Nice work on debugging - I think maybe your heatsink spreader may be touching one of the SMT caps? Are you using one of the machined ones with a mesa to keep the block away from the nearby parts?
Nice build, btw. A big linear PSU like this will give you some delicious bass. This amp with a MicroAudio SMPS630SO has already been credited with crushing other amps in the bass department. Most of it is having a good PSU and thick wiring and the board layout with double traces and 27A capable output inductors. Having a linear supply with a massive cap bank like yours will sound amazing.
Nice build, btw. A big linear PSU like this will give you some delicious bass. This amp with a MicroAudio SMPS630SO has already been credited with crushing other amps in the bass department. Most of it is having a good PSU and thick wiring and the board layout with double traces and 27A capable output inductors. Having a linear supply with a massive cap bank like yours will sound amazing.
Is it really debugging if just taking it apart and putting it back together fixes it? I'd love to know the root cause but for now I'm preceding on the layout.
I don't think my heatsink was touching anything. I'm using one of your heat transfer spaces along with a heat sink I had in my stash from who knows where. I didn't have a proper standoff to mount it but I found that 5 of the M3 washers from this kit were exactly correct along with a 1 mm thermal pad, just squishing it almost to the point of the heatsink touching the spacer. The touchy part was how to assemble the heatsink and spacer assembly to the board while keeping the stack of 5 washers in place. I ended up stacking the 5 washers on one of the longest screws in the referenced kit, then screwing it into the heatsink assembly. I used a small bit of tape to hold the washers in place. Then I took the screws out and used them to carefully screw through the washer stack and into the tapped hole in the heat spreader, then removed the tape. Of course these screws were much too long, but I could push the heatsink assembly flush against the TPA3255 and carefully take one of the long screws out while the other one held things together. While holding everything as steady as possible, using my 3rd hand I threaded an 8 mm long M3 screw into the board and the heat spreader.
I'm intending to use this amp to replace my NAD 5400 I used to drive my desktop computer Mini Maggies and DWM. The combination extends down to about 50 Hz, below that there's an SVS sub, so I may not get all the benefit of using all this cap reservoir but it was fun working on it!
I don't think my heatsink was touching anything. I'm using one of your heat transfer spaces along with a heat sink I had in my stash from who knows where. I didn't have a proper standoff to mount it but I found that 5 of the M3 washers from this kit were exactly correct along with a 1 mm thermal pad, just squishing it almost to the point of the heatsink touching the spacer. The touchy part was how to assemble the heatsink and spacer assembly to the board while keeping the stack of 5 washers in place. I ended up stacking the 5 washers on one of the longest screws in the referenced kit, then screwing it into the heatsink assembly. I used a small bit of tape to hold the washers in place. Then I took the screws out and used them to carefully screw through the washer stack and into the tapped hole in the heat spreader, then removed the tape. Of course these screws were much too long, but I could push the heatsink assembly flush against the TPA3255 and carefully take one of the long screws out while the other one held things together. While holding everything as steady as possible, using my 3rd hand I threaded an 8 mm long M3 screw into the board and the heat spreader.
I'm intending to use this amp to replace my NAD 5400 I used to drive my desktop computer Mini Maggies and DWM. The combination extends down to about 50 Hz, below that there's an SVS sub, so I may not get all the benefit of using all this cap reservoir but it was fun working on it!
I think you made it more difficult than it needed to be. Bolt the thermal spacer block to the amp with two M3 (at least 6mm long) screws, no washers needed and use thermal paste between the thermal block and chip. Then mount the amp with a 1mm silicone spacer sheet between the thermal block and the amp chassis floor and that’s all that is needed. Use 10mm long M3 brass standoffs on the four corners. The stack up height was designed to be perfect to clamp the thermal block to the floor plate when the standoffs are snugged down.
OK I'm almost ready to put the amp into service. Unfortunately the amp didn't come with the screws required to put the top on the case, luckily I had some sheet metal screws that will work but I had to spray paint them black.
I can hear a very small amount of noise from the outputs if I put my ear right up to the little test speaker. If I unplug the RCA input's there's nothing. I used an old Monster cable I found, I wonder if this could be an issue. When I moved the RCA cables away from the TPA3255 board the noise definitely went down.
Anyway let me know what you think, inputs are welcome.
I can hear a very small amount of noise from the outputs if I put my ear right up to the little test speaker. If I unplug the RCA input's there's nothing. I used an old Monster cable I found, I wonder if this could be an issue. When I moved the RCA cables away from the TPA3255 board the noise definitely went down.
Anyway let me know what you think, inputs are welcome.
Attachments
Yes I should have mentioned, I jumpered pins 1 and 3 on the bottom side of the board with a very short piece of bus wire. I originally had a jumper on the female connector that was about 3/8 of an inch long, but I found that bringing my finger close to it caused hum.
Maybe move your input wires farther away from the AC mains on the soft start.
If you grounded your RCA inputs the amp should be so quiet that you almost cannot hear anything. I see that you are using the BTSB gain set at 20dB. That’s 35dB total gain and might be another reason noise is picked up. Set it to 14dB for 29dB overall amp gain.
Also, I see the corner ground screw on the BTSB seems to be connected to the shield of the RCA jacks? That might be a ground loop source. I would remove that and only connect the corner ground screw to the chassis via a ground loop breaker (10ohm resistor and anti parallel diodes or a 1ohm NTC). Maybe use a nylon stand-off there to isolate chassis ground from the BTSB.
If you grounded your RCA inputs the amp should be so quiet that you almost cannot hear anything. I see that you are using the BTSB gain set at 20dB. That’s 35dB total gain and might be another reason noise is picked up. Set it to 14dB for 29dB overall amp gain.
Also, I see the corner ground screw on the BTSB seems to be connected to the shield of the RCA jacks? That might be a ground loop source. I would remove that and only connect the corner ground screw to the chassis via a ground loop breaker (10ohm resistor and anti parallel diodes or a 1ohm NTC). Maybe use a nylon stand-off there to isolate chassis ground from the BTSB.
Last edited:
Actually the amp got quieter when I moved the input wires away from the 3255 board and grounded the RCA shield to the chassis. I used the corner post of the BTSB board for chassis ground since it was close and it meant I didn't have to take anything else apart.
Actually thinking about this now the speaker I used for listening has essentially no 60 Hz response so I might have made that worse without realizing it.
Good point about the gain, I'll reset that and retest.
OK I lowered the gain to 14 dB and shorted the inputs, at this point I couldn't really hear anything except a slight hum from the full(er) range test speakers I hooked up. As I moved back from the speakers to the amp to consider my next move the hum stopped... turned out it was from the HVAC unit. Back to the speakers and silence! Time to take the Adcom 5400 out of the desktop system, hook the TPA3255 up to the Mini Maggies & DWM/SVS sub & go!
I did a quick sweep of 20 to 500 Hz with a phone spectrum analyzer app to see how the gain matched up to the subwoofer which crosses over at about 55 Hz. It was pretty much spot on as you can see below. I may adjust the sub a bit in the future but for now it's pretty close.
So on to the listening tests and I have to say I am very impressed. The first song sitting in my queue from last night was Supertramp's The Logical Song, and in the first few seconds there's a castanet riff in the right channel - but then at the end of the castanets there's a really low sound in the left channel - maybe a cymbal. I never heard that before! Then Roger Hogdson's vocals jumped forward dramatically, rather than being in a flat plane with the rest of the instruments. All I can say is wow, the Adcom was no slouch but this amp is clearly better.
So for now more listening is in order, but if my first impression holds up all I can say is build this amp, you won't be disappointed.
I did a quick sweep of 20 to 500 Hz with a phone spectrum analyzer app to see how the gain matched up to the subwoofer which crosses over at about 55 Hz. It was pretty much spot on as you can see below. I may adjust the sub a bit in the future but for now it's pretty close.
So on to the listening tests and I have to say I am very impressed. The first song sitting in my queue from last night was Supertramp's The Logical Song, and in the first few seconds there's a castanet riff in the right channel - but then at the end of the castanets there's a really low sound in the left channel - maybe a cymbal. I never heard that before! Then Roger Hogdson's vocals jumped forward dramatically, rather than being in a flat plane with the rest of the instruments. All I can say is wow, the Adcom was no slouch but this amp is clearly better.
So for now more listening is in order, but if my first impression holds up all I can say is build this amp, you won't be disappointed.
Nice work! Glad you found the source of the hum. So you are saying it is coming through the air as sound waves from the HVAC, but you mistook it for coming from the speakers? Or is the hum from the HVAC making the mains have more hum in the amp?
In any case, looks like the 14dB gain setting did the trick to match to the sub.
Glad you like how it sounds. It really has a very low noise baseline which gives great dynamic range and huge bass authority.
In any case, looks like the 14dB gain setting did the trick to match to the sub.
Glad you like how it sounds. It really has a very low noise baseline which gives great dynamic range and huge bass authority.
Apparently there was never any hum. I thought that perhaps my little test speaker not being able to reproduce low enough was fooling me into thinking there wasn't any hum when I just couldn't hear it, so I spent some time this AM digging through my old speakers and selected a pair of 6 inch driver full range speakers (XAM 5Es from EJ Korvette's if anyone remembers that store) and when I hooked them up and put my ear next to them I could hear some hum. But then the hum stopped, so I spent some time listening for exactly what was causing the hum. I was wrong saying it was the HVAC, it turned out to be a dehumidifier next to the HVAC, I caught the little sucker in the act when it came on.
So long story short there never was any hum from the TPA3255 amp. The noise I was getting with my initial layout was broadband noise which decreased when I moved the input cables away from the TPA3255 board and grounded the RCA shield to chassis ground.
Psst! You like a lot of caps, right? Look how much can be squeezed from that Adcom 5400:
https://www.diyaudio.com/community/threads/ps-hack-for-adcom-gfa-5400.361693/
Hmmm... potential next project on desktop Mini Maggie system is replacing Schitt DAC with MiniDSP Flex and blaming. I'd use the 5400 to drive the DWM so now, unless I do this, I'm always going to wonder how much better it would sound!
I'd consider it. If I did that I would sell the 5400 to help finance the second 3255. But I do like the 5400.
Can anyone say whether the BOM at Mouser is pretty accurate. Call it lazy, but I'd like to avoid line by line review. I am spending all the time I have trying to get various BOMs filled before the shortages get worse. I will do it and report back if nobody else can say. Maybe update it with the approval of XRK and/or the community. Thankyou
The BOM is pretty good but you might find that N131 for the LM5010ASD DCDC buck converter is all out. You can DNP this whole group:
And connect +15vdc from external source with flying leads to right side pin of L131. You need to use a 200mA 15v external supply. This feeds the 12V LDO for the opamps and internal TPA3255 analog circuit voltage Plus the 3.3v LDO for digital signals and warning LEDs.
The amp needs 44mA for gate driver, 30mA for Vdd, and 9mA for Vd logic signals, each 5532 opamp needs 16mA x 4 so about 150mA total. Add inefficiency of LDO’s and 200mA would be a good minimum with some margin. Also, add 2200uF of bulk capacitance on the 15v external supply at L131 flying lead to allow it to keep voltage high longer than the main amp power rail post shutoff so that there is no turn off thump.
And connect +15vdc from external source with flying leads to right side pin of L131. You need to use a 200mA 15v external supply. This feeds the 12V LDO for the opamps and internal TPA3255 analog circuit voltage Plus the 3.3v LDO for digital signals and warning LEDs.
The amp needs 44mA for gate driver, 30mA for Vdd, and 9mA for Vd logic signals, each 5532 opamp needs 16mA x 4 so about 150mA total. Add inefficiency of LDO’s and 200mA would be a good minimum with some margin. Also, add 2200uF of bulk capacitance on the 15v external supply at L131 flying lead to allow it to keep voltage high longer than the main amp power rail post shutoff so that there is no turn off thump.
Last edited:
Hi Folks,
I have been getting a lot of PM’s asking for ready to run fully assembled TPA3255 amps. Maybe I should start another GB interest list. We need interest for at least 15 and I can get another manufacturing run going. The boards will be revised to hold an optional 2pin Molex KK where L131 is to allow external 15v supply to be connected. If you can find the LM5010ASD at a later point, you are welcome to install it yourself and not require a separate external PSU.
If you are interested please add your name, number of units and country below.
TPA3255 Amp GB 3 Interest List
Name / Units / Country
————————————————
JohnDoe / 2 amps / USA
I have been getting a lot of PM’s asking for ready to run fully assembled TPA3255 amps. Maybe I should start another GB interest list. We need interest for at least 15 and I can get another manufacturing run going. The boards will be revised to hold an optional 2pin Molex KK where L131 is to allow external 15v supply to be connected. If you can find the LM5010ASD at a later point, you are welcome to install it yourself and not require a separate external PSU.
If you are interested please add your name, number of units and country below.
TPA3255 Amp GB 3 Interest List
Name / Units / Country
————————————————
JohnDoe / 2 amps / USA
