In my Service Workshop Room at Klarion Enterprises, the 'standard shelf amp' for testing line-level equipment was a Sansui 🙂
Some of those are extremely good. Others really not worth listening to. What model are you using there?
I can't actually remember the model number, but it was older than the models you mentioned > a very popular 70's unit that had
rather ugly large toggle switches > nothing super, but reliable 'in its time'.
The newer models you mention ARE far superior.
rather ugly large toggle switches > nothing super, but reliable 'in its time'.
The newer models you mention ARE far superior.
Let me know. Some of those were very good. I'm curious now.
Yeah, all those were extremely reliable. Good choice for a bench amp.
Yeah, all those were extremely reliable. Good choice for a bench amp.
After a bit of 'googling' it turns out that it was a 1969 AU555 > IMO, nothing special.
I actually remember replacing 2x front-end transistors after one went noisy.
In all the nostalgia it seems crossover distortion, plagueing the majority of power amps anno 1970, is overlooked. Quad 303 was one of the few power amps of that era that performed flawlessly even when put on the torture table.
Hi Mister Audio,
Yeah, that one is pretty grim, but better than most from that era. We still didn't know how to design audio with solid state. Tube thinking was still strong.
Hi Boden,
Yup, I remember the ads with a string of 303's linked to drive each one in a long line. Stupid concept. Those were not good amplifiers by the way, and they did not have crossover distortion licked. Not even close.
There were a couple products that were ahead of their time. Look up a Marantz 500. But like I said, design with solid state wasn't well developed back then. Many still used single supplies and no diff pair. You'll see many designs share much in common as the state of the art progressed.
Yeah, that one is pretty grim, but better than most from that era. We still didn't know how to design audio with solid state. Tube thinking was still strong.
Hi Boden,
Yup, I remember the ads with a string of 303's linked to drive each one in a long line. Stupid concept. Those were not good amplifiers by the way, and they did not have crossover distortion licked. Not even close.
There were a couple products that were ahead of their time. Look up a Marantz 500. But like I said, design with solid state wasn't well developed back then. Many still used single supplies and no diff pair. You'll see many designs share much in common as the state of the art progressed.
With all due respect: I beg to differ. Peter Walker knew pretty good what he was doing.
Can you show my any measurement or test report of a 303 displaying crossover distortion?
I've repaired enough of them over the years, and their distortion was higher than many. I do look at crossover distortion, but I don't record the scope images. It's a diagnostic procedure, not something anyone has ever asked for.
Tell you what, next time I see one of these I'll record the scope display of crossover distortion for you. I normally look at this if distortion measurements are too high at the manufacturer's recommended bias level setting, or when deciding on bias current for a new design.
This is no slight against Peter Walker at all, but we only knew so much back then, and we couldn't measure as well either. Now, the way you look at crossover distortion is to measure the amplifier on an 8R load, or 4R. Run the output into a THD meter and connect the monitor output into Y2 of a scope. Y1 gets the signal so you can sync. You will clearly see crossover notches plus noise. Most modern good amplifiers have any crossover notch buried in noise. I use 10 KHz as a test signal, very low distortion.
You can forget test reports. If you knew anything about this industry at all, you would know test reports show positive attributes and downplay negative characteristics. Plus as I mentioned, our test equipment back then wasn't nearly as good as today's equipment. So reading a test report from back in the day is next to useless compared to a current test report. That and you have to know how to set up the test. An oscilloscope alone will only show the most gross amounts of distortion by the way.
Tell you what, next time I see one of these I'll record the scope display of crossover distortion for you. I normally look at this if distortion measurements are too high at the manufacturer's recommended bias level setting, or when deciding on bias current for a new design.
This is no slight against Peter Walker at all, but we only knew so much back then, and we couldn't measure as well either. Now, the way you look at crossover distortion is to measure the amplifier on an 8R load, or 4R. Run the output into a THD meter and connect the monitor output into Y2 of a scope. Y1 gets the signal so you can sync. You will clearly see crossover notches plus noise. Most modern good amplifiers have any crossover notch buried in noise. I use 10 KHz as a test signal, very low distortion.
You can forget test reports. If you knew anything about this industry at all, you would know test reports show positive attributes and downplay negative characteristics. Plus as I mentioned, our test equipment back then wasn't nearly as good as today's equipment. So reading a test report from back in the day is next to useless compared to a current test report. That and you have to know how to set up the test. An oscilloscope alone will only show the most gross amounts of distortion by the way.
The people at Acoustical Manufacturing Company/QUAD Electroacoustics were not as obsessed with low distortion figures as some people are nowadays anyway.
Yes. That's it. Well summarized @WhiteDragon...
Old doesn't mean good.
Modern doesn't imply better.
Old are maintenable.
Modern are recyclable.
Choose your weapon !
T
For sure, there is a happy balance. Once your distortion numbers (THD and IMD) are low enough, the equipment will probably sound good to most people. This is marketing and sales at each price point.
Getting really low distortion and great performance + reliability is expensive. Most manufacturers then resort to a good story and downplay their weaknesses. When you add trading on a stock exchange or having investors, emphasis shifts too far towards high profits. Guess what suffers?
Getting really low distortion and great performance + reliability is expensive. Most manufacturers then resort to a good story and downplay their weaknesses. When you add trading on a stock exchange or having investors, emphasis shifts too far towards high profits. Guess what suffers?
anatech: your point about "low enough" is well taken and I believe that measurements are a modest part of the total story. Recently I rebuilt my entire chain of amplification from phono pre through preamp, through monoblocks so that each L/R PCB was separate AND each was powered by its own discrete PSU. This made a far more dramatic improvement than would moving from 0.1 to 0.01 in THD. Engineers used this discrete approach back in the 50s when many of the great analog recordings were made and now media and equipment developers keep trying to reach up to that original level.
The big picture includes sufficient speaker size to power your room to required listening levels, quality media types (yes, analog reel to reel is best but 45 rpm vinyl can't be discounted nor blu-ray digital), room size and treatment, and a minimum of conversions in the wave form throughout the reproductive cycle. And as you note these are all expensive when done right.
In my world, these kind of large-scale choices turn music reproduction into an emotional event that keeps your interest devoid of visual need. The greatest music has the enduring quality of repeated listening pleasure. You can listen to Little Feat's "Let It Roll" album a bunch of times just as you can Bartok's Concerto for Orchestra. The same musical content falls on the ears with recognized pleasure. Because instrumental music is abstract it falls on the ears directly and, as the 19th-century romanticists stressed, creates emotion without the need for interpretation.
With the best sound equipment you have a lifetime of listening pleasure for these repeated listenings. And if you pursue hard-copy media, you don't care if the Internet goes down. Or if your streaming service goes on a hard-core advertising binge.
Just listen to quality equipment of any type as music lovers have done for 100+ years. The equipment method changes but the best continues to produce real emotional value.
The big picture includes sufficient speaker size to power your room to required listening levels, quality media types (yes, analog reel to reel is best but 45 rpm vinyl can't be discounted nor blu-ray digital), room size and treatment, and a minimum of conversions in the wave form throughout the reproductive cycle. And as you note these are all expensive when done right.
In my world, these kind of large-scale choices turn music reproduction into an emotional event that keeps your interest devoid of visual need. The greatest music has the enduring quality of repeated listening pleasure. You can listen to Little Feat's "Let It Roll" album a bunch of times just as you can Bartok's Concerto for Orchestra. The same musical content falls on the ears with recognized pleasure. Because instrumental music is abstract it falls on the ears directly and, as the 19th-century romanticists stressed, creates emotion without the need for interpretation.
With the best sound equipment you have a lifetime of listening pleasure for these repeated listenings. And if you pursue hard-copy media, you don't care if the Internet goes down. Or if your streaming service goes on a hard-core advertising binge.
Just listen to quality equipment of any type as music lovers have done for 100+ years. The equipment method changes but the best continues to produce real emotional value.
Agreed. Your system should reproduce whatever music you feed it equally well. I've had some of the same equipment since the early 1980's, I serviced and upgraded these as time progressed. My newest stuff are streamers and DACs. But my CD and tuner hail from the 1990's (I think). Denon DCD-S10 and Revox B126s.
Great equipment can stand the test of time with some TLC and minor improvements.
Great equipment can stand the test of time with some TLC and minor improvements.
A well designed singleton can often beat a poorly matched diff pair. Back in the 70’s and early 80’s the diff pairs were often so far out of balance that even with infinite loop gain there was a limit on how much distortion could be reduced. Feedback theory only works so far as A-B is really A-B and not some other quantity. EM induction put further limits on the distortion “floor” in many old designs. Get these right, then crossover distortion becomes the most obvious artifact - and it can then be attacked. Decent PNP transistors went a long way.
Hi wg_ski,
Well, you know I've always talked about the matching in diff pairs. When I was trained at Marantz, they knew about this and made certain they trained their techs to match those transistors. They sold them as sets. Same for drivers (compliments) and output transistors. Look up the Marantz 500. For the time it is extremely advanced. Tell me what you think.
Better PNPs, yes. Better transistors in general made a huge difference. We had to learn how to design with transistors. Yes, as the noise floor dropped we learned about induced signals and pickup in other areas of circuitry. One improvement drives another. Absolutely right too. Crossover distortion was major problem back then,. but there were worse issues to deal with at the same time. We have learned so much since.
Singleton inputs normally do not have stable DC offset. They are also class A and have no cancellation for that.
Well, you know I've always talked about the matching in diff pairs. When I was trained at Marantz, they knew about this and made certain they trained their techs to match those transistors. They sold them as sets. Same for drivers (compliments) and output transistors. Look up the Marantz 500. For the time it is extremely advanced. Tell me what you think.
Better PNPs, yes. Better transistors in general made a huge difference. We had to learn how to design with transistors. Yes, as the noise floor dropped we learned about induced signals and pickup in other areas of circuitry. One improvement drives another. Absolutely right too. Crossover distortion was major problem back then,. but there were worse issues to deal with at the same time. We have learned so much since.
Singleton inputs normally do not have stable DC offset. They are also class A and have no cancellation for that.
Hi Mister Audio,
Yes. The 510, 510M. I'm very familiar with them and was authorized to repair those amps along with the 500. Those were "factory service only" products, the normal service network wasn't approved for those. Same for the 2500 receiver. It used the tunnel but the outputs were not stacked (newer design). I hear the 2600 was a 2500 with increased rails for more power. I don't think they were sold in Canada but I could be wrong.
Yes. The 510, 510M. I'm very familiar with them and was authorized to repair those amps along with the 500. Those were "factory service only" products, the normal service network wasn't approved for those. Same for the 2500 receiver. It used the tunnel but the outputs were not stacked (newer design). I hear the 2600 was a 2500 with increased rails for more power. I don't think they were sold in Canada but I could be wrong.
The device matching is necessary to get past that .05% hurdle. But I’ve worked on designs that even used matched (6-legged, now unobtainium) diff pairs that had no provision for enforcing DC balance and some operated with a 20-30% difference in current. A lot of these stupid circuits floating around on the internet are just as bad - and the poor transistor choices make it that much worse. Performance there, even with a monolithic matched set is going to be worse than just pulling two random KSA992’s out of the bin and putting them into a circuit which DOES enforce balance. Again, one improvement drives another. The guys at Marantz obviously knew this. If the front end is working properly, ALL forms of distortion are reduced.
The need for matched outputs/drivers can be designed out, at least for the most part. To get past the .005% hurdle maybe it can’t - it all depends on where you really need to be. When it matters to me, I just go to EF3, where the gain of the VAS is more dependent on the VAS’s Early voltage than on the gain of the output followers. When eking out the same residuals out of EF2 or equivalent, you need to go pining over which output/driver transistors are used. Result: choose carefully.
I don’t mess with singleton inputs unless the output is cap coupled. DC problem solved. I’ll be damned if I hear the effect of a 6800 uF output cap when using an 8 ohm bookshelf or desktop speaker. I’m sure it can be measured, but it’s in the “so what” category. As an exercise a few years back I wanted to see what could be done with that old classic Philips topology. You know, with the PNP input, boot strapped VAS and darlington outputs. Just apply the same attention to detail: CCS-load both gain stages, use as much beta as you can muster in the singleton input, cascode the VAS, go to EF3 output and optimally bias, watch that ground return current from the power stage. The difference is shocking. One does not need to use 252 current mirrors and design the output capacitor out. For bass duty or a lot of wattage, yeah, I’d want to get rid of the cap. But I did mention driving a small pair of near field speakers….
Hi wg_ski,
Yup, completely agree with you. I worked on many English designs, like early Creek and others. Singleton city!
When I upgrade a design I do many things. For example, a Marantz 300DC I just did now has 0.003% THD on the bad channel, 0.0022% on the better one. Many things matter at that point. Same ballpark as the Yamaha PC-2002 when done. If I went nuts and matched outputs and drivers it would be better, it is using the original outputs. Interestingly, it uses a dual JFet diff pair. The old 2SK109.
The jig I designed and has been published here can match signal transistors far better than dual transistors on a single die parts. I can't match the thermal coupling, but way better than say a 2SA979 or similar. I still have some of those other six legged critters on hand. I'll be honest, using a well matched monolithic pair is less expensive than hand matching, and modern ones can be had with closer matches. Even up here in the frozen north where exchange, duties and shipping dampens my enthusiasm.
I miss the 2SA970 and similar parts.
Yup, completely agree with you. I worked on many English designs, like early Creek and others. Singleton city!
When I upgrade a design I do many things. For example, a Marantz 300DC I just did now has 0.003% THD on the bad channel, 0.0022% on the better one. Many things matter at that point. Same ballpark as the Yamaha PC-2002 when done. If I went nuts and matched outputs and drivers it would be better, it is using the original outputs. Interestingly, it uses a dual JFet diff pair. The old 2SK109.
The jig I designed and has been published here can match signal transistors far better than dual transistors on a single die parts. I can't match the thermal coupling, but way better than say a 2SA979 or similar. I still have some of those other six legged critters on hand. I'll be honest, using a well matched monolithic pair is less expensive than hand matching, and modern ones can be had with closer matches. Even up here in the frozen north where exchange, duties and shipping dampens my enthusiasm.
I miss the 2SA970 and similar parts.