I thought I'd give an update on my DIY cutterhead experiments. I took a fresh look at my original design and made some major changes. I decided to mount the head on my 6N rather than my DIY lathe. So I spent a fair amount of time coming up with a decent mount. I had to remove all of the standard 6N hardware and roll my own. The goal was to make sure the stylus tracked to the center of the record. I also wanted to make it easy to adjust the height. Finally, I did not want to do any mod to the 6N that could not be easily restored.
Previously, I coupled the drivers to the stylus holder via short lengths of wire. I did some modeling and found this was creating a secondary compliance that tends to produce an unwanted resonance. So, in this pass I am coupling the driver cones directly to the stylus holder.
I also went to greater lengths to reduce mass. The cones are now paper rather than copper. I added stiffnes by dipping in lacquer several times. My stylus mount is re-designed for lower mass and I'm now using a taper mounted stylus. To increase sensitivity, I changed the torque bar from a steel screw to nylon (reduced compliance). So, with the reduced mass and compliance of the torque bar, my system resonance is now in the 850hz range. Not up to the Neumann standards, but still respectable. My total moving mass is a bit over 1.5 grams. I made some response measurements on the bench using a Kaman Eddy current position sensor I got cheaply from eBay. You can see the results below.
This is showing displacement vs. frequency, so its flat up to the system resonance and then falls off at -12db/oct. I added a dashot for damping and a simple lift using fishing line. If you convert the displacement response to velocity (ds/dt), you get the typical open loop moving coil head response that looks like an inverted V.
I did some test cuts using white noise and then developed a correction filter using the Adobe Audition FFT filter and Har-Bal. My first test cut using program material is here:
http://home.comcast.net/~markrob1066/docs/BC_playback.wav
The cut is about -8db below the level recorded on the vinyl release on ABC records (Steely Dan Black Cow). My cut was bandlimted to about 12Khz to reduce the amount of power used. It was cut with a 60 Watt per channel Dynaco 120 and it used all of the voltage swing available (+/- 35 Volts). The noise is due to the lower cutting level and I did not optimise the heater current. You can hear some breakup in a couple of spots. This was in my source. I had to do some extreme EQ to correct for both the RIAA and the open loop head response (nearly +40db at 12Khz!). Someplace in the process, I must have caused some clipping to occur.
Overall, its a good start. I need to drive with a higher powered amp (I have a Phase Linear 400 set to go). Also need to work on the signal processing to make sure its clean. I did some bench experiments with ferrofluid to see if I can improve the driver power handling. I got great results and plan to try adding the fluid to the drivers. I'd be glad to answer any questions.
Mark