I'm having some low-level distortion problems with my CMoy, but they only occur with two of the handful of op-amps I've tried: OPA2228 and LM4562. Both cause a slight pop when the amp is turned off, which leads me to believe that this may be a DC issue. None of the op-amps seem to be drawing an exceptional amount of current (checked with a DMM), but I do hear what sounds like ground hum when the amp is turned way up with no input connected.
It's strange... The distortion with LM4562 and OPA2228 made the music completely unlistenable with a pair of crappy Koss TD-60 headphones (32 ohm impedance, 95 dB sensitivity), but on my Sennheiser HD-497s (32 ohm, 112 dB), it's only noticeable in the background. Still, I'm worried about the potential of DC offset to damage the headphones, if that is indeed the problem. I really want to use these op-amps, because they sound very good apart from the distortion. 😎
Background info:
The CMoy was built using a PCB from Head-Fi member Joshatdot that basically mimics Tangent's popular schematic, but with a single power capacitor and TLE railsplitter instead of one capacitor per channel as in Tangent's design.
I have chosen to use resistors in place of input capacitors for the sake of sound quality (all devices feeding the amplifier should have output capacitors, in theory). I've also chosen to use amb's Mini3 resistor values to achieve a gain of roughly 5.5 instead of 11.
The only hole not soldered on the PCB (shown below) is between V+ and V-; it's a VIA that links the top and bottom of the board. I was told that this wasn't essential to bridge.
---------------------------------------------
Here is my parts list:
Resistors
2 - R2, RES 100K OHM 1% 50PPM 1/2W, Vishay Dale [CMF100KHFCT-ND]
2 - R3, RES 332 OHM 1% 50PPM 1/2W, Vishay Dale [CMF332HFCT-ND]
2 - R4, RES 1.5K OHM 1% 50PPM 1/2W, Vishay Dale [CMF1.50KHFCT-ND]
Power Section
1 - C1, CAP 470UF 35V ELECT FM RADIAL, Panasonic FM PSU capacitor [P12415-ND]
2 - C2, RES 332 OHM 1% 50PPM 1/2W, Vishay Dale [CMF332HFCT-ND]
1 - TLE2426CLP virtual ground rail splitter chip (TO-92) [296-1994-ND]
Op-amps
1 - IC DUAL PRECISION OPAMP 8-DIP, OPA2228P Op-amp [OPA2228P-ND]
1 - LM4562
1 - OPA2107
1 - OPA2132
1 - AD823
1 - IC SOCKET 8 PIN .300 GOLD, DIP-8 Socket [ED90032-ND]
Extra
1 - Alps RK0971221Z05 pot and switch, 10KΩ, from amb
2 - STRAP BATT ECON 9V I STYLE 4"LD, 9V Battery Straps [232K-ND]
2 - CONN JACK STEREO R/A 3PIN 3.5MM, In-out jacks [CP1-3533-ND]
1 - LED 3MM BLUE DIFF, Blue LED [350-1560-ND]
1 - RES 10.0K OHM 1% 50PPM 1/2W, Vishay Dale resistor for LED [CMF10.0KHFCT-ND]
Here's a picture of the completed PCB:
Since Head-Fi is down at the moment, I can't pull up Joshatdot's picture of the bare board itself to show the labels underneath the parts. I'll post it as soon as possible. The TLE2496 is underneath that huge 470 uF capacitor.
It's strange... The distortion with LM4562 and OPA2228 made the music completely unlistenable with a pair of crappy Koss TD-60 headphones (32 ohm impedance, 95 dB sensitivity), but on my Sennheiser HD-497s (32 ohm, 112 dB), it's only noticeable in the background. Still, I'm worried about the potential of DC offset to damage the headphones, if that is indeed the problem. I really want to use these op-amps, because they sound very good apart from the distortion. 😎
Background info:
The CMoy was built using a PCB from Head-Fi member Joshatdot that basically mimics Tangent's popular schematic, but with a single power capacitor and TLE railsplitter instead of one capacitor per channel as in Tangent's design.
I have chosen to use resistors in place of input capacitors for the sake of sound quality (all devices feeding the amplifier should have output capacitors, in theory). I've also chosen to use amb's Mini3 resistor values to achieve a gain of roughly 5.5 instead of 11.
The only hole not soldered on the PCB (shown below) is between V+ and V-; it's a VIA that links the top and bottom of the board. I was told that this wasn't essential to bridge.
---------------------------------------------
Here is my parts list:
Resistors
2 - R2, RES 100K OHM 1% 50PPM 1/2W, Vishay Dale [CMF100KHFCT-ND]
2 - R3, RES 332 OHM 1% 50PPM 1/2W, Vishay Dale [CMF332HFCT-ND]
2 - R4, RES 1.5K OHM 1% 50PPM 1/2W, Vishay Dale [CMF1.50KHFCT-ND]
Power Section
1 - C1, CAP 470UF 35V ELECT FM RADIAL, Panasonic FM PSU capacitor [P12415-ND]
2 - C2, RES 332 OHM 1% 50PPM 1/2W, Vishay Dale [CMF332HFCT-ND]
1 - TLE2426CLP virtual ground rail splitter chip (TO-92) [296-1994-ND]
Op-amps
1 - IC DUAL PRECISION OPAMP 8-DIP, OPA2228P Op-amp [OPA2228P-ND]
1 - LM4562
1 - OPA2107
1 - OPA2132
1 - AD823
1 - IC SOCKET 8 PIN .300 GOLD, DIP-8 Socket [ED90032-ND]
Extra
1 - Alps RK0971221Z05 pot and switch, 10KΩ, from amb
2 - STRAP BATT ECON 9V I STYLE 4"LD, 9V Battery Straps [232K-ND]
2 - CONN JACK STEREO R/A 3PIN 3.5MM, In-out jacks [CP1-3533-ND]
1 - LED 3MM BLUE DIFF, Blue LED [350-1560-ND]
1 - RES 10.0K OHM 1% 50PPM 1/2W, Vishay Dale resistor for LED [CMF10.0KHFCT-ND]
Here's a picture of the completed PCB:
An externally hosted image should be here but it was not working when we last tested it.
Since Head-Fi is down at the moment, I can't pull up Joshatdot's picture of the bare board itself to show the labels underneath the parts. I'll post it as soon as possible. The TLE2496 is underneath that huge 470 uF capacitor.
Do you have any compensation on the feedback resistor with the opa2228?, those chips are unstable as hell without. Also if memory serves me correctly the resistors on input is specificaly advised against in datasheet.
google for the datasheet, and you will see why its not good for just dropping into a general cmoy topology/pcb.
good chips for that pcb are NE5532, or opa2132 and 2134.
Implemented correctly it is my favourite chip so far... I have no experience with the National chip, it still has to make its way across the globe.
google for the datasheet, and you will see why its not good for just dropping into a general cmoy topology/pcb.
good chips for that pcb are NE5532, or opa2132 and 2134.
Implemented correctly it is my favourite chip so far... I have no experience with the National chip, it still has to make its way across the globe.
Thanks for the reply. I did notice that the input bias current for the OPA2228 was in the 2.5 nA range, as opposed to the OPA2132's 5 pA, so I thought it might have something to do with balancing the input bias currents, which I've never done before. 😀
I don't have anything in the R5 position--it's jumpered at the moment. I'll consider adding input capacitors instead of resistors if they're necessary. What sort of modifications would I need to make this work with the OPA2228, and would those modifications affect the stability or sound quality of other less-cranky chips?
I don't have anything in the R5 position--it's jumpered at the moment. I'll consider adding input capacitors instead of resistors if they're necessary. What sort of modifications would I need to make this work with the OPA2228, and would those modifications affect the stability or sound quality of other less-cranky chips?
Look at the datasheet, it has some suggestions... a small cap in parallel to fb resistor should keep things tight... also look at schematic to see which changes to hardwire into your board.
Make sure the chip is shielded from drafts and moveing air.
Make sure the chip is shielded from drafts and moveing air.
infinitesymph said:
Try connect pot shaft to ground.
Try connect small film caps between opamp's +power leg and ground, and between -power leg and ground right under chip socket on boards other side.
Thanks again Nordic, and welcome Zigis. I'm still new to electronics, so even simple schematics seem a little confusing to me--sorry if I need some spelling-out from time to time. I did read the datasheet and saw that it recommended 0.1 uF decoupling capacitors, just as you both said.
Sounds good to me... Can either of you recommend some good capacitors for this purpose, and are films better than ceramics for this application? I'll need to start a new order with Digi-Key, since I don't have any spares. I may add the NE5532 op-amp to my order, too, since I hadn't heard of that one before.
Will adding these 0.1 uF decoupling capacitors negatively affect the performance with any of the other op-amps?
Edit: I just noticed that the OPA2228 datasheet does say, "A resistor added to cancel the effect of the input bias current...may actually increase offset and noise and is therefore not recommended." Does this mean that I should jumper the input capacitor pads instead of using the resistors (and is this okay for other op-amps)?
Sounds good to me... Can either of you recommend some good capacitors for this purpose, and are films better than ceramics for this application? I'll need to start a new order with Digi-Key, since I don't have any spares. I may add the NE5532 op-amp to my order, too, since I hadn't heard of that one before.
Will adding these 0.1 uF decoupling capacitors negatively affect the performance with any of the other op-amps?
Edit: I just noticed that the OPA2228 datasheet does say, "A resistor added to cancel the effect of the input bias current...may actually increase offset and noise and is therefore not recommended." Does this mean that I should jumper the input capacitor pads instead of using the resistors (and is this okay for other op-amps)?
For audio filmcaps is better. In this place poliester is good enough.
If you don't use input cap, just replace it with jumper.
Solder resistor leg or wire to pot's ground pin and connect with nut to pot's shaft. This big aluminium shaft sometime work like antena.
If you don't use input cap, just replace it with jumper.
Solder resistor leg or wire to pot's ground pin and connect with nut to pot's shaft. This big aluminium shaft sometime work like antena.
Thanks again for the advice. I'll jumper the input capacitor pads and run a wire from the Alps pot shaft to ground.
Could you recommend any small film capacitors available on Digi-Key? The typical Panasonic and Vishday/Dale film capacitors seem a little large to fit on the underside of the board.
Could you recommend any small film capacitors available on Digi-Key? The typical Panasonic and Vishday/Dale film capacitors seem a little large to fit on the underside of the board.
I am not familiar with Digi-Key, I bay in Europe, maybe 3013PH ( from Tangent tutorial ), anyway, any metalized poliester box cap for 63V or 100V. Smaller voltage- smaller size. I use Wima.
In my CMoy boards I place El caps close to opamp power legs, opamp like it. I learn this from CarlosFM, ex- member.
If space allow, better way is bypassing with two 10mf X 16V under PCB.
Are shaft grounding help with noise problem ?
In my CMoy boards I place El caps close to opamp power legs, opamp like it. I learn this from CarlosFM, ex- member.
If space allow, better way is bypassing with two 10mf X 16V under PCB.
Are shaft grounding help with noise problem ?
For the decoupling caps that connect the opamp V+ and V- to ground, ceramic capacitors are best. They have lower impedance at high frequencies. Using decoupling caps are recommended, generally essential, for all opamps. Opamps may oscillate without these caps, which you may not be able to hear directly but could distort the audio.
Any of digikey's small ceramics are fine. You can also add larger caps (film or electrolytic) in parallel with them if you wish, though I don't think you'll hear any difference.
Film caps are better for input coupling capacitors.
Any of digikey's small ceramics are fine. You can also add larger caps (film or electrolytic) in parallel with them if you wish, though I don't think you'll hear any difference.
Film caps are better for input coupling capacitors.
After searching Digi-Key, I think I'll use these for decoupling capacitors:
Vishay/BC CAP .10UF 100V CERAMIC X7R 10% [part number: BC1123CT-ND]
Good choice? Then again, maybe axial would be more ergonomic (ex. part 1109PHCT-ND)... I'm assuming that a chip MLCC ceramic capacitor couldn't be used in this situation, since they don't appear to have any leads.
Zigis, I haven't had the opportunity to make any modifications yet, but I'll report back about the ground wire's effectiveness after it's installed.
Newbie question... When you guys say "ground," which pin/location is that on the PCB? On my Tangent DIY CMoy, the ground location was obvious, but I'm not as familiar with this board.
Vishay/BC CAP .10UF 100V CERAMIC X7R 10% [part number: BC1123CT-ND]
Good choice? Then again, maybe axial would be more ergonomic (ex. part 1109PHCT-ND)... I'm assuming that a chip MLCC ceramic capacitor couldn't be used in this situation, since they don't appear to have any leads.
Zigis, I haven't had the opportunity to make any modifications yet, but I'll report back about the ground wire's effectiveness after it's installed.
Newbie question... When you guys say "ground," which pin/location is that on the PCB? On my Tangent DIY CMoy, the ground location was obvious, but I'm not as familiar with this board.
With a battery as power source, I use 10nf Wima caps (those little red ones) my headphone amps are dead quiet...just solder them very close, or on power pins...
10 nF, eh? Will using a higher-valued capacitor be more beneficial?
I'm still not sure I'll be able to fit a full-size film cap on the bottom of the board, since I'm planning to mount it in a mint tin.
Besides the capacitance, are there other values I should be watching, ex. voltage, tolerance, temperature coefficient, etc.?
Sorry to agonize over such a small part of the build... I'm planning on stocking up with whatever I choose, so I want to get it right the first time. 🙂
I'm still not sure I'll be able to fit a full-size film cap on the bottom of the board, since I'm planning to mount it in a mint tin.
Besides the capacitance, are there other values I should be watching, ex. voltage, tolerance, temperature coefficient, etc.?
Sorry to agonize over such a small part of the build... I'm planning on stocking up with whatever I choose, so I want to get it right the first time. 🙂
too tired to give usefull answers, but while you are buying stuff, get a small breadboard for testing circuits without all the soldering...
Are you useing an Output cap?? Because if you are running off of a Single rail supply with a Virtual ground then you will have a large DC offset at the output......
No, there are no capacitors in the signal path at the moment. I was under the impression that the DC offset would be small enough not to matter, assuming everything was balanced (ex. input bias, etc.).
Here are the board layout and schematic, straight from Joshatdot:
CMoy v1.03 board layout
CMoy v1.03 schematic
Here are the board layout and schematic, straight from Joshatdot:
CMoy v1.03 board layout
CMoy v1.03 schematic
If you are useing a Virtual Ground (useing half Supply as Ground) then the input/output is referanced to Half supply which means that there could be half of the Supply of DC offset at the output so if you are useing a 12v single rail supply you can have up to 6v of DC offset at the output....
Most all Single supply circuits use DC blocking caps for this very reason.....
😀
Most all Single supply circuits use DC blocking caps for this very reason.....
😀
By DC blocking caps, do you mean input capacitors?
I'm using two 9V batteries (~18V), and the highest DC offset I measured was in the 1 mV range with the volume all the way up. I tested using mini-mini cables plugged into both the input and output. With TRS shorted together on the other end of the input cable, I turned on the amp and measured from tip to ground and ring to ground on the output cable.
It was my understanding that the purpose of the TLE2496 railsplitter was to evenly distribute the power to both channels... I was told that the biggest source of DC offset would be from the playback device attached to the input, assuming it doesn't have output capacitors. So, I had assumed that DC offset wouldn't be enough of an issue to matter.
With 18V, you say that the circuit might feed up to 9V into the output. Wouldn't this have fried my headphones almost instantly? 😛
I'm using two 9V batteries (~18V), and the highest DC offset I measured was in the 1 mV range with the volume all the way up. I tested using mini-mini cables plugged into both the input and output. With TRS shorted together on the other end of the input cable, I turned on the amp and measured from tip to ground and ring to ground on the output cable.
It was my understanding that the purpose of the TLE2496 railsplitter was to evenly distribute the power to both channels... I was told that the biggest source of DC offset would be from the playback device attached to the input, assuming it doesn't have output capacitors. So, I had assumed that DC offset wouldn't be enough of an issue to matter.
With 18V, you say that the circuit might feed up to 9V into the output. Wouldn't this have fried my headphones almost instantly? 😛
A virtual ground in effect is at half supply with relation to the power source... i.e., your "ground" in your case is at 9V relative to the batteries +/-. This is why when using a virtual ground, you must use an insulated power jack if using a wall supply and the supply must be floating. Since you are using batteries, it is a non-issue. From your headphone's perspective, the virtual ground is 0V.
- Status
- Not open for further replies.