How did you probe this? Two probes with a common ground on eiother side of the R?
BTW It would help a lot if you post scope shots to include the X- and Y-sclaes. A wiggle picture on itself is not very clarifying.
Jan
BTW It would help a lot if you post scope shots to include the X- and Y-sclaes. A wiggle picture on itself is not very clarifying.
Jan
I simply connected the RTX balanced input across the resistor in the C-R-C splitter.
Sorry for cutting of the axis scales, that was a stupid oooopsie. The wiggle frequency is 1 kHz, same as the frequency of the sine test signal.
Sorry for cutting of the axis scales, that was a stupid oooopsie. The wiggle frequency is 1 kHz, same as the frequency of the sine test signal.
Thats not a usefull test and wont tell anything about the bal performance. There's no particular spec for se use of a bal connection. It is only defined as the diff between them. Hence my suggestion to send the same signal to both phases and check the cm performance to judge that wiggle graph.
Jan
Jan
@jan.didden your test makes sense to me. I will take a look once I am back at the workbench (next week?).
Hi Jan,
Yes, I am well aware of that. I'm trying to break it down to find the issue, not prove balanced performance.
We know the balanced performance of the RTX is very good, I know from personal experience from doing my own balanced measurements. What is at question is the divider and the method I would think. I would think that first, we prove one leg of a divider works properly, troubleshoot if it doesn't. Then we can move onto getting the two legs balanced.
Yes, I am well aware of that. I'm trying to break it down to find the issue, not prove balanced performance.
We know the balanced performance of the RTX is very good, I know from personal experience from doing my own balanced measurements. What is at question is the divider and the method I would think. I would think that first, we prove one leg of a divider works properly, troubleshoot if it doesn't. Then we can move onto getting the two legs balanced.
Hi Jan,
I'm troubleshooting. The first step is to make sure the divider is linear and low distortion as a result.
I'm troubleshooting. The first step is to make sure the divider is linear and low distortion as a result.
But low distortion of a leg is useless. The bal output should be low distortion. A diff thing.
Jan
Jan
Hi Jan,
I agree completely. But to have a low distortion differential, each leg must be linear. Then they must be matched. You have to prove it is before trying to troubleshoot beyond that.
I agree completely. But to have a low distortion differential, each leg must be linear. Then they must be matched. You have to prove it is before trying to troubleshoot beyond that.
I decided to verify what I proposed. First the input C of the RTX is around 40 pF and doesn't change much with different attenuation. This makes it all pretty straightforward. The 100 pF of cap needs to withstand the full voltage with some headroom (or it could be expensive or worst lethal). its critical that the ground is connected. Also there is an input cap on the RTX which prevents a DC path to ground on the divider so any DC will transfer directly to the input. A 1 meg resistor at the output of the probe would alter the divider a little but would fix that issue.
Attachments
That should work.
I have the same darned capacitance meter - lol! Two versions of the same meter, and the calibration box.
I have the same darned capacitance meter - lol! Two versions of the same meter, and the calibration box.
Your circuit shows GND on one end of the amp output. However, that's unfortunately not how estat headphone amps work. These amps have a bipolar output, so both output pins are different from GND. Its a bit like a bridged amp.
Exactly, but the signal is referenced to ground with two phases output. Not unless it is with a transformer output (then you could in theory ground one side).
Hmm, this does not have the amp outputs at GND, but the schematic still shows a GND. Which GND is that? GND = RTX chassis?
Also, I believe your new schematic uses two inputs to the RTX. Wouldn't it be better to use just one input in balanced/differential mode? I guess this would mean to do a C1-C2-C1 voltage splitter with C1 = 200 pF and C2 = 200 nF, and probe across C2. Similar to my C-R-C splitter idea.
Yeah that's what you get when you insist on re-introducing gnd as a reference in a bal connection.
People still confuse balanced with symmetrical and still think that balanced means two opposite phase equal magnitude signals referred to ground.
You get measurements that may or may not have any meaning for the bal case.
I wish you guys good luck.
Jan
People still confuse balanced with symmetrical and still think that balanced means two opposite phase equal magnitude signals referred to ground.
You get measurements that may or may not have any meaning for the bal case.
I wish you guys good luck.
Jan
Hi Jan,
At the moment I was trying to solve the measurement problem to find the cause. Period.
You are correct that a balanced signal does not require a ground, however since the input of any device that isn't 100% isolated from "ground" may have a common mode problem, you cannot ignore the possibility.
At the moment I was trying to solve the measurement problem to find the cause. Period.
You are correct that a balanced signal does not require a ground, however since the input of any device that isn't 100% isolated from "ground" may have a common mode problem, you cannot ignore the possibility.