I have a Chinese JDS6600. And though it's a fine machine for above cartridge voltages, it's awful at MM and MC levels.
Excuse me, but if you're going to scan and reproduce copyrighted material without permission you might at least acknowledge the source as being Morgan Jones.
There is no point in testing a MC phono stage with a signal generator. At that low level every MC amplifier has low distortion and linear frequency response. The only parameter worth testing is noise.
A phono preamplifier with RIAA frequency correction and typically ~5 mV sensitivity is another matter. Unless the two function is built in one.
A phono preamplifier with RIAA frequency correction and typically ~5 mV sensitivity is another matter. Unless the two function is built in one.
Sorry didn’t know it was an issue , this diagram has been posted by yourself while ago and I just copied it
Where is the requirement to attribute Morgan Jones in your post? And what did he "scan"? He just posted the same .gif you did in 2006. 19 years ago.
I sit corrected; no problem with copyright on that diagram (which I recognized immediately). You'd be amazed how many people do scan copyrighted stuff without attribution. Worse, others link to sites with pirated pdfs of entire books.
That 3.18 us should not be there. It is presumably based on the so-called Neumann pole, which apparently doesn't exist.
https://en.m.wikipedia.org/wiki/RIAA_equalization#The_Mythical_"Neumann_pole"
https://en.m.wikipedia.org/wiki/RIAA_equalization#The_Mythical_"Neumann_pole"
Owon generator arrived , I’ll do some basic tests with scope but there seems to be no white or pink noise option only Gaussian white noise. I wonder if other cheap generators are also lacking that option ?
Sure, but why not put the undesired pole as high as you can? (That is, time constant as small as you can?) You are now approximating a curve that is at -0.6434 dB at 20 kHz with respect to the ideal RIAA recording curve to within 0.1 dB.
By the way, when you allow active circuits and/or inductors, you can approximate the theoretical RIAA curve very well up to 20 or 30 kHz without having to let the curve increase up to very high frequencies. Just multiply the theoretical RIAA recording curve with an odd-order Butterworth low-pass response with an order greater than 1, and use the active circuits and/or inductors to realize the complex pole pairs.
For example, you could make a 50 kHz, third-order Butterworth roll-off with respect to the ideal RIAA recording curve by combining your network with a second-order, 50 kHz Sallen and Key section with a Q of 1 and a natural frequency of 50 kHz. The response at 20 kHz then becomes -0.01765 dB with respect to the ideal RIAA recording curve, at 30 kHz, it would be -0.1969 dB.
Last edited:
Putting the pole higher requires more loss in the equaliser and it's useful to be able to test overload of an RIAA stage, which means you don't want too much loss. But I rather suspect that the reason I put it where I did is that the component values fell out nicely. 😉
If you want pink noise it's easy enough to build a pink noise filter and stick it on the output. Back when 1/3 octave analyzers were a thing, you wanted pink noise to produce a flat response curve. Those analyzers are what's called a constant-percentage-bandwidth analyzer (CPB). With proper spectrum and FFT analyzers (fixed percentage bandwidth analyzer) you want white noise for a flat response curve. Long ago, National Instruments made a white noise generator, the MM5437. At the end of the data sheet they show a pink noise filter- https://w140.com/tekwiki/images/2/29/MM5437.pdf
I built a pink noise generator with the MM5437 and post RC filter. It is good for measurements, but audibly there is a bump every few seconds. I still have a surplus MM5437 if someone interested.
Why would one want to measure RIAA accuracy with noise rather than with sine waves?
I doubt that approximating pink noise with white noise and a fourth-order low-pass filter with three zeros will be very accurate. In principle you need either a half-order filter, which you can't make with lumped capacitors and resistors, or a very high order filter with many zeros that approximates a half-order filter well.
I doubt that approximating pink noise with white noise and a fourth-order low-pass filter with three zeros will be very accurate. In principle you need either a half-order filter, which you can't make with lumped capacitors and resistors, or a very high order filter with many zeros that approximates a half-order filter well.
No, it make little sense to measure RIAA or other amplifiers with noise, but hey, the question was asked. I use white or pink noise mostly for room or speaker measurements. An interesting piece of trivia- you'll get different answers if you measure a tape deck with noise vs. sine waves.
If you take enough measurements with noise you'll eventually get the right answer, but you will also die waiting! Nice sine wave with 0.01% or less distortion will give you an answer with many sigmas in short order!
Iphone signal gen produces perfect sinewaves from 2 to 20khz from the headphones jack 2v pp is plenty to drive the phono pre to 8mv at 1khz through reverse RIAA. A/B tests are difficult to distinguish from each other using mp4 files.