SPEC 1 Frequency Response Guidance Required

Recently picked up a piece of equipment that will allow me to plot frequency response versus amplitude (bode plot).

I noticed the output of my SPEC 1 (recapped last winter) was about .5 dB down at about 20 kHz and trailed off from there. According to the specs, it isn't supposed to be .5 dB down until about 70 kHz.

The audio signal is flat out to 100 kHz from the input until where it leaves the Filter Amplifier Assembly (FAA) on pins 26 and 27. From there, it goes through the volume attenuator to pins 1 and 22 on the Tone Amplifier Assembly (TAA). On pins 1 and 22 of the TAA, I can see the roll off at 20 kHz. Now here's the strange thing. If I turn the volume control up to max, the roll off is gone and the bode plot is flat to 100 kHz. The only thing I'm aware of between pins 26 and 27 of the FAA and pins 1 and 22 of the TAA is the volume attenuator.

Nothing seems abnormal when I operate the volume control. No scratchiness or anything unusual that would lead me to believe there is something wrong with the attenuator or that it's dirty.

So, what gives? Do I need to open the attenuator and clean everything and reassemble or is there something I'm missing? Is it possible Pioneer obtained the frequency response specs for the SPEC 1 by having the volume control at max?

Oh, and by the way, I'm not looking for a response of: "Why do you care? You can't hear it anyway!". I care.

The response of preamps is always measured with the volume control at max.

You have your order mixed up a little.
The order is Tone, Filter, Buffer stages. The volume control is a double ganged four element unit and the volume control is before the tone and an in between the filter and buffer stages. The balance control is between the tone and filter stages.

I would expect you are seeing normal behaviour, as the filter assembly would be loaded down by the volume control at its low level settings.

That said, I agree with you that a frequency response should be the same across all volume settings. Sadly, this is rarely the case and the primary reason frequency response is pretty much always measured at volume max. The same happens with input level controls on power amps- they affect the HF response as soon as you pad them down in level.

Thanks!!! Exactly the info I was looking for. I know the SPEC 1s volume control is a multi position switch with carbon traces printed on multiple wafers. You wouldn't think a carbon trace woud have that much affect on upper frequency response. A wire wound resistor perhaps because of the impedance but a simple carbon trace? Go figure.

To keep the sweep range of the volume knob reasonable and improve the signal to noise ratio, I typically have the pots on the SPEC 4 at half volume and the attenuator switch on the SPEC 1 at -15 dB. If I have the pots on the SPEC 4 at max and the attenuator switch on the SPEC 1 set to -30 dB then high end frequency response will be a bit better but that would also likely increase the noise level.

I'm still hoping AK member laatsch55 has a chance to run a bode plot on one of his restored SPEC 1s. Lee, if you get a chance to do this, please run the bode plot with the volume control at about half and let me know what you get. I see a .5 dB drop at 20 kHz (approximate) and then it trails off more or less logarithmically after that.

Like it says in your signature: "Some things are better left unsaid..."

I happen to have a SPEC-1, and I love it. And now I wish I hadn't read this thread!

Guess I'll just have to remember that I can't hear it anyway ;-)

Have you run a test of the set-up, substitute a double female adapter for the spec-1 and see what your set-up reports.

When you are getting this stringent/picky you CAN start seeing test equipment limitations. So you control for them.

Then also, source impedance and load capacitance can also affect the readings.

exactly mtf, probably measurement error. as it is, if you are going to measure a bode plot, you need to know the amplitude & phase differences between the input and output signals. what is your instrumentation? and setup?
If you are going to test against the Pioneer spec's you better figure out their test setup first or you maybe chasing your tail.

markthefixer said:

Have you run a test of the set-up, substitute a double female adapter for the spec-1 and see what your set-up reports.

Click to expand...

Yes. Flat to 100 kHz.

rcs16 said:

exactly mtf, probably measurement error. as it is, if you are going to measure a bode plot, you need to know the amplitude & phase differences between the input and output signals. what is your instrumentation? and setup?
If you are going to test against the Pioneer spec's you better figure out their test setup first or you maybe chasing your tail.

Click to expand...

That's why I want Lee to use his very fancy tester to see if he gets the same results that I'm getting.

This is what I have:

http://www.vellemanusa.com/products/view/?country=us&lang=enu&id=524708

Just to confirm, you tested the response (-0.5dB@20kHz) after the buffer stage (at the RCA jacks) initially didn't you?

Why don't you run several sweeps at various volume settings and stack them so we can see the result.

Also, are you putting 150mV or so into the line inputs and using the 3v scale on your bode software?

i do not mean to poop on your instrument but i did notice a few things that limit its capabilities to test a spec1 pre-amp out.
I do not see a flatness spec in the instrument specs, so it should be tested with a accurate ac voltmeter. you can not trust its own voltmeter, esp when you do not know its specs.
do a plot of itself and see what the flatness deviation shows.
1% thd is of no use for thd/fft work
note, scopes are very inaccurate measurement tools both in frequency and voltage!

down the road from you, in newmarket, toronto surplus (on vivian side road/kennedy rd) is selling an old Amber 3501, for a decent price. i have two of them now, i bought a spares/extra unit from them a month or so ago, mine works too. amber 3501 does -100 dB (0.001%) thd+noise and it has a very accurate freq meter option too. it is currently listed ebay and their web site. on par with a hp 339a but much better build quality, check it out.
if you went in them with $400 cdn cash, i am sure they would go for it.
amber3501 flatness spec is, +/-0.2dB 1-100KHz, seems about what i have measured.

hey john, what instrument(s) do you use for your audio testing?

xlbee said:

Thanks!!! Exactly the info I was looking for. I know the SPEC 1s volume control is a multi position switch with carbon traces printed on multiple wafers. You wouldn't think a carbon trace woud have that much affect on upper frequency response. A wire wound resistor perhaps because of the impedance but a simple carbon trace? Go figure.

To keep the sweep range of the volume knob reasonable and improve the signal to noise ratio, I typically have the pots on the SPEC 4 at half volume and the attenuator switch on the SPEC 1 at -15 dB. If I have the pots on the SPEC 4 at max and the attenuator switch on the SPEC 1 set to -30 dB then high end frequency response will be a bit better but that would also likely increase the noise level.

I'm still hoping AK member laatsch55 has a chance to run a bode plot on one of his restored SPEC 1s. Lee, if you get a chance to do this, please run the bode plot with the volume control at about half and let me know what you get. I see a .5 dB drop at 20 kHz (approximate) and then it trails off more or less logarithmically after that.

Click to expand...

XL, I'll run 4 FR sweeps at different volume levels....not a problem...give me a few more days, gotta get it outta the rack..

rcs16 said:

i do not mean to poop on your instrument but i did notice a few things that limit its capabilities to test a spec1 pre-amp out.
I do not see a flatness spec in the instrument specs, so it should be tested with a accurate ac voltmeter. you can not trust its own voltmeter, esp when you do not know its specs.
do a plot of itself and see what the flatness deviation shows.
1% thd is of no use for thd/fft work
note, scopes are very inaccurate measurement tools both in frequency and voltage!

down the road from you, in newmarket, toronto surplus (on vivian side road/kennedy rd) is selling an old Amber 3501, for a decent price. i have two of them now, i bought a spares/extra unit from them a month or so ago, mine works too. amber 3501 does -100 dB (0.001%) thd+noise and it has a very accurate freq meter option too. it is currently listed ebay and their web site. on par with a hp 339a but much better build quality, check it out.
if you went in them with $400 cdn cash, i am sure they would go for it.
amber3501 flatness spec is, +/-0.2dB 1-100KHz, seems about what i have measured.

hey john, what instrument(s) do you use for your audio testing?

Click to expand...

Poop all you want. $ is very tight for me and there isn't an opportunity to spend more. Again, that is why I am asking another member with more pro equipment to post his findings. So far I am seeing lots of speculation in this thread and not much in the line of actual test results. There was a member of AK (RIP) who's tag line was "More data. Less wank." So, either:
- there is something wrong with my device
- there is something wrong with my method
- there is something wrong with my SPEC 1
- there is nothing wrong and other SPEC 1s measure the same

So, which is it?

And don't you dare tempt me to go to Toronto Surplus! Too many things there I want!

The plot of "itself" or just an audio cable or just a connector is flat to 100 kHz.

restorer-john said:

Just to confirm, you tested the response (-0.5dB@20kHz) after the buffer stage (at the RCA jacks) initially didn't you?

Why don't you run several sweeps at various volume settings and stack them so we can see the result.

Also, are you putting 150mV or so into the line inputs and using the 3v scale on your bode software?

Click to expand...

The frequency response was flat to 100kHz up until where the signal leaves AWM-082 on pins 26 and 27.

Once the signal enters AWG-035 on pins 1 and 22: if the volume control is at maximum, the response is flat to 100kHz, if the volume control is turned down, the response is .5dB down at 20kHz.

Posting some sweeps with settings is a good idea. I will try to get some time tonight or tomorrow...

laatsch55 said:

XL, I'll run 4 FR sweeps at different volume levels....not a problem...give me a few more days, gotta get it outta the rack..

Click to expand...

Awesome Lee. You rock! Looking forward to it.

So, which is it?

Click to expand...

from what I have been reading, I suspect, one or two of the following
- there is something wrong with my instrumentation
- there is something wrong with my method(s)

your TE must be more accurate (they say 10X) than the unit under test or your results will be wrong or erroneous.

Yes. Flat to 100 kHz.

Click to expand...

When you say this, it is meaningless, there must be a min and max level, it depends on the resolution and accuracy of the measuring instrument however.

The plot of "itself" or just an audio cable or just a connector is flat to 100 kHz.

Click to expand...

Loop the sine wave signal source back to itself(input), using one of your test cables. if the i/o's are bnc, then use a 50 or 75 ohm bnc cable.
There is a flatness spec for your instrument, you just have to measure it correctly, using the right instrumentation.

I suspect the flatness spec of your instrumentation is much worse than the spec1, so you will not be able to measure it correctly.

150 millivolts RMS output from function generator in all cases:

Function Generator into CH1 input on PCSGU250


Pin 27 AWM-082: Two sweeps. Volume min and max.


Pin 1 AWG-035: Three sweeps starting with max volume.

A Panasonic VP7723A (to 100kHz), Leader LDM-171 (for S/N to -80dB and THD to 0.002%), AudioLab (sweeps/true RMS VM/impedance/capacitance/inductance etc- out to 50kHz/line and mic front end (for speakers)- PC interfaced, but I have to maintain a Win95 machine just for that, as it needs direct serial port access), 3 cros, a Kenwood FL-140 W&F meter, 2.7GHZ counter, audio sig gen, a pile of true RMS meters and a stack of dummy loads. I can do IM with a custom front end for VA, but I find it too fiddly and results are not consistent. Phew. And a Leader curve tracer.

rcs16 said:

hey john, what instrument(s) do you use for your audio testing?

Click to expand...

the top trace of the generator loop backed to itself looks like +/-0.5dB, which is fine.
second traces, buffer amp, looks better than the generator itself, I wonder why?
third trace with the tone control/filters, in circuit, lower volume settings have reduced hi-freq response. Stuff they do not document maybe?
so freq resp spec, assumes volume set at near max. i guess this is your concern?
did you measure gain/phase?
the design is unique, 2-stage volume control, signal passes through the tone and filter ckts, no bypassing
with graphs it is hard to see mins and maxs, i guess you have to zoom in or maybe the software reports this information?
the schematic is a bit hard to follow without printing it out.
i'd say your unit looks okay except for the hi-freq roll-off in the tone circuit.
easy for you to test the filters/tone stuff too.

thx john, did not expect that much typing = thx

I would focus your attention on the buffer amplifier. Clearly it doesn't like its front end being dragged close to ground (low volume setting) and that is affecting its HF response. You definitely replaced the two 3.3uF non polarized caps in each line buffer amp with the right value when you recapped?

Actually, I have noticed something rather strange. The block diagram and the circuit diagram show two completely different wiring setups for the volume pot. On the block diagram the pot is used to short the signal to ground with a variable load on the previous stage, on the schematic, the pot is used as a constant load for the previous stage, with the next stage tapping off signal.

If it is wired as per the block diagram, the fall off in HF will occur at low volume settings.

block diag

schem

Basically the only difference between the second and third plot is the volume control. Plot 2 is at the input to the volume control. Plot 3 is the output of the volume control. I'm just really surprised at how much the SPEC 1 volume control seems to roll off the high frequencies.