500C Loudness Contour – Modeling and Choosing Component Values

[daylite]

-The Fisher 500C loudness contour circuit was modeled in SPICE.
-Choices for loudness circuit components are discussed.

For my 500C project I am considering modifying the Fisher 500C loudness switch circuit. It is a passive filter with three* components: C63, C64, and R91. To get a better idea of what the filter might be capable of, I modeled the circuit with SPICE. *(The volume pot with tap is a fourth component not modeled here other than to employ resistors representing a single volume setting.)

The three figures below show the effects of about 24 combinations of component values. For each figure one of the components is held constant, while four values of a second component are tried with each of two choices of the third component. This gives eight component combinations in each figure.

I. Effects of C63 (C64=0.02uF, R91=22K and 68K)
II. Effects of R91 (C63=120p, C64=0.02uF and 0.082uF)
III. Effects of C64 (C63=180p, R91=22K and 68K)​

I.

II.

III.

The observed effects are in line with what Dave described in this 2016 post. A calculated curve for a set of components may have systematic inaccuracies - there has been no instrument-based testing of the model.

I'm posting these results in case other readers may find them useful. They'll be useful if the observable trends help us predict the effect of a component substitution and estimate a magnitude for that effect. With these results as a guide I’m experimenting with loudness contour component changes. -d (Saturday tomorrow...back to the bench! )

 

[dcgillespie]

Day -- What a super job you did in generating and presenting this information!! Thanks so much for the effort -- and sharing it!

Dave

 

[Hannes123]

very nice and interesting although i don´t understand how to read the diagrams. ^^
Is there a combination that would be the most "desirable"?

 

[daylite]

I hope to get back to you by the weekend with more specific information. This is to let you know I see your question and I’m thinking about the answer(s). Two things occur to me as an introduction:

First – What’s ‘desirable’ or ‘sounds good’ is really a personal choice. I’m at the early stages in exploring loudness contour adjustments, but I’m willing to share what’s sounding good to me at present (and why) in upcoming posts.

Second –I confess I overused equalizers and tone controls in the past. I think I’ve gotten over that and have come to think the less of that, the better. (I love the ‘tone control bypass’ often discussed in this forum.) Loudness, too, can be overused. Sometimes a complaint that loudness creates ‘muddiness’ or ‘boominess’ is due to misuse – for example by applying a loudness contour when the volume is too high. If ‘loudness on’ sounds muddy, you can turn it off. But first try turning down the volume. The sound may get better. The loudness control and the volume interact and must be adjusted together. There is feedback from the volume to the loudness circuit through the extra connection (the tap on the volume pot) but it’s crude. It won’t perfectly limit the effect to only the volume range where it sounds good.

There is no doubt a given loudness contour can be wrong for a listener even when used as well as possible. I am sure an unoptimized loudness contour can be improved to better work for the person, the room, the speaker setup, and the preferred listening volume. That’s what may have led Yamaha to offer a variable loudness contour feature. For more on this from others, look near this post. The entire thread (‘Loudness Controls’ in the General Audio Discussion forum) is a good source for background and provides some of the ideas, likes, and dislikes of a group of audio listeners.

I intend to write more later in response to your question, but in the meantime, others may decide to contribute their own thoughts. Thanks for writing. -d

 

[daylite]

Is there a combination that would be the most "desirable"?

Short answer:
Try C63/C106=180 pF, C64/C65=0.02 uF, R91/R92=47K. Remember, these are personal choices. There’s no ‘right’ value. I think C63/C106 at 180-220 pF boosts the treble properly. It’s not too much boost for me, though 220 pF might be approaching my ‘too bright’ territory.

Take-home advice for modders: first try increased capacitance for C63/C106 without any other component changed. I recommend trying 120 pF and then 180 pF in place of the 68 pF stock caps. And also try 220 pF if you’re an old guy (like me). Among those, choose the value that sounds best to you. Remember you can simply add 120 pFs parallel to the 68 pFs to get close enough to 180 pF to hear the effect. I found the capacitance increase greatly reduced what I think of as muddy/muffled sound. [I don’t know the vocabulary used for hi-fi sound. What I’m saying is “I like it and I think some of you would, too”.] -d

EDIT:
P.S. - Hannes123, thanks again for writing. I have changed the thread title to include the topic of your question.

 

[daylite]

Is there a combination that would be the most "desirable"?

Longer answer:
The loudness feature arose from the work on Equal Loudness Level Contours (ELLCs) pioneered at Bell Labs by Fletcher and Munson in the 1930s and redetermined by Robinson and Dadson in 1956. To create an ELLC test subjects are treated to a 1 KHz tone and a second tone of another frequency (say 10 KHz), both with the same ‘sound pressure level’ (aka ‘SPL’ – a physics sound wave measurement unaffected by our ears and brains) and are asked to turn up or down the 10 KHz tone volume until it ‘sounds as loud’ as the 1 KHz tone. The dB amount of change needed for equal subjective loudness is noted. Do that for lots of frequencies vs 1Kz, plot the results, and you have an ELLC (examples below).

In principle a ‘loudness contour’ is meant to cancel out the ELLC. For example if the ELLC tells us the average person needs an 18 dB boost at 8KHz, the loudness contour is meant to provide that boost. But it’s all based on averages and will change with changes in the test subjects that are chosen.

What about those of us who are non-average? For example, we all know many older adults (males especially) have diminished high frequency hearing. Wouldn’t they have a different ELLC compared to 30-somethings?

It turns out there’s a good professional report on that and it’s available for free download. Here’s a figure from it comparing younger vs older group averages for ELLCs (this also shows the industry-standard ISO ELLC curves):


A lot could be said about this besides “Sure enough – older folks need more treble boost”. I'll say only that IMHO the stock Fisher 68 pF value for C63 is probably not providing the amount of treble boost needed by many of us who listen at low to moderate levels. To see that, compare the stock curve as calculated (the lowest curve in the first graph of the first post in this thread) to the 30-60 dB ELLCs above. In case you're wondering - I looked into all this after I found more than 68 pF at C63/C106 sounded better to me (see previous post), and I wondered if there could be any reason behind the choice that could apply generally.

[BTW - given the changes that come with age, a contour that was set up to counteract old-age hearing changes might be a popular feature. But if we want that, we better build it into the loudness contour ourselves – no manufacturer is going to market a receiver with an ‘oldness contour’ button!! ] -d

 

[dcgillespie]

Great job Day. I also found one other aspect that influences all of this, and that's the recording medium used. Let's face it, conventional vinyl recordings made back in the day of (say) the "1812 Overture" have (by comparison) precious little LF information in the recording (at the places where it counts!), compared to a CD (or even a vinyl copy of which I have) of Telarc's recording of the same Overture. Point being that some efforts to install a loudness function were SOOOO heavy handed back in the day, that they are virtually unusable today -- this because the recorded LF information of yesteryear simply wasn't even in the same universe as the amount of recorded LF information today. The first stereo preamp of Fisher's -- the 400C was/is a perfect example of this. In my own 400C, I removed the old two-step loudness circuit, and replaced it with the single step circuit as found in the 400CX-2, which is eminently more enjoyable -- and usable!!

Dave

 

[daylite]

As explained above, for the 500C loudness contour I like C63/C106=180-220 pF, C64/C65 = 0.01 or 0.02 uF, and R91/R92 = 47K.

How was 47K chosen? For testing the effect of R91/R92, I replaced the 22K stock resistors with two small Amphenol Piher 100K pots (PT10LV10-104A2020-S):

With 220pF caps at C63/C106, with 10K resistance from the pot the music sounded muffled, and was still noticeably muffled at 22K-33K. The 47K value was chosen because with 220pF for C63/C106, when the resistance of the pot was dialed from low to high, 47K was about the point where things just became clear (for me). The higher resistance moves the minimum in the contour lower in the low midrange (too much energy in this range can cause muddiness and muffle high-frequency instruments), and raises 5 KHz compared to 1 KHz by 2 dB, possibly accounting for greater presence and enhanced clarity.

Here for comparison are three simulation plots. Left, R91=22K; Center, R91=47K (C63=200 pF, C64=0.02 uF for both resistance values); Right, the Fisher 500C stock loudness contour:

Fisher did seem to be most interested in the low frequency end. Now I can see that my personal choice does less of that low frequency boost, more for treble, and very little to the low midrange. -d

 

[daylite]

Dave - Thanks for your note. It was a good reminder for me that the source is the first link in the chain. I'll have to think about what source to use when playing with tone. I can see I might end up optimizing tone for something off-base.
I have to apologize - after you wrote I was reading some of my AK notes and rediscovered this post from you. In item #3 there you encapsulate most what I had to say above, even down to advice on low mids and mud. I guess if my post above has any remaining value it is only that it illustrates the point you made and offers some comment on the resistor effects. I'm sorry I didn't recall your comments when putting together the above. [P.S. Your post appeared Jan 10, 2020. Little did we know how the world was about to change...]

 

[fred soop]

The amount of compensation needed is not the actual curve but rather the DIFFERENCE between the curves as level is reduced. This has been a common error for many years. Very little, if any, compensation is needed at high frequencies. Of course, one may want it due to hearing problems.

Some amplifiers have a separate control, allowing for variable loudness. This is more flexible than the fixed circuit tied to the volume control.

 

[daylite]

Update: I completed configuring the tone control bypass circuit for this unit. What a revelation! I see my choices for boost/cut from the LC were being influenced by what may be a poor-functioning circuit around the Baxandall PEC boards. I'll have to test more, but in the meantime I have dropped C63/C106 to 150 pF and with Baxandall's bypassed the 500C is sounding great. So improved I want to re-listen to everything I have!

 

[BlueBlaze]

Update: I completed configuring the tone control bypass circuit for this unit. What a revelation! I see my choices for boost/cut from the LC were being influenced by what may be a poor-functioning circuit around the Baxandall PEC boards. I'll have to test more, but in the meantime I have dropped C63/C106 to 150 pF and with Baxandall's bypassed the 500C is sounding great. So improved I want to re-listen to everything I have!

Hi daylite. I'd like to try this mod on my 500b. The schematic shows the same original values being used in the B & C models (400 too). I could use some help, basic novice here.
Do you mind sharing exactly what replacement parts I'd need?

Thanks!
Steve

 

[daylite]

BlueBlaze - I did a post on the bypass that can be found here. Parts needed: an 8 pF capacitor and 1.2M resistor for each channel – so four parts total.

Here’s what I remember: To prepare for the mod, I disconnected the two caps connected to the low filter switch and joined them together. Do that for each side of the switch – left and right are the same. The low filter disconnects are illustrated in the post above. Then, disconnect what’s on the high filter switch, both sides. On the high filter switch the center had been connected to ground and one end (the ‘ON’ end) had been connected to terminal ‘2’ of the ‘high filter PCS’. The high filter PCS is a rectangular “big chiclet” with numbers 1-2-3 on it; in the 500C it was close to the high filter switch; there is one of them for each side of the high filter switch. When the lead to terminal ‘2’ is disconnected the terminal can be simply sealed off, or you can remove the entire PCS and connect what was connected to ‘1’ to what was connected to ‘3’.

That clears both switches. Both filters are now forever ‘off’. Check that you can still hear music! After that follow the ‘step-by-step’ in the above post to turn the high filter switch into a tone bypass switch. Best of luck! -d

 

[BlueBlaze]

BlueBlaze - I did a post on the bypass that can be found here. Parts needed: an 8 pF capacitor and 1.2M resistor for each channel – so four parts total.

Here’s what I remember: To prepare for the mod, I disconnected the two caps connected to the low filter switch and joined them together. Do that for each side of the switch – left and right are the same. The low filter disconnects are illustrated in the post above. Then, disconnect what’s on the high filter switch, both sides. On the high filter switch the center had been connected to ground and one end (the ‘ON’ end) had been connected to terminal ‘2’ of the ‘high filter PCS’. The high filter PCS is a rectangular “big chiclet” with numbers 1-2-3 on it; in the 500C it was close to the high filter switch; there is one of them for each side of the high filter switch. When the lead to terminal ‘2’ is disconnected the terminal can be simply sealed off, or you can remove the entire PCS and connect what was connected to ‘1’ to what was connected to ‘3’.

That clears both switches. Both filters are now forever ‘off’. Check that you can still hear music! After that follow the ‘step-by-step’ in the above post to turn the high filter switch into a tone bypass switch. Best of luck! -d

Thanks for the info daylite. I guess I should've replied to an earlier post in the thread. I was thinking I'd start with the loudness contour mod & go from there. Can you tell me what components you settled on? Are ceramic caps pretty generic, or do you have a preferred make? Size, voltages, tolerances?

Thanks again,
Steve

 

[daylite]

Steve: I am using 150pF. I recommend you get 100pF,120,150,180,220pF and that way you can find what you like most. Most ceramic caps would give identical results. Here's an example from Mouser and here's another example. 5% tolerance is fine. Here's a link to the 50K pots I used. -d

 

[BlueBlaze]

Steve: I am using 150pF. I recommend you get 100pF,120,150,180,220pF and that way you can find what you like most. Most ceramic caps would give identical results. Here's an example from Mouser and here's another example. 5% tolerance is fine. Here's a link to the 50K pots I used. -d

It's much appreciated daylite.

 

[thornev]

This is interesting. I was trying out 50k pots on the Loudness switch like Daylite did. I disconnected R91/92 from ground so I could get a pure result from just the 50k pot. What became apparent is that, with all the tone and volume controls the same, tone is clearer - treble is more crisp and bass is more prominent - and volume is noticeably louder when the Loudness switch resistors are out of circuit. Just doing that made an improvement in sound. In other words when the Loudness switch is off and in circuit, the sound is affected negatively. So now I have to decide if I can make the Loudness switch, when on, sound better with different resistors and caps than when the resistors are out of circuit. (I do however, with resistors out of circuit, have to turn up the bass higher to get the bass I like at lower volume. That's one thing I was hoping I could improve with different component values.).

EDIT: Above too wordy. What I did was sever R91 and R92 from ground to produce the improvement in sound. Doing so deactivates the Loudness switch, but I like it.

EDIT2: So question - Is there a way to affect bass and treble boost and bass frequency range outside of the loudness switch but for FM? If so, where in the circuitry would that be?

EDIT3: I thought I read that bridging a cap across C64/65 would affect the frequency range of the bass. Raising the total cap value would make lower bass notes louder. So I was expecting to hear lower/deeper bass notes when I bridged a 0.1uF across C64 which is .02uF. But what happened is that the bass boost was decreased.

I also tried the 50k trimpot bridged across C64 and got the same results - lower bass volume.

 

[baconbadge]

This is interesting. I was trying out 50k pots on the Loudness switch like Daylite did. I disconnected R91/92 from ground so I could get a pure result from just the 50k pot. What became apparent is that, with all the tone and volume controls the same, tone is clearer - treble is more crisp and bass is more prominent - and volume is noticeably louder when the Loudness switch resistors are out of circuit. Just doing that made an improvement in sound. In other words when the Loudness switch is off and in circuit, the sound is affected negatively. So now I have to decide if I can make the Loudness switch, when on, sound better with different resistors and caps than when the resistors are out of circuit. (I do however, with resistors out of circuit, have to turn up the bass higher to get the bass I like at lower volume. That's one thing I was hoping I could improve with different component values.).

EDIT: Above too wordy. What I did was sever R91 and R92 from ground to produce the improvement in sound. Doing so deactivates the Loudness switch, but I like it.

EDIT2: So question - Is there a way to affect bass and treble boost and bass frequency range outside of the loudness switch but for FM? If so, where in the circuitry would that be?

EDIT3: I thought I read that bridging a cap across C64/65 would affect the frequency range of the bass. Raising the total cap value would make lower bass notes louder. So I was expecting to hear lower/deeper bass notes when I bridged a 0.1uF across C64 which is .02uF. But what happened is that the bass boost was decreased.

I also tried the 50k trimpot bridged across C64 and got the same results - lower bass volume.

Thorne,
Remember that increasing the value of C64 (.02uf) lowers the frequency range, or window, that is boosted. So you might have pushed the window too low to hear what was being boosted (or what you were playing didn’t contain any of those low frequencies). So it sounded like the bass was being reduced because all the low mids that you’re used to hearing aren’t being boosted anymore. It’s a little counterintuitive.

You’d be able to see what’s actually happening using your scope to look at various low frequencies and plotting out how they’re affected by changing the values in the loudness circuit.

 

[thornev]

Thorne,
Remember that increasing the value of C64 (.02uf) lowers the frequency range, or window, that is boosted. So you might have pushed the window too low to hear what was being boosted (or what you were playing didn’t contain any of those low frequencies). So it sounded like the bass was being reduced because all the low mids that you’re used to hearing aren’t being boosted anymore. It’s a little counterintuitive.

You’d be able to see what’s actually happening using your scope to look at various low frequencies and plotting out how they’re affected by changing the values in the loudness circuit.

I tried different component values and I did get the result you mention when I used a value, something around 380uF I forget the exact value, that was too high. In that case the bass all but disappeared. But when I used a 0.1uF value, that was not the case because the bass was still audible, just not as loud.

If I can figure out how, I'll try using a scope to see if I can further prove my case.

 

[thornev]

OK. I used a scope. I sent sine and square waves to my 500-C, all at 1V, 50% duty, tested one channel with the bass and treble controls at 12 o'clock, Loudness control on. I tested 5 different bass frequencies. I tested with a 0.1uF film cap, 5%, 100V. That cap was bridged across C64 (or C65). Also, R91/92 have a 5.6k resistor in series making R91/92 = 28k instead of the stock 22k. Those 5.6k resistors serve to lower the amount of bass boost about 3dB.

I measured peak-to-peak voltage and RMS voltage. It isn't the case that any frequencies were cut out by the 0.1 cap with which I tested. The data shows a gradual decrease in amplitude as the frequencies tested are increased which I interpret to mean that the cap works well in decreasing the lowest bass frequencies and not as much on the higher bass frequencies. The 0.1uF cap does not cut out the lowest frequencies but does reduce their amplitude.

NOTE: Those cells that show a range of voltage values are because the scope could not settle on a single value. In those cases I recorded the lowest and highest values the scope displayed.

Anybody else care to add your analyses?