Sansui AU-9500: Phono Stage RIAA Equalization Error
I ran these through my RIAA network calculator and here's what I came up with. RIAA spec is 50Hz bass shelf, 500Hz turnover and 2,122Hz treble cut.
The AU-999 is an outlier in that there's no resistor in parallel with the capacitor for the 50Hz time constant. The 20M is a placeholder for the formula. With a turnover frequency of 589Hz and no 50Hz bass shelf it would be expected to be Mid & Bass heavy. The schematic's a bit complicated so there must be further filtering going on downstream.
The AU-999 is an outlier in that there's no resistor in parallel with the capacitor for the 50Hz time constant. The 20M is a placeholder for the formula. With a turnover frequency of 589Hz and no 50Hz bass shelf it would be expected to be Mid & Bass heavy. The schematic's a bit complicated so there must be further filtering going on downstream.
Turns out the AU-999 (at least the early models?) have a sub-sonic filter installed between the pre and power amp boards. There's references to removing or bypassing this filter but for phono use I would probably leave in place or install a switch to bypass. One could also modify the RIAA network and leave out the sub-sonic filter.
Leestereo
Super Member
The lack of a resistor for the 3180µs (50Hz) transition point is equivalent to a very high resistance; hence a 20Mohm value for the RIAA calculations is acceptable. The measured RIAA response for the AU-999 phono stage from https://www.audiovintage.fr/2021/12/09/sansui-au-999-8/ confirms the mid and bass emphasis predicted by @WizardWG4.
stereofun
Super Member
The most competent criticism I've heard about the 9500, comes from kevzep.
If I recall, his issue was not the low-end, but the treble response, that seemed coarse or not "right" to his ear.
I've never heard one, but I've had a 7500 on my bench, for just a tune-up. I too had issues with it's top end, liked the clarity at first, but then after 5 min found it tiring.
------ aside from that, another fantastic thread by Leestereo. Not only identifying an issue, but showing owners how to get a near perfect curve. Damn, we are lucky to have experts like that, there ain't many left.
If I recall, his issue was not the low-end, but the treble response, that seemed coarse or not "right" to his ear.
I've never heard one, but I've had a 7500 on my bench, for just a tune-up. I too had issues with it's top end, liked the clarity at first, but then after 5 min found it tiring.
------ aside from that, another fantastic thread by Leestereo. Not only identifying an issue, but showing owners how to get a near perfect curve. Damn, we are lucky to have experts like that, there ain't many left.
Note: lengthy post follows.
I find it difficult to believe these were bonafide design errors. Not that the engineering staff at Sansui were infallible, but even without the aid of computer design and perhaps a slightly different understanding of the math involved then versus now (as with speaker design predating development and widespread adoption of Thiel-Small parameters) I somehow think these were intentional design decisions. Likely not ideal ones as we consider them from a contemporary point of view, admittedly.
Are there any known running changes to the RIAA EQ circuit in the AU-9500? It is the most glaringly unusual EQ curve of the bunch in that era. Enough so that it almost looks like a possible design error. If there were changes, that would support the idea that the original circuit was either done incorrectly, or that later units were at least changed in response to customer feedback over it "not sounding right". In searching about, I find no indication that running changes were made, FWIW.
In either above case, I expect a TSB would have been issued to document any such changes. I've never seen one, nor do I have any knowledge of whether there are or aren't, just throwing out ideas here. Obviously the phono EQ on the 9500 and it's smaller sibling models can be pretty easily modified for a more uniformly flat response curve as already demonstrated upthread.
Then too, is it certain the RIAA spec has never changed at all since it was first established? Not being a hardcore vinyl guy, I've never done a deep dive into the minutia of that. I know early on (1950's) there were competing phono EQ curves, have even seen at least one old Sansui tube era preamp with multiple labeled settings on a rotary switch for different curves that apparently were still in use. Clearly RIAA prevailed in that regard, essentially becoming an international standard as all the others look to have been abandoned.
I also wonder if these were intentional design choices since some of the early solid state integrated amps have overall frequency response curves that are clearly far from flat, from say 20Hz to 20kHz. There was a thread some time back here where this was discussed regarding (I think) the AU-505, or perhaps it was the 101. If not mistaken, I think @kevzep posted a graph in that thread of the frequency response profile, and it was no where near flat. Such a design now would be frowned upon by those desiring the 'wire with gain' discipline, but in the context of the time, may have been thought perfectly normal.
There appears to be some precedent for Sansui intentionally designing circuits that tend to favor an enhanced low to mid-bass response. We know this is a fairly well noted but subtle characteristic of the cap coupled era amp designs, and is part of what gives them their unique sound. It may not be accurate, but it is musical.
Might this have been done to compensate for typical deficiencies in the common speakers systems of the day? Or just an arbitrary decision made because those in charge of design and testing simply perceived it as sounding better at the time of product development?
It wasn't until the debut of the AU-919 that the company design philosophy clearly shifted to a more purist approach of endeavoring not to color the sound of the source signal from input to output. Subsequent topologies and designs by Sansui explicitly pursued this objective, culminating with the Alpha and it's slight sub variants. This was a mindset that took hold among most of the Japanese majors around that same time frame, though one might well argue Yamaha had already be striving for that previously with the "natural sound" designation on their product line.
I find it difficult to believe these were bonafide design errors. Not that the engineering staff at Sansui were infallible, but even without the aid of computer design and perhaps a slightly different understanding of the math involved then versus now (as with speaker design predating development and widespread adoption of Thiel-Small parameters) I somehow think these were intentional design decisions. Likely not ideal ones as we consider them from a contemporary point of view, admittedly.
Are there any known running changes to the RIAA EQ circuit in the AU-9500? It is the most glaringly unusual EQ curve of the bunch in that era. Enough so that it almost looks like a possible design error. If there were changes, that would support the idea that the original circuit was either done incorrectly, or that later units were at least changed in response to customer feedback over it "not sounding right". In searching about, I find no indication that running changes were made, FWIW.
In either above case, I expect a TSB would have been issued to document any such changes. I've never seen one, nor do I have any knowledge of whether there are or aren't, just throwing out ideas here. Obviously the phono EQ on the 9500 and it's smaller sibling models can be pretty easily modified for a more uniformly flat response curve as already demonstrated upthread.
Then too, is it certain the RIAA spec has never changed at all since it was first established? Not being a hardcore vinyl guy, I've never done a deep dive into the minutia of that. I know early on (1950's) there were competing phono EQ curves, have even seen at least one old Sansui tube era preamp with multiple labeled settings on a rotary switch for different curves that apparently were still in use. Clearly RIAA prevailed in that regard, essentially becoming an international standard as all the others look to have been abandoned.
I also wonder if these were intentional design choices since some of the early solid state integrated amps have overall frequency response curves that are clearly far from flat, from say 20Hz to 20kHz. There was a thread some time back here where this was discussed regarding (I think) the AU-505, or perhaps it was the 101. If not mistaken, I think @kevzep posted a graph in that thread of the frequency response profile, and it was no where near flat. Such a design now would be frowned upon by those desiring the 'wire with gain' discipline, but in the context of the time, may have been thought perfectly normal.
There appears to be some precedent for Sansui intentionally designing circuits that tend to favor an enhanced low to mid-bass response. We know this is a fairly well noted but subtle characteristic of the cap coupled era amp designs, and is part of what gives them their unique sound. It may not be accurate, but it is musical.
Might this have been done to compensate for typical deficiencies in the common speakers systems of the day? Or just an arbitrary decision made because those in charge of design and testing simply perceived it as sounding better at the time of product development?
It wasn't until the debut of the AU-919 that the company design philosophy clearly shifted to a more purist approach of endeavoring not to color the sound of the source signal from input to output. Subsequent topologies and designs by Sansui explicitly pursued this objective, culminating with the Alpha and it's slight sub variants. This was a mindset that took hold among most of the Japanese majors around that same time frame, though one might well argue Yamaha had already be striving for that previously with the "natural sound" designation on their product line.
I think this nugget of information is particularly telling, and may be the root of this conundrum.
After studying a bunch of eq networks they range from spot on to RIAA spec to pretty far off target. Generally speaking the Pioneer networks were some of the more accurate I've found. I guess it all depends on the skill of the engineer designing the phono stage. Also keep in mind they were using 5% or maybe even 10% tolerance parts. They probably figured close enough is good enough and moved on.
Hard to say since there are many other variables to consider upstream and downstream (turntable, cartridge loading, speaker response and room acoustics). Also, some may even prefer that mid bass boost. The way I see it is if we can mod the RIAA network to spec with 1% parts at least that leaves one less variable.
Leestereo
Super Member
I have had 4 different AU-9500 amplifiers on my bench and although there were some differences in their power supply configurations, all had the same phono stage RIAA equalization components. This is consistent with @Sansuiman's conclusion, but there may have been changes made that have yet to be found "in the wild" and/or that weren't included in documents made available publicly (e.g., TSB, revised schematic). Nevertheless, there is precedence for Sansui making changes to the RIAA equalization, the Sansui AU-777 variants:
Now that is interesting. I have an AU-777, but as of yet, have not specifically compared the phono stage to my other Sansui. Will put that on the to do list.
bolero
Super Member
I think a lot of ppl aren't even using the phono inputs: it's primarily line/aux/tape ins with media/cd sources. Although I certainly do use phono a lot.
So phono RIAA may not be a general stat to apply across the board, to each amps sonic signature? I guess also people who've been commenting on sound may not be using phono
So phono RIAA may not be a general stat to apply across the board, to each amps sonic signature? I guess also people who've been commenting on sound may not be using phono
Definitely one of the most profound posts/threads I've read on AK. As I listen mostly to vinyl. And I like to know the how as much as the why.
What struck me about this thread was it seemed to explain, at the theory and circuit leves, what is typically the attractive and nostalgic feature of vinyl - "the warmth". And as @Leestereo posted earlier in the thread "These RIAA equalization errors are not limited to Sansui phono stages, but rather are all too common in vintage phono stages designed prior to the late 1970s". This seems to suggest that on a mass scale across most hi-fi systems from this period the RIAA designs conditioned our ears to deviate towards "warmth" for vinyl playback.
For anyone at the 101 level of Hi-Fi, like myself, I highly recommend this very informative RIAA equalization Wikipedia page. It pairs well with a strong coffee
. And it explains the why-
"RIAA equalization is a form of pre-emphasis on recording and de-emphasis on playback. A recording is made with the low frequencies reduced and the high frequencies boosted, and on playback, the opposite occurs. The net result is a flat frequency response, but with attenuation of high-frequency noise such as hiss and clicks that arise from the recording medium."
The intention of the RIAA curve for playback was to tune the playback to the inverse of the recording to produce a flat frequency response. And in some cases, like the AU9500, there was an over correction in the low frequencies.
So it seems this "error" in many 70's era designs, due to "perpetuation in print of incorrect formulae for the design of the RIAA equalization networks", resulted in, for some and maybe most, the fond experience of warmth from vinyl.
Lingering question, as some have alluded too - Did Sansui and other designers "tune" for this warm sound signature as well?
Great thread. Thanks @Leestereo and all.
What struck me about this thread was it seemed to explain, at the theory and circuit leves, what is typically the attractive and nostalgic feature of vinyl - "the warmth". And as @Leestereo posted earlier in the thread "These RIAA equalization errors are not limited to Sansui phono stages, but rather are all too common in vintage phono stages designed prior to the late 1970s". This seems to suggest that on a mass scale across most hi-fi systems from this period the RIAA designs conditioned our ears to deviate towards "warmth" for vinyl playback.
For anyone at the 101 level of Hi-Fi, like myself, I highly recommend this very informative RIAA equalization Wikipedia page. It pairs well with a strong coffee
. And it explains the why-"RIAA equalization is a form of pre-emphasis on recording and de-emphasis on playback. A recording is made with the low frequencies reduced and the high frequencies boosted, and on playback, the opposite occurs. The net result is a flat frequency response, but with attenuation of high-frequency noise such as hiss and clicks that arise from the recording medium."
The intention of the RIAA curve for playback was to tune the playback to the inverse of the recording to produce a flat frequency response. And in some cases, like the AU9500, there was an over correction in the low frequencies.
So it seems this "error" in many 70's era designs, due to "perpetuation in print of incorrect formulae for the design of the RIAA equalization networks", resulted in, for some and maybe most, the fond experience of warmth from vinyl.
Lingering question, as some have alluded too - Did Sansui and other designers "tune" for this warm sound signature as well?
Great thread. Thanks @Leestereo and all.
Last edited:
bolero
Super Member
Yes, I think RIAA serves to increase the volume of music content vs surface noise that is inherent in vinyl playback too. Oh, I see that is listed above!
As well as allowing bass content that would otherwise limit one side of an LP to a few minutes, without RIAA since the grooves would be wider.
As well as allowing bass content that would otherwise limit one side of an LP to a few minutes, without RIAA since the grooves would be wider.
stereofun
Super Member
Whats funny here is that sansui went in the complete opposite direction with the 999, No loudness for one, but also a devilish sub-sonic filter on some models, which really restrict bass sub 100 hz.
Then in the successor 9500, they design lots of low-end including in the phono stage???
It's hard for me to believe in errors when it comes to the sound - I am with Sansuiman on this, especially since Sansui over and over tout their "listening tests" or "show their listening rooms". They must have heard the "bassiness" and said "yep, this is what we want".
If this is true - the deliberateness - then why the stark difference between two models so close in time.
Were they searching for a new sound - leaving the era of tubes behind? ( so well expressed by their cap-coupled designs)
...First going for a balanced lean "professional" 999, then finding it a bit "straight", before pivoting the other direction?
I dunno.
But I love the fact that we can have these discussions and that we have - not just a brand on our hands - but a living breathing company of people who over a generation of time build music pieces with different ideas and philosophies - characteristics which they expressed in their physical designs as well as internal designs. As such they are much more than just amplifiers.
When tech's today device modifications, it's like a chef revisiting an old recipe saying "hey guys, how'd you like this take - better eh?"
Right in the spirit of the original engineers.
Then in the successor 9500, they design lots of low-end including in the phono stage???
It's hard for me to believe in errors when it comes to the sound - I am with Sansuiman on this, especially since Sansui over and over tout their "listening tests" or "show their listening rooms". They must have heard the "bassiness" and said "yep, this is what we want".
If this is true - the deliberateness - then why the stark difference between two models so close in time.
Were they searching for a new sound - leaving the era of tubes behind? ( so well expressed by their cap-coupled designs)
...First going for a balanced lean "professional" 999, then finding it a bit "straight", before pivoting the other direction?
I dunno.
But I love the fact that we can have these discussions and that we have - not just a brand on our hands - but a living breathing company of people who over a generation of time build music pieces with different ideas and philosophies - characteristics which they expressed in their physical designs as well as internal designs. As such they are much more than just amplifiers.
When tech's today device modifications, it's like a chef revisiting an old recipe saying "hey guys, how'd you like this take - better eh?"
Right in the spirit of the original engineers.
I'm surprised the eq differences between the AU777 and AU777a are so great. Having much personal experience with the 777a, I considered it a flat sounding amp with nearly no deviation from 0. I have never heard or seen a 777 in action. Until now, I thought the two amps were more similar.
My guess is they were engineering their sound for market taste rather than audio accuracy. If they perfected the RIAA curve in 1970 with the 777a/d, then that bridge had been crossed early.
My guess is they were engineering their sound for market taste rather than audio accuracy. If they perfected the RIAA curve in 1970 with the 777a/d, then that bridge had been crossed early.
Last edited:
If listening tests had something to do with this, it makes you wonder if the listening rooms were 'treated acoustically' neutralised? Was the room a typical Japanese room, European room, US room ? - and whose speakers did they use? - AFAIK Sansui speakers were not known for their neutrality. So if they made tweaks to the amplifiers having made mistaken conclusions about how the amplifier sounded in the listening room, perhaps that was it. 
Cartridge resistive load (typically 47K for MM) and capacitance load (tonearm wire + interconnects + preamp load capacitor) can have a significant effect on frequency response. The AU-999 for example offered selectable input impedance on the Phono 2 input. Higher end gear would sometimes also offer selectable input capacitance. There are many pieces to the puzzle and they're all synergistic.
From the Pioneer C-21 Preamp manual:
