TI TPA3255EVM

Sings with Mods !

An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


I would like to share my recent experience with the TPA3255EVM board from Digi Key / Texas Instruments. I sincerely hope this is the most appropriate location to convey my thoughts.

Thanks to all the gentlemen previously who posted their thoughts and comments!

I suppose the bottom line is… Wow, with some output capacitor mods, this amp really sings! :D I really didn’t expect this level of sound quality. The results this amplifier conveys are worthy of me taking the time to post comments. I haven’t done this for a while.

Please read my comments understanding they are not intended to empirical evidence with distortion numbers, scope ripple, double blind testing. Etc etc. I am just a speaker guy with something hopefully helpful to share.

My objective in this project was an energy efficient 4 channel amplifier to power my son’s park blaster music box. The primary associated components were/are:

Pioneer DEH 80PRS – source, DAC, DSP, with measurement microphone.

Hiquphon OW1 – High frequency driver

Accuton C95 – mid-woofer current model number C180-6-191

Accuton P220 – passive radiator

DROK Micro DC Converter Power Transformer 120W 12V/24V/48V 10-32V to 35-60V Volt Regulator - purchased from Amazon

Chronological evolution of the project / amplifier.

The first amplifier in my son’s music box was an older Class D Zed designed USA acoustic amplifier. This amp was warm and kind, but soaked 3-5 amps at 12 watts and we killed a fairly substantial AGM wheelchair battery. Overall, the setup was tolerable, but heavy and… dead.

We purchased a Lithium Iron battery and an ebay Mosconi Gladen AS 100.4. Unfortunately the Mosconi only had 3 channels working and the Mosconi factory tech couldn’t repair it :( (despite gladly accepting payment:mad:).

Eventually my anger subsided and I decided that slightly home brew Class D would probably work acceptably. I decided this because some fine gentlemen owning Pass First Watt gear was very happy with the 3255 and found it easily livable. Also, many of the new fancy car audio amplifiers are indeed class D. The best built board appeared to be the TPA3255EVM from digikey.

Initial testing involved using the power supply with 2 channels in use in single ended mode in my living room. I used a dragonfly DAC and my 1801C loudspeakers. I was pleasantly surprised! The amp was very dynamic, much better than the old Zed USA Acoustic amp, and roughly on par with the Mosconi Gladen AS 100.4 (used in 2 channel mode). The Mosconi Gladen wa slightly smoother/quieter. The TPA3255 was slightly more dynamic/punchy. This is commendable considering the pedigree and cost of the Mosconi amplifier.

After pondering the schematic for some time and looking at the physical space on the board, I decided to accomplish the following chronologically:

Removed the outer 1500uf output coupling capacitors and install 2 x 8.2uf Sonicap Gen I & 1 Sonicap Platinum in the same location. These are for the tweeters that have a 6.2 ohm DCR. I needed to pass at least 2khz (or lower) and didn’t want any significant phase shift. I figured the smaller and higher quality capacitors would be better. Wow! This really worked. The noise floor / grit really dissipated significantly. This was very worthwhile!

Remove the inner 1500uf output coupling capacitors and install 470uf Slimic Series II from Digikey. This was for the woofer circuit at 6 ohms DCR and I wanted to pass at least 50hz. This was positive.

Added a 5.1uf Sonicap Bypass to the Silmic extended leads on the bottom of the board. This was again very positive.

Added a .1uf Sonicap Platinum bypass around the 5.1uf (and 470uf Silmic). I was skeptical about this since the application was a midrange driver, but again the results were very positive.

Experimented with adding a 63v 1000UF Nichicon Gold at before (12v) or after (56v) the power supply. Neither location proved to add any audible impact. Apparently, there is already significant power supply capacitance.

Experimented with my power supply voltage. This can easily be adjusted via potentiometer on the power supply. I learned a few things. First, at lower voltages under about 40v there was a buzz / humm from the speakers. Applying a metallized propylene 3.3uf capacitor would not eliminate the buzz / humm. Second, I used the music box on 48v for a few weeks and wanted to be safe. I found the amplifier clip light would illuminate with about the same SPL as a 30 watt push pull EL34 tube amp. Third, I discovered that with a 6 ohm DCR load that 56v was optimal and I believe this is correct. I did obtain about 5db more spl from the speakers before the clip light illuminated. Also, there might have been a very minor improvement in sound quality at higher spl on 56vdc , but I am not certain in this regard. Fourth, I attempted to physically cool the heat sink of the TPA3255EVM with a fan. This didn’t work. The ambient temperature of the heat sink dropped 10 degrees as measured with an infrared temp gun, but the clip light illuminated at the same SPL.

Other notes…

I have not tried this amplifer in BTL mode and cannot provide any comment therein.

This amplifier “drinks” about .5 amp of 13v batter juice when playing music at Comfortably high SPL.

In SE mode at 56v this amplifier plays comfortably too loud for my relatively large living room with 87db speakers. But, don’t expect rock concert SPL.

Other amplifiers… some old solid state stuff, a Tubes4Hifi Octal Dynaco ST 70, a Heavily Modified Golden Tube SE40SE, a Aspen Soraya CB 105, and a Cary 572SE MkII. In my opinion, my mildly modified TPA3255EVM won't displace the Apsen, Modded Dynaco or Golden Tube, but my modded TPA3255EVM could easily and smartly be the last amplifier a gentlemen ever purchases.
 
Sings with Mods !

Would it have a place in a domestic audio system?!

I figured that either the pioneer head unit or the TPA3255 would be a decent step in sound quality below home audio stuff, but this isn't really the case in my opinion. Actually, my wife (the house member with the really good ears) did a short listening test from a 30' distance. She thought the music box had more detail, and should be better for listening to music. She thought my home system sounded more warm and would be better for watching TV. The same drivers are used in the speakers. Only the source "gear" is different. So, the "difference" is extremely close.

I do think there remains very minor grit / noise from either the Pioneer 80 PRS or the TPA3255. But, it's VERY minor. On balance, the automated in room crossover function of the 80PRS is the cats-meow :cool: and really seems to work terrific.

A couple years ago I searched for some processor that would do a DAC function and microphone based DSP (automated). I wanted a DSP that would accomplish a 2 way or 3 way stereo active setup. I found the 80 PRS, Mini DSP, and DEQX. DEQX was/is $$$. MiniDsp was slightly more expensive. The 80 PRS seemed was portable and 12v friendly

Hopefully this helps somewhat.

Subjective comments are always so... subjective :confused:
 
Would it have a place in a domestic audio system?!

My first power supply was an AGM wheelchair battery and a battery tender Jr. It would be very easy to use a plain lawnmower battery from the local farm store (less expensive) and a battery tender consistently applied.

The only "bugger" with the 80PRS is the memory power is a slow consistent trickle. The 80PRS will kill a battery if the Battery Tender Jr is not plugged in.
 
Interesting: only ~7w used for this 315w. While there are transient peaks, it is still kinda-of overkill.

Well... this is on the fringe of my understanding. The Clip light on the PCB illuminates when a sensor reaches 125 degrees. I think this is 125 degrees Fahrenheit. This doesn't necessarily mean the amplifier IS clipping the peaks of the waveform.

I was puzzled when my application of airflow dropped the heatsink temp about 10 degrees, but didn't seem to alter the spl when the amplifer clipped. Maybe the heat sink compound isn't so good ???

Someone else might have to chime-in with more experience / insight.

I am perfectly happy with the SPL the TPA3255EVM produces. I didn't want or need a blasting loud amplifier capable of stretching the suspension of my woofers.
 
Member
Joined 2008
Paid Member
Other amplifiers… some old solid state stuff, a Tubes4Hifi Octal Dynaco ST 70, a Heavily Modified Golden Tube SE40SE, a Aspen Soraya CB 105, and a Cary 572SE MkII. In my opinion, my mildly modified TPA3255EVM won't displace the Apsen, Modded Dynaco or Golden Tube, but my modded TPA3255EVM could easily and smartly be the last amplifier a gentlemen ever purchases.

I assume your build passed the necessary product testing and evaluation? :D
 
I'm sorry to have to post this but I think buyers should be aware of my experience in the last couple of weeks with Connex, I have sent them emails regarding info of the aux pin-outs on both amp and smps and they just don't reply or they say that my emails go to their spam, their products are good but if you have any problems you are on your own.
 
Member
Joined 2008
Paid Member
That's been an issue since they, well he, started. Though I haven't heard any complaints in a long while. Their products are usually well documented and well designed enough to not need support. But yeah, he used to have these periods where it took a long time to get replies. It's a one-man operation far far away so support is the weakest link..
 
Has anyone tried running a pair of EVM boards as PBTL monoblocks with higher impedance speakers? The datasheet doesn't cover PBTL for 8 ohm loads, and extrapolating from 4 ohm I'd guess there probably is little to no difference from BTL, Maybe splitting the output current across 4 inductors rather than 2 has some minor benefits, but I guess I'm also curious about there being a penalty due to mismatches between channels.

I'm considering this since I'm still not sure whether I want to go active on my desktop monitors or not. A 2 EVM setup would allow me to do PBTL monoblocks, but if I decided to go active I could revert to BTL for 4 total channels. Even at full price for the EVMs, this isn't an overly expensive settup.
 
As I understand it, and someone please correct me, is that in PBTL, the load on each amplifier/lc filter is identical to SE, not BTL. In other words, when wired BTL, each amp sees half the load, when wired parallel each amp sees twice the load, when wired PBTL you are right back to the speaker load.

Mike
 
Hi.
I'm having some issues with my TPA3245EVM and I figured this thread would be most appropriate to post in as the board is quite similar to the TPA3255EVM.

I decided to change the 4 caps in the output filter to better fit my two 8 ohm speakers which I am running in PBTL mode. The TPA3245EVM LC output filter has 7uH dual inductor and 0.68uF caps, which I've changed to 0.22uF after plotting the values in the TI's class D filter design spreadsheet.
I didn't really expect any improvement in the sound but I certainly didn't expect one speaker (OUTC,OUTD) to drop 7dB in volume.
I've been trying to troubleshoot it with the scope and DMM but had no luck in finding what could be wrong. As it happened right after changing the caps I was thinking the 0.22uF capacitance is too small. I've done RC snubber mods and output filter changes before on tpa3116 and tpa3118 boards so I have some basic soldering experience and the EVM is quite a large board so I've ruled out any soldering errors for now.

The only thing I really know at the moment is, that I have to up the volume by 7dB for one stereo channel, to get the same volume from both speakers. I've tried changing the RCA cables, swapped L and R speaker, tried different DAC's and the problem persists on the OUTC, OUTD channel. The power supply (LRS-150) is also brand new and I'm running the board with 29V.

Can anyone offer some advice, what would be the first step in troubleshooting this? As far as measurement tools go, I have a DMM, UR12 sound interface and a basic USB scope. I would appreciate any help. Thanks!
 
As I understand it, and someone please correct me, is that in PBTL, the load on each amplifier/lc filter is identical to SE, not BTL. In other words, when wired BTL, each amp sees half the load, when wired parallel each amp sees twice the load, when wired PBTL you are right back to the speaker load.

Mike

Can confirm just that.

Snuggie -- I'm at a loss at the moment, but I'll try to have a look at the datasheet to see if it accidentally kicked off a "limp home" mode.
 
Snuggie -- I'm at a loss at the moment, but I'll try to have a look at the datasheet to see if it accidentally kicked off a "limp home" mode.
Thank you!
If your loss was 6dB I would guess that here only one input is driven - hence only one leg of PBTL puts out sound - the other is static DC. Could you check this please?
Could be 6dB, I forgot to account for room correction and found now it was -0.8 db on that channel and judging by ear, the stereo image is now centered.

How would I best measure if only one leg outputs sound? Scope on each of the outputs or PWM measuring points before the inductors?
There is 15V DC on each of the outputs (running on 30V now, not 29V like I said in my previous post, sorry) so 1/2 of the PVDD.
 
Voltwide's suggestion is a very good start, the 7db threw me for a loop.

DC checks out, as it should. Octuple check through the jumpers to make sure they're set correctly. Are you using single ended or differential input?
http://www.ti.com/lit/ug/slvuat6a/slvuat6a.pdf pg 7
Have rechecked the jumpers now, everything configured for single ended inputs (RCA).
Haven't really touched the jumpers since I got the board, only the reset rocker switch.