Robert,
First let me say that I am sorry that you have access to such excellent equipment and I don't (envy envy envy 🙂)
Being in part responsible for the SHM, I have a few suggestions which may assist you in finding the problem. It seems strange that you should be getting 1K and 10K interference. That seems like too much of a coincidence -- why exactly these values???? Since you are in the US, you should not be getting these fromt he wall unless they are artifacts of a switched mode wall-wart. Perhaps they are byproducts of something else. On the SHM, there is a microcontroller witch a resonator. I cannot remember which frequency that switches at, but it might be the culprit (even though I doubdt it. You chould attempt to shield this area, but first consider trying this:
1. Battery power to SHM (benchmarkish performance and easy to do, perhaps easier than moving digital power supply. You appear to put most of the "crap" onto the signal with all Apox boards off ...
2. Your output impedance of the preamp is 10 Ohm. No matter which attenuator you use, you will get much higher impedance, and the noise will not be "masked" so well. Try soldering in a few resistors for an absolute attenuator and see what that does. That will be your benchmark of obtainable result.
2.5 Consider soldering in resistors on the SHM to set up a fixed volume - similar to the above and see what happens with/without SHM power.
3. Grounding: How fat is the ground wire? What ground topology are you using? Are you using star grounding given you have various Apox units, and how does this hook up to the rest of the analog circuitry (if at all?).
However, the most likely culprit is:
"Interestingly, the 5V digital supply even without a load does contaminate the signal as seen on the THD plot. Thus, despite my trying to minimize garbage getting into the gain stages power supply, it still gets in. I may either need to re-locate it from the chassis, or go back and change it to a linear style supply".
There has been significant discussion of grounding techniques - but I think you probably have a pretty poor digital supply. Perhaps it is switched mode radiating all kinds of crap. You could try moving it far away and have som ferrite on the way in, or better still do half as good a job as what you already did on the analog supply.
It would be nice if you were able to post pictures of your unit - it might be easier to provide more suggestions with such material.
Petter
First let me say that I am sorry that you have access to such excellent equipment and I don't (envy envy envy 🙂)
Being in part responsible for the SHM, I have a few suggestions which may assist you in finding the problem. It seems strange that you should be getting 1K and 10K interference. That seems like too much of a coincidence -- why exactly these values???? Since you are in the US, you should not be getting these fromt he wall unless they are artifacts of a switched mode wall-wart. Perhaps they are byproducts of something else. On the SHM, there is a microcontroller witch a resonator. I cannot remember which frequency that switches at, but it might be the culprit (even though I doubdt it. You chould attempt to shield this area, but first consider trying this:
1. Battery power to SHM (benchmarkish performance and easy to do, perhaps easier than moving digital power supply. You appear to put most of the "crap" onto the signal with all Apox boards off ...
2. Your output impedance of the preamp is 10 Ohm. No matter which attenuator you use, you will get much higher impedance, and the noise will not be "masked" so well. Try soldering in a few resistors for an absolute attenuator and see what that does. That will be your benchmark of obtainable result.
2.5 Consider soldering in resistors on the SHM to set up a fixed volume - similar to the above and see what happens with/without SHM power.
3. Grounding: How fat is the ground wire? What ground topology are you using? Are you using star grounding given you have various Apox units, and how does this hook up to the rest of the analog circuitry (if at all?).
However, the most likely culprit is:
"Interestingly, the 5V digital supply even without a load does contaminate the signal as seen on the THD plot. Thus, despite my trying to minimize garbage getting into the gain stages power supply, it still gets in. I may either need to re-locate it from the chassis, or go back and change it to a linear style supply".
There has been significant discussion of grounding techniques - but I think you probably have a pretty poor digital supply. Perhaps it is switched mode radiating all kinds of crap. You could try moving it far away and have som ferrite on the way in, or better still do half as good a job as what you already did on the analog supply.
It would be nice if you were able to post pictures of your unit - it might be easier to provide more suggestions with such material.
Petter
there should be a new thread started and it all moved out of the trading post😎
Cheers!!The DIRT®
Cheers!!The DIRT®
Petter,
Thanks for the reply.
First, the 1K / 10K is simply the signal frequency that I'm sending into the unit. This is not the value of any distortions, etc. The distortion artifacts are probably 100x greater. I displayed the 10K values as these were easier to visualize since the THD at 1Khz of the gain stage is very low. Only the upper left image is at 1kHz; all the rest are at 10kHz. That is, the bottom tracing is a 10kHz sine wave; the upper tracing the THD at that frequency.
Second, when the SHM was used, I set the volume to 255: full on. On all other tests, the signal in was without attentuation. Since the gain stage is the last stage (see my previous message), the output impedance of 10R is non-contributory to this discussion.
The fact that there are glitches on the THD tracing when the SHM is used implies that the SHM is contriuting something above and beyond the digital niose that is observed in the middle right image.
I admit that using a battery to provide power would be a good check: re-run the APOX portions with a 5V battery supply for the APOX boards.
I only have 12V lead acid batteries on hand and the current demand is too great for AA types, so coming up with 5V is going to take some effort unless I use a regulator from the 12V battery. Is this what you have in mind?
I truly don't think grounding is an issue since the signal is quite clean without a the digital power supply. Turning it on simply adds noise to the THD (no hum component is added). The grounding is Star: the digital ground and signal ground are only at one point. (I've checked; if they are separated at this spot, there is no commonality.)
Regards, Robert
Thanks for the reply.
First, the 1K / 10K is simply the signal frequency that I'm sending into the unit. This is not the value of any distortions, etc. The distortion artifacts are probably 100x greater. I displayed the 10K values as these were easier to visualize since the THD at 1Khz of the gain stage is very low. Only the upper left image is at 1kHz; all the rest are at 10kHz. That is, the bottom tracing is a 10kHz sine wave; the upper tracing the THD at that frequency.
Second, when the SHM was used, I set the volume to 255: full on. On all other tests, the signal in was without attentuation. Since the gain stage is the last stage (see my previous message), the output impedance of 10R is non-contributory to this discussion.
The fact that there are glitches on the THD tracing when the SHM is used implies that the SHM is contriuting something above and beyond the digital niose that is observed in the middle right image.
I admit that using a battery to provide power would be a good check: re-run the APOX portions with a 5V battery supply for the APOX boards.
I only have 12V lead acid batteries on hand and the current demand is too great for AA types, so coming up with 5V is going to take some effort unless I use a regulator from the 12V battery. Is this what you have in mind?
I truly don't think grounding is an issue since the signal is quite clean without a the digital power supply. Turning it on simply adds noise to the THD (no hum component is added). The grounding is Star: the digital ground and signal ground are only at one point. (I've checked; if they are separated at this spot, there is no commonality.)
Regards, Robert
I just finished discussing this matter with a friend of mine who used to be in the industry (Julius Siksnius of Audire).
He felt that the glitches in the middle right trace were indeed from the switching 5V PS, showing up at about 220 to 250kHz frequency (determined by counting the glitches per 10kHz sine wave). These he thougth could be removed either with brute force filtering (RC filtering: 0.1R + 10,000uF + ceramic bypass), by removing the digital supply from the same chassis, or by using a linear power supply.
The greatest difficulty he thought, might be getting rid of the other distortions.
He agreed that the much larger glitches in the bottom right image are due to digital contamination in the SHM. These are occuring at about 60kHz and are superimposed on the 250kHz PS glitches.
Julius felt that the 60kHz (or so) frequency was probably due to the microprocessor. However, since he is not familiar with the circuitry he could not be certain. He did suggest that since both the lower left and right images have the same pattern and both have the same microprocessors that this was likely. (The left lower image has the signal going through the IS1 board, but the SHM was not used.) The source could either be close proximity of the digial signal from the microprocessor to the analog signal, or due to inductive pulses from the relays (he imaged these pulses could be much larger and what we're seeing is only a small residual of what is being emitted). He was wondering if the microprocessor is sending 60kHz pulses to keep the relays engaged. Is this the way it works?
If the latter were the problem, would bypassing each relay PS pins with a 0.1 uF ceramic cap help?
If it is the former (that is, a proximity effect), then is some sort of shiedling possible as a retro-fit?
Regards, Robert
He felt that the glitches in the middle right trace were indeed from the switching 5V PS, showing up at about 220 to 250kHz frequency (determined by counting the glitches per 10kHz sine wave). These he thougth could be removed either with brute force filtering (RC filtering: 0.1R + 10,000uF + ceramic bypass), by removing the digital supply from the same chassis, or by using a linear power supply.
The greatest difficulty he thought, might be getting rid of the other distortions.
He agreed that the much larger glitches in the bottom right image are due to digital contamination in the SHM. These are occuring at about 60kHz and are superimposed on the 250kHz PS glitches.
Julius felt that the 60kHz (or so) frequency was probably due to the microprocessor. However, since he is not familiar with the circuitry he could not be certain. He did suggest that since both the lower left and right images have the same pattern and both have the same microprocessors that this was likely. (The left lower image has the signal going through the IS1 board, but the SHM was not used.) The source could either be close proximity of the digial signal from the microprocessor to the analog signal, or due to inductive pulses from the relays (he imaged these pulses could be much larger and what we're seeing is only a small residual of what is being emitted). He was wondering if the microprocessor is sending 60kHz pulses to keep the relays engaged. Is this the way it works?
If the latter were the problem, would bypassing each relay PS pins with a 0.1 uF ceramic cap help?
If it is the former (that is, a proximity effect), then is some sort of shiedling possible as a retro-fit?
Regards, Robert
Hi,
I did not keep up with all posts here (i think i posted somewhere in the beginning myself).
And don't know your setup, but if you have that display, try removing it for a moment. Hope the processor is not waiting for ack or other signals from it.
There's a processor on board which drives the digits, dont think those things are static = noise.
Don't know the microcontroller's programing, but i guess using sleepmode is best: only start it up when needed. Remote = interrupt = changing channel or whatever and then sleep again.
Can't the digital gnd be completely separated from analog gnd anyway ?
Sorry to just bump in, but maybe something above helps.
GuidoB
I did not keep up with all posts here (i think i posted somewhere in the beginning myself).
And don't know your setup, but if you have that display, try removing it for a moment. Hope the processor is not waiting for ack or other signals from it.
There's a processor on board which drives the digits, dont think those things are static = noise.
Don't know the microcontroller's programing, but i guess using sleepmode is best: only start it up when needed. Remote = interrupt = changing channel or whatever and then sleep again.
Can't the digital gnd be completely separated from analog gnd anyway ?
Sorry to just bump in, but maybe something above helps.
GuidoB
E.g.
The normal lcd's (4/8 bit standard interface) work with 250kHz clock to the lcd controller. From there you get refresh rates for the digits, line etc.
GuidoB
The normal lcd's (4/8 bit standard interface) work with 250kHz clock to the lcd controller. From there you get refresh rates for the digits, line etc.
GuidoB
guido,
Thanks. I don't believe the IR1 (display board with the main processor) is the culprit. This is because the IR1 was active for both bottom images, so the added distortion on the lower left is from the SHM itself.
In fact, the bottom left, where the IS1 is used, the SHM was powered up and receiving signals from the IR1; I had just routed the analog signal around it. Thus, the IR1, or communications from the iR1 to SHM, did not add the extra distortions to the lower right had image. At least, if it does, it is not sending into the ground lead. This must be a consequence of the analog signal going through the SHM.
Petter,
This extra distortion from the SHM (and maybe IS1) is also why using a battery won't help: the microprocessors might get a cleaner supply voltage, but they'll still be functioning in the digital domain, and the extra glitches at 60kHz will most likely still be present. The only distortions that a battery will remove is the 250kHz saw tooth in the middle right image.
Regards, Robert
BTW, if the moderator wants to move this discussion to the main forum is OK by me, but it should probably start with the THD image post.
Thanks. I don't believe the IR1 (display board with the main processor) is the culprit. This is because the IR1 was active for both bottom images, so the added distortion on the lower left is from the SHM itself.
In fact, the bottom left, where the IS1 is used, the SHM was powered up and receiving signals from the IR1; I had just routed the analog signal around it. Thus, the IR1, or communications from the iR1 to SHM, did not add the extra distortions to the lower right had image. At least, if it does, it is not sending into the ground lead. This must be a consequence of the analog signal going through the SHM.
Petter,
This extra distortion from the SHM (and maybe IS1) is also why using a battery won't help: the microprocessors might get a cleaner supply voltage, but they'll still be functioning in the digital domain, and the extra glitches at 60kHz will most likely still be present. The only distortions that a battery will remove is the 250kHz saw tooth in the middle right image.
Regards, Robert
BTW, if the moderator wants to move this discussion to the main forum is OK by me, but it should probably start with the THD image post.
Robert,
I know this is getting painful, here are some friendly suggestions which might be of use to you.
Can you post photographs of your setup?
Are you using two SHM boards or just one (stereo or mono)?
Are you using balanced or single ended?
Is there ANYTHING that is different when the SHM is not in the circuit (such as connecting digital ground to analog -- open when SHM out?
Do you have any ground loops (for example SHM connected to analog ground on board + also connected to analog ground. If you have more than one board, you will typically get some strange behaviour.
What is your grounding topology?
What have you done with the two ground planes for the SHM, particularly digital?
Have you tried varying the display settings (intensity) as these are PWM'ed.
Have you checked that you are using nice fat grounding wires?
When I suggested batteries, I ment batteries, but I know it is not always so practical. You could make 4.5V out of 3 series connected large round batteries
Removing sawtooth will be a good thing. It would make it easier to isolate other issues.
Petter
I know this is getting painful, here are some friendly suggestions which might be of use to you.
Can you post photographs of your setup?
Are you using two SHM boards or just one (stereo or mono)?
Are you using balanced or single ended?
Is there ANYTHING that is different when the SHM is not in the circuit (such as connecting digital ground to analog -- open when SHM out?
Do you have any ground loops (for example SHM connected to analog ground on board + also connected to analog ground. If you have more than one board, you will typically get some strange behaviour.
What is your grounding topology?
What have you done with the two ground planes for the SHM, particularly digital?
Have you tried varying the display settings (intensity) as these are PWM'ed.
Have you checked that you are using nice fat grounding wires?
When I suggested batteries, I ment batteries, but I know it is not always so practical. You could make 4.5V out of 3 series connected large round batteries
Removing sawtooth will be a good thing. It would make it easier to isolate other issues.
Petter
Petter,
I posted images earlier of my preamp (the 3 levels chassis, yes?), so I don't think more images too helpful.
I have 6 SHM for 6 active channels. To keep things simple, all are presently being run single ended. (I believe I described all of this at lenth in earlier posts.)
This SHMs should be perfect; they were all checked by Dale after I built them.
The grounds are kept different. I did try using a 10R at the "FB" point that connects digital and analog grounds, if I recall correctly, and no relief of distortions with this measure.
There are no ground loops. As I mentioned in previous post: I use star ground.
I have not varied the display; it is set at intensity level 3.
Digital voltage wires are 18G, and run back to output from digital supply. With all of my APOX boards, this requires about 1.8 amps, which why a small battery won't survive.
Regards, Robert
I posted images earlier of my preamp (the 3 levels chassis, yes?), so I don't think more images too helpful.
I have 6 SHM for 6 active channels. To keep things simple, all are presently being run single ended. (I believe I described all of this at lenth in earlier posts.)
This SHMs should be perfect; they were all checked by Dale after I built them.
The grounds are kept different. I did try using a 10R at the "FB" point that connects digital and analog grounds, if I recall correctly, and no relief of distortions with this measure.
There are no ground loops. As I mentioned in previous post: I use star ground.
I have not varied the display; it is set at intensity level 3.
Digital voltage wires are 18G, and run back to output from digital supply. With all of my APOX boards, this requires about 1.8 amps, which why a small battery won't survive.
Regards, Robert
Petter,
Julius was over for dinner tonight. We discussed the matter some more. He says that it is impossible that ground looping would look like image 5. It must be digital contamination. Period.
As an aside, I am not inexperienced building analog circuitry. After several rounds of problems already, first with malfunctioning and misprogrammed chips and now with whatever is happening, I am frankly quite tired of being on the defensive explaining why my build job is not a problem.
Furthermore, while you may be envious of my equipment (I bought it used on Ebay for $300), I am disturbed to hear that circuitry is designed and sold without proper test equipment.
Robert
Julius was over for dinner tonight. We discussed the matter some more. He says that it is impossible that ground looping would look like image 5. It must be digital contamination. Period.
As an aside, I am not inexperienced building analog circuitry. After several rounds of problems already, first with malfunctioning and misprogrammed chips and now with whatever is happening, I am frankly quite tired of being on the defensive explaining why my build job is not a problem.
Furthermore, while you may be envious of my equipment (I bought it used on Ebay for $300), I am disturbed to hear that circuitry is designed and sold without proper test equipment.
Robert
Robert,
My questions and notes were intended as accusations of your building prowess which I thought would be clear based on the following:
"I know this is getting painful, here are some friendly suggestions which might be of use to you." from my last post.
and
"Mr rljones, we are in awe. You and I are doing similar things - except you have already done it!"
from 6-22-2003
I think you should get rid of the sawtooths first, then go after the other problems - that should be the easiest route. However, in light of recent postings it seems I won't be able to contribute to your efforts, but I am glad that you have an expert locally who is assisting.
Petter
I have access to test equipment - but don't own such equipment. I think you will find that the parts have been designed AND tested and that the after sales support (not me) has been very good.
I don't sell anything but admit to have done the analog layout of the SHM which I am quite proud of 🙂
And I am good friends with the Apox guys even though we have never met given the physical distance.
My questions and notes were intended as accusations of your building prowess which I thought would be clear based on the following:
"I know this is getting painful, here are some friendly suggestions which might be of use to you." from my last post.
and
"Mr rljones, we are in awe. You and I are doing similar things - except you have already done it!"
from 6-22-2003
I think you should get rid of the sawtooths first, then go after the other problems - that should be the easiest route. However, in light of recent postings it seems I won't be able to contribute to your efforts, but I am glad that you have an expert locally who is assisting.
Petter
I have access to test equipment - but don't own such equipment. I think you will find that the parts have been designed AND tested and that the after sales support (not me) has been very good.
I don't sell anything but admit to have done the analog layout of the SHM which I am quite proud of 🙂
And I am good friends with the Apox guys even though we have never met given the physical distance.
ack!
Hi Robert,
I just started to read all of the recent posts.
Maybe I missed it when I read the posts, but
in your oscillocope images. Where exactly are you probing?
Is that a picture of the power supply, and
the output signals???
The only pulsing we do is with the LEDS at about 2 Hz.
There is no relay driver pulsing. We run the internal
oscillator at 4MHz.
Just as an experiment. Ground your oscilloscope probe to the
circuit, and then without touching any traces,
move your oscilloscope probe tip in the air around
the microcontroller and around the relay driver chips.
I've found this to be a quick check for emitted radiation.
And you might be able to hunt down where the most radiation
is coming from. The high input imedance of the scope allows you to do this. And now that I think about it, the 1Meg input impedance of your pre-amp might be helping it pick up digital noise as well? (maybe?)
What kind of THD are you actually getting?
What is the amplitude of the signals in your photos.
Thanks,
Craig Beiferman
Hi Robert,
I just started to read all of the recent posts.
Maybe I missed it when I read the posts, but
in your oscillocope images. Where exactly are you probing?
Is that a picture of the power supply, and
the output signals???
The only pulsing we do is with the LEDS at about 2 Hz.
There is no relay driver pulsing. We run the internal
oscillator at 4MHz.
Just as an experiment. Ground your oscilloscope probe to the
circuit, and then without touching any traces,
move your oscilloscope probe tip in the air around
the microcontroller and around the relay driver chips.
I've found this to be a quick check for emitted radiation.
And you might be able to hunt down where the most radiation
is coming from. The high input imedance of the scope allows you to do this. And now that I think about it, the 1Meg input impedance of your pre-amp might be helping it pick up digital noise as well? (maybe?)
What kind of THD are you actually getting?
What is the amplitude of the signals in your photos.
Thanks,
Craig Beiferman
Readings are not from a probe, but using a shielded cable from RCA output jacks.
That is, signal sent from HP339 to DUT via RCA shielded cable to input RCA jacks on IS1; this is for images 4 and 5. For images 1-3, when APOX boards are inactivated, this same signal is connect to RCA jacks directly soldered via shielded cable to gain stage board. Output for images 4 and 5 is again via the RCA jacks on the IS1; and for images 1-3 taken via another set of RCA jacks coming directly off of the gain stage.
Input signal is 1V in at 1kHz for image 1 and 1V in at 10kHz for images 2-5. Output is approximately 2V due to 2x gain setting of gain stage. The output goes from the DUT to HP339. The THD and output voltage is monitored off of the HP339 by the oscilloscope.
THD levels are 0.007% for image 1; 0.015% for image 2; 0.020% for image 3; 0.025% for image 4; and 0.030% for image 5.
Looked at another way, in the THD once the SHM is working (image 5) is 0.025% from 100Hz to 5kHz or so, raising to 0.030% at 10kHz. In other words, there is a constant 0.02% being added by the SHM which seems to be frequency independent. (The gain stage alone has <0.01%THD below about 5kHz.)
It is pretty clear that the signal going through the SHM board is having significant distortion added. Nothing else changes from set-up to set-up as far as cabling, etc. This is what is termed a controlled experiment, which is why I presented the data as I have. (Not to bog all done with excess details, but I have tried other gain stage boards and other SHM boards along with other inputs jacks on the various IS1s that I have. All all the same.)
That is, signal sent from HP339 to DUT via RCA shielded cable to input RCA jacks on IS1; this is for images 4 and 5. For images 1-3, when APOX boards are inactivated, this same signal is connect to RCA jacks directly soldered via shielded cable to gain stage board. Output for images 4 and 5 is again via the RCA jacks on the IS1; and for images 1-3 taken via another set of RCA jacks coming directly off of the gain stage.
Input signal is 1V in at 1kHz for image 1 and 1V in at 10kHz for images 2-5. Output is approximately 2V due to 2x gain setting of gain stage. The output goes from the DUT to HP339. The THD and output voltage is monitored off of the HP339 by the oscilloscope.
THD levels are 0.007% for image 1; 0.015% for image 2; 0.020% for image 3; 0.025% for image 4; and 0.030% for image 5.
Looked at another way, in the THD once the SHM is working (image 5) is 0.025% from 100Hz to 5kHz or so, raising to 0.030% at 10kHz. In other words, there is a constant 0.02% being added by the SHM which seems to be frequency independent. (The gain stage alone has <0.01%THD below about 5kHz.)
It is pretty clear that the signal going through the SHM board is having significant distortion added. Nothing else changes from set-up to set-up as far as cabling, etc. This is what is termed a controlled experiment, which is why I presented the data as I have. (Not to bog all done with excess details, but I have tried other gain stage boards and other SHM boards along with other inputs jacks on the various IS1s that I have. All all the same.)
THD
Robert,
Please excuse my THD ignorance.
I just did some reading about the HP339, and THD measurements
in general.
On this audiovideo
web site
They suggest that a THD of below 0.5% is adequate.
and your reporting a value that is more than ten times better.
What sort of THD are you expecting?
Does the audio sound bad? Is there a constant hum?
Also if you didn't know, the HP339 is only capable of measuring
down to .01% THD.
-Craig Beiferman
Robert,
Please excuse my THD ignorance.
I just did some reading about the HP339, and THD measurements
in general.
On this audiovideo
web site
They suggest that a THD of below 0.5% is adequate.
and your reporting a value that is more than ten times better.
What sort of THD are you expecting?
Does the audio sound bad? Is there a constant hum?
Also if you didn't know, the HP339 is only capable of measuring
down to .01% THD.
-Craig Beiferman
HP / Agilent 339A Distortion Analyzer (image below)
Officical Specs:
*10 Hz to 110 kHz
* Measures THD to 0.01% full scale
* Internal 10 Hz to 110 kHz oscillator with 0.0022% THD to 20 kHz
The unit can easily read THD to 0.001% (0.01% is 'full scale', I do know; and full scales means you can read a bit lower, yes?) and self-test noise is about 5 microvolts wtih a full scale of 1,000 microvolts; self test THD is < 0.001 on my unit.
Now as to 'adequate', well 0.5% is maybe good for a tube circuit, but look at any decent specs, such as products by Mr. Pass, and you'll see much lower values. THD is a function of frequency, which is why I showed in image 1 and described the frequency effects.
If you check with others on this forum (ask Hugh Dean), most amplifiers (preamps too) have a low THD that begins rising at 2kHz to 10kHz and continues upwards, depending upon basic topology and the amount of feedback.
I certainly don't want the garbage that I see on image 5. A company that sold something like that would get trashed in the market place and by reviewers. The problem here is that the active gain stage has a certain amount of inherent distortion; that's a given. What I don't understand, is why the 'passive' volume component should be adding more.
Besides jumping over my equipment or techniques or whether I use monster sized ground cables, is there constructive self-examination of this circuitry? Could there possibly be a design flaw?
I don't think my initial question concerning these distortions are so unreasonble, especially after have spent several hundred dollars and probably bought more parts from you than just about any other single person.
Officical Specs:
*10 Hz to 110 kHz
* Measures THD to 0.01% full scale
* Internal 10 Hz to 110 kHz oscillator with 0.0022% THD to 20 kHz
The unit can easily read THD to 0.001% (0.01% is 'full scale', I do know; and full scales means you can read a bit lower, yes?) and self-test noise is about 5 microvolts wtih a full scale of 1,000 microvolts; self test THD is < 0.001 on my unit.
Now as to 'adequate', well 0.5% is maybe good for a tube circuit, but look at any decent specs, such as products by Mr. Pass, and you'll see much lower values. THD is a function of frequency, which is why I showed in image 1 and described the frequency effects.
If you check with others on this forum (ask Hugh Dean), most amplifiers (preamps too) have a low THD that begins rising at 2kHz to 10kHz and continues upwards, depending upon basic topology and the amount of feedback.
I certainly don't want the garbage that I see on image 5. A company that sold something like that would get trashed in the market place and by reviewers. The problem here is that the active gain stage has a certain amount of inherent distortion; that's a given. What I don't understand, is why the 'passive' volume component should be adding more.
Besides jumping over my equipment or techniques or whether I use monster sized ground cables, is there constructive self-examination of this circuitry? Could there possibly be a design flaw?
I don't think my initial question concerning these distortions are so unreasonble, especially after have spent several hundred dollars and probably bought more parts from you than just about any other single person.
Attachments
Robert,
I'll start looking into this problem right away
First I want to narrow down how this noise is being created.
I want to eliminate the SHM microcontroller from the
picture. So I'll remove the microcontroller, and hardwire the highest volume setting by tying high pin 2 and 3, and pins 5 and 6 of IC3. (Which is volume setting 255).
If you look at your +5V power supply on the SHM board
with AC coupling and high magnification,
Am I correct to assume that you can see the same
signal noise pattern on +5V power supply, as you do on the distortion output?
I remember a similiar noise problem back in my college
days. I was designing an audio to digital recorder that
transmitted the recording with AM modulation.
But the noise on the audio was terrible. I then placed a IN4004
diode in series with the power supply, and the noise went away
completely. I still never figured out what that diode was doing
that improved the noise so much.
-Craig
I'll start looking into this problem right away
First I want to narrow down how this noise is being created.
I want to eliminate the SHM microcontroller from the
picture. So I'll remove the microcontroller, and hardwire the highest volume setting by tying high pin 2 and 3, and pins 5 and 6 of IC3. (Which is volume setting 255).
If you look at your +5V power supply on the SHM board
with AC coupling and high magnification,
Am I correct to assume that you can see the same
signal noise pattern on +5V power supply, as you do on the distortion output?
I remember a similiar noise problem back in my college
days. I was designing an audio to digital recorder that
transmitted the recording with AM modulation.
But the noise on the audio was terrible. I then placed a IN4004
diode in series with the power supply, and the noise went away
completely. I still never figured out what that diode was doing
that improved the noise so much.
-Craig
Problem Solved!
First, the APOX products are not at fault.
I now have the image 5 set-up results looking like image 2. (BTW, the sine waves in the images are 1V out, not 2V as written.) THD is now 0.013% with about 50 uV of residual noise. This means that the S/N is about 110 dB (17V/50uV). Decidedly, no hum.
How?
Back when the APOX-2 came out, there was a hum problem that was seemingly solved by connecting the signal ground to the digital ground. I followed this same toplogy, using a central star for connecting the two grounds at only one location.
This morning, I severed the digital ground from this star, effectively floating the digital supply, and all distortions disappeared. Therefore, the two grounds should not be connected (at least with the SHM; I have not re-tested the APOX-2). I have left the wall wart digital supply in the bottom layer of the 3 layer chassis, so its proximity to the other supplies is not a problem.
After the above was resolved, I further tested R3 on the IS1 board and the FB1 on the SHM. R3 had no effect. Jumpering FB1, lowered the THD, but insignificantly.
Bottom line: keep the digital supply totally separate from the analog supply and float it.
On another matter, that is less important, the above fiddling did not help a pop issue. I seem have what I imagine is a software-treatable pop when changing input selections in BAL mode. That is, when using balanced (XLR) inputs on one IS1 and single ended (RCA) on another. If the XLR inputs are assigned to be 'SE', no pops. If set to 'BAL', pops occur when selecting inputs.
Regards, Robert
First, the APOX products are not at fault.
I now have the image 5 set-up results looking like image 2. (BTW, the sine waves in the images are 1V out, not 2V as written.) THD is now 0.013% with about 50 uV of residual noise. This means that the S/N is about 110 dB (17V/50uV). Decidedly, no hum.
How?
Back when the APOX-2 came out, there was a hum problem that was seemingly solved by connecting the signal ground to the digital ground. I followed this same toplogy, using a central star for connecting the two grounds at only one location.
This morning, I severed the digital ground from this star, effectively floating the digital supply, and all distortions disappeared. Therefore, the two grounds should not be connected (at least with the SHM; I have not re-tested the APOX-2). I have left the wall wart digital supply in the bottom layer of the 3 layer chassis, so its proximity to the other supplies is not a problem.
After the above was resolved, I further tested R3 on the IS1 board and the FB1 on the SHM. R3 had no effect. Jumpering FB1, lowered the THD, but insignificantly.
Bottom line: keep the digital supply totally separate from the analog supply and float it.
On another matter, that is less important, the above fiddling did not help a pop issue. I seem have what I imagine is a software-treatable pop when changing input selections in BAL mode. That is, when using balanced (XLR) inputs on one IS1 and single ended (RCA) on another. If the XLR inputs are assigned to be 'SE', no pops. If set to 'BAL', pops occur when selecting inputs.
Regards, Robert
Robert,
That is great news!
The sweat was starting to bead up on my forehead.
I'll have Dale look at that popping problem.
S/N of 110dB though the entire pre-amp
is that pretty darn good isn't it?
-Craig
That is great news!
The sweat was starting to bead up on my forehead.
I'll have Dale look at that popping problem.
S/N of 110dB though the entire pre-amp
is that pretty darn good isn't it?
-Craig
Robert,
I am so pleased that this worked out for you!!!!
I have one suggestion:
It is not usually optimal to have digital ground floating - there is potential (sic!) for the voltage to stray quite far from cabinet or other grounds.
For the best of your equipment, may I politely suggest that you consider some form of connection which might be someting like:
1. A couple of large resistors in series - perhaps 5-6 1-100K each.
2. A stack of diodes in anti-parallell (perhaps 2-3 in series for each leg) + a resistor (diodes have very high resistance to low current - you should have zero current there at steady state. Resistor should be closest to the star ground (or so says my theory).
3. Any other fun ideas perhaps involving high frequency inductors and regular inductors?
If you decide to look into this, you should try placing any such assemblies close to the star point.
Good luck!
Petter
I am so pleased that this worked out for you!!!!
I have one suggestion:
It is not usually optimal to have digital ground floating - there is potential (sic!) for the voltage to stray quite far from cabinet or other grounds.
For the best of your equipment, may I politely suggest that you consider some form of connection which might be someting like:
1. A couple of large resistors in series - perhaps 5-6 1-100K each.
2. A stack of diodes in anti-parallell (perhaps 2-3 in series for each leg) + a resistor (diodes have very high resistance to low current - you should have zero current there at steady state. Resistor should be closest to the star ground (or so says my theory).
3. Any other fun ideas perhaps involving high frequency inductors and regular inductors?
If you decide to look into this, you should try placing any such assemblies close to the star point.
Good luck!
Petter
- Status
- Not open for further replies.