I am working on a run of 200 Warp-1’s (which are ready to run with PSU and chassis 
). I wanted to make more amp modules but the shortage of parts has been making it very hard. You can’t imagine the level of effort required to get the parts for the last batch. Even with all the work we put into it, the DCDC buck converter is no longer available so the amp will use external 15v from the main SMPS via a linear regulator and 2 pin Molex KK.
If you were to build an amp yourself with a nice chassis, 600w PSU, BTSB panel mount, connectors, and amp module will probably cost you more than $1k. But the joy of DIY is priceless.
). I wanted to make more amp modules but the shortage of parts has been making it very hard. You can’t imagine the level of effort required to get the parts for the last batch. Even with all the work we put into it, the DCDC buck converter is no longer available so the amp will use external 15v from the main SMPS via a linear regulator and 2 pin Molex KK.If you were to build an amp yourself with a nice chassis, 600w PSU, BTSB panel mount, connectors, and amp module will probably cost you more than $1k. But the joy of DIY is priceless.
Last edited:
- Hello, I recently read a very good article about the advantages of using good Linear Power Supplies in Class D Amplifiers and I was wondering if anyone has tried a good quality Linear Power Supply on your TPA3255 Reference Amplifier?
- I realize they are heavier and run warmer, and can be more expensive but I read an interesting article that said that a well designed linear power supply could reward you with :
- A lower noise floor
- Tighter outlines and a more fleshed out sound
- A much smoother, refined, and gradational sound
- A more natural soundstage in both size and atmosphere
- Improved timbre
- Here is a link if you want to check out the article :
https://audiobacon.net/2018/06/29/linear-power-supplies-for-audiophiles-getting-closer-to-live/
- I realize they are heavier and run warmer, and can be more expensive but I read an interesting article that said that a well designed linear power supply could reward you with :
- A lower noise floor
- Tighter outlines and a more fleshed out sound
- A much smoother, refined, and gradational sound
- A more natural soundstage in both size and atmosphere
- Improved timbre
- Here is a link if you want to check out the article :
https://audiobacon.net/2018/06/29/linear-power-supplies-for-audiophiles-getting-closer-to-live/
Hi hallcon83,
I think Vunce has tried both and hope he can chime in.
Some of the generalizations about SMPS not having as good transient response might be true if there is a smaller bulk capacitance reserve between the PSU and the amp. With the Micro-Audio SMPS, they use a topology that permits large value cap bank in between the PSU and the amp. I think up to 40,000uF can be used and that should give quite a stiff rail for better bass authority. This would be true for Class AB as well as Class D amps.
SMPS may have more RFI emission but it would be out of the audio band and a lot less in the mains 50/60Hz band which are very audible as “hum”.
Here is an FFT from my amp powered by a low cost Chinese LED lighting SMPS set to 48v and the amp is putting out 20w into 3.3ohm load. Note the absence of any mains hum at 60Hz/120Hz/180Hz. The noise floor is about 1uV rms. Inaudible with no music playing and ear pressed to the speaker cone.
I have not been able to achieve this low of a noise floor with a linear supply, even with a shunt regulator or cap multiplier. If we think about it, it’s very hard to quash a 0dB 120vac signal at 60Hz by -120dB with any filtering.
I think Vunce has tried both and hope he can chime in.
Some of the generalizations about SMPS not having as good transient response might be true if there is a smaller bulk capacitance reserve between the PSU and the amp. With the Micro-Audio SMPS, they use a topology that permits large value cap bank in between the PSU and the amp. I think up to 40,000uF can be used and that should give quite a stiff rail for better bass authority. This would be true for Class AB as well as Class D amps.
SMPS may have more RFI emission but it would be out of the audio band and a lot less in the mains 50/60Hz band which are very audible as “hum”.
Here is an FFT from my amp powered by a low cost Chinese LED lighting SMPS set to 48v and the amp is putting out 20w into 3.3ohm load. Note the absence of any mains hum at 60Hz/120Hz/180Hz. The noise floor is about 1uV rms. Inaudible with no music playing and ear pressed to the speaker cone.
I have not been able to achieve this low of a noise floor with a linear supply, even with a shunt regulator or cap multiplier. If we think about it, it’s very hard to quash a 0dB 120vac signal at 60Hz by -120dB with any filtering.
Last edited:
Hi Hallcon83,
Yep, I did use a linear supply to power my TPA3255.
-400va trafo with 2 x 37v secondaries,
-LT4320 ideal bridge,
-Pair of 22,000uF Mundorf M-Lytic HC caps.
The amp was dead silent, couldn’t hear anything at the speakers with music paused and sounded wonderful. Much better than I anticipated for a Class D amp.
I had very good results using Cresnet/MicroAudio SMPS units in the past and was curious about how the TPA3255 would be powered by one. So I swapped out the linear supply for a MicroAudio SMPS630-44v unit. To my ears, just as good as the linear psu. My setup added 20,000uF of capacitance at the SMPS output. Sami’s power supplies can run with an incredible amount of added capacitance without ever a problem. Most smps have a problem and enter ‘hiccup’ mode, and shut down.
Since sound quality was essentially the same the reason I kept the MicroAudio unit as the power supply came down to the creature comforts and auxiliary features like softstart, able to use a low voltage pushbutton power switch, 12v aux power for BTSB board that handles SE—>Bal conversion and gain adjustment and a smaller overall package.
Yep, I did use a linear supply to power my TPA3255.
-400va trafo with 2 x 37v secondaries,
-LT4320 ideal bridge,
-Pair of 22,000uF Mundorf M-Lytic HC caps.
The amp was dead silent, couldn’t hear anything at the speakers with music paused and sounded wonderful. Much better than I anticipated for a Class D amp.
I had very good results using Cresnet/MicroAudio SMPS units in the past and was curious about how the TPA3255 would be powered by one. So I swapped out the linear supply for a MicroAudio SMPS630-44v unit. To my ears, just as good as the linear psu. My setup added 20,000uF of capacitance at the SMPS output. Sami’s power supplies can run with an incredible amount of added capacitance without ever a problem. Most smps have a problem and enter ‘hiccup’ mode, and shut down.
Since sound quality was essentially the same the reason I kept the MicroAudio unit as the power supply came down to the creature comforts and auxiliary features like softstart, able to use a low voltage pushbutton power switch, 12v aux power for BTSB board that handles SE—>Bal conversion and gain adjustment and a smaller overall package.
Attachments
Last edited:
Sorry if this is there in some obvious place and I missed it, but is there a latest updated BoM and Mouser list for the amp? I haven't read each post in the thread, I admit, but I poked around a fair bit. How much is the parts cost for one PCB, excluding the PSU etc?
Is there a list of recommended PSU modules one can use to power one board?
Second: if I use this amp, do I need additional circuits for turn-on delay and speaker DC detect? I read in one of the posts here that if the amp is used in BTL stereo mode or PBTL mode, then such detection is done by the amp chip itself. Is that good enough? What would you guys do?
Is there a list of recommended PSU modules one can use to power one board?
Second: if I use this amp, do I need additional circuits for turn-on delay and speaker DC detect? I read in one of the posts here that if the amp is used in BTL stereo mode or PBTL mode, then such detection is done by the amp chip itself. Is that good enough? What would you guys do?
Last edited:
I’ll send you the final schematics and BOM once you order the PCB. The latest has some small but critical changes. Or PM me and I’ll get it to you in advance. Cost of BOM was $125 a year ago. There are several parts that are hard to get. The DCDC buck converter being one. I would power the 15v from an external supply from the SMPS.
No need for DC protect or turn on delay other than to make sure your balanced buffer opamps are on first so you don’t get thump. Same goes for not hot plugging. A large thump will trigger auto self protect from clip or DC.
The PSU has been discussed in thread but briefly, my suggestion is to use Micro-Audio SMPS630SO set for 51 to 53vdc. It has all the needed delay logic and remote turn on and auxilary 12v or +/-15v that you might need.
Alternatively, you can use 48v SMPS supplies from China designed for LED lighting. Circa $70 but they have a noisy fan you need to replace with a Noctua or else intolerable noise. Or defeat the built in fan as it’s not needed at normal music playing levels. Get the 800w version so it’s not working hard.
On AliExpress search for “800w 48v SMPS LED”. Lately I noticed 800w is sold out. You should get 1000w. About $57.
My recommendation is still Micro-Audio though for ease of use and extra features. You get all the added features like extra PSU’s included. Sami will build you a custom unit with the exact voltages and auxiliary supplies you need. (I don’t get anything from Micro-Audio for recommending them - just the best product for this use as I have bought hundreds of units, IMO). Another very good one if you plan on powering a multi channel system is the 1000w SMPS1KPFC.
https://micro-audio.com/store/product/smps630-g/
No need for DC protect or turn on delay other than to make sure your balanced buffer opamps are on first so you don’t get thump. Same goes for not hot plugging. A large thump will trigger auto self protect from clip or DC.
The PSU has been discussed in thread but briefly, my suggestion is to use Micro-Audio SMPS630SO set for 51 to 53vdc. It has all the needed delay logic and remote turn on and auxilary 12v or +/-15v that you might need.
Alternatively, you can use 48v SMPS supplies from China designed for LED lighting. Circa $70 but they have a noisy fan you need to replace with a Noctua or else intolerable noise. Or defeat the built in fan as it’s not needed at normal music playing levels. Get the 800w version so it’s not working hard.
On AliExpress search for “800w 48v SMPS LED”. Lately I noticed 800w is sold out. You should get 1000w. About $57.
My recommendation is still Micro-Audio though for ease of use and extra features. You get all the added features like extra PSU’s included. Sami will build you a custom unit with the exact voltages and auxiliary supplies you need. (I don’t get anything from Micro-Audio for recommending them - just the best product for this use as I have bought hundreds of units, IMO). Another very good one if you plan on powering a multi channel system is the 1000w SMPS1KPFC.
https://micro-audio.com/store/product/smps630-g/
Last edited:
Since I've already ordered a PCB and it's already been delivered to Colorado, I guess the schematics and BOM are on their way?
Wow. $125. That's really amazing value. You're almost competing against China at that price.
I remember your first edition of your RTR solid-state speaker protect module too -- unbelievable value at $40.Thanks for all these answers. Yes, I too was seriously thinking that there's no point building an array of these amps if I'm going to scrimp on the PSU, and I was thinking Micro-Audio or Hypex. (I am thinking of building a pair of 4-driver active speakers, with 4 channels of your amps inside each speaker. Two of your PCBs and one PSU.)
Isn't the Hypex a good PSU? Not that there is a very strong reason for me to insist on Hypex. Just asking.
Last edited:
I have not used Hypex so can’t say. I know Micro-Audio is one of few smps that can drive a huge bulk cap bank without going into hiccup mode. Plus it is very quiet. Quieter than Connex. I’ll send you BOM and schematic via email soon. Sorry, I forgot that you ordered and received the boards already. Most people will ping me and ask for the BOM.
Chinese sell fully built TPA3255 for $70. Their BOM must be circa $30 to do that. The 4 coil craft inductors alone on my BOM blow that price already.
But you are paying for the key part that allows low distortion at higher power plus the fact that the inductors are rated at 27A means that they will be superior for bass authority.
Chinese sell fully built TPA3255 for $70. Their BOM must be circa $30 to do that. The 4 coil craft inductors alone on my BOM blow that price already.
But you are paying for the key part that allows low distortion at higher power plus the fact that the inductors are rated at 27A means that they will be superior for bass authority.
Hi Folks,
For those of you who are building the TPA3255, you may have some trouble finding a critical part - the LM5010ASD DC/DC buck converter chip is very hard to find. My temporary solution is to use the 15v from an external PSU (if you are using the MicroAudio SMSP630SO you are in luck). Tie the external 15vdc supply with a flying lead to the W141 solder bridge that isolates the 15v supply from the 12v LDO. Add a 2200uF bulk rail cap to your 15vdc supply to make sure it stays on longer than the time the TPA3255 takes to power down to avoid turn off thump. Any decent 15v >50mA supply can work. There is a 12v LDO after this that can drop the voltage to the required 12v for the analiog circuits in the amp.
For those of you who are building the TPA3255, you may have some trouble finding a critical part - the LM5010ASD DC/DC buck converter chip is very hard to find. My temporary solution is to use the 15v from an external PSU (if you are using the MicroAudio SMSP630SO you are in luck). Tie the external 15vdc supply with a flying lead to the W141 solder bridge that isolates the 15v supply from the 12v LDO. Add a 2200uF bulk rail cap to your 15vdc supply to make sure it stays on longer than the time the TPA3255 takes to power down to avoid turn off thump. Any decent 15v >50mA supply can work. There is a 12v LDO after this that can drop the voltage to the required 12v for the analiog circuits in the amp.
You may be interested to hear about a subjective comparison shootout between the TPA3255 as implemented in the LSA Warp-1 and the Benchmark AHB-2, several Audiophonics Purifi Class D amps, and a DIY Tripath amp by member RichiNL here:
https://www.diyaudio.com/community/...ign-class-d-amp-with-pffb.372379/post-7004767
https://www.diyaudio.com/community/...ign-class-d-amp-with-pffb.372379/post-7004767
Congrats X! Pretty impressive for all your efforts. I may have get a pair of your 3255 amp boards if you ever have another GB.
Pete
Pete
After buying a TPA3255 board back on January 2020... then getting unexpectedly laid off, selling my house and moving I am FINALLY building my amp. I did my first attempted power up today and all is working well with the exception of OUTB. It just sits at +20 V (Vcc = 45 V) no matter what the input does. I'm driving it with a BTSB board driven by a function generator on the unbalanced input, and I have checked that the input signal is getting onto INB. I swapped the input cables, both work fine on OUTC and OUTD, and OUTA works with both of them, but OUTB is always a straight DC. Anybody got any insights? Thanks in advance.
Hi Tommost,
Sorry for your job and stuff. You issues are odd - I have not heard of this behavior ever with all the amps out there over 170 amps in the wild thus far so that’s a strange one. So you are driving with balanced output from BTSB and A, C, D show responsive outputs but B just sits at 20v?
Check the output pin of 5532 opamp driving Ch B to see if it is output correct waveform.
Lack of output on a single channel of the TPA3255 is a new one. I can only think that maybe there is a bad solder joint, a bad or broken leg on the 10uF Silmic coupling caps, or a bad bootstrap cap, or worst case, a bad TPA3255 chip. Those usually fail all channels and not just 1.
If you don’t have luck debugging it, contact me via Etsy messaging on my shop or a PM and I’ll see if we can work out a repair for you.
Thanks,
X
Sorry for your job and stuff. You issues are odd - I have not heard of this behavior ever with all the amps out there over 170 amps in the wild thus far so that’s a strange one. So you are driving with balanced output from BTSB and A, C, D show responsive outputs but B just sits at 20v?
Check the output pin of 5532 opamp driving Ch B to see if it is output correct waveform.
Lack of output on a single channel of the TPA3255 is a new one. I can only think that maybe there is a bad solder joint, a bad or broken leg on the 10uF Silmic coupling caps, or a bad bootstrap cap, or worst case, a bad TPA3255 chip. Those usually fail all channels and not just 1.
If you don’t have luck debugging it, contact me via Etsy messaging on my shop or a PM and I’ll see if we can work out a repair for you.
Thanks,
X
I hung it up for the night, time to watch Formula 1 qualifying. I'll check tomorrow.
Well I took the heat sink off, probed the traces for continuity and carefully examined all the traces and joints, both channels looked fine. So I powered the amp with no heat sink with the supply turned down to 25 V and 100 ohm load to keep the thermals manageable. I traced the signal through the op amps, all four channels were fine. So I looked at the speaker outputs and all four were working! Hmmm.... I wonder what could be going on here. So I put the heat sink back on and it still works.
So this bothers me a bit but I guess could have just done something dumb. Oh well time to start assembling the chassis and laying out the front and back panels.
Here's my preliminary layout in the tray for the Pesante 3U chassis. I'm using a CZH Labs soft start module and bipolar cap bank with the minus and plus sides joined together, and an Antek PS-5N48R12 linear supply. Together they give me about 70K uF of capacitance, I don't have 2 load banks to test both channels at once but driving one channel to clipping I only saw about 100 mV pk-pk ripple on the 48 V supply. I don't love the fact that the mains transformer is in the front but this will allow me to keep the audio leads as short as possible so I think this is what I'm going with.