LC AUDIO MILLENNIUM XP - HELP if possible from Lars Clause!

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
AndrewT said:
Hi,
If the Dact is a conventional stepped attenuator, then the output impedance varies between 0ohm and [10k+Rs]/4~=2k5.
The minimum load for that range of source impedance is 5times i.e Zin>=12k5
I would try for 20times i.e Zin=50K.
But, only change the 2k0 resistors, not the 301e.

Hi Andrew,

what do you base that rule of thumb on?


/Peter
 
Hi,
most folk say Zin~=10times Rsource.
I have found that you can get only a slight reduction in performance/interaction by going as low as 5times.
Equally, I think there is a small gain in performance, by going up to 20times.

I cannot recall any/many builders saying how well their systems worked when outside this range. But, I regularly hear of complaints of performance when straying outside this range.
With two major exceptions:
Balanced interconnections as described by Walt Jung where he shows that CMRR improves by increasing this ratio to around 100,000times.
And for maximum signal power transfer for very low level signals, eg at RF inputs.

Overall, I'd say anecdotal.
 
Ok!

But let's look at it from what's really happening.

The attenuator is a voltage divider. Let's say we have a 5k series resistor and the setting that gives aprox. -6dB (assuming low drive impedance from source and highish load impedance at amp input) which means a 5k to GND.

Now all that is happening if the amp input impedance is taken down is that the attenuator shunt resistor to GND in needs to be bigger to give the same gain reduction as we started out with.

As long as the source can handle the attenuator input without overloading (causing increased distortion) then it should not matter if the attenuation takes place in the shunt/ground resistor or in the amps input resistor. Of course cablecapacitance may enter the equation but only for long runs and high resistance attenuators.

One positive thing with a low resistance/impedance between attenuator and amp input is lower hum and noise for unbalanced connection.

To sum it up. I see no reason that the load impedance that the attenuator sees is about the same value as the attenuatorn itself.

Indeed I have had excellent results with similar solutions myself.

The old 10x may be valid in power transmission and if the active circuit driving the load have problems handling a low impedance load.



Just my thoughts on the subject! :)


edit:
about Walt Jung. Do you mean that CMRR improves by going even further in balanced lines.. like 1k source impedance and 100k or 1M load?




/Peter
 
Pan said:
about Walt Jung. Do you mean that CMRR improves by going even further in balanced lines.. like 1k source impedance and 100k or 1M load?
the clue of his name should have been enough. Go read what he has to say about 1M5 for the input resistors and 10r for the source resistors and why this bridge imbalance improves CMRR, but still requires resistor matching of better than 0.1%.
 
rellum said:
Hi Olsen,

I'm a "millenium XP" owner and I don't have an end 3.1 schematic. Are those 301R the serial input base resistors ? If so, don't replace them. It will be of no help for gain setting. You should rather look at two resistors wich are in // with the inputs. Or simply measure resistance between each input and ground with a dmm. If close to 2K, those will have to be increased.

Best regards,
Francis

Hi again!
Yes it is The End Mk3.1. I dont know what serial input base resistors are :xeye: But here is a schematic for an Mk3 and from what i can see, the input is the same as Mk3.1. The schematic is not complete as you can see. I have e-mailed LC Audio today, since i could not find schematics on the internet.
Thank you all for your quick response :)

Olsen
 

Attachments

  • the end mk3 schematic_.jpg
    the end mk3 schematic_.jpg
    71 KB · Views: 334
Hi Olsen

I just found the attached schema wich seems to be the one you refer to. And it's way different from what I expected. It has no differential input stage and isn't suited to be feeded passively on it's xlr input. I can understand your feeling of lack of gain. I guess you want to increase R27/R28. You can try it, it will help but 10K won't suffice and will alter the log behaviour of your attenuator. Use rather Andrew's recommended value : 47K. Offset shouldn't become an issue if loose connection, as the base currents should cancel nicely.
Last thing, the way it's done on the End MKIII, driving the xlr input without a low impedance buffer, you will loose most of the common mode noise rejection provided by the use of R24/R25.

Best regards,
Francis
 

Attachments

  • end_mkiii.jpg
    end_mkiii.jpg
    56.2 KB · Views: 357
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.