PS any advantage of pulling resistors 30,31 and 32 off the board as I have no use for this VU meter circuit ? - it's not connected to anything
Could you reply to the 4 points? I changed the sequence.
1: did it occur to you to pull the fan and see if that makes a difference? Yes/No
2: what is the input voltage of that 7815 in reality? ...V
3: that 7815 might be oscillating. It has a pretty dumb voltage dropping pre regulator with a 1k2 resistor/24V Zener diode but no decoupling input cap. Add a 1 µF 100V or higher input cap. Is there a difference? Yes/No
4: are the main filter caps or even all electrolytic caps still the original ones? Samwha electrolytic caps were notorious in the past and these seem to be from the past. Yes/No
1: did it occur to you to pull the fan and see if that makes a difference? Yes/No
2: what is the input voltage of that 7815 in reality? ...V
3: that 7815 might be oscillating. It has a pretty dumb voltage dropping pre regulator with a 1k2 resistor/24V Zener diode but no decoupling input cap. Add a 1 µF 100V or higher input cap. Is there a difference? Yes/No
4: are the main filter caps or even all electrolytic caps still the original ones? Samwha electrolytic caps were notorious in the past and these seem to be from the past. Yes/No
Last edited:
1. I removed the 230V fan (very noisy) and fitted a 12v PC fan with a small linear power supply that is triggered by temperature, the noise is the same with it connected or disconnected so not fan related
2,3. I will do that shortly
4. The build is 2005 and I don't think any caps have been replaced, where I have been able to measure them they seem fine and in spec with expected uf and ESR, but I haven't been super diligent about this, I also dont know how much use the amps had. They don't look heavily used and no obvious signed of over heating. Perhaps I should replace the smaller electrolytic as that will be cheap ?
2,3. I will do that shortly
4. The build is 2005 and I don't think any caps have been replaced, where I have been able to measure them they seem fine and in spec with expected uf and ESR, but I haven't been super diligent about this, I also dont know how much use the amps had. They don't look heavily used and no obvious signed of over heating. Perhaps I should replace the smaller electrolytic as that will be cheap ?
1: OK
2: TBD
3: TBD
4: 2005, OK. With Samhwa caps that is a high chance of high ESR/ being dried out. Of course replace 20 year old electrolytic caps, it is the first thing to do when buying such old stuff. I would pick a bipolar MUSE ES 100 µF for C3.
2: TBD
3: TBD
4: 2005, OK. With Samhwa caps that is a high chance of high ESR/ being dried out. Of course replace 20 year old electrolytic caps, it is the first thing to do when buying such old stuff. I would pick a bipolar MUSE ES 100 µF for C3.
47 to 220uF is fine.
follow + from main cap to first output fet and put it there, repeat for - rail
A lot is going on here, and maybe retracing your steps to confirm you are headed in the right direction would be a good idea.
1. When you removed the input signal ground connection to the chassis, did it improve the noise YES/NO? (This is a ground loop issue)
2. When the inputs are shorted, is the amplifier quiet? YES/NO? If no, this looks like a common impedance noise problem
3. Did Teeing off the amplifier 0V return from the power 0V connection reduce noise YES/NO? (This is a common impedance issue - do not mix power 0V, decoupling 0V and signal 0V)
4. Is the input signal ground Teed off the amplifier 0V connection and not just going back to the power 0V YES/NO (this is a common impedance issue arising from mixing the order of the ground connections)
5. Is the feedback resistor network (feedback resistor and gain setting resistor) routed correctly, and specifically, is the gain setting resistor connected to a clean 0V point YES/NO? Taking the gain setting resistor to the power 0V will result in serious noise issues
6. Is the loop area between the feedback resistor and the front-end signal 0V small YES/NO?
What is the ripple on your supply rails with no input signal? Adding 100nF here or there or a few 10s of microfarad will make no difference to the noise if it is arising from any of the above mechanisms. As discussed earlier, you should try to confirm that your amp does not have an HF oscillation problem. Just a few hundred mV is enough to cause havoc noise-wise. do you have an oscilloscope?
1. When you removed the input signal ground connection to the chassis, did it improve the noise YES/NO? (This is a ground loop issue)
2. When the inputs are shorted, is the amplifier quiet? YES/NO? If no, this looks like a common impedance noise problem
3. Did Teeing off the amplifier 0V return from the power 0V connection reduce noise YES/NO? (This is a common impedance issue - do not mix power 0V, decoupling 0V and signal 0V)
4. Is the input signal ground Teed off the amplifier 0V connection and not just going back to the power 0V YES/NO (this is a common impedance issue arising from mixing the order of the ground connections)
5. Is the feedback resistor network (feedback resistor and gain setting resistor) routed correctly, and specifically, is the gain setting resistor connected to a clean 0V point YES/NO? Taking the gain setting resistor to the power 0V will result in serious noise issues
6. Is the loop area between the feedback resistor and the front-end signal 0V small YES/NO?
What is the ripple on your supply rails with no input signal? Adding 100nF here or there or a few 10s of microfarad will make no difference to the noise if it is arising from any of the above mechanisms. As discussed earlier, you should try to confirm that your amp does not have an HF oscillation problem. Just a few hundred mV is enough to cause havoc noise-wise. do you have an oscilloscope?
Jean- Paul, hi Looks like I will need to build one with a remote transformer as no room to measure the voltage that part is hidden under the transformer, so probably 2moro now as I'm on log cutting duties my fiancé's Mums today
Ctrix, thanks so something like this, I can only have small <10mm diaper caps is located this close to the MosFets, also need to think through the best route for tracks. I could locate the caps nearer the GND but suspect that defeats the point ?
Andrew, hi
Here goes
1. when I had no 0.01uf from 0v to case the most 'buzz' and GND loop was with the case earthed at the local plug and the unit connected to the DSP with no connection between 0v and the case. in this configuration (no 0.01uf) connecting the 0v to the chassis earth was better then disconnected but not as good as no connections and the input shorted. However with no connections to the input and shorted still some noise
2. When the inputs are showed it's reasonably quiet but still a little audible 'hiss' and a very low background 'buzz' this is not loud but not good or for me good enough
3. The relocating the 0v from the transformer, adding to the 0v/GND and moving the speaker out 0v did seem to make a difference (now taking measurements and seeing an improvement, I can confirm) at the time as it wasn't transformational I didn't stand back and notice the marginal improvement at the time - it wasn't quiet like the 500 is
4. The input GND has the 0.001uf cap to chassis as noted above and a single line from the connection to the board (bottom left above), where I tacked in some more copper to reduce the impedance to the centre 0v. I will tidy up these connections, I'm just waiting for a 200W iron to be delivered my 85W Weller station is marvellous for 90% of things but these ugly solders were right on the limit of my iron. I did have a connection wire added and it racked though the hole I drilled in the board but anywhere I located it to (RCA, Earth, chassis no earth etc) it just increased the noise, so I removed it
5. The gain resistor (now 390R) and 100uf cap are close to the input, blue circle board input green circle feedback cap and resistor photo of the tracks (I have removed the 100uf an it measures fine but will replace as suggested by Jean
6. Not sure if you would say small or large with loop area I do get a little confused by what is included and what isn't (less important)
Alas I don't have a scope
Ctrix, thanks so something like this, I can only have small <10mm diaper caps is located this close to the MosFets, also need to think through the best route for tracks. I could locate the caps nearer the GND but suspect that defeats the point ?
Andrew, hi
Here goes
1. when I had no 0.01uf from 0v to case the most 'buzz' and GND loop was with the case earthed at the local plug and the unit connected to the DSP with no connection between 0v and the case. in this configuration (no 0.01uf) connecting the 0v to the chassis earth was better then disconnected but not as good as no connections and the input shorted. However with no connections to the input and shorted still some noise
2. When the inputs are showed it's reasonably quiet but still a little audible 'hiss' and a very low background 'buzz' this is not loud but not good or for me good enough
3. The relocating the 0v from the transformer, adding to the 0v/GND and moving the speaker out 0v did seem to make a difference (now taking measurements and seeing an improvement, I can confirm) at the time as it wasn't transformational I didn't stand back and notice the marginal improvement at the time - it wasn't quiet like the 500 is
4. The input GND has the 0.001uf cap to chassis as noted above and a single line from the connection to the board (bottom left above), where I tacked in some more copper to reduce the impedance to the centre 0v. I will tidy up these connections, I'm just waiting for a 200W iron to be delivered my 85W Weller station is marvellous for 90% of things but these ugly solders were right on the limit of my iron. I did have a connection wire added and it racked though the hole I drilled in the board but anywhere I located it to (RCA, Earth, chassis no earth etc) it just increased the noise, so I removed it
5. The gain resistor (now 390R) and 100uf cap are close to the input, blue circle board input green circle feedback cap and resistor photo of the tracks (I have removed the 100uf an it measures fine but will replace as suggested by Jean
6. Not sure if you would say small or large with loop area I do get a little confused by what is included and what isn't (less important)
Alas I don't have a scope
It seems the 7815 PCB pads are in plain sight and adding an input cap to the 7815 just to exclude it is no problem either. A remote transformer is again a new variable that would not make anything more clear. Stop adding new variables, it just won’t help you.
Many words but still 20 year old Samwha filter caps. If you have a few fresh 2200 …..10,000 uF 100V caps you can clip those with crocodile clips in parallel to see if hum disappears. You should also have measured the current in both + and -. I have had amplifiers where the filter caps became a load themselves.
Why? I simply do not believe this product was sold humming. It should not require a redesign. At all. If it is the Fiat Ritmo of amplifiers it would be known. With 20 year old stuff a refresh is much more likely a necessity.
Many words but still 20 year old Samwha filter caps. If you have a few fresh 2200 …..10,000 uF 100V caps you can clip those with crocodile clips in parallel to see if hum disappears. You should also have measured the current in both + and -. I have had amplifiers where the filter caps became a load themselves.
Why? I simply do not believe this product was sold humming. It should not require a redesign. At all. If it is the Fiat Ritmo of amplifiers it would be known. With 20 year old stuff a refresh is much more likely a necessity.
Last edited:
Jean, hi the 7815 is over the other side I think and under the transformer when built on the chassis, looking at the board this is actually a 7812 which is shown on the layout -so I guess they changed the design slightly.
If the caps are dried out and there is a 1k2 plus Zener diode but no decoupling cap that 7812 might oscillate and cause hiss. If that Zener diode is old it may be off its value. That is why you should measure. Like the current in both + and - in idle. The simplest of basic measurements.
It is old stuff with age sensitive parts. When I repaired such stuff sometimes old transistors and diodes had become noisy because of age. Literally anything old can break down. Imagine having more than 2 channels with such old stuff and that many parts of that age. A repair mans wet dream.
Not a fan of again a variable but that input cap is normally required anyway and will not harm.
If the amplifier does need a complete redesign then it IS the Fiat Ritmo of amplifiers and better not causing anymore time loss.
It is old stuff with age sensitive parts. When I repaired such stuff sometimes old transistors and diodes had become noisy because of age. Literally anything old can break down. Imagine having more than 2 channels with such old stuff and that many parts of that age. A repair mans wet dream.
Not a fan of again a variable but that input cap is normally required anyway and will not harm.
If the amplifier does need a complete redesign then it IS the Fiat Ritmo of amplifiers and better not causing anymore time loss.
Last edited:
I hear you with old stuff but I did have one of these as the processor (£14k new) and removed the older kit and the 'hiss' from this unit was simply too much, I do not find the idea that you can sit in the listening position with a normal relatively high volume and hear anything. It may well be that some thing the BK MF1000's are quiet but they aren't to me
With the RCA and 0.01uf cap in place the hum is as good as gone BUT the amplifier is audible at the seating position hence my continued pursuit too resolve, as said none of the other amps have this issue
If I could buy another two 500 BPSP I would have done so as they work well, they are not quite powerful enough for the 18" drivers and the BK1000's should be ideal !
The loud - Buzz is now gone but the hiss noise and it 'sounds' like a sensitive speaker with a 'cheap' amp noise is still there.
With the RCA and 0.01uf cap in place the hum is as good as gone BUT the amplifier is audible at the seating position hence my continued pursuit too resolve, as said none of the other amps have this issue
If I could buy another two 500 BPSP I would have done so as they work well, they are not quite powerful enough for the 18" drivers and the BK1000's should be ideal !
The loud - Buzz is now gone but the hiss noise and it 'sounds' like a sensitive speaker with a 'cheap' amp noise is still there.
I don't think adding small caps like this will solve anything. The small cap I mentioned earlier on from the signal ground at the connector to the chassis is there to provide RFI mitigation. But you can't determine if you have an RFI or HF oscillation problem unless you can look at the output with a scope. I would urge you to buy a scope, or borrow one, to look at the amlifier output and your LM7815 regulator. Without this, you are flying blind. I would not go changing out capacitors until you were sure they were not working. It's pretty easy to determine if they are with a scope - you can look at the ripple on the rails and make a guess as to what state the reservoirs are in, and any decouplers. My bet is these are good enough to use as you work to reduce the noise by improving the layout - you can always circle back later to work on these.
https://www.amazon.co.uk/HANMATEK-HO11-Handheld-Digital-Oscilloscope/dp/B0DHL6LJ88/ref=sr_1_18?crid=2SKC2EIJRQQO6&dib=eyJ2IjoiMSJ9._ahLQ1AZquX2j17D_bgxGOvA-sGqFlRQGX0Qly_GBWcxYBMfzVlaHMBtTUCUFgqGDpcI8Xp86C1Mxw0eKnz42eRLKudzlKACAsM0DN-SJgahP1ty9_uPVFcMv4nhOn8HTQdlcd9hLRAGHiQo-eqxIkn0YlIiIbFcI5CrJsqu_mg5TxvwxWuBZiLyEDm4v1yeGxs7_h_Edn_I5rP3NCjpJtLvT9w-4t3esMzSkdc2ajPFPkTExORB3Zp2-1nTjELtMUE9YzHxNUdn-AZcKG0091riNwQHc91ZIXXyNYLIEHgCaZTKEmiqsc_x7GoAeGyToFk0kpgshIbh-pZ7PAuKZgVkcZnU2Xt4US6RBAD0rXqUWghAFqV5V9OnzHccYupiDrg_Tclsih79eH73K9ytZx-B6rX1r86E044ZVBn0BK2EzChU5rEvoloZFe5cmrCU.kLO46_gGNZQpguYZbpuTHkeaFvszVgU3Jl7lS3YEoVY&dib_tag=se&keywords=oscilloscope&qid=1742047374&sprefix=oscill,aps,83&sr=8-18
If you want to build a quiet amplifier, you have to get the PSU and module wiring right and the PCB layout right.
https://www.amazon.co.uk/HANMATEK-HO11-Handheld-Digital-Oscilloscope/dp/B0DHL6LJ88/ref=sr_1_18?crid=2SKC2EIJRQQO6&dib=eyJ2IjoiMSJ9._ahLQ1AZquX2j17D_bgxGOvA-sGqFlRQGX0Qly_GBWcxYBMfzVlaHMBtTUCUFgqGDpcI8Xp86C1Mxw0eKnz42eRLKudzlKACAsM0DN-SJgahP1ty9_uPVFcMv4nhOn8HTQdlcd9hLRAGHiQo-eqxIkn0YlIiIbFcI5CrJsqu_mg5TxvwxWuBZiLyEDm4v1yeGxs7_h_Edn_I5rP3NCjpJtLvT9w-4t3esMzSkdc2ajPFPkTExORB3Zp2-1nTjELtMUE9YzHxNUdn-AZcKG0091riNwQHc91ZIXXyNYLIEHgCaZTKEmiqsc_x7GoAeGyToFk0kpgshIbh-pZ7PAuKZgVkcZnU2Xt4US6RBAD0rXqUWghAFqV5V9OnzHccYupiDrg_Tclsih79eH73K9ytZx-B6rX1r86E044ZVBn0BK2EzChU5rEvoloZFe5cmrCU.kLO46_gGNZQpguYZbpuTHkeaFvszVgU3Jl7lS3YEoVY&dib_tag=se&keywords=oscilloscope&qid=1742047374&sprefix=oscill,aps,83&sr=8-18
If you want to build a quiet amplifier, you have to get the PSU and module wiring right and the PCB layout right.
Last edited:
Tks , I guess this would be useful for a DIY guy then. Is this an okay model looks almost affordable !
Yes! It's all you need for basic fault-finding. If you are prepared to spend a bit more, you can get a pretty decent 2-channel scope nowadays for £150-200 that will do 100MHz+.
I've updated the 'Amplifier Wiring' presentation I posted earlier - please see the last two slides where the HBR is shown for your amp. An HBR is a very effective technique for isolating the signal grounds which absolutely cannot be mixed up with the 'dirty' ground on the amplifier PCB.
I've updated the 'Amplifier Wiring' presentation I posted earlier - please see the last two slides where the HBR is shown for your amp. An HBR is a very effective technique for isolating the signal grounds which absolutely cannot be mixed up with the 'dirty' ground on the amplifier PCB.
My scope has arrived, so I'll try and work out how to use this, the manual isn't super helpful !
Thinking about this HBR I assume the far left (390R/100uf) the 100k bottom and the right (4K7, 22nf) are disconnected from the local 0v and joined to a HBR then joined to the 'T'
If so doesn't that mean that 0v is no longer 0v for these circuits due to the resistance ?
I also assume that the 12K in the bottom left corner is left alone or would this be better relocated to the 'T' as well direct and not via the HBR
Thinking about this HBR I assume the far left (390R/100uf) the 100k bottom and the right (4K7, 22nf) are disconnected from the local 0v and joined to a HBR then joined to the 'T'
If so doesn't that mean that 0v is no longer 0v for these circuits due to the resistance ?
I also assume that the 12K in the bottom left corner is left alone or would this be better relocated to the 'T' as well direct and not via the HBR
I’d leave the 12k out for now. The main thing is to sort the big noise problem out.
‘Thinking about this HBR I assume the far left (390R/100uf) the 100k bottom and the right (4K7, 22nf) are disconnected from the local 0v and joined to a HBR then joined to the 'T'’
Yes. Run the clean ground back to the T so the loop area between the feedback resistors and the 0V is minimized. You do this by running it under the board beneath this resistors and then off to the T the other end.
‘Thinking about this HBR I assume the far left (390R/100uf) the 100k bottom and the right (4K7, 22nf) are disconnected from the local 0v and joined to a HBR then joined to the 'T'’
Yes. Run the clean ground back to the T so the loop area between the feedback resistors and the 0V is minimized. You do this by running it under the board beneath this resistors and then off to the T the other end.
Jean-Paul, hi
"3: that 7815 might be oscillating. It has a pretty dumb voltage dropping pre regulator with a 1k2 resistor/24V Zener diode but no decoupling input cap. Add a 1 µF 100V or higher input cap. Is there a difference? Yes/No"
Is this just a cap to GND after the resistor before the Diode ? Could I use a 22uf 100V cap ?
"3: that 7815 might be oscillating. It has a pretty dumb voltage dropping pre regulator with a 1k2 resistor/24V Zener diode but no decoupling input cap. Add a 1 µF 100V or higher input cap. Is there a difference? Yes/No"
Is this just a cap to GND after the resistor before the Diode ? Could I use a 22uf 100V cap ?
From input pin to GND pin at the 78xx. You know the standard input cap and output cap for 78xx/79xx/regulators?!
A 0.33…1 uF 100V film input cap is OK. If you have 100 or 220 nF then use one of those. Anything is better than nothing.
https://www.ti.com/lit/ds/symlink/ua78.pdf?ts=1742359174394
A 0.33…1 uF 100V film input cap is OK. If you have 100 or 220 nF then use one of those. Anything is better than nothing.
https://www.ti.com/lit/ds/symlink/ua78.pdf?ts=1742359174394
Last edited:
Ah okay, I looked at the 7812 in the drawing and only saw the output 0.1 and missed the fact nothing on the input !