Hi guys,
I'm having issues with a buzzing noise on my self made pre amp.
I manufactured it (together with a power amp, see pictures) during my apprenticeship 22 years ago. I believe the noise was present since ever, that's why it did not use for a longer time.
Unfortunately I did not have an oscilloscope in place to track down the noise.
However, I can state the following:
Thanks in advance, Michael
I'm having issues with a buzzing noise on my self made pre amp.
I manufactured it (together with a power amp, see pictures) during my apprenticeship 22 years ago. I believe the noise was present since ever, that's why it did not use for a longer time.
Unfortunately I did not have an oscilloscope in place to track down the noise.
However, I can state the following:
- The buzzing noise is constantly present and not depending on any switching to different inputs or connecting/unconnecting inputs.
- As well it appears when the volume is down.
- I used the spectroid app on the handy to record the noise in front of the speaker. There are obvious peaks at 50Hz, 100Hz, 200Hz and harmonics. See attachment for details of the spectragram.
- The root cause must be somewhere in or before the power supply (see schematics attached). Btw. my supply is +/12V instead of 15V, using LM7812/7912 from Motorola (output noise voltage Vn = 10uV/Vo).
As well the filter before the voltage regulator is modified with an RC-circuit (10R and 4700uF Elko instead of 2200uF)) - When disconnecting the internal power supply +12/-12V and using an external device, the noise dissapears.
- Exchanging all capacitors on the power supply did not solve the problem
- Exchanging the rectifier did not solve the problem
- Exchanging to different voltage regulators LM7812/7912 from On Semi did not help. It even amplified to problem (higher Vn = 40uV/Vo might be responsible).
- I checked the ground routing with an ohm-meter, every assembly seems well connected to ground and the housing.
- Exchanging the transformer does not solve the problem.
Thanks in advance, Michael
Attachments
Hi! There is a fault in the system. For starters please use a 3 pin IEC inlet so with PE as the casing is made from metal and we like safety. You will love the challenge as it means drilling and filing. Best would be to use an IEC inlet that has PE combined with a quality mains filter so HF/RF will be filtered. That types are somewhat more easy to fit in general as the hole often is rectangular. Connect PE directly to the casing with the well-known yellow/green flexible wire in the specified way so with bolt, nut, washer. Please measure continuity.
Disconnect Audio GND (Masse) from the casing and connect with a 1W 100 Ohm resistor again. So now Audio GND/Masse is only connected to the casing/PE with a 100 Ohm resistor. The idea should be that the casing is not a part of the circuit. The casing should only be directly connected to Sicherheitserde/Protective Earth (PE) for safety and shielding.
Also check if, for instance, the headphones connector GND does not make contact with the casing. There should only be 1 point where Audio GND and PE meet. I am pretty sure your issues will be gone then.
In the power amplifier also a 3 pin IEC inlet with PE should be used, again for safety but there you may connect Audio GND to PE directly (this in no other sources device!).
Disconnect Audio GND (Masse) from the casing and connect with a 1W 100 Ohm resistor again. So now Audio GND/Masse is only connected to the casing/PE with a 100 Ohm resistor. The idea should be that the casing is not a part of the circuit. The casing should only be directly connected to Sicherheitserde/Protective Earth (PE) for safety and shielding.
Also check if, for instance, the headphones connector GND does not make contact with the casing. There should only be 1 point where Audio GND and PE meet. I am pretty sure your issues will be gone then.
In the power amplifier also a 3 pin IEC inlet with PE should be used, again for safety but there you may connect Audio GND to PE directly (this in no other sources device!).
Last edited:
Another obvious problem is D1 & D2 have no RF snubbers on them. Put .01 uf across them or between the output rails.
2nd non-obvious problem, the .047 uf caps across the bridge rectifiers. Are these ceramic disks? That number is too big to be economic for ceramic disks, but wound plastic caps would have too much inductance to reduce the transmission of RF by the shut off of the diodes in the bridge. Most preamps I've used had one .01 uf ceramic disk from + to - of the bridge, not 4 of them.
3rd non-obvious problem, is that small board on the left the rectifier diodes & bridge? you need to put a grounded steel box around them. Safety ground, not analog ground. Both my dynaco PAS2 and my Peavey MMA mixer-amps have grounded steel walls between the rectification section and the high gain audio sections.
2nd non-obvious problem, the .047 uf caps across the bridge rectifiers. Are these ceramic disks? That number is too big to be economic for ceramic disks, but wound plastic caps would have too much inductance to reduce the transmission of RF by the shut off of the diodes in the bridge. Most preamps I've used had one .01 uf ceramic disk from + to - of the bridge, not 4 of them.
3rd non-obvious problem, is that small board on the left the rectifier diodes & bridge? you need to put a grounded steel box around them. Safety ground, not analog ground. Both my dynaco PAS2 and my Peavey MMA mixer-amps have grounded steel walls between the rectification section and the high gain audio sections.
Last edited:
Thanks for your inputs Jean Paul.
I understand the concerns with regards to safety and earthing of the housing (btw. it's anodized aluminum).
But can you eloborate a bit more into detail why you believe the earthing and 100 Ohm separation of the Audio GND might be a good thing and pervent the occuring noise?
Whatsoever, I can try to make a mockup with earthing and 100 resistor at the weekend to try before modifying anything on the housing.
The amp is (in this color) unique. We did everything on our own from PCB design, etching the PCB, to CNC milling the housing, turning the knobs and assemble everything - so there is quite an emotional value in this device and I like to prevent any hasty actions...
Thanks for your feedback.
I'm quite sure that there is not the issue as I exchanged components (or left some away to try out) on the power supply and tried to opimize the filtering.
Non the attempts have been successful.
Actually I found this same circuitry for the power supply on the web Power Supply for Preamps - I have hard time to believe that the design of the supply it's the general issue of sthe noise I'm observing.
50 hz, 100 hz, 200 hz? serious indicators to switching noise.
I doubt if you have equipment necessary to detect the radio frequency interference. Optimize the design how? by looking at the DC output with a scope? Radio interference transmits through space, not the wire.
My dead quiet PAS2 predates all safety grounding of cases. 1959 design with 2 pin power cord. The case is connected to the shell of the turntable arm by a wire. That's it. How does it not rattle, hum & buzz? One indicator, vacuum tube rectification that doesn't shut off in the radio band. Other is complete separation of the AC parts & wiring from the high gain parts & wiring, by steel. I'm not suggesting lifting the safety ground, that is modern & safer. But safety grounded case has nothing to do with quiet design.
@Scientek: it is not hasty when it hummed for 22 years 😉
Just try it and see. It is required by law (the device is NOT safe) and Audio logic, you’ll see.
The PSU by Rod Elliot (only wisdom from this master!) is not the same as it does not connect Audio GND to the casing as that is wrong. Your PSU is fine too except for the GND to casing connection.
Just try it and see. It is required by law (the device is NOT safe) and Audio logic, you’ll see.
The PSU by Rod Elliot (only wisdom from this master!) is not the same as it does not connect Audio GND to the casing as that is wrong. Your PSU is fine too except for the GND to casing connection.
Last edited:
First of all, a 3-pin mains connection is absolutely not required for a device like this, and in fact you will find that a majority a big-name hi-fi components have 2-prong mains leads, and for good reason - it is pretty much necessary to avoid trouble when using unbalanced audio connections. However, internal construction must conform to the higher standards of an IEC Class II (double insulated) device. In a nutshell, no single insulation fault must be able to make the chassis go live. Every mains conductor must have two layers of insulation. The mains switch wiring looks to be single insulated in this case, so that would be a violation - I would suggest using lamp cord (ordinary 2-wire double insulated mains cable) for this run instead. Using heatshrink on soldered connections was a good idea, keep doing that - heatshrink + several mm of air is double insulated, too. (*)
The mains switch would no doubt also appreciate a small R-C snubber in parallel for arcing suppression, you can buy these as RF suppression capacitors. Nothing too big for this device, maybe 10-22 nF and 100 ohms or so?
Incidentally, I don't have a problem with the 4 caps across the rectifier diodes - this is standard practice for keeping them from interfering with AM radio. These are not a snubber per se, so adding one across the secondary may still be beneficial.
There is a good amount of 50 Hz in the offending buzz, which suggests either of two sources:
1. Electrostatic coupling of high-impedance nodes to transformer E field. Now there is what appears to be a shielding can over the transformer, so this becomes rather unlikely even with the ribbon cables running close by. A look at ribbon cable pinout would still be good. I may be inclined to wrap them in copper tape connected to chassis if need be. I take it that all chassis parts have good electrical connection between them, and chassis connects to circuit ground somewhere? (Anodised surfaces do not seem to be very conductive at all.)
2. Magnetic coupling of internal ground loops to transformer H field. This pretty much requires finding and eliminating the offending ground loops. We would need to make a plan of the entire grounding scheme of this amplifier, which will take about all the details you could provide us with, OP - notably board layouts and a wiring plan. I bet the problem is somewhere in there.
For one thing, I am not liking the individual wires connecting the RCA jacks very much, it's not exactly the epitome of shielding. Also, it is generally more advantageous to be using the metal backplate as a (sub-)star ground, rather than trying to painstakingly insulate all the RCA jacks from it as seems to be the case here (or am I seeing that wrong?). It means much less effort required in keeping RF out if nothing else. This is standard practice in basically anything Japanese and mass-market with lots of RCA jacks.
I am guessing the sub PCBs are single layer affairs, while the main board may be dual layer? It's all looking very '90s for sure.
*) Note: A 3-prong mains connection may still be beneficial when your amplifier is nominally IEC Class II. All mains transformers have a bit of capacitive coupling from primary to secondary (several hundred pF), which tends to elevate circuit ground above PE potential, potentially causing open inputs to hum. The bigger the transformer the worse it is - I hope yours isn't more than 10-20 VA tops. (This is why hi-fi integrated or power amplifiers often have transformers with a shield winding, which cuts down on capacitive coupling dramatically.) You can counteract this leakage by using a Y2 rated safety capacitor to connect the incoming PE to chassis ground (and nothing else). 10 nF should be plenty here. Use heatshrink for insulation.
The mains switch would no doubt also appreciate a small R-C snubber in parallel for arcing suppression, you can buy these as RF suppression capacitors. Nothing too big for this device, maybe 10-22 nF and 100 ohms or so?
Incidentally, I don't have a problem with the 4 caps across the rectifier diodes - this is standard practice for keeping them from interfering with AM radio. These are not a snubber per se, so adding one across the secondary may still be beneficial.
There is a good amount of 50 Hz in the offending buzz, which suggests either of two sources:
1. Electrostatic coupling of high-impedance nodes to transformer E field. Now there is what appears to be a shielding can over the transformer, so this becomes rather unlikely even with the ribbon cables running close by. A look at ribbon cable pinout would still be good. I may be inclined to wrap them in copper tape connected to chassis if need be. I take it that all chassis parts have good electrical connection between them, and chassis connects to circuit ground somewhere? (Anodised surfaces do not seem to be very conductive at all.)
2. Magnetic coupling of internal ground loops to transformer H field. This pretty much requires finding and eliminating the offending ground loops. We would need to make a plan of the entire grounding scheme of this amplifier, which will take about all the details you could provide us with, OP - notably board layouts and a wiring plan. I bet the problem is somewhere in there.
For one thing, I am not liking the individual wires connecting the RCA jacks very much, it's not exactly the epitome of shielding. Also, it is generally more advantageous to be using the metal backplate as a (sub-)star ground, rather than trying to painstakingly insulate all the RCA jacks from it as seems to be the case here (or am I seeing that wrong?). It means much less effort required in keeping RF out if nothing else. This is standard practice in basically anything Japanese and mass-market with lots of RCA jacks.
I am guessing the sub PCBs are single layer affairs, while the main board may be dual layer? It's all looking very '90s for sure.
*) Note: A 3-prong mains connection may still be beneficial when your amplifier is nominally IEC Class II. All mains transformers have a bit of capacitive coupling from primary to secondary (several hundred pF), which tends to elevate circuit ground above PE potential, potentially causing open inputs to hum. The bigger the transformer the worse it is - I hope yours isn't more than 10-20 VA tops. (This is why hi-fi integrated or power amplifiers often have transformers with a shield winding, which cuts down on capacitive coupling dramatically.) You can counteract this leakage by using a Y2 rated safety capacitor to connect the incoming PE to chassis ground (and nothing else). 10 nF should be plenty here. Use heatshrink for insulation.
Last edited:
@sgrossklass:
Thanks for this in-depth information with very reasonable items.
- I agree on the 2- and 3-pin mains connection topic, I quickly overflew the requirements on the web.
- Yes, it's very 90's. As said we designed and manufactured it during our apprenticeship as electrotechnician (Elektoniker) in 1998. Unfortunately I was not responsible for the circuitry, so I cannot tell from where we adopted it.
To be clear, it's not based on hundred of hours of engineering craftmanship.
- All PCB's do have only one layer. We even exposed and etched them ourselves as we had the devices/baths in our apprentice workshop.
-All plates of the housing are well connected together (anodises surfaces have been grinded where the screw are. I checked it with an ohm-meter.
- I could only organize paperform information of schematics, BOM and PCB assembly. Unfortunately not Gerber files and stuff like that.
I can attach the schematic of the pre amp as additional information.
@Jean Paul and all:
I tried out yesterday the ground-lift fix (using 3-pin mains with PE (on housing) and separating the audio GND with a 100 Ohm resistor.
It helps to remove the hum coming from the transformer/supply side!
So I will definitively follow up on a IEC inlet to have a proper 3-pin mains solution.
However, I found out, that there is a generic noise (not low freq hum) present, even when using DC-voltages from an external supply and internal supply side switched off.
I changed all the electrolyte capacitors although they still looked well and measured the remaining caps. All good there.
I assume I have to live with noise appearing as it may be based on the design/circuitry and components chosen - my colleagues at work don't think it's worth to mention, as disappears quickly in sound once musis plays,
But I'm a bit more picky there...
I will follow up when I find again time for it.
So if anyone would observe some critical stuff in the pre amp circuitry, any further inputs are much appreciated.
Anyway thanks a lot for your help so far!
Thanks for this in-depth information with very reasonable items.
- I agree on the 2- and 3-pin mains connection topic, I quickly overflew the requirements on the web.
- Yes, it's very 90's. As said we designed and manufactured it during our apprenticeship as electrotechnician (Elektoniker) in 1998. Unfortunately I was not responsible for the circuitry, so I cannot tell from where we adopted it.
To be clear, it's not based on hundred of hours of engineering craftmanship.
- All PCB's do have only one layer. We even exposed and etched them ourselves as we had the devices/baths in our apprentice workshop.
-All plates of the housing are well connected together (anodises surfaces have been grinded where the screw are. I checked it with an ohm-meter.
- I could only organize paperform information of schematics, BOM and PCB assembly. Unfortunately not Gerber files and stuff like that.
I can attach the schematic of the pre amp as additional information.
@Jean Paul and all:
I tried out yesterday the ground-lift fix (using 3-pin mains with PE (on housing) and separating the audio GND with a 100 Ohm resistor.
It helps to remove the hum coming from the transformer/supply side!
So I will definitively follow up on a IEC inlet to have a proper 3-pin mains solution.
However, I found out, that there is a generic noise (not low freq hum) present, even when using DC-voltages from an external supply and internal supply side switched off.
I changed all the electrolyte capacitors although they still looked well and measured the remaining caps. All good there.
I assume I have to live with noise appearing as it may be based on the design/circuitry and components chosen - my colleagues at work don't think it's worth to mention, as disappears quickly in sound once musis plays,
But I'm a bit more picky there...
I will follow up when I find again time for it.
So if anyone would observe some critical stuff in the pre amp circuitry, any further inputs are much appreciated.
Anyway thanks a lot for your help so far!
As predicted.... But ... how did it end?
(Forget about DIYing a class II device, make that class I with chassis/casing connected to PE)
Last edited:
Current cannot flow without voltage drop. Ground wiring is not equipotential if supply current is flowing in it. The supply current will have 100Hz and higher harmonics in it. This noise voltage can easily contaminate signals if any flows in ground wires associated with inputs, either internally or through ground loops with externally connected equipment.
For example, the 12V supply return wire to transformer centre tap will have the capacitor charging ripple current on it. You must keep this current out of the other ground connection to the ±15V supplies by taking the 12V supply return current directly to the transformer centre tap with its own conductor not shared with the ±15V section. You have enough pins in you connector to keep them seperate. This may not be the source of your problem, however it is the kind of poor circuit technique that does introduce ripple current noise.
For example, the 12V supply return wire to transformer centre tap will have the capacitor charging ripple current on it. You must keep this current out of the other ground connection to the ±15V supplies by taking the 12V supply return current directly to the transformer centre tap with its own conductor not shared with the ±15V section. You have enough pins in you connector to keep them seperate. This may not be the source of your problem, however it is the kind of poor circuit technique that does introduce ripple current noise.
Sorry for not feedbacking on this topic.
Fix 1.
The low buzz was gone after implementing your input Jean-Paul.
Fix 2:
Actually, I was chasing ghosts.
The generic noise present is circuit/component related and due to suboptimal signal-to-noise ratio.
As a fix I have installed RCA-attenuators (10dB) inbetween the pre-amp and the power-amp, which work very well and satisfactory for the daily use in the living room.
This helps to minimize the noise present (non volume related) and that I can use the volume knob more conveniently (no high volume already at 8 o'clock).
Thanks again for your inputs.
Is there any possibility to close threads?