Ultima Thule - The design of this headshell , and the way it is mounted in the tonearm tube , is a source of additional resonances . It 's a good idea . But the manufacturing method is bad .
And again - this is a non - removable headshell . And this means that it is no longer possible to connect any other phono preamp to this turntable .
And again - this is a non - removable headshell . And this means that it is no longer possible to connect any other phono preamp to this turntable .
Last edited:
You 're putting on public display a very funny and flimsy headshell . It is mounted very illiterate . But you must really like it 😉 .
But I like to use good headshells .
And in the headshell that you showed - there are more disadvantages than advantages .
If you make such a headshell at the factory , then it should be like a good factory product ....
Nothing personal .
Last edited:
Yamaha's HA-2 from 1979
https://audio-heritage.jp/YAMAHA/etc/ha-2.html
and HA-3 from 1982
https://audio-heritage.jp/YAMAHA/etc/ha-3.html
are both of this type (headshell-mounted JFETs). There is a general schematic on the HA-2 webpage.
Due to the super short wiring between cartridge and JFET, Yamaha judged that there was no need for any resistive termination at the JFET gates, therefore the input impedance of the HA-2 was specified as being "in excess of 1MΩ".
I subsequently heard that a key reason why Yamaha stopped using this technology in their follow-up phono stages was electric corrosion of the contact points, due to the tonearm leadwires having to supply DC power to the JFETs as well as handing the AC signal.
This discussion is rather amusing. In the 1970s I considered putting two emitter follower transistors in the head shell (Holman idea) to isolate the cartridge inductance and decided that it would be more trouble than it was worth.
JFET buffers like Marcel described are built into essentially all electret microphone capsules, even very famous ones used for recording. For simple speech applications this is called "plug-in power". Signal voltage levels from an electret are likely comparable to or larger than a MM phono cartridge, so linearity can likely be made good enough. Scott Wurcer has published a lot of work on elaborated versions for DIYers.
Seems no real reason that plug-in power inputs couldn't be supplied along with conventional 47K inputs.
Electret phono cartridges like the MicroAcoustics have capsules loaded with a coupla-K Ohm resistors to differentiate the signal (to simulate the velocity response of magnetic cartridges) and reduce level (again, to simulate a MM).
All good fortune,
Chris
Seems no real reason that plug-in power inputs couldn't be supplied along with conventional 47K inputs.
Electret phono cartridges like the MicroAcoustics have capsules loaded with a coupla-K Ohm resistors to differentiate the signal (to simulate the velocity response of magnetic cartridges) and reduce level (again, to simulate a MM).
All good fortune,
Chris
Yes I can see electrolytic corrosion being an issue given the pogo-pin style contact interface of most headshells - but that's really an argument for better connector tech to be used I think - even without DC there will be issues of contact reliability that could be fixed - say by taking a leaf from the SpeakOn style connector but dine in miniature.
The opamp U6 ("synthesized" input resistance amplifier) uses the current of the first cascade feedback divider. Will the input noise current of U6 add to a noise current of U3? There is another way - you can take a signal from the output of U3 and make a mirror frequency responce.
I built a phono stage using the AD797 op amp, with the input going to a low noise JFET bootstrapped by the first AD797. Dead quiet.
The noise current of U6 will have very little impact because of the low impedance driving U6 (namely R2), but its noise voltage contributes directly to the input noise voltage of the amplifier.
You can also add a FET op-amp connected as a voltage follower or voltage amplifier, just to drive U6.
Yes, right, here you must first talk about noise voltage. The S/N ratio will become worse by 3 dB, which will worsen the effectiveness of such a "synthetic" load of the cartridge.
For cartridges with inductances above about 250 mH, the overall RIAA- and A-weighted noise will still be less than without U6, with 47 kohm across the input. For lower cartridge inductances, it is the other way around.
The simplest way to reduce the effect of U6 on the noise voltage is to replace it with an ultralow voltage noise, ultrahigh current noise model.
The simplest way to reduce the effect of U6 on the noise voltage is to replace it with an ultralow voltage noise, ultrahigh current noise model.
Marcel, ultralow voltage noise is clear to me, but could you elaborate why you pronounce the ultrahigh current noise in this context?
Thanks, Boris
Thanks, Boris
When a semiconductor manufacturer makes an op-amp with a bipolar input stage with plenty of base contacts and a relatively high tail current, it will have a very low voltage noise as well as a relatively high current noise (due to shot noise and 1/f noise of the high base currents). The current noise will increase further when there is a base current compensation circuit added to it. For obvious reasons, these devices are usually marketed as ultralow voltage noise op-amps and not as ultrahigh current noise op-amps, but they are both.
For U6, ultralow voltage noise is desirable and ultrahigh current noise doesn't matter, as long as it stays below 5 pA/√Hz or so.
For U6, ultralow voltage noise is desirable and ultrahigh current noise doesn't matter, as long as it stays below 5 pA/√Hz or so.
LT1028 comes to mind here, with the bias compensation feature described as a nuisance by Douglas Self (but not in this context, IIRC).
Yes, the LT1028/LT1128 has a base current compensation circuit that produces lots of common-mode current noise, see https://audioxpress.com/article/boo...spice-models-to-simulate-vintage-op-amp-noise Still, even the 3.4 pA/√Hz from an LT1028 or LT1128 would be perfectly acceptable for U6.
I've just come across the ADA4625 which is a JFET opamp with 3.3nV/√Hz, unity-gain stable, 18MHz bandwidth. Not cheap but probably useful for MM preamp duty - the 1/f knee isn't too bad either, and there's a dual version to help spread the cost.
And he is right, - a DC input bias compensation is not a compensation of not correlated input noise currents (for example - not correlated Ibase of Q1 4.5 uA and Icollector of Q7 4.5 uA inside LT1028\1128). compensated DC difference is 0, and AC non correlated currents "difference" is 1.41 times bigger.
It's far worse than that in an LT1028, see the link of post #57.