GBW is 11MHz in my model.
Inserting 0.22pF brings no change at all.
Gain is a 30dB at 300KHz where your graph shows a dip, so everything seems well under control there ???
Hans
LD,
I know what caused the dip at 300Khz in your FR.
When you don't have a floating voltage source and when the 580mH is all on one side of the voltage source, one halve of the differential stage has no phase turn from the coil and the other has the full phase turn.
Either split the coil in two equal parts and place them on both sides of the voltage source or use a floating voltage source just like your Cart.
Even better, just to prevent any misinterpretations, is to have a floating voltage source in the middle and equal parts of Lcart, Rcart and R on both sides.
Hans
I know what caused the dip at 300Khz in your FR.
When you don't have a floating voltage source and when the 580mH is all on one side of the voltage source, one halve of the differential stage has no phase turn from the coil and the other has the full phase turn.
Either split the coil in two equal parts and place them on both sides of the voltage source or use a floating voltage source just like your Cart.
Even better, just to prevent any misinterpretations, is to have a floating voltage source in the middle and equal parts of Lcart, Rcart and R on both sides.
Hans
Thank you Lucky for commenting and for the sim.
I have to work with what is in my parts bin (and dual PDIPs when possible, for dead bug construction).
The NE5532 (and the LM833) would be good candidates even for IC1 but I think FETs at the input are a protection against RF noise ingress, therefore a JFET IC has a place there.
Good job there Ricardo !
George
Attachments
Yes, agreed that is the cause, and a balanced voltage source with split impedances avoids the problem. Then I obtain similar plots/stability as you do, Hans.
I think in reality this means, for the balanced Aurak versions, the cart coil should be floating, ie one side not grounded, which is worth noting. This might interact with/degrade ground screening of environmental sources of noise, relying instead on CMRR, I suppose. But at least transimpedance input impedance is low.
LD
Last edited:
well I have never liked having cart -ve strapped to the body and prefer cart grounded to tonearm*, then twisted pair wiring with good screening. As this method is immune to capacitance you could even go the whole hog and use star quad for total overkill.
If you swap carts and have different headshells for each then there is the possibility of the series R being at the headshell end. Slightly tricky soldering job but convenient.
*As shown by George on one of his threads.
If you swap carts and have different headshells for each then there is the possibility of the series R being at the headshell end. Slightly tricky soldering job but convenient.
*As shown by George on one of his threads.
Starting with 10nF and 330nF for the two important capacitors, I have recalculated all resistors to existing E192 values.
Most resistors still should be 0.01% to keep CMRR below 60dB.
A 330nF 1% cap is much more affordable then 1% values of 1uF or above.
Going down with the cap value from 2.25uF to 330nF causes an equal increase in surrounding resistor values leading to a slight loss in S/N, but 79dBA is still more than good enough IMO.
Bill, this is not the final solution for the 3 Carts that you have, but when I have your final and confirmed values for Lcart and Rcart, I will do another sim for you as promised.
Hans
Most resistors still should be 0.01% to keep CMRR below 60dB.
A 330nF 1% cap is much more affordable then 1% values of 1uF or above.
Going down with the cap value from 2.25uF to 330nF causes an equal increase in surrounding resistor values leading to a slight loss in S/N, but 79dBA is still more than good enough IMO.
Bill, this is not the final solution for the 3 Carts that you have, but when I have your final and confirmed values for Lcart and Rcart, I will do another sim for you as promised.
Hans
Hi Bill,
What do you mean with cartridge gnd ?
I think that all TT metal including tonearm should be connected to the TT chassis gnd, at least that is what I did with trouble-free results.
Hans
Hi George, yes if that's what you have in your toolbox they are OK IMO, good enough for prototyping and trying out the principle in the single ended Aurak 3.0.
To avoid doubt, the single ended version is fine with standard tonearm wiring and doesn't need a floating cartridge coil.
LD
Last edited:
There is often a separate ground connection within the cartridge to internal components and screening, and sometimes the cantilever ground. It is sometimes hard-wired to one of the cart ground pins, sometimes it is produced at the top plate of the cartridge for pressure connection to the tonearm/headshell.
LD
What is important for CMRR is that all pairs on both differential sides are within 0.01% to each other like R4,R6 and R3,R5 etc.
In absolute value 0.1% is good enough.
So you are right, you could select from 0.1% resistors pairs that are within 0.01% to each other. I do not think you even need 10 pcs for this selection, probably 5 will be enough.
Hans
Two Vbe's and two Vcesat's have to fit in your 1.5V so hence my comment to scan first for ridiculously low Vcesat in choosing devices to test for low rbb (it's easier to measure and the manu's actually get it right).
I think you mean this
http://www.diyaudio.com/forums/analogue-source/100408-results-m97xe-mod-2.html#post1190696
That’s a very clever idea Bill.
LuckyThis is correct but please consider the detail that single ended Aurak 3.0 is inverting the signal.
Thus with a standard tonearm wiring and a non floating cartridge coil, signal polarity is reversed.
If one wants to preserve the absolute polarity when using single ended Aurak 3.0, the cartridge wiring has to be reversed (+ for – as shown in the hand drawing in my previous post #403) and consequently the cartridge has to be (or made to be) floating.
Personally, I am not able to hear the difference btn proper and reverse absolute polarity.
Hans, although I will not build the balanced version, I want to thank you for all the effort you have put in this.
George
I'd suggest, if absolute polarity is important to anyone, to re-do the 2nd stage so as to be inverting, rather than use a floating cartridge connection.
But it's not important to me at all, I can't perceive any difference either, George.
One practical point for the balanced version using a floating coil: consider the possibility of static charge accumulation or discharge to the floating coil. Not sure if the 1632 has internal clamps, but probably a good idea to protect the input pins with a network in any event, methinks. Also consider the condition when power is off and discharge might occur.
HTH !
LD
Hmm... that seems skinny in the context here, esp since that is HBM. Plus the input diff voltage spec suggests protection might have deferred to speed perhaps?
For the balanced version, some external protection by way of diode clamps to both rails for both inputs seems sensible.....and external cross-clamps for +/- inputs? It's harmless and cheap and prob a good idea methinks, esp if the layout is not finalised.
Also, the 1632 has layout guidelines which seem worth following, though anyone designing with it prob already knows this.
HTH !
LD