Bridged T Attenuator Simulation Problem

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

voitto

Well-known member
Joined
Dec 19, 2012
Messages
45
Location
Janakkala - Finland
Hi there,

I'm attempting to design a 2.5k bridged-T attenuator to act as a volume control, I made an excel spreadsheet based on the equations presented here - https://www.electronics-tutorials.ws/attenuators/bridged-t-attenuator.html

The problem I'm having is that while the figures from this spreadsheet seem to match up with other online bridged-T calculators, when I ran it through a spice simulation the attenuation figures don't match.

I've attached the spreadsheet and a spice screenshot. If anyone has the time to look this over and point out where I'm going wrong i'd be very grateful.

The final volume control will be either a rotary switch or a relay stepped thing, so the values used in series will differ, this excel sheet is just an attempt to understand it.

Excel file and spice screenshot here - https://www.dropbox.com/sh/foebqw6y2oxw6yj/AACJ4I-lAhPawuILHLIV_CWYa?dl=0

Best,
Robin
 

Attachments

  • attenuator schem.PNG
    attenuator schem.PNG
    26.8 KB · Views: 26
voitto said:
I think I've answered my own question, I put a 2.5k resistor in series at the input and the values matched.

Yes, the equations assume equal source and load resistances.

When turned right up the loss is 6dB

Cheers

Ian
 
Cool!
So if i was to implement this as a volume control, would I have to adjust that resistor value based on the output impedance of the thing that's driving it? How would I go about that, use a rheostat and measure the resistance when the attenuation matches?
 
voitto said:
Cool!
So if i was to implement this as a volume control, would I have to adjust that resistor value based on the output impedance of the thing that's driving it? How would I go about that, use a rheostat and measure the resistance when the attenuation matches?

Bridged T attenuators are really designed to be used in systems where the impedance is well defined. In the old days input and output impedance were 600 ohms so this way easy. Today we use bridging loads so this is not the case. The question is, why do you want to used a bridged T attenuator?

Cheers

ian
 
Yes, why a 'T'? 

Loaded with 2K5 it will be pretty close to proper operation.  Into a bridging load it will not attenuate as much as expected, the curve will be greatly shifted, but still work. 
 
ruffrecords said:
The question is, why do you want to used a bridged T attenuator?

Ultimately I hope to end up with a volume control between my Dac (Benchmark DAC1) and my loudspeakers  (Meyer Hd1)
I've been using a pot until recently, and then a relay stepped attenuator, but with both I can hear differences in the high end at different levels. I was looking to the bridged-t attenuator as a way round this.
 
I frequently measure differences in treble with old Daven and new continuous T's at the extremes of attenuation.  Even greater with ladders.  It could be helpful to design for a smaller range if it suits. 
 
voitto said:
Ultimately I hope to end up with a volume control between my Dac (Benchmark DAC1) and my loudspeakers  (Meyer Hd1)
I've been using a pot until recently, and then a relay stepped attenuator, but with both I can hear differences in the high end at different levels. I was looking to the bridged-t attenuator as a way round this.

OK, what impedance were the pot and relay attenuators?

Cheers

Ian
 
If you get response errors with both a pot and a relay attenuator, maybe the wiring to the volume control is the problem. If there's leakage between the send and return, then HF feedthrough will couple to the output at a greater proportion at high attenuator settings, making the signal brighter at quieter levels.
 
Monte McGuire said:
If you get response errors with both a pot and a relay attenuator, maybe the wiring to the volume control is the problem. If there's leakage between the send and return, then HF feedthrough will couple to the output at a greater proportion at high attenuator settings, making the signal brighter at quieter levels.

I'm with Monte here but there are all sorts of other variables.

One thing to note is that "relay attenuator" means very little.  Was it a 5k ladder or 100k?  Was it buffered on the input, the output, both or neither? 

Cable capacitance can be an issue if you have a long run and sufficiently high output impedance on your attenuator.


 
Calculations and sims do not take into account parasitics; with a characteristic impedance of 2.5kohm, the capacitance of a few meters of shielded cable can have a significant effect.
With modern equipment, we are not anymore in a matched-Z world; let's take advantage of their low source impedance and drive capability. A 2.2k pot (simple L-attenuator) presents a worst-case impedance of about 600 ohms to the cable, allowing 4x the length compared to 2.5k.
 
ruairioflaherty said:
Cable capacitance can be an issue if you have a long run and sufficiently high output impedance on your attenuator.

Cable capacitance and a poor remote ground to the attenuator could also be the cause of capacitive HF feedthrough around the attenuator.
 
ruffrecords said:
OK, what impedance were the pot and relay attenuators?
ruairioflaherty said:
One thing to note is that "relay attenuator" means very little.  Was it a 5k ladder or 100k?  Was it buffered on the input, the output, both or neither? 

The pot was 10k,

I made two attempts at the relay version;  ladder type  600ohm and 10k. Both were balanced and unbuffered. there's about 3meters between attenuator and amplifier input. The differences in HF were noticeable below 45dB of attenuation.

emrr said:
I frequently measure differences in treble with old Daven and new continuous T's at the extremes of attenuation.  Even greater with ladders.  It could be helpful to design for a smaller range if it suits. 

I'm sure this is an 'it depends' answer, but for this kind of application what sort of usable range can you expect to from a bridged-T?

I have been wondering if an active solution might be more suited. Perhaps something like the circuit presented in this article, replacing the pot with relays. - http://www.edn.com/Pdf/ViewPdf?contentItemId=4430977

I'll also test the wiring!
 
voitto said:
Not entirely,  I’ll do some measurements with a mic to verify.
What do you mean "measurements with a mic"? Any measurement of the effect you describe should not involve any transducer, just electronics. Some of your earlier comments make me think you are trying to find a solution in a non-scientific way; you should change only one parameter at a time. Comparing a 600 ohm ladder attenuator with a 10k pot (L-pad) is like comparing oranges and onions.
 
abbey road d enfer said:
What do you mean "measurements with a mic"? Any measurement of the effect you describe should not involve any transducer, just electronics. Some of your earlier comments make me think you are trying to find a solution in a non-scientific way; you should change only one parameter at a time. Comparing a 600 ohm ladder attenuator with a 10k pot (L-pad) is like comparing oranges and onions.

I had also built a 10k ladder type and found the same issue.

I mentioned the mic as a means to verify if my ears were playing tricks on me at lower levels, but if you you can point me toward a method to emulate the load of the following stage and measure this electronically I’ll give it a go :)
 
Measure the voltage at the amplifier end of the 3m cable, 100Hz, 1KHz, 10KHz, 0dB and -45dB.

However 3m of cable is only 100pFd and assuming 2.5K impedance in a 10K pad that's 600KHz. Very possibly there is 470pFd inside the amplifier input making 110KHz hi-cut.

45dB is a LOT of change in sound level and WILL induce significant change in ear perception.

Measuring with a mike ALSO measures all the errors in speaker and room; -45dB may put you down near background noise. You are cluttering the test with things not related to the attenuator. If the mike moves a few inches that will be far more change than the pad and cable.
 
ruffrecords said:
Are you sure this a real effect and not simply the change in response of the ear at low levels?

You beat me to it

Also, if you're going to use a relay-operated attenuator, maybe you could locate the attenuator next to the monitors and keep the control close by? That should solve the problem

Nick Froome
 

Latest posts

Back
Top