I was thinking this would be a precision attenuator
as opposed to a load.
So If I'm thinking correctly (sometimes that hard for me to do)
to measure a complete stereo system (less speakers for the moment)
I can run the amplifier right into a clarostat decade power resistor
box set to 8 ohms (240 watts) with the HP350D input in parallel
and the HP350D output to the HP339a or Shibasoku 725D or
QA400 for FFT use.
On the input side of the system, I can then do a couple of things
now. Use the Leader 1300s for its automated frequency sweep.
Or for more precision
HP339a for low distortion oscillator with the BK1607 for
the variable notch.
Then I can compare that part of the system with the
whole system with speakers and mic them
or
have the mic setting where the listener would be.
I have a pair of AT mics there were recommended to me
that are excellent for the money. I've also found another
funky JVC stereo mic that isn't too bad either.
Yes, I know there are other mics and specialized software
etc., just for this purpose, but I'm not there.
Interesting thing about the mics though,
I was going to buy some pick ups from
"The King of the Texas Blues and His Big-Band Horns".
But, when I talked with Ezra Charles,
the inventor of the piano pickup mic that Elton John and others
use, he recommended that I save my money and buy a pair
of Audio-Technica mics. They are non-stocked, special-order
only Audio-Technica mics and very good.
Link for the good ol' "King of the Texas Blues"
Hmm anyone have any recommendations for a low noise
phantom power source for mics? I have an old analog
tape system that I used to use for recording but not sure
where that is hiding....The other power supplies I have
for them would be great...but they don't work they are
on my to fix list. ; )
Hmmm, I'm always open to suggestions because this
is research for me; (i.e., I don't know what I'm doing).
... and the beat goes on.
as opposed to a load.
So If I'm thinking correctly (sometimes that hard for me to do)
to measure a complete stereo system (less speakers for the moment)
I can run the amplifier right into a clarostat decade power resistor
box set to 8 ohms (240 watts) with the HP350D input in parallel
and the HP350D output to the HP339a or Shibasoku 725D or
QA400 for FFT use.
On the input side of the system, I can then do a couple of things
now. Use the Leader 1300s for its automated frequency sweep.
Or for more precision
HP339a for low distortion oscillator with the BK1607 for
the variable notch.
Then I can compare that part of the system with the
whole system with speakers and mic them
or
have the mic setting where the listener would be.
I have a pair of AT mics there were recommended to me
that are excellent for the money. I've also found another
funky JVC stereo mic that isn't too bad either.
Yes, I know there are other mics and specialized software
etc., just for this purpose, but I'm not there.
Interesting thing about the mics though,
I was going to buy some pick ups from
"The King of the Texas Blues and His Big-Band Horns".
But, when I talked with Ezra Charles,
the inventor of the piano pickup mic that Elton John and others
use, he recommended that I save my money and buy a pair
of Audio-Technica mics. They are non-stocked, special-order
only Audio-Technica mics and very good.
Link for the good ol' "King of the Texas Blues"
Hmm anyone have any recommendations for a low noise
phantom power source for mics? I have an old analog
tape system that I used to use for recording but not sure
where that is hiding....The other power supplies I have
for them would be great...but they don't work they are
on my to fix list. ; )
Hmmm, I'm always open to suggestions because this
is research for me; (i.e., I don't know what I'm doing).
... and the beat goes on.
Last edited:
Here is the link to
Ezra Charles
Alumnus Spotlight: 'Ezra' Charles Helpinstill : Rice University George R. Brown School of Engineering
Ezra Charles
Alumnus Spotlight: 'Ezra' Charles Helpinstill : Rice University George R. Brown School of Engineering
If the internals are based on a "T" attenuator then Source impedance (Rs) does not affect the attenuation.
Rs does affect the impedance seen by the load.
If the LOAD requires to see an effective impedance of 600ohms, then the source impedance (or a dummy source across the input) at the INPUT must be 600ohms.
The "T" attenuation accuracy only requires the Load to be equal to the attenuator's impedance of 600ohms.
eg.
set the series resistors along the top of the T, to 100r
set the shunt resistor from the middle of the T to return to 1750r
add on the 600r load.
Vi = source emf.
Rs = source impedance.
V1 = input voltage at the attenuator Input.
V2 = voltage at the middle of the T.
V3 = voltage at the attenuator output = Vload
V1 = Vi * 600/(600+Rs)
V2 = V1 * 500/600
V3 = V2 * 600/700 ~ 0.714286*V1 ~ -2.9226dB
V3 = 500/700*V1 = (600-100)/(600+100)*V1 and does not depend on Rs.
Insert either 0r0 or 600r for Rs and the attenuation is always -2.9226dB
Last edited:
Did someone check out this notch filter?
Three-op-amp state-variable filter perfects the notch | EDN
Three-op-amp state-variable filter perfects the notch | EDN
Note the low input Z with values as shown; higher values may bring higher noise. He also notes that notch depth is limited by opamp characteristics, although it is theoretically infinite...
See Bob Cordell's state-variable notch filter in his outstanding distortion analyzer. The reprint is available on his web site at CordellAudio -- google it. As a side note, Tektronix used the Bainter filter this guy references in the AA501.... I am about to borrow an AA501 from an online friend to evaluate depth, residuals, noise, etc...
See Bob Cordell's state-variable notch filter in his outstanding distortion analyzer. The reprint is available on his web site at CordellAudio -- google it. As a side note, Tektronix used the Bainter filter this guy references in the AA501.... I am about to borrow an AA501 from an online friend to evaluate depth, residuals, noise, etc...
The Tektronix AA501 manual shows, page 3-2, a notch filter which looks more as a Multiple Feedback Bandpass filter than a Bainter filter.
There seems to have been a AA501 spanish clone named PROMAX DA-523,
it was shown and detailed recently in this french video-blog :
https://jipihorn.wordpress.com/2015...de-dun-distorsiometre-promax-da-523/#comments
I suspect that the SVF notch filter that I used is easier to auto-tune because with a SVF, Q and center frequency can be relatively independently set (corresponding to amplitude and phase for the notch cancellation). This permits in-phase and quadrature auto-tune detectors to be used whose outputs can be used to directly control variable gain JFET elements in the SVF.
I am not sure if the original AP used this same approach.
BTW, the variable gain JFET elements use the 50% gate feedback technique to minimize distortion, but I later discovered that 50% is not quite always optimal. I subsequently added trimmers for the gate feedback percentage.
Cheers,
Bob
I am not sure if the original AP used this same approach.
BTW, the variable gain JFET elements use the 50% gate feedback technique to minimize distortion, but I later discovered that 50% is not quite always optimal. I subsequently added trimmers for the gate feedback percentage.
Cheers,
Bob
It's been quite awhile, probably at least 2 years, and I don't recall definitely, but it was enough to make me go to the trouble of putting in the trim. Probably at least 3 dB. I also don't remember by what percentage I had to go away from 50% to get to the optimum, but I'm pretty sure it was less than 10%.
I think I first tried it on the oscillator agc JFET.
Cheers,
Bob
Hi Bob,
Its all on the 339A analyzer upgrade forum......... you said there you tried it after i did it to lower distortion of the 339A oscillator. I got a lot of canellation of the 2nd..... you get to almost zero (noise) but then 3rd would be too high. So, I usually balance the 2nd and 3rd levels and let it go at that.
THx-RNMarsh
Its all on the 339A analyzer upgrade forum......... you said there you tried it after i did it to lower distortion of the 339A oscillator. I got a lot of canellation of the 2nd..... you get to almost zero (noise) but then 3rd would be too high. So, I usually balance the 2nd and 3rd levels and let it go at that.
THx-RNMarsh
I don't recall seeing the third go up as I got to the optimum for the JFET's overall contribution to THD. Are you sure that the 3rd was not coming up for some other reason, even perhaps becoming more evident as the second disappeared? Did you see the actual magnitude of the 3rd increase on a spectrum analyzer?
Cheers,
Bob
There are a couple of possibilities that may depend on the design of the oscillator itself. IIRC, the 339 is not a state variable oscillator. Moreover, the amount of nominal signal amplitude impressed across the JFET may not be the same as in my oscillator.
The JFET control circuit will create 3H distortion that is of course not reduced by the gate feedback technique. However, the 3H amplitude may be influenced in some way by the amount of gate feedback. I also note that the amount of 2H percentage (whatever amount there is after reduction by the gate feedback) will go up linearly with signal amplitude, while the amount of 3H percentage will go up as the square of amplitude. This means that in an agc circuit where a larger signal is present across the JFET (as perhaps in the 339A circuit), 3H will be a relatively larger issue.
One possibility is that third harmonic cancellation is at work in the 339A. If so, a slight change in the JFET gate feedback that makes a slight change in the 3H contributed by the agc circuit could reduce or increase 3H of the overall oscillator by optimizing or degrading the 3H cancellation that may be occurring. For example, it is possible that 3H is reduced somewhat when the gate feedback is optimal for 2H reduction, but this reduced 3H may reduce some 3H cancellation taking place that actually was reducing overall oscillator 3H.
Cheers,
Bob
The 339 osc. is a single-opamp Bridged-T and the AC level across the FET is quite low, comparable to the FET circuit in your osc., Bob. Your osc. has an issue of the control opamp switching polarity through zero, where things briefly get a little unsettled. Optimally, the 339 osc. has slightly lower distortion than your SV design, at least here in my shop.