DIYHiFi.org

Mulling over a little bit different MC phono stage, a concern arised. In case of not so tight matching of components or thermal drift, a certain dc current could be impressed on the coil of the connected MC pickup.

Does anyone know what the current range would be above which a low output MC coil could be damaged? I really have no idea.....and am not keen to check it out.


Thank you
Tino

Exactly what MC phono stage is it?
Do you have a schematic?
Other than using step up transformers I would only consider a single ended low noise JFet for the first stage. Like a 2SK170, 2SK369 or BF862.

_________________
Carlos Filipe

"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius - and a lot of courage - to move in the opposite direction." Albert Einstein

Err... not yet. I'll draw one (a conceptual one), it's just in my head, nothing more.

Later
Tino

OK, here is the thing I was mulling over during my car travels.

The input is at the emitters, hence a lowish input impedance.
It could be made also with J-Fets, making things even more simple, as they are self biasing.

An MC pickup will not care about the low impedance (or maybe it would even like it?), as the whole generated current will serve to convert into voltage.

This could even go as an I/V for a DAC...

But back to the question, what if the nodes will have different dc voltages? How much is too much? 100µV? 500µV? I really have no idea what voltage/current such a coil could stand.

Ciao
Tino

I would use JFETs anyway. Solves all your problems.

Jocko

the question is not if but when...
I don't think an MC will like it so use FET (common source) or do as some have done use a large cap - but a large electrolytic is not what you want connected to a cartridge IMHO.




probably a better use

Bill

I think there's a more subtle problem that takes place long before the cartridge might go up in smoke. Doesn't any "static" current through the coil cause deflection of the cantilever, upsetting the linearity?

I seem to recall reading some information about this from either Jonathan Carr or John Curl about this sometime in the past decade. It may have been over at diyaudio... Some designs used common base input stages in the past, but this fell out of favor with the newer generations of MC cartridges.

JFET's solve this and other problems in any case.

Hmmm...I wonder why. I have some new search words now.

Thank you anyway.

Tino

Tino,

Some suggestions... To minimize thermal drift, you should try integrated bipolar transistor quads for the input and input bias transistors, such as THAT340P, or at least duals such as SSM2210 / SSM2220. Make sure the integrated transistors are arranged symmetric across the inputs so as to provide balanced temperature drifts which cancel. For example: one integrated quad for the inputs Q2, Q3, Q9, Q10 and second quad for the bias Q's. That will keep the junctions thermally matched as well as anyone could hope for. Low TC or matched TC resistors for R1, R3, R5, R8 will help ensure consistent reference currents.

Then you have to deal with whatever fixed offset remains. A little trim network could be used to replace the ground connection from the emitters of Q1 / Q4. Something like a high-ratio resistor divider from a potentiometer tied between the rails might work.

Then you'll need a sensitive voltmeter to measure the offset and characterize the temperature drift.

You could also ovenize the circuit... sounds crazy, but it's really not that hard: http://www.techlib.com/electronics/ovenckts.htm (gotta love Wenzel's personal site - some real jewels in there).

I think if you really wanna do it, super low DC offset can be done.

_________________
- Chad.
__________________________

"Fashion is mistaken for good design; moral fashion is mistaken for good." -- Paul Graham

Hi Chad

Thanks for your suggestions!
2SK389/2SJ109 would also be something which is matched...although they're Fet's.

Tino

True, although if you intend to use FETs at the input in the same way as the bipolar schematic you started with (common gate, sources tied to inputs), I don't think you'll get as good DC performance... off the top of my head, it's the relatively constant Vbe and high gm of the bipolars which would facilitate a low, stable offset across the input. FETs may do the same, if operated near Idss in their low-tempco region, but I have less confidence in that.

I guess one important thing to know is the DC resistance of the cartridge. I wouldn't try measuring it, as your multi-meter is likely to screw it up, delicate things that they are. But, the DCR info is probably out there somewhere if you search for it. I'm guessing it's probably quite low, which of course only shrinks the DC offset target.

Indeed, a common-source JFET input is the safe / easy way to go... then the coil only ever sees the miniscule DC appearing at the gate. But who ever said the easiest way is always the best way?

_________________
- Chad.
__________________________

"Fashion is mistaken for good design; moral fashion is mistaken for good." -- Paul Graham

Agreed about what you are saying about J-FET's dropping into this circuit.
And I have only some 2SJ109's, no 2SK389, and they are hard to find, as anyone here might have found out....maybe the LSK389 would fit too, if it's really a drop in for the Toshiba?

So, a Bipolar version may work better...

Btw, this is the information on my system:

Ortofon MC 15 SUPER MK II
Output voltage at 1000 Hz, 5cm/sec: 340 µV
Internal impedance, DC resistance: 7 Ohm
Recommended load impedance > 20 Ohm

I wonder what diameter might have the wire....
At work, we are using fine resistance wires, as thin as 16µm.

Tino

Guys, I hate to rain on your parade, but this design is not really appropriate for quality MC input.
First, it is not really low noise, no matter what input transistors that you use, because the bias transistors will also add in to the noise.
Second, very low input inpedance is not necessarly the best way to load a MC cartridge. Most MC cartridges like 50-50,000 ohms, these days. There are a few rare exceptions.
Third, you don't HAVE to measure your MC cartridge. What you have, you have, as far as DC resistance. Most modern digital meters will not destroy a MC cartridge, but it could magnetize it.
Fourth, you really want an input stage with virtually no input current. The best way to do this is a fet in a common source configuration.
Complementary fet input is a quality option.
The best parts are made by Toshiba, but many discussed here are now discontinued and difficult to find. Still, some are still available.

John, your input is always valuable information for me, thank you.

The initial idea for such an input was that I found my cartridge sounding less "crispy" with lower impedance input, but I didn't try less than 100 Ohms, as my actual phono pre (a design similar to an old Elector one, with complementary differentials, 4 transistors paralleled and common emitter output, active EQ) is at the limit of gain (but pretty low noise).

So I thought to make a "current" input, not wasting output "energy" from the cartridge with low ohm parallel termination resistors....

Tino

Been there, done that, but it doesn't work very well.