I have largely moved away from using SMPS power supplies for my ACA builds, but I'm still thinking about trying one again. This is partly due to my wanting to use some of my nice 300VA Antek toroids on other projects.
When I did my first set of ACA mods on the original kits, as powered by the 24V SMPS bricks, I changed C4 from the ubiquitous 10 uF Elna Silmic to a 100 uF Nichicon Aluminum Organic Polymer. This seems to help supply clean power to the front end JFet, Q4. I can't claim that it helps by eliminating SMPS switching noise, but perhaps by reducing the interaction between audio frequency noise on the local power rail (yes, even class A amps have this), and said switching noise. Anyway, it does seem to help.
For builds with higher voltage linear (or SMPS) supplies, I keep C4 at 100 uF, and add what I have designated R16. The value of R16 may be increased to protect Q4 from excess voltage, but it also helps supply filter noise at the Drain of Q4, whatever the source may be.
When I did my first set of ACA mods on the original kits, as powered by the 24V SMPS bricks, I changed C4 from the ubiquitous 10 uF Elna Silmic to a 100 uF Nichicon Aluminum Organic Polymer. This seems to help supply clean power to the front end JFet, Q4. I can't claim that it helps by eliminating SMPS switching noise, but perhaps by reducing the interaction between audio frequency noise on the local power rail (yes, even class A amps have this), and said switching noise. Anyway, it does seem to help.
For builds with higher voltage linear (or SMPS) supplies, I keep C4 at 100 uF, and add what I have designated R16. The value of R16 may be increased to protect Q4 from excess voltage, but it also helps supply filter noise at the Drain of Q4, whatever the source may be.
This is what want to see. I'll check interconnected equipment's ground. Thank you!
PS: How do you ground Meanwell itself? This is what I do: Meanwell chassis to safety ground (amp chassis), One output to signal ground, then connect chassis ground to signal ground (I tried this both with/without Thermister, and it seems no difference).
Last edited:
I use the switchmode bricks on the ACA.
I've used both earthed bricks and non earthed bricks.
I haven't opened them up but I would suspect on the non earthed switchmode supply the chassis of the supply is connected to the neutral wire on the other type the chassis is connected to earth and isolated from the neutral pin.
Instead of pulling your hair out, try both types of bricks and see if there is a difference.
This might help solve any possible grounding problems with how you have wired up your switcher.
If there is still no difference, then it might be the transformer radiating 60Hz hum or AC lines radiating 60Hz noise.
If the 60Hz noise is at -100dB, I feel it is likely radiated 60Hz noise and maybe not a grounding problem.
Try those things mentioned above, surely you can find some bricks even the 19V ones just for comparison.
I've used both earthed bricks and non earthed bricks.
I haven't opened them up but I would suspect on the non earthed switchmode supply the chassis of the supply is connected to the neutral wire on the other type the chassis is connected to earth and isolated from the neutral pin.
Instead of pulling your hair out, try both types of bricks and see if there is a difference.
This might help solve any possible grounding problems with how you have wired up your switcher.
If there is still no difference, then it might be the transformer radiating 60Hz hum or AC lines radiating 60Hz noise.
If the 60Hz noise is at -100dB, I feel it is likely radiated 60Hz noise and maybe not a grounding problem.
Try those things mentioned above, surely you can find some bricks even the 19V ones just for comparison.
This is what want to see. I'll check interconnected equipment's ground. Thank you!
PS: How do you ground Meanwell itself? This is what I do: Meanwell chassis to safety ground (amp chassis), One output to signal ground, then connect chassis ground to signal ground (I tried this both with/without Thermister, and it seems no difference).
This is how I would do it.
Meanwell chassis physically and electrically isolated from the amp chassis. Mains earth connected to the meanwell chassis.
Powersupply neutral is connected to the amp chassis via the standoffs on amp pcb.
On my ACA pcb, 0V is connected to the heatsink via pcb through the standoffs. The heatsink should then be electrically connected to the rest of the chassis.
Mains earth and PSU 0V are completely isolated (no connection).
I did find one potential issue on my amp chassis. I have the original monoblock chassis.
The small issue is that the chassis is painted steel, in order to get shielding of the whole chassis you need to remove paint on the chassis where the bolts connect the chassis to the heatsink.
So you could be either picking up radiated mains noise because of poor shielding of the chassis or it could be conducted via the interconnected equipment through interconnect cables then amplified through the amp.
My gut says if it's -100dB it probably isn't a ground loop problem but it's worth making everything that should be electrically connected nice and tight (also removing paint or using star washers to achieve this)
Last edited:
I agree with the above.
When using an SMPS (external brick or internal chassis mount), I leave the audio ground on the channel boards separate from the mains ground. This requires that the RCA input connectors are isolated from the chassis. If a connection needs to be made between the audio ground and the chassis, then this may be done with a CL-60 thermistor or a 10 Ohm resistor in parallel with back-to-back diodes. I've found that the GBPC35xx bridge rectifiers to be useful for this function, as they may be connected in a way that supports two channels, which remain isolated from each other.
I apply these same principles when using a linear power supply.
When using an SMPS (external brick or internal chassis mount), I leave the audio ground on the channel boards separate from the mains ground. This requires that the RCA input connectors are isolated from the chassis. If a connection needs to be made between the audio ground and the chassis, then this may be done with a CL-60 thermistor or a 10 Ohm resistor in parallel with back-to-back diodes. I've found that the GBPC35xx bridge rectifiers to be useful for this function, as they may be connected in a way that supports two channels, which remain isolated from each other.
I apply these same principles when using a linear power supply.
I'm going to build and ACA with my 10 year old son. Will be my first amp and obviously his too. Planning to make a small unassuming system at our getaway cabin, low level listening. My friend built these Fostex FE206En Back-Loaded Horn Loudspeaker Cabinets a few years ago, but has anyone tried paring them with the ACA?
amp and speaker project
Hi Andrew,
I have not heard the Fostex speakers in the horn enclosures, but
I use the amp camp with a couple open baffle 12 inch full range speakers, and I have plenty of volume using a unity gain buffer to control sound level.
The Fostex drivers are quite sensitive, and I am sure more so than my speakers, and the plans look quite doable if you have the woodworking tools and skills.
All I can say while you are waiting for someone with the Fostex speakers to chime in, is that it sounds like a Fantastic fun project for both you and your son,
and you will be making bonds and great memories for you both!
Best wishes to you and your son in your endeavor
Den
Hi Andrew,
I have not heard the Fostex speakers in the horn enclosures, but
I use the amp camp with a couple open baffle 12 inch full range speakers, and I have plenty of volume using a unity gain buffer to control sound level.
The Fostex drivers are quite sensitive, and I am sure more so than my speakers, and the plans look quite doable if you have the woodworking tools and skills.
All I can say while you are waiting for someone with the Fostex speakers to chime in, is that it sounds like a Fantastic fun project for both you and your son,
and you will be making bonds and great memories for you both!
Best wishes to you and your son in your endeavor
Den
I am running ACA with fullrange with great results, but my ACA is an small variation without feedback (current source amplifier) which is easy to implement since actually the circuit turn to less components. Is need to delete R12 feedback resistor, then delete R11 input resistor, and leave C3 as input but raise 10uF to only 3,3uF. That will extend bass and high frequencies, much in need using fullrange drivers.
Then would be nice add a passive filter to tame unbalanced response of this particular driver:
Accordingly to the article from Nelson Pass in their website firstwatt.com "current source amplifiers and fullrange drivers". Passive eq. values recommended for Fostex FE206N is a 47ohm resistor, and then parallel a 15uF in series to 4R7. That crossover is suitable only in case you succeed in current source modification (without feedback).
Then would be nice add a passive filter to tame unbalanced response of this particular driver:
Accordingly to the article from Nelson Pass in their website firstwatt.com "current source amplifiers and fullrange drivers". Passive eq. values recommended for Fostex FE206N is a 47ohm resistor, and then parallel a 15uF in series to 4R7. That crossover is suitable only in case you succeed in current source modification (without feedback).
Hi Andrew,
I have not heard the Fostex speakers in the horn enclosures, but
I use the amp camp with a couple open baffle 12 inch full range speakers, and I have plenty of volume using a unity gain buffer to control sound level.
The Fostex drivers are quite sensitive, and I am sure more so than my speakers, and the plans look quite doable if you have the woodworking tools and skills.
All I can say while you are waiting for someone with the Fostex speakers to chime in, is that it sounds like a Fantastic fun project for both you and your son,
and you will be making bonds and great memories for you both!
Best wishes to you and your son in your endeavor
Den
Thank you for the info! Yeah I agree, its a nice idea I think, I've been wanting to try something like this and with it being friendly on the costs I think its a perfect idea.
I'm not totally set on the Fostex either, but seemed like a kind of smart idea since its part of this forum which asking for advice nice. My friend built a few pairs of them for friends a while back, not sure its the exact same plan but looks like, he also recently added a Fostex supertweeter on top but I haven't been round to hear that yet.
Fostex 206 with back loaded horn design Cabinet
That was the first full range speaker that I built when I bought my first amp camp amp version 1.1. With 19 V power supply I believe was rated at 5 W of power ? Then I later discovered the power supply I built was actually only putting out 15.5 V supply.
Sounds beautiful one thing I can definitely tell you with those speakers and their efficiency you will probably never use more than one watt. They will be too loud for a normal listening
That was the first full range speaker that I built when I bought my first amp camp amp version 1.1. With 19 V power supply I believe was rated at 5 W of power ? Then I later discovered the power supply I built was actually only putting out 15.5 V supply.
Sounds beautiful one thing I can definitely tell you with those speakers and their efficiency you will probably never use more than one watt. They will be too loud for a normal listening
I have Jericho Horns as my main speakers. They use the old Fostex FE208 Sigma (not EZ), but have frequently been built with FE206 / FE206En as well after the 208Sigmas were not available anymore.
My 12 year old sons and myself have built an ACA as well, it was a great project for us - still need to do a write-up of that as well ...
Long story short, the Fostex 8 inchers in backloaded Horns sound great with the ACA, and with enough volume as well.
The ACA is now playing in the boys' basement music / TV / "recording studio" room on a pair of Fostex FE126 bass reflex bookshelf fullrange speakers (Fostex Recommended Enclosure (bass reflex) for these speakers). It sounds great as well, and easily enough volume for the 18 square meters room.
Regards, Claas
My 12 year old sons and myself have built an ACA as well, it was a great project for us
- still need to do a write-up of that as well ... Long story short, the Fostex 8 inchers in backloaded Horns sound great with the ACA, and with enough volume as well.
The ACA is now playing in the boys' basement music / TV / "recording studio" room on a pair of Fostex FE126 bass reflex bookshelf fullrange speakers (Fostex Recommended Enclosure (bass reflex) for these speakers). It sounds great as well, and easily enough volume for the 18 square meters room.
Regards, Claas
ACA with CREE C3M0160120D
Still cooking these CREE C3M0160120D silicon parts, but sounding promising. I am using them for both Q1 & Q2. Very clear sound, still working on getting a way of describing the difference from the IRFP044's that I had previously. The bass seems to be less than the 044's, but well articulated. The different sounds seem to have greater separation than before, no mushing of the individual component sound boundaries if that makes any sense? Mid-range and highs crystalline without harshness. Have lowered my voltage rails to 24.4 volts from the previous 29+.
I have the bias on the drain of Q1 set to 13.3 volts. Presently I am using 500 ohms for R5 & R6, but may lower them to 300 ohms to see if it remains stable. The amp biases much easier with the silicon parts than it did with the 044's, less jumpy on the meter.
The amp gain is less than it was with the 044's, maybe lost a quarter of the gain, but I noticed this loss even on the 29+ rails, so it a factor apart from the lower running voltage. I have less feedback, using a 332K resistor for R12.
Will see how the sound changes as the parts get broken in, and let you know
Still cooking these CREE C3M0160120D silicon parts, but sounding promising. I am using them for both Q1 & Q2. Very clear sound, still working on getting a way of describing the difference from the IRFP044's that I had previously. The bass seems to be less than the 044's, but well articulated. The different sounds seem to have greater separation than before, no mushing of the individual component sound boundaries if that makes any sense? Mid-range and highs crystalline without harshness. Have lowered my voltage rails to 24.4 volts from the previous 29+.
I have the bias on the drain of Q1 set to 13.3 volts. Presently I am using 500 ohms for R5 & R6, but may lower them to 300 ohms to see if it remains stable. The amp biases much easier with the silicon parts than it did with the 044's, less jumpy on the meter.
The amp gain is less than it was with the 044's, maybe lost a quarter of the gain, but I noticed this loss even on the 29+ rails, so it a factor apart from the lower running voltage. I have less feedback, using a 332K resistor for R12.
Will see how the sound changes as the parts get broken in, and let you know
I have Jericho Horns as my main speakers. They use the old Fostex FE208 Sigma (not EZ), but have frequently been built with FE206 / FE206En as well after the 208Sigmas were not available anymore.
My 12 year old sons and myself have built an ACA as well, it was a great project for us
Long story short, the Fostex 8 inchers in backloaded Horns sound great with the ACA, and with enough volume as well.
The ACA is now playing in the boys' basement music / TV / "recording studio" room on a pair of Fostex FE126 bass reflex bookshelf fullrange speakers (Fostex Recommended Enclosure (bass reflex) for these speakers). It sounds great as well, and easily enough volume for the 18 square meters room.
Regards, Claas
Great to know! Looks like project #2 is sorted then!
re-visit of CREE C3M0160120D
Still cooking these CREE C3M0160120D silicon parts, but sounding promising. I am using them for both Q1 & Q2. Very clear sound, still working on getting a way of describing the difference from the IRFP044's that I had previously. The bass seems to be less than the 044's, but well articulated. The different sounds seem to have greater separation than before, no mushing of the individual component sound boundaries if that makes any sense? Mid-range and highs crystalline without harshness. Have lowered my voltage rails to 24.4 volts from the previous 29+.
Final combo is CREE for Q1 and using FQH44N10 for Q2. Have 300 ohm gate stoppers.
Running 24.4 volts on rail, 13.4 volts on Q1 drain. Sound retaining the crystal clarity that I had with the CREE in both Q 1 & 2, but the 44N10 has restored lower spectrum to a level that I am happy with. The fact that I am using compensated open baffle speakers, that tend to roll off the lows somewhat
prompted me to try the 44N10 substitution for Q 2. Otherwise, I might have been content with the CREE's in both positions, so your setup may be to your liking with the CREE in both places. The CREE does have a clear and relaxed sound, and great resolution. There you have it. I believe that I will rest my solder iron on this project, and say that I will listen for a while, like I have said before! Maybe this time I will
Great to hear about your success! I was also considering the FQH44N10 for the Q2 position. I am planning to build a pair of ACA monoblocks with two boards wired in parallel for each channel. I think that the CREE silicon carbide parts will work well in tandem. I'll probably go for about 26V on the local power rail, after some RC filtering from an internal chassis mount SMPS.
I have liked the IRFQ044's also, but decided to go ahead with the 44N10's since they had a bit more Forward Transconductance, and thought it would help the low end of the CREE. Also the 044's are only 55 volt parts, not a killer, but the 100 volt rating on the 44N10 and higher dissipation means less long term stress. I think it is a nice compromise, and blend. I might up the rail a bit, but will need to get a smaller power resistor to replace the ones I have. Presently have a 1 ohm 50 watt job in the first part of the power supply, and another 1.3 ohm one before the final two 33,000uf filter caps, one for each channel. Probably a .5 ohm replacement would raise the voltage back in the 26 volt area.
Really glad you mentioned that CREE to try. I had been looking at Newark, and they did not carry that one, but Mouser did. There was another one that looked interesting that had higher capacitance in and out, but it was listed as only being available later on. Who wants to wait when the iron is up to temp
Here is to your waiting projects success
Really glad you mentioned that CREE to try. I had been looking at Newark, and they did not carry that one, but Mouser did. There was another one that looked interesting that had higher capacitance in and out, but it was listed as only being available later on. Who wants to wait when the iron is up to temp
Here is to your waiting projects success
