Balanced High Pass Filter

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

earthsled

Well-known member
Joined
Dec 1, 2009
Messages
405
Location
Asheville, NC USA
At the risk of bringing up an elementary design topic...

I'm needing to build a simple, passive, in-line, balanced high pass (low cut) filter to use with a TLM103.

Something similar to the Shure A15HP (w/ a 100Hz roll-off):
http://store.shure.com/store/shure/en_US/pd/productID.104210700

I imagine this circuit would use series caps on the hot and cold of the balanced line - thereby preventing phantom power from passing through. So, I assume the filter would need to be inserted post-preamp (line level).

Does anyone have a simple schematic for something like this? Or, perhaps there is a circuit that will pass phantom?

Thanks for your help!

 
What is the box your preamp plugs into? What is its input impedance? Calculate the necessary coupling capacitors using the formula:

C = 2 / (2 x pi x Zin x 100)

where Zin is the input impedance of the next stage. The 2 isn't a misprint; because you're putting the caps in the two legs, effectively in series, each one has to be twice as big as it would be in an unbalanced circuit.

Peace,
Paul
 
Thanks Paul,

So the circuit would simply be 2 caps? Would I need to add any resistors? I'm asking because the "textbook" high pass circuits I've seen use an resistor-capacitor circuit (for unbalanced).

Best,
Seth
 
I recently wanted a line level balanced high pass filter for a compressor sidechain. Surprisingly google gave me a link to gearslutz:

http://www.gearslutz.com/board/geekslutz-forum/356482-easy-diy-passive-filter.html

Third post has all the instructions you need.
 
earthsled said:
Thanks Paul,

So the circuit would simply be 2 caps? Would I need to add any resistors? I'm asking because the "textbook" high pass circuits I've seen use an resistor-capacitor circuit (for unbalanced).

Right. But the input impedance of your next stage is the equivalent of the resistor.

Peace,
Paul
 
Please be aware that such a simple circuit may have serious drawbacks when used with a transformer-balanced input. The combination of C and L introduces a bump in the frequency response that may be worse than the problem you want to solve. Also, the source impedance presented to the xfmr may affect its THD performance adversely.
 
Thank you for the gearslutz link. It had a lot of good information in it!

I suppose I have the advantage of dedicating this filter for one signal chain (rather than trying to get multiple uses out of it). The mic pre is a 1272 clone, followed by an LA-2A clone, then into an A to D (a Digi 002).

Ideally, I'd like to insert the filter before the LA-2A. For this compressor, I calculate an input impedance of about 821 ohms. (That's 68K divided by the transformer ratio (9.1) squared). The gearslutz post suggests a 30K shunt resistance for a line-level filter. Using this resistance and a cut-off frequency of 100Hz, I get a capacitor value of about 4 uF. This isn't a standard value, so I could adjust the resistance based on 3.3 uF or 4.7 uF caps.

Now, I'm concerned about placing the filter before a transformer-balanced input (as D Enfer suggests). Is there a better circuit to use? Or, should I opt to have the filter precede the A to D?
 
Abbey's concern is justified. If you need to do the filtering before the compression (a good idea), it may be worth considering a buffered circuit.

Peace,
Paul
 
Okay. Maybe the buffered circuit is the best solution. Does anyone have resources on building such a thing?

I appreciate everyone's comments!

 
Unbalance your signal,

stick one of these

http://www.discovercircuits.com/DJ-Circuits/hipass.htm

and balance it back again.
 
Thanks for the link!

In addition, I found this PDF which has some simple schematics for unbalancing and balancing a signal.
http://www.jensen-transformers.com/an/an003.pdf

I was hoping to not have to deal with power supplies if possible. So, the idea of building a passive filter to precede the A to D seems more appealing (even if it won't keep the LF artifacts from pumping the comp).

The spec for the Digi 002 claims that the input impedance is 10Mohms:
http://www.digidesign.com/index.cfm?langid=100&navid=54&itemid=19443
In another post, Abbey pointed out that this spec seems to be incorrect.
If I were to design a passive filter to precede the Digi 002, what input impedance should I use? Maybe 10K?
 
To be able to go passive without any interaction you have to know your source and load impedances well. Even then it requires some work.

However, if you go active then the balance to unbalance schematic on Jensen's link is fine. However, when balancing it back again do not use the configuration that uses the inverting amp. Use the non inverting half in unity gain and invert the ouput.

By the way for your concern of a non standard 4uF capacitor, nothing stops you from paralelling 4 x 1uF. Also its tolerance will swamp most of it and to get it really accurate you will need to fiddle with the component values with your final circuit anyway.
 
Thank you, this is great info.

I'm going to stick with a passive design.

Something like this:
gadget-2.jpg


If I use a 30k shunt (as opposed to the 2k in the diagram), I can get close to an 85Hz cut-off frequency with 4.7uF caps. This is based on an input impedance of about 821 ohms. I can be flexible with the cut-off frequency. Somewhere between 80 and 100Hz ought to take care of the LF artifacts I'm experiencing. The 30k shunt value was suggested somewhat arbitrarily in the gearslutz post (mentioned earlier).

Could someone let me know if 30k is indeed an appropriate shut value? Or, is there a better value to use between a 1272 and an LA-2A?
 
Abbey brings up a good point about the "bump" in frequency (resonance freq.) when using a simple passive filter like this. I'm noticing that the Shure A15HP's spec sheet shows a minimal rise at the resonance frequency.
http://www.shure.com/ProAudio/Products/Accessories/us_pro_a15hp_ug

Is there a formula I can use to calculate the volume of the resonance frequency? Are there design techniques to minimize this effect in a passive RC high-pass?
 
Is anyone familiar with a filter design like this one?

xformer-hi-pass.jpg


It's from the Audio Cyclopedia, and intended to be placed in-line with a microphone, but I wonder if it could be adapted for line-level use.

The original text reads: A rather simple 100-Hz filter for use with a microphone to reduce the effects of wind noise is shown in Fig. 7-113, part (a). The device consists of a pi section, high-pass filter connected between two impedance-matching transformers. This filter is generally made and packaged in a metal box with standard microphone connectors at each end, for connecting directly into the microphone cable. The taps on the transformers permit the filter to be used for different impedances. The coils are toroidal wound, high-Q, and are encased in a mu-metal box to eliminate the effect of stray magnetic fields. The frequency response is shown in Fig. 7-113, part (b).

Or perhaps something like this?

l-high-pass.jpg


Thanks for your help!
 
Yes,l they all can be used as long as you maintain the source and load impedances that they are based on. Also if you can find suitable inductors.

Is there a formula I can use to calculate the volume of the resonance frequency? Are there design techniques to minimize this effect in a passive RC high-pass?

Regrettably not a simple one. I never liked equations at college (thirty years ago) and I still don't like them. I already asked what turned out to be a simple question to PRR and his reply revealed my secret that I slept thorough all my trigonometry classes. However, if you are so keen, then google for transfer functions for series and parallel resonant circuits.

What abbey points out is that when an rc circuit is loaded with a transformer it is no longer a simple rc but becomes rlc. The inductance of the transformer will interact with the capacitor. If the capacitance is very high and the inductance very low then the resistive part of the load has very little loading effect on the circuit. With circuit being poorly damped and Q becoming high the signal frequencies near resonance will start to peak.



 
I know this is a really old topic but I thought the attached schematic might be helpful for someone. The tricky part seems to be the inductors. They are a non-standard value and Shure must have them made very small to fit inside the XLR shell. I imagine a circuit using the caps alone might work with a reduced slope.

Enjoy!
 

Attachments

  • a15hp_4.gif
    a15hp_4.gif
    6.7 KB · Views: 164
> caps alone might work with a reduced slope.

No, the cutoff frequency becomes totally dependent on the load resistance which is typically not well known.

Only one coil is needed. The primary of the Radio Shack 1K:8 transformer may be very near this value.
 
According to sim, the A15 seems to be designed to work with low-ish impedance on the preamp side.
See attached sim'd FR for 2k (cyan), 1k (green) and 500 ohms (red) input impedance.
 

Attachments

  • Shure A15.jpg
    Shure A15.jpg
    73.9 KB · Views: 89
abbey road d enfer said:
According to sim, the A15 seems to be designed to work with low-ish impedance on the preamp side.
See attached sim'd FR for 2k (cyan), 1k (green) and 500 ohms (red) input impedance.

The attached datasheet notes a design load impedance of 150R
 

Attachments

  • us_pro_a15hp_ug.pdf
    146.3 KB · Views: 62

Latest posts

Back
Top