SPEC 1 Frequency Response Guidance Required

AP Self Test

Here are the Audio Precision self test results.
 

Attachments

  • Self-Test, 1 001.jpg
    Self-Test, 1 001.jpg
    81.6 KB · Views: 14
  • Self-Test, 1.jpg
    Self-Test, 1.jpg
    141.7 KB · Views: 14
Forgive me if this has already been said earlier in the thread (because I haven't read all five pages), but a conventional volume control has an output impedance that varies with the position of the wiper. The output impedance is equal to the upper half of the track in parallel with the lower half, so the maximum output impedance occurs when the two halves are the same, ie with the wiper at the electrical half-way point. So, for a 10k pot the maximum output impedance is 2.5k. The minimum output impedance is zero.
If the performance of the stage following the volume control is dependant on a particular source impedance, then the performance will vary with the position of the volume control. This is quite common.
 
Forgive me if this has already been said earlier in the thread (because I haven't read all five pages), but a conventional volume control has an output impedance that varies with the position of the wiper. The output impedance is equal to the upper half of the track in parallel with the lower half, so the maximum output impedance occurs when the two halves are the same, ie with the wiper at the electrical half-way point. So, for a 10k pot the maximum output impedance is 2.5k. The minimum output impedance is zero.
If the performance of the stage following the volume control is dependant on a particular source impedance, then the performance will vary with the position of the volume control. This is quite common.

Sorry, when you write "impedance", do you mean "resistance"?
 
For those of you who have been following this thread, a brief summary of events:

  • Established that my fully recapped SPEC 1's frequency response is only to spec when at full volume.
  • AK member laatsch55 confirmed that a completely different and also fully recapped SPEC 1 in his possession has a frequency response just like mine.
  • I traced the signal from the input to where it goes through the volume control and found that is where the frequency response deteriorates.

FINALLY had time to see if my suspicions from the above tests were correct.

Here is a pic of the volume control:
attachment.php

Pin 2 is confirmed as the wiper.

  • I disconnected the other ends of the wires that go to pins 2 and 3.
  • I fed the function generator signal into the wire that goes to pin 3.
  • I connected the scope to the wire that goes to pin 2.

Here's what happened:
attachment.php

Black - frequency generator into scope
Blue - frequency generator through volume control. Max volume (0 ohms)
Red, Green, Purple - frequency generator through volume control. Turned down (resistance increased) one notch per color.

When I started this journey, I really didn't expect this result. I figured that perhaps I had done something wrong during my recap but it seems that for some reason the volume control in the SPEC 1 introduces this "flaw" where you only get maximum frequency response at full volume. I guess there must be some capacitance in the SPEC 1's volume control.

How about that? I suppose I'll put it back together and go back to enjoying it.
 

Attachments

  • SPEC 1 Volume Control.jpg
    SPEC 1 Volume Control.jpg
    114.8 KB · Views: 53
  • Volume Control - Segment Closest to Shaft.jpg
    Volume Control - Segment Closest to Shaft.jpg
    70 KB · Views: 54
Last edited:
Do the same with a conventional pot for that wafer. See if it's any different. Same end to end resistance - just different element type.
 
Do the same with a conventional pot for that wafer. See if it's any different. Same end to end resistance - just different element type.

Mark, I tried two 50k pots I had "laying around". The results were fairly similar to the SPEC 1 volume control.

Just for fun, I tried two resistors in series (10k and 47k) like this:

gnd-47k-S-10k-signal generator

I connected the scope where the "S" is. There was still some high frequency rolloff but not as much as a pot.

Just for fun, I also tried a single 1uf cap between the signal generator and the scope and there was no rolloff at high frequency.

Seems every pot your audio signal travels through eats away at the high frequencies a bit.

Anyway. Turkey season is on it's way and the garden and yard are demanding my attention so I buttoned everything back up and put it back into the rack. Won't have much more time for electronics till winter...
 
If you want to straighten out the design, so it functions correctly, look me up next round.
happy farming.
 
The problems are many in the Spec1, not the least of which are about a dozen caps from one end to the other- the phase shift was hideous on the two examples I had. Put a square wave into it and weep.

It's a collector's piece IMO, not a good example of design.

The four ganged volume pots were IMO only done to make the residual noise less intrusive at volume settings in the normal range. IIRC, Yamaha did it in the C-80 too, as did a bunch of integrated amps in the 1970s.


If you want to straighten out the design, so it functions correctly, look me up next round.
happy farming.
 
Mark, I tried two 50k pots I had "laying around". The results were fairly similar to the SPEC 1 volume control.

Seems every pot your audio signal travels through eats away at the high frequencies a bit.

Not quite true. It's the interaction between the potentiometer and the rest of the circuit that causes the high frequency roll-off. A lower value of pot would help, but the previous bit of circuit would have to be able to drive it, so changing it might cause other problems.
 
The most interesting thing about this thread is thinking on the best way to fix it.

I'm thinking there are several ways it could be tackled. I'm tending towards using the front end gain switch settings to allow the most rotation of the volume pot to be the easiest and only non-intrusive mod. That would allow the volume pot to enjoy maximum rotation, minimal noise, maximum dynamic range, and flat frequency response.

All my other ideas involve additional buffer stages or design changes.

I also believe the high (2v) output of a CD player (or modern line source) contributes to the issue by forcing the user to have the volume control at 8-11 o'clock vs the 2-4 O'clock you'd get from a standard (150mV or less) input source.

I say, use the gain control switch to give you volume pot range. It only attentuates after the initial buffer stage and won't affect FR.
 
the spec1 has a huge amount of dynamic range with its high supply voltages, so over load is a non issue.
like i said earlier, try to use only one section of the quad pot/attn. disable/pass through one of the sections, by lifting ground and shorting the wiper to the cw terminal. so have to figure what section is causing the early roll off.
i say to leave the input section as is and disable the section that is used later in the chain.
did anyone do a freq response test with the gain set to max and stepping through the input attn switch settings?
 
You missed my point I think. The switch-able gain will allow the volume pot to be used in a position where it doesn't adversely affect the FR.

With a high level input, like a CD player, and not using the gain switch to attenuate the front end, will result in the volume pot being at the lower end and he will get more roll-off. He needs to use more rotational range to minimise the roll-off.

Overload was not mentioned or inferred.

We know which section is causing the roll-off- that has been established. It is not the volume pot gangs placed after the input buffer stage, it is the gangs placed between the filter stage and the output buffer stage.

Note, the schematic for the final buffer stage also is different to the component values on the PCB. R65 and R60 are 110k ohms whereas on the large schematic, they are listed as 220k. The input SEPP buffer has matching 220k pairs. (R7/12, R8/11)

This seems very strange to me...






the spec1 has a huge amount of dynamic range with its high supply voltages, so over load is a non issue.
like i said earlier, try to use only one section of the quad pot/attn. disable/pass through one of the sections, by lifting ground and shorting the wiper to the cw terminal. so have to figure what section is causing the early roll off.
i say to leave the input section as is and disable the section that is used later in the chain.
did anyone do a freq response test with the gain set to max and stepping through the input attn switch settings?
 
You missed my point I think. The switch-able gain will allow the volume pot to be used in a position where it doesn't adversely affect the FR.
sure i get you now, i agree as a quick fix, use more input attn so that the volume setting can be set higher or as i suggested, bypass/defeat (VR1c,d)
it is the gangs placed between the filter stage and the output buffer stage.
Note, the schematic for the final buffer stage also is different to the component values on the PCB. R65 and R60 are 110k ohms whereas on the large schematic, they are listed as 220k. The input SEPP buffer has matching 220k pairs. (R7/12, R8/11
i will have to look at that again to see what's up.
okay looks to be a mistake with 110K instead of 220K, but in reality it does not make much diff since the diodes & r61(r62,2k4) set the bias or class "A" range. having 110K will lower buffer stage input Z. will it affect hi freq resp, I doubt it. none the less those R's should be symmetrical.
 
The parts list cross reference and PCB overlay also list them as 110k. The full schematic shows them as 220k. I'd like to know what is actually in the Spec.

I'm getting a Spec1 down to play with in the next few weeks as I love a good mystery.
 
I know this is an old thread, but I've just tested another Spec1 (original, untouched) and yes, it exhibits the same poor HF response. In fact, I'd call The Spec1 a flawed design by Pioneer. Remember they rated this Spec 1 as 10Hz-70kHz +0/-0.5dB and sold it as TOTL! (see specs)

With 150mV in the line (aux) input (rated input for 2.0v output), the only positions where the response is flat to 20KHz are "16,18,20" which correspond to between 12:30 and 2:00pm on the volume dial. This 150mV input however, results in only 174mV out, hardly the position where a typical power amp would be driven to full power. Consider Pioneer rated the unit at 150mV sensitivity for its rated output. Trouble is, at rated output and 150mV in, its high end response is dreadful, rolling off above 13kHz.

At full volume, the response is still rolled off by approximately 0.4dB at 20kHz and continues to dive until at 30KHz it is almost 3dB down.

At detent '22' (12o'clock high), we have a nice 2v out (unity gain for 2.0V in), but we are still 0.4dB down at 20kHz.

With 2v in (typical of modern CD/line sources at 0dB) there is only one position where a flat response can be obtained. (see below)

With 2 volt in (aux) we have the following:

Position 33, 0.576V out and -0.3dB at 20kHz
Position 37, ruler flat (!)
Position 30, -0.35dB at 20kHz
Position 42, +0.8dB at 20kHz (yes it is now rising at HF)

These are all with a 50k ohm load as Pioneer specified, although the output impedance is rated around 600 ohms.

By comparison, an inexpensive Yamaha CX-600 preamplifier from the 1980s gave me a response out to well above 50KHz before any roll-off was evident and a Sony TAE77es preamplifier was flat from 0.5Hz to well above 100kHz (it's rated to 300kHz).

I will say however, the tone control action (if anyone ever uses them) is the best I have seen with gorgeous curves not extending into places where they shouldn't. The low end response is not flat, but pretty good considering the massive number of caps the signal goes through.

None of that helps however, the preamplifier is not worthy of its reputation which is most likely based on build quality rather than actual performance. I've had three of them now and measured two- but they both exhibited the same issues.
spec 1 specs.JPG
 
Last edited:
In earlier threads, I proposed a fix. It is the two stage volume control which I believe (almost certain at the time) is causing the problem. One section needs to be shorted out or disabled. IIRC it is the back section. I do not have a spec-1 to test, so i can not physically test my fix.
I guess one could sim up the design, interesting exercise, it is a fair bit of work, i'd be doing it if I owned one. That will shed more light on the problem.
Now that I read
At full volume, the response is still rolled off by approximately 0.4dB at 20kHz and continues to dive until at 30KHz it is almost 3dB down.
Need to trace the signal through the circuitry and determine where this hi freq attenuation is happening. Compare again simulation.
I agree, the design as it stands is flawed!!
I just put a new power amp together, LME49830 with Semelab Lateral Mosfet ALF16N(P)16. It uses a SX-1050 power supply. I am getting -3dB at around 600KHz and that is because I have a RFI filter at the input.
 
Back
Top Bottom