I modded and listen to a Sure TK2050 board for a while. The tank caps on the board was replaced with 2x2200uF ROE gold caps with 2.2uF film cap in parallel. The SMPS was a 19v 2.4A Dell Laptop power supply and I knew from day one it is underrated. I have also added another 15000uF stiffener cap external to the board. It was OK overall, only thing I wanted more is bass due to the weak laptop supply.
I finally did the last thing on the list, which is to put a much bigger SMPS. Got a 24V 10A SMPS from the amazon. To my big surprise, first thing I noticed after the PS swap was obvious stronger bass with great low frequency extension. But the Tripath magic was gone. Resolution, depth and sound stage somehow dropped. The amp sounded like some kind of low distortion class AB transistor amp. I did read the threads about preference on stiffener caps or not so I yanked the 15000uF cap. The magic came back but the bass extension has also been dropped back to the previous level similar to with the laptop supply.
I'd wonder the stiffener cap is not the cause of the problem but, rather, a combination of the power supply's internal resistance, the ESR of the stiffener cap, tank caps on the board and even the circuit board layout on the power trunk.
Any thoughts ? ideas ?
I finally did the last thing on the list, which is to put a much bigger SMPS. Got a 24V 10A SMPS from the amazon. To my big surprise, first thing I noticed after the PS swap was obvious stronger bass with great low frequency extension. But the Tripath magic was gone. Resolution, depth and sound stage somehow dropped. The amp sounded like some kind of low distortion class AB transistor amp. I did read the threads about preference on stiffener caps or not so I yanked the 15000uF cap. The magic came back but the bass extension has also been dropped back to the previous level similar to with the laptop supply.
I'd wonder the stiffener cap is not the cause of the problem but, rather, a combination of the power supply's internal resistance, the ESR of the stiffener cap, tank caps on the board and even the circuit board layout on the power trunk.
Any thoughts ? ideas ?
Please don't randomly connect Capacitors to the SMPS. It involves lots of calculations. Also the laptop charger you used is designed totally different, that can not cater the need of audio amplifier. I would request you to read the basic of SMPS designs.
You haven't mentioned what speakers are you using, ie single driver or two way or three way. with increase in low frequency two way and single driver speakers tent to lose the details on sound.
To understand this I would say. Use car battery in series to generate 24 V (for testing) . Use thick power cable. and keep it short as possible.
You haven't mentioned what speakers are you using, ie single driver or two way or three way. with increase in low frequency two way and single driver speakers tent to lose the details on sound.
To understand this I would say. Use car battery in series to generate 24 V (for testing) . Use thick power cable. and keep it short as possible.
I reckon the bigger SMPSU is a stronger source of common-mode noise. Sure, its giving you stiffer rails but the CM noise is polluting your audio. You need some way to break the noise loop - perhaps an input trafo or a much more effective CM choke on the mains input.
Yes to using the car battery, a great solution. But not in conjunction with a SMPSU, on its own. Isolating the amp supply from the mains will clean up the noise issues considerably.
Yes to using the car battery, a great solution. But not in conjunction with a SMPSU, on its own. Isolating the amp supply from the mains will clean up the noise issues considerably.
Sorry, didn't put it correctly,
I meant use two car batteries in series with each other to get 24V. Not to connect it to SMPS when testing with battery.
Moreover if the SMPS is regulated. it has its own correction oscillations, that varies with load. although in most cases it hardly matters.
I meant use two car batteries in series with each other to get 24V. Not to connect it to SMPS when testing with battery.
Moreover if the SMPS is regulated. it has its own correction oscillations, that varies with load. although in most cases it hardly matters.
Last edited:
Thanks for all ideas.
I'm well awared of the SMPS design and the effects of adding a stiffener cap can throw off the PS. But I tend to agree that the increased in bass has degrade the details in sound. I'm well awared of the insufficient of the power rating of the laptop supply that's why I'm saying I just changed it out.
I like the idea of increasing the filtering mechanism from SMPS to the amp board, and try out batteries. I used to run a TA2024 off a car battery but I didn't notice much difference with using SMPS. I may be getting a toroid tranny to build a linear supply too.
I do want to try out a meanwell SMPS since for some reasons they are used for class D amps a lot.
I'm well awared of the SMPS design and the effects of adding a stiffener cap can throw off the PS. But I tend to agree that the increased in bass has degrade the details in sound. I'm well awared of the insufficient of the power rating of the laptop supply that's why I'm saying I just changed it out.
I like the idea of increasing the filtering mechanism from SMPS to the amp board, and try out batteries. I used to run a TA2024 off a car battery but I didn't notice much difference with using SMPS. I may be getting a toroid tranny to build a linear supply too.
I do want to try out a meanwell SMPS since for some reasons they are used for class D amps a lot.
I had a little bit of time yesterday so I played around in between the SMPS and the amp board. I think I might have identified the issue of being too much EMI or noise coming from the no name SMPS. The smaller laptop SMPS seems giving less noise therefore not affecting the sound of the amp as much as the bigger and stronger SMPS.
I did two things that seems to have bring back the magic:-
1) I added a 2.2uF file cap between v+/v- right at the output of the SMPS.
2) I twisted together the + and - wires going from the SMPS to the amp board. This pair of wires is just about 6 inches or so.
I keep the 15000uF stiffener cap at the location where the twisted power wires connects to the amp board.
It seems like filtering is the key here. Next I will try to add a C-L-C pi filtering network coming out of the SMPS to the amp.....
I did two things that seems to have bring back the magic:-
1) I added a 2.2uF file cap between v+/v- right at the output of the SMPS.
2) I twisted together the + and - wires going from the SMPS to the amp board. This pair of wires is just about 6 inches or so.
I keep the 15000uF stiffener cap at the location where the twisted power wires connects to the amp board.
It seems like filtering is the key here. Next I will try to add a C-L-C pi filtering network coming out of the SMPS to the amp.....
Ok... Try this instead because at least half of the options will screw up some other aspect to good sound.
Use more, smaller power caps.2x2200uf isn't sufficient. Whats the wattage? Try like 9000-10,000uf of high quality fast low ESR caps, but several. Like 6-10 caps. If you go over 10,000uf it's no big deal. Skip the big crappy 15kuf cap. Don't use the 2.2uf film. That's for 15w a channel or so, if you've got higher up it to around the same ratio.
Now for my latest trick... connect 6-10 low noise diodes across V+ and ground (you have V+ and ground, not V-, by the way). The should be in parrallel, and cathode to the + DC.
Don't use inductors, don't use big slow high ESR caps, don't use small caps that cause high frequency ringing, etc etc etc.
What makes class D sound good.
1. Fast, immediate current creates refinement, good high frequencies, timber, etc.
2. Constant voltage, so you have to use what looks like an excessive amount of caps, but it isn't for D. This is what gives you bass and even more importantly impressive body to all the sounds (insturments and voices). No I don't recommend a voltage regulator yet (maybe some day I'll figure how one can work right with class D).
3. As low noise possible on ground that can feed into the unit itself, or rather any sort of common-mode/ac/emi/rf/whatever, hence the use of making a dividing network for all non-forward noise on the DC. Noise on the positive isn't actually as destructive by comparrison (largely depending on frequency).
4. Power filtration before your SMPS to reduce the noise it puts out, and also reduce the noise it feeds back into the rest of your system. A few X rated caps is good too. There is a design refered to as the "Felix" that works well.
Use more, smaller power caps.2x2200uf isn't sufficient. Whats the wattage? Try like 9000-10,000uf of high quality fast low ESR caps, but several. Like 6-10 caps. If you go over 10,000uf it's no big deal. Skip the big crappy 15kuf cap. Don't use the 2.2uf film. That's for 15w a channel or so, if you've got higher up it to around the same ratio.
Now for my latest trick... connect 6-10 low noise diodes across V+ and ground (you have V+ and ground, not V-, by the way). The should be in parrallel, and cathode to the + DC.
Don't use inductors, don't use big slow high ESR caps, don't use small caps that cause high frequency ringing, etc etc etc.
What makes class D sound good.
1. Fast, immediate current creates refinement, good high frequencies, timber, etc.
2. Constant voltage, so you have to use what looks like an excessive amount of caps, but it isn't for D. This is what gives you bass and even more importantly impressive body to all the sounds (insturments and voices). No I don't recommend a voltage regulator yet (maybe some day I'll figure how one can work right with class D).
3. As low noise possible on ground that can feed into the unit itself, or rather any sort of common-mode/ac/emi/rf/whatever, hence the use of making a dividing network for all non-forward noise on the DC. Noise on the positive isn't actually as destructive by comparrison (largely depending on frequency).
4. Power filtration before your SMPS to reduce the noise it puts out, and also reduce the noise it feeds back into the rest of your system. A few X rated caps is good too. There is a design refered to as the "Felix" that works well.
Member
Joined 2009
Paid Member
Interesting. Is it the reverse bias capacitance of the diodes making them act like snubbers ???
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.