For the record -- here's the LM317 PSRR measured with and without 10uF on the adjust pin and the data sheet -- note that the y axes are in different directions:
Attachments
Eagle Computers - they made pretty decent CP/M86 minicomputers. The CEO's name was Barnhart, IIRC. He was worth $9M on the day of his death, a sizable sum in that era. The Ferrari landed upside-down in the Lexington reservoir or a creek feeding the reservoir somewhere off Route 9, IIRC.
The co-passenger, who may have actually been the driver at the time of the accident, was the owner of a Yacht outlet:
CORPORATE TRIUMPH, THEN DEATH IN A FERRARI - NYTimes.com
1. I would imagine so...a bit more gain
2. I added a 1K, no difference...
3. Copied that right off Mooly's schematic, I think it keeps stray capacitance away from the 713
4. It was terrible, almost rail voltage on both channels -7 VDC, passed right on through the 713, so this 604 buffer came out and was eliminated. After this, the circuit behaves great! Output offset 1-3 mV and that is it. No noise of any kind on the volume control, no hiss, hum, or crackling. The sound is clear and the op amps are fast, very fast. The sound is totally high-end hi-fi, even with these few components, no output caps or anything. You have to try it!
So attached is a working basic circuit for the LME49713, perhaps some of you with experience can dress it up and improve it. It is basically a copy of Mooly's work referenced to another thread earlier, with Ri and Rf changed for about 4X gain. The sound being produced is definitely worth the effort and the cost of the LME49713HA (metal can) op amps at Mouser, at this writing is $12.37 each.
Thanks for all the help.
Attachments
Last edited:
The gain/noise problem comes from the volume control configuration. perversly that configuration increases the gain of the opamp as the control goes to ground. The input signal is reduced by the pot but the gain increases and the noise increases as the gain is increased. The pot output drives the 330 Ohm resistor which is connected to a virtual ground.
In short, it won't work. You could use a 30 Ohm pot, not practical or follow Marks advice and put the pot on the output. Again use a low value pot. The input impedance of the gain stage is 330 Ohms, very low, so not much beyond a headphone output can drive it.
I don't think the 82 Ohm resistor on the - input does anything useful either. That node is low impedance and the resistor will degrade the operation in several ways.
What is the source driving this?
In short, it won't work. You could use a 30 Ohm pot, not practical or follow Marks advice and put the pot on the output. Again use a low value pot. The input impedance of the gain stage is 330 Ohms, very low, so not much beyond a headphone output can drive it.
I don't think the 82 Ohm resistor on the - input does anything useful either. That node is low impedance and the resistor will degrade the operation in several ways.
What is the source driving this?
Whaleman,
did you try adding 100nF ceramics between pins 4 and 7 of each opamp? I don't see any caps on your photos and all the wires are looooong....
I ran into something similar just yesterday. I built a little circuit utilizing a TL074 on a small piece of veroboard, powered by a lab supply (leads about 60cm). I measured THD at 1kHz and about 6Vrms output into 1kOhm and it was several percents, which is way too much even for an 'old' opamp. Fitted the cap and everything was fine!
did you try adding 100nF ceramics between pins 4 and 7 of each opamp? I don't see any caps on your photos and all the wires are looooong....
I ran into something similar just yesterday. I built a little circuit utilizing a TL074 on a small piece of veroboard, powered by a lab supply (leads about 60cm). I measured THD at 1kHz and about 6Vrms output into 1kOhm and it was several percents, which is way too much even for an 'old' opamp. Fitted the cap and everything was fine!
Sort the 604 stage, not by removing it, but by getting the inputs such that the output stays at near zero output offset.
Really nice! Would love to have the skill to end up something like this.
Yes, that was the behavior experienced earlier.
The source is a Parasound ZDAC, thick black interconnects coming from it in the photo.
I see clearly, a buffer is a theme.
Simple enough, I'll give it a try, one cap across pins 4-7 on each 713 op amp.
OK, I tried to get away with it and got caught
I do hear a bit of slight trim of the very high frequencies probably caused by the unfriendly input impedance, so the input buffer seems necessary.
So, the offset output of this buffer is -7 VDC (rails about 9 VDC), obviously something is wrong in the schema attached. Does ground go anywhere on the buffer?
Attachments
What's the DC at the INPUT? Try a coupling cap to be sure. Did you put both opamps in parallel, like the pin numbers on the schematic imply? If so, don't do it!
I did not measure at the input, but there was nothing connected to the circuit at the input. You mean a coupling cap between the buffer and the 713s?
The pin numbers are just for reference, the 713s are wired individually.
You seem to have misunderstood Mooly's schematic. The 600 Ohm resistor on his diagram is not part of the amplifier, it is the load driven by the amplifier. The 39R resistor is part of the amplifier.
IOW, there should only be one resistor at the output.
The input impedance at the 713 is exceedingly low.
If you used a coupling capacitor here to block DC it would have to be enormous.
Forget that as a solution.
Instead block DC at the input and arrange your component values to end up with near zero output offset after the 604.
A further alternative:
Provide both AC and DC inputs at the input.
The DC input socket would have no coupling capacitor.
The other input socket would have the capacitor.
As far as I have seen, every opamp datasheet specifies decoupling capacitors.
Each manufacturer specifies a mix of MF and HF decoupling, some on both power pins to power ground, some between the power pins, some from one power pin to power ground.
Initially you MUST follow the manufacturers advice. With experience (sometimes from others) you may decide to vary the decoupling to slightly different from the manufacturer specified values and types. Many manufacturers specify Tantalum caps for the MF decoupling. This Forum tend to shy away from Tantalum.
I used to use Tantalum, then I stopped, recently I have started again but at <50% of voltage rating and where impulsive interference cannot damage the cap.
If you used a coupling capacitor here to block DC it would have to be enormous.
Forget that as a solution.
Instead block DC at the input and arrange your component values to end up with near zero output offset after the 604.
A further alternative:
Provide both AC and DC inputs at the input.
The DC input socket would have no coupling capacitor.
The other input socket would have the capacitor.
As far as I have seen, every opamp datasheet specifies decoupling capacitors.
Each manufacturer specifies a mix of MF and HF decoupling, some on both power pins to power ground, some between the power pins, some from one power pin to power ground.
Initially you MUST follow the manufacturers advice. With experience (sometimes from others) you may decide to vary the decoupling to slightly different from the manufacturer specified values and types. Many manufacturers specify Tantalum caps for the MF decoupling. This Forum tend to shy away from Tantalum.
I used to use Tantalum, then I stopped, recently I have started again but at <50% of voltage rating and where impulsive interference cannot damage the cap.
Last edited:
I think this is the last step that remains, there is some DC bleed through on the sound.
I was able to get the 604 offset to virtually zero, I had forgotten to tie pins 1 and 2 on it.
Offset then at the 713 output was about 800 mV with the volume all the way down and less as you turned the volume up. I added a 32.5K resistor to get this steady then a trim circuit on the 713 to get the output offset down to zero. Now it is starting to get much better.
Very faint DC hum in sound and the 1K scope follows, input wave at bottom, output wave above it. Then wave at buffer output, top second scope photo. This remains to be fixed. Not sure how to interpret the "sea" wave.
Sound is quite worth the effort and I need to perfect it. I know, wires going in every direction
See pics.Attachments
Last edited:
That sawtooth is the output from the buffer, while the sinewave below is its input? What am I missing?
You don't have to use such huge caps, btw. Use X7R for best results. Small foil caps like WIMA MKS2 will do just fine, too. Best is to make the leads as short as possible, ie. place the cap on the bottom of the board and solder it right between the pins of the socket. Please try this with the buffer, too. A buffer should not produce a sawtooth, output should look like input...
EDIT: Guess I got the scope shots confused. But still, that "sea" wave looks strange and seems to be present at the input already.
You don't have to use such huge caps, btw. Use X7R for best results. Small foil caps like WIMA MKS2 will do just fine, too. Best is to make the leads as short as possible, ie. place the cap on the bottom of the board and solder it right between the pins of the socket. Please try this with the buffer, too. A buffer should not produce a sawtooth, output should look like input...
EDIT: Guess I got the scope shots confused. But still, that "sea" wave looks strange and seems to be present at the input already.
Last edited:
I think this is the last step that remains, there is some DC bleed through on the sound.
Sound is quite worth the effort and I need to perfect it. I know, wires going in every direction
We are losing sight of the goal here. The 604 (or 134 or 627 etc.) are more than adequate buffers in their own right. The CF stage may be able to drive more current but all you will get is the output of the existing 604. If you don't need the current its just more monkey motion. It won't bring any magic and probably will oscillate like crazy. You will need a good ground plane implementation to keep it stable. Possibly surface mount .01 uF MLCC at each power supply pin.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.