What is the residual output noise level of this amplifier with power supply filter and Meanwell power supply?
Thanks Nelson. Since I'd be driving 107dB/W drivers, would like to confirm if power supply filtering is the best way to keep this as low as possible.
I don't know about the measurment and even have only the 1st edition (P and not N), so not this VFET version, but for whatever it is worth...
To start with, I built it exactly as intended by Papa and it was dead silent. ABSOLUTELY dead silent. More so than any amp we tried, including Class D, which really amazed us (we thought initialy the amp wasn't working). That was with Klipsch speakers advertised as 98dB/W/m, but which are less efficient in reality, admitely not with your super efficient speakers. But we heard nothing having our ears on the speakers and up to a volume position that is higher than any listening level!
For the sake of it, at a later point, we cranked the volume to max, which would result in an impossible listening level and started hearing a very faint noise from the tweeter, again ears on LS.
I went then through numerous filtering options, adding PS capacity, adding 2 SMPS filters on top of the existing one and while all these steps brought TO MY EARS quite a few sonic improvements over what was already a very engaging unit, it didn't change anything to the experience listed above. So still dead silent... and still the same very very faint noise when having volume at max and ears directly on the tweeter.
Bottom line: I am not sure in your case there will be a noise issue to start with, only experience in your very specific case can probably tell, but Papa designed this to be really dead silent. Now, if there is one in your case (it is the N version, so possibly noisier than mine), then I am not sure that filters will address it. Perhaps... but perhaps we start just hearing the VFET's electron noise LOL
Al IME and again with another VFET version
Claude
To start with, I built it exactly as intended by Papa and it was dead silent. ABSOLUTELY dead silent. More so than any amp we tried, including Class D, which really amazed us (we thought initialy the amp wasn't working). That was with Klipsch speakers advertised as 98dB/W/m, but which are less efficient in reality, admitely not with your super efficient speakers. But we heard nothing having our ears on the speakers and up to a volume position that is higher than any listening level!
For the sake of it, at a later point, we cranked the volume to max, which would result in an impossible listening level and started hearing a very faint noise from the tweeter, again ears on LS.
I went then through numerous filtering options, adding PS capacity, adding 2 SMPS filters on top of the existing one and while all these steps brought TO MY EARS quite a few sonic improvements over what was already a very engaging unit, it didn't change anything to the experience listed above. So still dead silent... and still the same very very faint noise when having volume at max and ears directly on the tweeter.
Bottom line: I am not sure in your case there will be a noise issue to start with, only experience in your very specific case can probably tell, but Papa designed this to be really dead silent. Now, if there is one in your case (it is the N version, so possibly noisier than mine), then I am not sure that filters will address it. Perhaps... but perhaps we start just hearing the VFET's electron noise LOL
Al IME and again with another VFET version
Claude
Just to be more specific with values... in my very specific case, the difference between unreasonably loudest level in real listening conditions, and volume to absolute max position was around 15dB.
That could cover the efficiency difference between our speakers and give you some guideline re noise... hopefully
Enjoy music with this fab amp
Claude
That could cover the efficiency difference between our speakers and give you some guideline re noise... hopefully
Enjoy music with this fab amp
Claude
I can confirm the approx. 100 uV with my N-Channel build. It is mostly some remains of 100 Hz ripple.
With my Jericho Horns (as seen in my Avatar), which are 100 dB/W efficient and feature the old Fostex FE208 Sigma drivers, I can just hear the faintest hum with my ear next to the driver in a very quiet room.
I have, in addition to the N-Channel kit power supply filter, put in to PO89ZB filters, one for each channel between the power suply filter PCB and the output boards.
Regards, Claas
If you want to reduce 100 Hz ripple, the PO89ZB filter boards sold in the diyAudio Store won't help very much. They are designed to eliminate SMPS high frequency (30KHz - 3 MHz) switching noise, not mains ripple.
To reduce 100 Hz ripple, you could HOPE that increasing the filter capacitance on the Nch VFET PSU filter card, will not cause the Mean Well brick to enter "Hiccup mode" at startup. The Nch VFET PSU filter card is laid out with capacitor footprints 12.5mm in diameter, so you could search around to find caps bigger than 1000uF/50V with that diameter. The biggest I found this morning were 1300uF/50V/12.5mm but maybe you'll have better success than me.
The THESEUS-PSFILT card is laid out with capacitor footprints 16.0mm in diameter, so you could search around to find caps bigger than 1000uF/50V with that diameter. The biggest I could find this morning were 2200uF/50V/16mm but maybe you'll have better success than me.
The spare-no-expense approach to reducing 100 Hz ripple, is probably (the Tuba filter) . It's a pair of 5 ampere voltage regulators, one for the Left channel and one for the Right channel. Datasheet claimed ripple reduction at 100 Hz is 3100x (70 dB).
_
To reduce 100 Hz ripple, you could HOPE that increasing the filter capacitance on the Nch VFET PSU filter card, will not cause the Mean Well brick to enter "Hiccup mode" at startup. The Nch VFET PSU filter card is laid out with capacitor footprints 12.5mm in diameter, so you could search around to find caps bigger than 1000uF/50V with that diameter. The biggest I found this morning were 1300uF/50V/12.5mm but maybe you'll have better success than me.
The THESEUS-PSFILT card is laid out with capacitor footprints 16.0mm in diameter, so you could search around to find caps bigger than 1000uF/50V with that diameter. The biggest I could find this morning were 2200uF/50V/16mm but maybe you'll have better success than me.
The spare-no-expense approach to reducing 100 Hz ripple, is probably (the Tuba filter) . It's a pair of 5 ampere voltage regulators, one for the Left channel and one for the Right channel. Datasheet claimed ripple reduction at 100 Hz is 3100x (70 dB).
_
Attachments
Hi Mark,
thank you for your reply and your suggestions for even better filtering - I have no need to further reduce ripple on my VFET, but JoeLam might with his 107 dB/W drivers.
I put in the PO89ZB exactly for the switching noise, because I had good experiences with them at my ACP+, and several Raspberry / DAC builds. They seem to be very good in filtering out the switching noise.
Also, the PO89ZB brings additional 940 uF in filtering capacity, and then I have another 1000 uF Nichicon KW each at the B+ entry of my output boards ...
All that said, I will probably go for your Theseus filter in my upcoming Round 3 build, because I need to build one from scratch anyway - and I want to bring my 2SJ18 to use and try a P-Channel VFET ...
Regards, Claas
thank you for your reply and your suggestions for even better filtering - I have no need to further reduce ripple on my VFET, but JoeLam might with his 107 dB/W drivers.
I put in the PO89ZB exactly for the switching noise, because I had good experiences with them at my ACP+, and several Raspberry / DAC builds. They seem to be very good in filtering out the switching noise.
Also, the PO89ZB brings additional 940 uF in filtering capacity, and then I have another 1000 uF Nichicon KW each at the B+ entry of my output boards ...
All that said, I will probably go for your Theseus filter in my upcoming Round 3 build, because I need to build one from scratch anyway - and I want to bring my 2SJ18 to use and try a P-Channel VFET ...
Regards, Claas
Tuba is a medicine which treats the disease called FFT Mains Spike Neurosis.
Thanks for the feedback. Will start with passive supply filter first to see if regulator is necessary.
156 up and running!
finally finished it. I went with Mark’s Marauder Front End, instead of the standard Edcor version.
Biased up at 14V and it’s looking good on the oscilloscope.
Check out these square waves at ~1W into a 8ohm load. 20kHz looks great. I don’t see the 100khz ring as Papa pointed out in his Part 1 article. Seems the Marauder front end is cleaner than the Edcor. Even 200khz looks pretty good.
Gotta hook it up to speakers next.
Thank you Nelson and DIY Audio!
Cheers
finally finished it. I went with Mark’s Marauder Front End, instead of the standard Edcor version.
Biased up at 14V and it’s looking good on the oscilloscope.
Check out these square waves at ~1W into a 8ohm load. 20kHz looks great. I don’t see the 100khz ring as Papa pointed out in his Part 1 article. Seems the Marauder front end is cleaner than the Edcor. Even 200khz looks pretty good.
Gotta hook it up to speakers next.
Thank you Nelson and DIY Audio!
Cheers
Attachments
Congratulations @danskercher !! Your build is super clean and your scope results look terrific. Good on ya.
BTW a third transformerless front end has been unveiled: In addition to Marauder and Dreadnought, there's now Lexington.
BTW a third transformerless front end has been unveiled: In addition to Marauder and Dreadnought, there's now Lexington.
Does anyone have the dimensions for a drop-in replacement DPDT switch for the old switch to stop the thump? Having trouble getting exact measurements and sourcing.
Thanks!!
Thanks!!
I think I used this one. Sorry I can't be more definitive.
https://www.parts-express.com/DPDT-Rocker-Switch-060-310
https://www.parts-express.com/DPDT-Rocker-Switch-060-310
Grab a tool like (this one) and measure the hole in your panel. Who cares what size hole / what size switch fits in somebody else's panel? It's your panel that matters.
If you can't find a switch that fits exactly, grab a flat file and make a (slightly) bigger hole. Then buy a slightly bigger switch.
Easy.
..... and, with time, after years of struggle to have all newest shiny toys, when you became a member of Old Farts Club, you'll be able to do both the measuring and the filing, solely with Eyeballs
Age brings independence, lesser costs of tools
ZM, having Swiss Knives behind Spectacles
Age brings independence, lesser costs of tools
ZM, having Swiss Knives behind Spectacles