The Korf Blog

In the previous post, I've lamented that the blog updates became a bit scarce. Well, what did I know? Since our last update, our reference playback system required maintenance, measurement apparatus became unserviceable, and lots of other unpleasant things happened. More than a month passed... Today, we are still spending 90% of our time putting out figurative fires, and only 10% moving forward.

But we do try to move forward. In the last post, I promised to do the laser azimuth "wobble" measurement. Alas, I cannot — the focussing laser I need for it is dead, and the replacement is taking its own sweet time to arrive. Critical listening? No can do, sorry. Our Micro Seiki RX-1500 is being serviced, because the speed controller stopped working properly. Practically only thing I could do was vibrometry.

A new arm presented a rather unexpected problem. I've made its back too heavy, and the arm wouldn't balance! None of the counterweights from various donor arms fit. You can see the temporary solution in the photo. Stopper rings balanced the arm, but I am quite positive this is an awful configuration when it comes to vibration.

As usual, we are comparing our arm to a similar Prototype 5 with regular bearings. We've also kept the vibrometry results from Prototype 5 with steel counterweight stub as our flexure bearing prototype uses steel stub too.

And aren't we in for a surprise?

At first look, our flexure bearing tonearm has same 6kHz resonance as the one with ball bearings. In fact, the version with ball bearings and steel counterweight stub has less of a peak there. Are we mistaken? Is something else resonating at this frequency, not the bearings? Is it a harmonic of something else, for example?

Yes and no. Let's take a closer look.

I've also moved from log scale to linear as it makes the picture less cluttered.

What we actually have here is an intersection of 2 resonances of different nature. The most prominent peak at 5900 Hz isn't only the bearings — I was a bit too sure on this one, and my arrogance came back and bit me. A large part of it is a subharmonic of a 12kHz joint accelerometer+headshell resonance. I have made a separate measurement (not shown) using a less rigid accelerometer mount. Lo and behold, the resonance moved down in frequency and lost its intensity. All as it should have.

But the ball bearing resonances stayed. You can see them on a zoomed chart, two peaks at 4900 and 5400 Hz. And there's another peak at 5800 Hz, masked by that prominent second harmonic. Our flexure bearing arm has none — success!

Figuring out what was going on required deeper understanding of ball bearing vibration profiles. I recommend the following 2 articles on the subject:

The first, published in "Motion and Control," is an excellent introduction and has a good chart showing the tell-tale double hump at 5-6 kHz

The second, from a 2014 European Conference on Non-Destructive Testing, perhaps asks more questions than it answers. But it's an excellent vibration technique primer, and clearly shows the effects of different accelerometer mounting.

One of the reasons I love doing all this is the daily dose of learning. Just when I become complacent and think I have an understanding, Mother Nature has her ways of humbling me. Picking today's vibrational profiles apart, we have learned a lot about accelerometer mounting and ball bearing resonance profiles. This new knowledge will perhaps call for a change in how we do vibrometry.

If we ever succeed in developing a new protocol that is more effective in suppressing artifacts of accelerometer's own resonance, I'll let you know first! In the meantime, we'll just have to be more vigilant and not rush to conclusions.

Don't forget to subscribe to get the latest blog updates!