That's **WAY** too much pop. That's all the way to Xmax. At least the DC servo is doing its job and brings the output back to zero. This is NOT normal behaviour.
If the amp does that when the input is either floating or shorted, something is seriously wrong. I suggest making a shorting plug (XLR with the three pins tied together), plugging it into the amp, and testing again.
Tom
If the amp does that when the input is either floating or shorted, something is seriously wrong. I suggest making a shorting plug (XLR with the three pins tied together), plugging it into the amp, and testing again.
Tom
As long as I see Pin1 being implemented incorrectly, I feel compelled to "regurgitate" the accepted method.
Pin1 goes to Chassis.
If one has RF filtering on board, then the LOW Impedance connection that takes the RF interference into the Chassis should be a short low impedance connection at the board, not a long wire back to the remotely located Pin1 at the entry.
I am inclined to agree. It seems unlikely that the chassis of source and receiver are at significantly different voltages, if the screen/shield interconnects them.
Last edited:
nycavsr2000 I have a passive Arduino LDR preamp and then ahead of that a RPi with BOSS DAC.
But that aside the power up pops also occur with just the pseudo-differential cable connected and the RCA disconnected. If I remove the cable (all XLR pins open or short all the XLR pins I don't get the issue. If I just ground pin 2 or 3 then I get the pop.
The particular XLR chassis connector I'm using has a chassis earth tab right besides pin 1 and the two are soldered together.
But that aside the power up pops also occur with just the pseudo-differential cable connected and the RCA disconnected. If I remove the cable (all XLR pins open or short all the XLR pins I don't get the issue. If I just ground pin 2 or 3 then I get the pop.
The particular XLR chassis connector I'm using has a chassis earth tab right besides pin 1 and the two are soldered together.
Is it possible that the turn-on thump is caused by one of the power supply rails not coming up (or going down on turn-off) at the same time/rate as the other? As has been discussed previously (I think in this thread or another Mod86 thread), the mute pin is reference to only one of the supplies.
I have (yet) no first-hand experience with a built-up Neurochrome amp, but I know that LM3886-based circuits can be made to thump if, after turning off the amp, you turn it on again before the supply caps have fully discharged -- a situation that can aggravate the asymmetric ramping of the pos and neg supply rails. It doesn't sound like the case in this situation.
I have (yet) no first-hand experience with a built-up Neurochrome amp, but I know that LM3886-based circuits can be made to thump if, after turning off the amp, you turn it on again before the supply caps have fully discharged -- a situation that can aggravate the asymmetric ramping of the pos and neg supply rails. It doesn't sound like the case in this situation.
I've now got it on the bench. And it still exibits the behaviour with pin1 & 2 shorted at the chassis XLR.
This is what it looks like on the scope across the speaker output with a dummy load connected. This is just an example as they do not happen every time, will be positive or negative and will vary in size. You can clearly see the servo doing its job though.
Also to park the connection issue to pin 1 for a while it does it with or without the connection made at J1 GND. In fact I can dissconnect everything at J2 except a sorting link between GND & IN- and it will do it.
I have checked the component placement and values and they seem correct. The only thing I cant be 100% sure of are the smaller capacitors
BTW I'm draining the PSU caps to GND before each test.
This is what it looks like on the scope across the speaker output with a dummy load connected. This is just an example as they do not happen every time, will be positive or negative and will vary in size. You can clearly see the servo doing its job though.
Also to park the connection issue to pin 1 for a while it does it with or without the connection made at J1 GND. In fact I can dissconnect everything at J2 except a sorting link between GND & IN- and it will do it.
I have checked the component placement and values and they seem correct. The only thing I cant be 100% sure of are the smaller capacitors
BTW I'm draining the PSU caps to GND before each test.
Attachments
The plop is caused by the input bias current of the THAT1200. I am now able to reproduce it in my setup too.
If you leave the pseudo-differential cable floating, you'll get the plop. If you connect the pseudo-differential cable to a DC coupled source, you do NOT get a plop. With an AC coupled source you DO get a plop.
If you use a differential source or leave the differential inputs floating, you do NOT get a plop.
This issue is caused by the rather large input bias current (1400 nA) and input offset current (±300 nA) of the THAT1200. You need to provide a path for this current to flow or it will dump into the input resistors within the THAT1200 and cause a large output voltage offset.
There are two ways to address this:
1) Use differential connections.
2) Use a pseudo-differential connection and make sure to use a DC coupled output (such as an opamp buffer).
Using a differential driver, such as my THAT Driver, will resolve the issue as well.
I'm surprised I didn't stumble upon this when I built my 4xMOD86 with MiniDSP 2x4 built in as I used pseudo-differential connections from the MiniDSP to the MOD86es. None of those four channels plop and I could swear the MiniDSP is AC coupled. Then again, maybe they saved the output cap and just used an opamp output.
Tom
If you leave the pseudo-differential cable floating, you'll get the plop. If you connect the pseudo-differential cable to a DC coupled source, you do NOT get a plop. With an AC coupled source you DO get a plop.
If you use a differential source or leave the differential inputs floating, you do NOT get a plop.
This issue is caused by the rather large input bias current (1400 nA) and input offset current (±300 nA) of the THAT1200. You need to provide a path for this current to flow or it will dump into the input resistors within the THAT1200 and cause a large output voltage offset.
There are two ways to address this:
1) Use differential connections.
2) Use a pseudo-differential connection and make sure to use a DC coupled output (such as an opamp buffer).
Using a differential driver, such as my THAT Driver, will resolve the issue as well.
I'm surprised I didn't stumble upon this when I built my 4xMOD86 with MiniDSP 2x4 built in as I used pseudo-differential connections from the MiniDSP to the MOD86es. None of those four channels plop and I could swear the MiniDSP is AC coupled. Then again, maybe they saved the output cap and just used an opamp output.
Tom
You're probably thinking of the degradation in CMRR caused by an imbalance in the series resistance (or driver output impedance). The InGenious input of the THAT1200 is designed to minimize the degradation in CMRR from such an imbalance.
Tom
Where else could you get product support on a Saturday guys? 
Tom your a star.
I think you know what I'm ordering next!.. Time to give my preamp a balanced line output.
You can also see why I have been keen on a power-up mute and speaker protection. I should have questioned the behaviour earlier.
I'm assuming that it would be best to power up the amp last? So back to one of my other questions does anyone see any issues with using a solid state relay to switch the supply via the pre-amp?
Tom your a star. I think you know what I'm ordering next!.. Time to give my preamp a balanced line output.
You can also see why I have been keen on a power-up mute and speaker protection. I should have questioned the behaviour earlier.
I'm assuming that it would be best to power up the amp last? So back to one of my other questions does anyone see any issues with using a solid state relay to switch the supply via the pre-amp?
Tom, I believe the miniDSP boards all run on a single supply rail for their output op-amps, so they should be AC coupled on the audio output pins.
Am I correct in pointing out that the Mod286 board doesn't have the THAT chip and the lower bias current of the OPA1612 input buffers eliminates the possibility of plop under similar conditions?
Could the THAT1200 problem with psuedo-diff be remedied with a resistor to ground at the input? Perhaps 10k? Or, if differential mode is to be maintained, a pair or well-matched 10k, one from each input pin to ground?
Am I correct in pointing out that the Mod286 board doesn't have the THAT chip and the lower bias current of the OPA1612 input buffers eliminates the possibility of plop under similar conditions?
Could the THAT1200 problem with psuedo-diff be remedied with a resistor to ground at the input? Perhaps 10k? Or, if differential mode is to be maintained, a pair or well-matched 10k, one from each input pin to ground?
I agree. It's the input bias current and input offset current into the input impedance creating the voltage. The reason for the transient behaviour on the output of the THAT1200 is that the common-mode cap on the THAT1200 (Cb in the data sheet) needs to charge.
Charging Cb generates a plop on the output of the THAT1200, which the MOD86 amplifier section happily amplifies, so you get a plop on the output.
Oh, you know you're right. It runs on a 3.3 V rail, so the outputs are AC coupled.
Yep. The significantly lower input bias current and input bias offset current of the OPA1612 will significantly reduce any plop. The input on the MOD286 also doesn't have the common-mode cap, so it won't have that transient behaviour either.
A 5 kΩ resistor from the input to ground would reduce the plop by a factor of 10.
Tom
I am confident that Tom will help with any turn-on/off issues, and personally I think that should be explored to it's natural conclusion.
But if anyone feels it's necessary, the somewhat unconventional genius who makes output transformerless amps and preamps over at Transcendent Sound offers a simple mute PCB that can be used if desired, for $35. He's been in business for a long time and has a good reputation.
Mute Circuit
But if anyone feels it's necessary, the somewhat unconventional genius who makes output transformerless amps and preamps over at Transcendent Sound offers a simple mute PCB that can be used if desired, for $35. He's been in business for a long time and has a good reputation.
Mute Circuit
