question on very high input impedance for B1

[wintermute]

Hi,

If I want to use a B1 as a load for a filter circuit that requires effectively an infinite input impedance, is it ok to use a 1M or higher resistor to ground at the input of the B1, or in fact to eliminate the resistor to ground altogether?

There have been some comments about the input impedance not being more than 25K, but I think that has more been to do with volume controls and the fact that the source won't necessarily like a higher impedance.

I can't see anything wrong with having a higher input impedance if that is what the thing driving the B1 wants but I thought I'd ask just the same

I did find that in LTSpice the circuit became unstable with no resistor (but was fine with a value of 10M) I'm not sure if that is just LTSpice or whether there really does need to be at least some reference to ground at the input.

Tony.

 

[Zen Mod]

you're fine with megagigatera input resistor .....

 

[wintermute]

Thanks Zen Mod

Tony.

 

[AndrewT]

the output impedance of the filter will be fairly low. This will give stability to the B1 buffer without needing another resistor to ground.

Very high impedances are more susceptible to picking up interference. This demands that the filter output and the B1 input be very close coupled and probably also screened (inside a grounded metal/conductive box).

 

[wintermute]

Thanks for the tip Andrew! I was planning on putting it all in a metal chasis, but I hadn't thought about the need for localized screening! The circuits I've been simulating seem fine with around 10M ohm impedance so maybe I will just put in a 10M resistor and be done with it, rather than temp fate by having the extremely high impedance of a non-terminated B1 I may well be able to drop that 10M a bit as well.

Tony.

 

[AndrewT]

if you have a passive filter in front of the B1, then it's either RC or CR.
Both have a relatively low impedance to ground seen as a lowish Rs by the B1.
Why add an extra 1M to 10M resistor in parallel?
Adding a non infinite load to the filter requires the filter to be adjusted to restore the F and Q values to the theoretical. I don't know how to do that for the RC version. Obviously in the CR version the R after the C is the total equivalent R that the cap feeds.

 

[wintermute]

Hi Andrew, it might not be quite what you are thinking Here is a sneak preview (of what I'm calling the synergy crossover)... This is the LP section which is pretty much sorted.

It is working very well in the sim, exactly 200Hz crossover and introduces virtually no distortion.

I'm currently re-visiting the HP section as I discovered my gyrator is introducing quite a bit of distortion... still not too bad at about -94db but no where near as good as a B1 by itself (which is around -104db in the sim)....

but your answer is relevant to the other area I need a high impedance and that is the BSC bit of the circuit so I'll take your advice and try it without the resistor I suspect you are right with regards to the FDNR as well but as it is rather different to a RC I'm not 100% certain

cheers,

Tony.

 

[AndrewT]

there are other threads discussing the Gyrator in lieu of an inductor.

What is the 10M doing to improve the circuit?

 

[wintermute]

At the moment the only function it is providing is to keep LTSpice happy. Without it there is some spruious noise/distortion on the waveforms when running a transient analysis. What I don't know is if that is just a limitation in spice, or whether it would be there in real life too

Tony.