The drawing is misleading. The top wire goes only to the wiper, not on the last position.
That's how many attenuators go, emulating potentiometers.
Design for 600R Bridged-T Pad Attenuator 24 Step Rotary Switch!
- Thread starter friesdan
- Start date
Help Support GroupDIY Audio Forum:
scott2000
Well-known member
Sorry, I'm not following... Are you saying there's a connection that shouldn't be there? Maybe I goofed in this area? I'll keep at it...
I meant that the 1st drawing may be confusing. It's easy to think that the upper wire goes to the last position as well as the wiper.
No there isn't. It's shunted fully to ground. There may be leakage that's audible as very low level audio though, depending on wiring and layout.
scott2000
Well-known member
I think I follow you. I'll check to see if that's what I did.
scott2000
Well-known member
So my drawing with hot and cold isn't correct? I am using it with balanced in/out...... I should tie ground to cold?(this seems to work) but at least one of my converters needs to be set up internally or I use special cables( disconnecting cold) if using it like this I think. Would like to avoid that. Need to look into it.
Unless I'm not understanding which it feels like...
Attachments
Last edited:
ruffrecords
Well-known member
I am trying to convince people not to call the third leg of a balanced connection "ground" because it isn't. The signal in a balanced connection exists only between the hot and the cold. The third pin does not carry signal. It is meant to act as a screen. It may well be connected to "ground" or "chassis" at one or both ends but its function is to screen the connection, no more no less.
Cheers
Ian
scott2000
Well-known member
I had to tie the screen to the box/chassis the switch is in because of noise. If I use the hot and cold pins as in the drawing, the switch works but I don't get silence at off. If I tie the cold to the chassis as well , the switch works with silence at off but, doesn't this unbalance this chain and some units that want pin 3 left unconnected if using unbalanced need to be considered? I still need to look at the wiring more for errors but, if it has to be shunted to ground in order to have silence at off,how does this happen without tying cold to ground? I know it'll click for me, just lost atm...
I know it takes a 3 pole switch to use for true balanced but, I was under the assumption I could use a 2 pole with balanced with no worries about anything.
Last edited:
Seems to indicate the CMRR of the receiver is too small.
scott2000
Well-known member
Pretty sure I've checked all my wiring.... I'm guessing the wipers are always connected to position 1 on deck 1 and 23 on deck 2? I just jumped the cold pins (3) input/output and they're tied to position 1 on deck 2. Pretty sure my drawing is right?
Don't understand how I could get complete silence unless grounding the cold pin? I guess I need to measure all the steps anyway to see what is going on in reality anyway. Sounds fine. 1dB steps are pretty small... 2dB would've been better for more attenuation.
Don't understand how I could get complete silence unless grounding the cold pin? I guess I need to measure all the steps anyway to see what is going on in reality anyway. Sounds fine. 1dB steps are pretty small... 2dB would've been better for more attenuation.
Attachments
radardoug
Well-known member
This will never work. You have a permanent short between the two series resistors.
You shouldn't guess. No, they must not be connected. As I mentioned earlier, teh drawing is a little confusing. The upper wire goes to the wiper only.
scott2000
Well-known member
Is this a valid way of checking the bridged T?
( can't get video to post for some reason so added a screenshot)
Is he just putting a load resistor (I used 2.5k) across hot and cold on the outputs and putting the meter across hot and cold on the input?
Is the meter showing what the switch impedance looks like to the source feeding it?
Could you just hook up the equipment as well? Like plug an amp into the switch outputs and read the input of the switch and see what it looks like? What about plugging in the source and load equipment and taking measurements? Any usefulness there?
I'm getting 2.5k across all the steps btw when testing like I mentioned above but not sure if that means anything.
( can't get video to post for some reason so added a screenshot)
Is he just putting a load resistor (I used 2.5k) across hot and cold on the outputs and putting the meter across hot and cold on the input?
Is the meter showing what the switch impedance looks like to the source feeding it?
Could you just hook up the equipment as well? Like plug an amp into the switch outputs and read the input of the switch and see what it looks like? What about plugging in the source and load equipment and taking measurements? Any usefulness there?
I'm getting 2.5k across all the steps btw when testing like I mentioned above but not sure if that means anything.
Attachments
Last edited:
ruffrecords
Well-known member
Yes it is valid. If the T is working properly, if you load it with 600 ohms it should look like 600 ohms across the input no matter what the position of the switch.
Cheers
Ian
Cheers
Ian
waltzingbear
Well-known member
a bridged T is meant to be used in a matched impedance signal distribution situation, and built for the specific impedance's involved. In the old days of matched distribution systems, 600 ohms was the most common impedance (although 150 ohms was also used commonly), hence 600 bridged T attenuators were common (ie Daven)
If you design to a different load and source value, as you have done, then you use that, ie 2.5K, as your load termination value. Assuming you have a bridging load, nominally greater than 10kohms. Although as Abbey has mentioned, you generally want greater than 10X the term resistance to qualify as a bridging load(25K here). This is because what you want in reality is the parallel load impedance of the term resistor and the input impedance of the following equipment to be the desired load impedance (2.5K) AND the output impedance of the driving equipment to be the desired impedance, (2.5K, generally using a build out resistor)
The bridged T will work fine without meeting these conditions, but as has been mentioned before, the step values will not be as calculated nor will the load value be constant. But they will not vary much and the errors are commonly ignored. You will still have repeatable and matching step response.
Cheers
ps
the errors are at the top of the range, if the attenuator is commonly used in the middle or below of the range the errors are small.
If you design to a different load and source value, as you have done, then you use that, ie 2.5K, as your load termination value. Assuming you have a bridging load, nominally greater than 10kohms. Although as Abbey has mentioned, you generally want greater than 10X the term resistance to qualify as a bridging load(25K here). This is because what you want in reality is the parallel load impedance of the term resistor and the input impedance of the following equipment to be the desired load impedance (2.5K) AND the output impedance of the driving equipment to be the desired impedance, (2.5K, generally using a build out resistor)
The bridged T will work fine without meeting these conditions, but as has been mentioned before, the step values will not be as calculated nor will the load value be constant. But they will not vary much and the errors are commonly ignored. You will still have repeatable and matching step response.
Cheers
ps
the errors are at the top of the range, if the attenuator is commonly used in the middle or below of the range the errors are small.
Last edited:
ruffrecords
Well-known member
If you load it with a value higher than 600 ohms all bets are off although looking into the input it will still be close to 600 a lot of the time. It should never be less than 600 but may be greater than it.
Cheers
Ian
waltzingbear
Well-known member
what do you want this attenuator to do?
what problems do you want it to solve?
when you have the specific answers that can help you decide what parts of the design are important.
A bridged T gives you constant input and output impedances, if implemented with designed source and termination resistances it will give expected step attenuation responses.
Deviations at the high attenuation end of the range are more from capacitance leakage than design or termination errors.
what problems do you want it to solve?
when you have the specific answers that can help you decide what parts of the design are important.
A bridged T gives you constant input and output impedances, if implemented with designed source and termination resistances it will give expected step attenuation responses.
Deviations at the high attenuation end of the range are more from capacitance leakage than design or termination errors.
ruffrecords
Well-known member
The thing about a bridged T is that with the designed source and load, it presents a constant load to the source and and a constant source to the load. if you do not need that particular property then there is really no reason to use it.
if, as it appears, you are feeding from a low source impedance and driving a much higher impedance load then you might as well use a simple balanced PI attenuator. The attached schematic is for a limited range trimer control but the principle can be extended to any number of switch positions.
Cheers
Ian
Attachments
Similar threads
