spent more human resources and more processing time then I have access to.


Greetings Mark,

 

I've found it before, remember it took  'days' for computers to model/analyze Jim's design. I will attempt to give you that URL.

post #205 of 267
No fair chuckling at the FAP-v1 model name. biggrin.gif
post #206 of 267

http://www.stereophile.com/content/jim-fosgate-surround-sound-mind-page-2

 

Fosgate: Actually, I built a solid-state breadboard patterned after the tube version, substituting operational amplifiers for the tubes. Dolby's computers crunched numbers around the clock for three months on the circuit, and they needed several more months to write digital code before we could audition the result

 

[Page 2]

post #207 of 267
@Glimmie -- FYI, a friend and I added a large capacitor bank to a DH220 many years ago and the result was substantially enhanced bass, measured and heard. Virtually no change in the rest of the spectrum. One of the times a mod did some real good. We changed the power bus wiring to the output modules at the same time, so it's possible that helped, but the DH220's power supply appeared to need some help...
post #208 of 267
I don't have any problem running the simulation (in my previous life I designed and simulated 10k+-transistor bipolar data converter designs, and now work at a company that does millions of transistors but most sims are RTL). But, I am not sure where to find models for all those components... And in the end you'll have macro models for the op-amps and have to decide if that's good enough.
post #209 of 267
Wait, this thread is about external DACs? What?!? smile.gif
post #210 of 267
Quote:
Originally Posted by DonH50 View Post

@Glimmie -- FYI, a friend and I added a large capacitor bank to a DH220 many years ago and the result was substantially enhanced bass, measured and heard. Virtually no change in the rest of the spectrum. One of the times a mod did some real good. We changed the power bus wiring to the output modules at the same time, so it's possible that helped, but the DH220's power supply appeared to need some help...

Given that the basic performance of the DH220 is quite good, what was the nature of the measured improvement?
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Audio theory, Setup and Chat
AVS › AVS Forum › Audio › Audio theory, Setup and Chat › External DAC Questions