Owen
Thanks for the response. BTW if I should be asking these questions in the other thread please let me know.
What I'm thinking about at the moment is to do the following:
- use two unused DAC channel outputs from my Theta Casablanca IV and duplicate the main L/R channels to these. The CB IV can then supply volume controlled balanced output .
- the DAC card has the following specs:
- the headphones I have desired for some time aren't cheap. I would be a bit insecure about pulling them to pieces. Some specs:
Given they are 'made to order' I have asked if they can be made with balanced connections.
My presumption is that I would leave the BAL-BAL gain at 1x.
What's the recommended PSU current capability (roughly, given 18R load, +/- 12V supplies)?
Sorry if these are basic questions, I'm relatively new to this stuff.
Regards
Steve
PS: for the connections to the output XLR connectors, I see +output and -output on the board. XLR pin 1?
EDIT: seems my DAC is remarkably similar to R1200CL
Thanks for the response. BTW if I should be asking these questions in the other thread please let me know.
What I'm thinking about at the moment is to do the following:
- use two unused DAC channel outputs from my Theta Casablanca IV and duplicate the main L/R channels to these. The CB IV can then supply volume controlled balanced output .
- the DAC card has the following specs:
- the headphones I have desired for some time aren't cheap. I would be a bit insecure about pulling them to pieces. Some specs:
Given they are 'made to order' I have asked if they can be made with balanced connections.
My presumption is that I would leave the BAL-BAL gain at 1x.
What's the recommended PSU current capability (roughly, given 18R load, +/- 12V supplies)?
Sorry if these are basic questions, I'm relatively new to this stuff.
Regards
Steve
PS: for the connections to the output XLR connectors, I see +output and -output on the board. XLR pin 1?
EDIT: seems my DAC is remarkably similar to R1200CL
Last edited:
NTD1 V.4 Polymer cap leakage, DC uh-oh!
Hi Owen,
I got the NTD1 and power supply mounted up to a 3/8"x15"x17" slab of 6061. This appears to be just adequate sinking. The 330µF 25V polymer caps are pretty leaky though? I see ~18 VDC at their input and over 1VDC leakage on each output.
I did listen using another amp I have which is cap coupled on its inputs, and the sound was quite good, but I cannot have over 1VDC at the inputs of my usual (direct coupled) amplifier.
Is this amount of leakage normal for these parts? I did allow for some charging/forming time, and still have more than 1 VDC at the outputs, any ideas?
Perhaps a higher Voltage rated part is needed here? I really like the conceptive using the polymer caps here, and was excited to not have to resort to enormous film caps, but the dc levels are not acceptable for my amplifier.
Hi Owen,
I got the NTD1 and power supply mounted up to a 3/8"x15"x17" slab of 6061. This appears to be just adequate sinking. The 330µF 25V polymer caps are pretty leaky though? I see ~18 VDC at their input and over 1VDC leakage on each output.
I did listen using another amp I have which is cap coupled on its inputs, and the sound was quite good, but I cannot have over 1VDC at the inputs of my usual (direct coupled) amplifier.
Is this amount of leakage normal for these parts? I did allow for some charging/forming time, and still have more than 1 VDC at the outputs, any ideas?
Perhaps a higher Voltage rated part is needed here? I really like the conceptive using the polymer caps here, and was excited to not have to resort to enormous film caps, but the dc levels are not acceptable for my amplifier.
Hi Barrows,
I had the same measurements and the DC was high enough to trigger my nCore400's DC protection. I used 33uF/250v Clarity Cap ESA and it not only works but sounds great. It seems that for a DC coupled preamp/amp, the polymer caps is too leaky. The only low leakage polymer caps available doesn't have a high enough voltage rating, unfortunately.
I had the same measurements and the DC was high enough to trigger my nCore400's DC protection. I used 33uF/250v Clarity Cap ESA and it not only works but sounds great. It seems that for a DC coupled preamp/amp, the polymer caps is too leaky. The only low leakage polymer caps available doesn't have a high enough voltage rating, unfortunately.
DC leakage...
I had some Blackgate NP 10µF 50 V parts laying around, so subbed those in for preliminary listening. Sounds fantastic, definitely better than my highly tweaked Legato 3 with all premium parts. Psyched. DC out is <4mV with BGs.
I will probably swap out to some high quality film/foil caps soon.
I had some Blackgate NP 10µF 50 V parts laying around, so subbed those in for preliminary listening. Sounds fantastic, definitely better than my highly tweaked Legato 3 with all premium parts. Psyched. DC out is <4mV with BGs.
I will probably swap out to some high quality film/foil caps soon.
Hi Owen,
Would you mind providing insight into the following two questions when you have a moment? I'd love to get my board order with you finalized and build underway.
Are the Texas Components TX220S resistors, with a 10 watt max power rating, sufficient for the NTD1, or is a higher power rating required? The 45R and 98R custom value TX220Ss are already pricy at ~$30 per resistor, so $240 for a complete circuit may approach the range of doing a custom run of PFS35s from Riedon.
It seems like a major roadblock in building the low-voltage NTD1 is finding a source for the 45R and 98R power resistors.
For maximum performance, would you recommend populating the BAL-BAL with a LME49724 op-amp and running it at 18v from the same pair of LT3042 regulators as the NTD1, or should I use the slightly better-spec'd OPA1632, run the BAL-BAL at 12v, and power it from a separate pair of LT3042 boards?
Would you mind providing insight into the following two questions when you have a moment? I'd love to get my board order with you finalized and build underway.
Are the Texas Components TX220S resistors, with a 10 watt max power rating, sufficient for the NTD1, or is a higher power rating required? The 45R and 98R custom value TX220Ss are already pricy at ~$30 per resistor, so $240 for a complete circuit may approach the range of doing a custom run of PFS35s from Riedon.
It seems like a major roadblock in building the low-voltage NTD1 is finding a source for the 45R and 98R power resistors.
For maximum performance, would you recommend populating the BAL-BAL with a LME49724 op-amp and running it at 18v from the same pair of LT3042 regulators as the NTD1, or should I use the slightly better-spec'd OPA1632, run the BAL-BAL at 12v, and power it from a separate pair of LT3042 boards?
Last edited:
Hey guys...
Maybe we need to start a new build thread for the new version NTD1?
I have 4 pieces of 3 µF 200V film foil caps coming from MusiCap for my build, psyched to hear it with those soon...
I have another question for OPC or anyone who understands circuit design better than I (which is many folks here). With the off board caps I will put 100K load resistors at the XLR jacks, hence I have removed the onboard 100K resistors.
Could I place some caps at the position the onboard load resistors were, to provide an analog low pass filter? I would assume these would operate with the output impedance (200R) to make the corner frequency, am I correct here? I do get some noise with my amp, a course noise floor if you will, which in my experience with adjusting the analog filter on other DACs is due to HF noise from the ESS 9018.
Maybe we need to start a new build thread for the new version NTD1?
I have 4 pieces of 3 µF 200V film foil caps coming from MusiCap for my build, psyched to hear it with those soon...
I have another question for OPC or anyone who understands circuit design better than I (which is many folks here). With the off board caps I will put 100K load resistors at the XLR jacks, hence I have removed the onboard 100K resistors.
Could I place some caps at the position the onboard load resistors were, to provide an analog low pass filter? I would assume these would operate with the output impedance (200R) to make the corner frequency, am I correct here? I do get some noise with my amp, a course noise floor if you will, which in my experience with adjusting the analog filter on other DACs is due to HF noise from the ESS 9018.
Barrows,
The original schematics for OPC's D1 showed filtering caps across the 200R & from the junction of the 400R & the MOSFET to ground. I've got a couple of schematics with those positions shown, but only '?' on the caps for values. Owen did publish values for those caps in one of his early schematics... shall I forward to you?
Good to hear your impressions of the NTD1 compared to your earlier I/V stages. Confirms other reports I've heard. Slowly working on my builds that will use this I/V... still need to work out a heatsinking scheme I like.
Gonna post updated pix on your CA thread?
Greg in Mississippi
The original schematics for OPC's D1 showed filtering caps across the 200R & from the junction of the 400R & the MOSFET to ground. I've got a couple of schematics with those positions shown, but only '?' on the caps for values. Owen did publish values for those caps in one of his early schematics... shall I forward to you?
Good to hear your impressions of the NTD1 compared to your earlier I/V stages. Confirms other reports I've heard. Slowly working on my builds that will use this I/V... still need to work out a heatsinking scheme I like.
Gonna post updated pix on your CA thread?
Greg in Mississippi
Greg...
Once i get the MusiCaps installed and do some more listening I'll post something at CA, this thing is still in the "experimental" stage for me. It will be a long while before the chassis gets built out, but everything should be pretty solidly mounted up on my aluminum slab by the end of this week...
I can figure the cap values for an analog filter in conjunction with the 200R output impedance, just kind of wanted to know from those who understand circuit design if putting filter caps at the output would work as I suspect it would. I'll probably shoot for around a 200 kHz corner frequency or so..., easy to try anyway.
Once i get the MusiCaps installed and do some more listening I'll post something at CA, this thing is still in the "experimental" stage for me. It will be a long while before the chassis gets built out, but everything should be pretty solidly mounted up on my aluminum slab by the end of this week...
I can figure the cap values for an analog filter in conjunction with the 200R output impedance, just kind of wanted to know from those who understand circuit design if putting filter caps at the output would work as I suspect it would. I'll probably shoot for around a 200 kHz corner frequency or so..., easy to try anyway.
Hi Owen,
I have some questions for NTD1 V3.
1. ADP7142ARDZ-5.0 in PSU BOM (U5-8), part name is correct? I think the part is for 5V fixed output.
2. I'm going to use a toroidal trans with (32v-0v 600mA) * 4 outputs. What do you think its VA?
3. Is it possible to use wirewound resistor (ex, A102433-ND, digikey) instead of Reidon?
TIA,
Sean
I have some questions for NTD1 V3.
1. ADP7142ARDZ-5.0 in PSU BOM (U5-8), part name is correct? I think the part is for 5V fixed output.
2. I'm going to use a toroidal trans with (32v-0v 600mA) * 4 outputs. What do you think its VA?
3. Is it possible to use wirewound resistor (ex, A102433-ND, digikey) instead of Reidon?
TIA,
Sean
Hi Sabrosa,
My apologies for leaving you hanging so long on this one.
The Texas Components parts will be just fine for this application, as the maximum power dissipation in the standard setup is just under 4W for the lower leg resistors, and just under 2W for the upper leg.
For the low gain version, you're also fine to use the Texas Components parts as the power dissipation is the same as above if you use 18V rails and 45R/98R resistors.
I should also mention that it's probably not worth spending a pile of money to get exactly 98R and 45R. If they have 100R and 50R as standard values for a lower cost, just go for those. Also, remember that it's easy to trim these values by soldering higher value 1210 SMD film resistors across the terminals of the D2PAK resistors. I did this on mine when I couldn't source the 180R/390R film parts. I bought 200R/400R parts and trimmed them to the correct values using the above technique. The gain drops a little, but there is otherwise no measurable performance improvement so it's probably not even worth doing.
It's also worth mentioning that those Texas Components resistors are TO-220 packs modified for SMD use, so you will need to cut off the upper portion of the tab (where the screw hole is) to get them to fit on the D2PAK footprints. This should be pretty easy to do, and a Dremel tool or a shear would work well. You might even get away with some side cutters and patience
For the power arrangement, I would suggest just using an LME49724 and running everything off the same pair of LT3042 regulator boards all set to +/-18V.
It's debatable which of the LME49724 or the OPA1632 are really better, and there isn't enough in the datasheet to give the whole picture. Noise is lower on the OPA1632, but otherwise it's about a wash.
If absolute maximum performance is your only objective, then it might be slightly advantageous to run the OPA1632 as you described below with 3 supplies, but I would bet money that there would be no measurable difference.
Regards,
Owen
Hi Barrows:
I double checked the DC leakage on those caps, and you are indeed correct, they seem to be quite leaky initially and then they settle down to much lower leakage after several hours of on-time. That's why I didn't catch it in my testing. I had already burned everything in before running any tests.
On my NTD1 with the same polymer caps as suggested in the BOM, after a few hundred hours, I'm seeing between 40mV and 80mV common mode DC, and about 2-3mV differential mode DC. This would never be enough to cause problems with a downstream amplifier, even if it is DC coupled.
I tossed in a brand new pair of caps and measured an initial value of 2.3VDC common mode at the output Luckily diff mode was still below 5mV so most proper differential inputs should cope with this just fine.
In the end though, I think it's probably best to use standard (non-polymer) SMD electrolytics and avoid this whole problem. I will update the BOM accordingly. Alternatively, you could use large film caps if you would like. The value of the output coupling caps depends entirely on the input impedance of the stage following the NTD1. I personally would aim for at least 10uF and preferably over 20uF, but that's because I tend to prefer lower input impedance for my amplifiers. If your amplifier has an input impedance of 100k, then 3uF is more than enough.
There should already be a build thread for the older version... I would use that thread instead of creating a new one.
Very curious to hear the results! Keep us posted.
All the above is exactly correct, and I do cover what value should be used at some point in the original NTD1 thread. I ran several simulations to get the correct values, and then I measured them in-circuit. It does exactly what you would expect, in that it attenuates above ~20kHz but there is also a tiny rolloff and associated phase shift in the pass band.
The original Pass Labs D1 circuit places small value caps across each of the power resistors (eg, input to V- and output to V+) to accomplish the same thing. This can easily be done by soldering 1206 C0G caps across the SMD resistor terminals. The details are shown in the original NTD1 V1 schematic that Greg refers to.
Finally, which amplifier are you using this with? Noise at the output is very odd, and indicates something might be adrift with the amp itself. Is it prone to oscillating? I've tried the NTD1 with dozens of amplifiers at this point, and have never heard anything at all at the output. I'm also curious why it would have struggled with 1V of common mode DC voltage at the input. An amplifier with a proper differential input stage should be fine with a few volts of common mode DC. It's diff mode DC that causes problems, and that should have only been a few mV worst case.
Regards,
Owen
I double checked the DC leakage on those caps, and you are indeed correct, they seem to be quite leaky initially and then they settle down to much lower leakage after several hours of on-time. That's why I didn't catch it in my testing. I had already burned everything in before running any tests.
On my NTD1 with the same polymer caps as suggested in the BOM, after a few hundred hours, I'm seeing between 40mV and 80mV common mode DC, and about 2-3mV differential mode DC. This would never be enough to cause problems with a downstream amplifier, even if it is DC coupled.
I tossed in a brand new pair of caps and measured an initial value of 2.3VDC common mode at the output
Luckily diff mode was still below 5mV so most proper differential inputs should cope with this just fine.In the end though, I think it's probably best to use standard (non-polymer) SMD electrolytics and avoid this whole problem. I will update the BOM accordingly. Alternatively, you could use large film caps if you would like. The value of the output coupling caps depends entirely on the input impedance of the stage following the NTD1. I personally would aim for at least 10uF and preferably over 20uF, but that's because I tend to prefer lower input impedance for my amplifiers. If your amplifier has an input impedance of 100k, then 3uF is more than enough.
There should already be a build thread for the older version... I would use that thread instead of creating a new one.
Very curious to hear the results! Keep us posted.
All the above is exactly correct, and I do cover what value should be used at some point in the original NTD1 thread. I ran several simulations to get the correct values, and then I measured them in-circuit. It does exactly what you would expect, in that it attenuates above ~20kHz but there is also a tiny rolloff and associated phase shift in the pass band.
The original Pass Labs D1 circuit places small value caps across each of the power resistors (eg, input to V- and output to V+) to accomplish the same thing. This can easily be done by soldering 1206 C0G caps across the SMD resistor terminals. The details are shown in the original NTD1 V1 schematic that Greg refers to.
Finally, which amplifier are you using this with? Noise at the output is very odd, and indicates something might be adrift with the amp itself. Is it prone to oscillating? I've tried the NTD1 with dozens of amplifiers at this point, and have never heard anything at all at the output. I'm also curious why it would have struggled with 1V of common mode DC voltage at the input. An amplifier with a proper differential input stage should be fine with a few volts of common mode DC. It's diff mode DC that causes problems, and that should have only been a few mV worst case.
Regards,
Owen
This is the correct part. There's a trick implemented here to get higher output voltage while maintaining low noise. It's outlined in the datasheet if you dig around for it. The supply as shown in the schematic only works with the 5V output part. If you use a lower voltage part, you need to re-calculate the values to get the correct output voltage.
You calculate total VA (Volt Amps) by doing exactly what the name suggests... multiply V and A!
In your case you have 32V and 600mA so you will get 19.2VA per winding for a total of 76.8VA for the whole transformer. This is pretty much perfect for an NTD1.
It's definitely possible, but you will need to wire them off-board, and it's probably overkill to use 50W parts when we're only dissipating about 4W. There's also the potential for increased noise with this setup as the leads will be quite long and the input nodes are sensitive.
In the end, it's probably better to stick with the SMD D2PAK parts.
Regards,
Owen
Would this work? Ive seen something similar in the front end of an amp with good results.
RHS0J102MCN1GS Nichicon | Mouser
RHS0J102MCN1GS Nichicon | Mouser
Nope... voltage is too low, and you might still have the same issue with leakage since it's a polymer type.
You need at least a 25VDC rating, and it would be preferable to use a non polymer type if you are concerned about leakage (Common mode DC at the outputs).
As an example, the polymer type suggested has a leakage value of about 1890uA where a standard electrolytic cap has a leakage of about 77uA for a similar value.
Regards,
Owen
You need at least a 25VDC rating, and it would be preferable to use a non polymer type if you are concerned about leakage (Common mode DC at the outputs).
As an example, the polymer type suggested has a leakage value of about 1890uA where a standard electrolytic cap has a leakage of about 77uA for a similar value.
Regards,
Owen
Just because I'm so damn inpatient I put some Silmic II 100uF 35V and hooked it up to a couple 26.5vac. With 32vdc on the rails I can only trim to 1.48v. So thats the bad news, as for the good news the thing seems to make music. Oh the glorious music, with only 4 hours of burning around 55*C, it makes wood instruments sound correct and even hits the metal notes too. I can't believe the improvement over the 4 lme49990 I was using prior. Now don't get me wrong the Opamps didn't make anything bad but the NTD1 is no doubt an improvement.
Thank you OPC this is a great addition to my collection.
Now its time to place an order with Antek....
Thank you OPC this is a great addition to my collection.
Now its time to place an order with Antek....
Attachments
Hi Dan,
Congrats on the build!
Double check the polarity of the voltage at the source of the FET (The input to the NTD1) as it needs to be +1.65 relative to GND. It's easy to get the polarity wrong and adjust this to be -1.65 which then triggers the body diodes inside the DAC and limits it to ~-1.4V and doesn't allow for further adjustment.
I have done this myself before and was baffled as to why it wouldn't adjust any further. A quick look at the leads on my DMM reveled that they were reversed and I was holding the red lead on GND and the black lead on the source.
This won't damage anything, but it does result in higher distortion and worse performance.
Regards,
Owen
Congrats on the build!
Double check the polarity of the voltage at the source of the FET (The input to the NTD1) as it needs to be +1.65 relative to GND. It's easy to get the polarity wrong and adjust this to be -1.65 which then triggers the body diodes inside the DAC and limits it to ~-1.4V and doesn't allow for further adjustment.
I have done this myself before and was baffled as to why it wouldn't adjust any further. A quick look at the leads on my DMM reveled that they were reversed and I was holding the red lead on GND and the black lead on the source.
This won't damage anything, but it does result in higher distortion and worse performance.
Regards,
Owen