I actually had my Pass DIY B1 Kong driving my TI TPA3255EVM before a lightning strike near my home damaged my B1 Korg. I am taking my time finding out where the damage occurred—either with the Korg NuTube or the JFETs. I’ll get it back up soon, because it is the best tube preamp for warm Summers, and coupled with the cool-running TPA325x Class D amps, the sound is very enjoyable. I took your advice and installed film caps on the inputs and a Tocos Cosmos volume pot. The sound quality was getting very close to my treasured Audio Research LS7 tube line stage preamp.
I brought that to the BAF and then connected with my audio buddies in the San Francisco Bay Area where I used to live. My friends were very impressed with the B1 Korg in their systems. They couldn’t believe how good the holographic soundstage was, and how smooth it sounded. It’s no surprise how quickly the kits sell out every time they are stocked.
I really like the OPA1656 op amps. They are really settling in. I keep forgetting that I have another classic amp to compare the Aiyima TPA3251 to — my Enhanced Fixed Bias-modified Dynakit Stereo 35 tube amp. This has the original transformers coupled with Dave Gillespie’s Enhanced Fixed Bias modification to produce incredible headroom. Once, I had Roger Modjeski of RAM Audio Labs over at my home with some of my friends, and Roger conducted a blind A/B test between the Dynakit Stereo 35 and his RM-10 MkII 6BQ5/EL84 amp. We could hardly tell the difference between his great amp and the Dynakit. Roger recently passed away, and I hope someone is carrying on his legacy. He was very passionate about his craft and he taught me a lot about audio circuits. Unfortunately, I missed the lesson about op amp orientation.
Last edited:
Thanks Rhing ) I did not know that.
Seems that the RTR TPA3255 by XRKaudio is benchmarking very well, do you know about this one ? I read that it was really a safe bet despite a really expensive price
Hi daniboun,
I have read the same about XRKaudio's TPA3255 amp as well, but I already have the "TPA3255 / TPA3251 / TPA3245 universal - one for all" boards from doctormord of 360Customs. Having used Class D amps since the Tripath days, doctormord has been a great resource on these Class D thread on optimizing Class D amps, especially those using the Texas Instruments TPA-3xxx amps.
I have been using the OPA1656 op amps in my Aiyima TPA3251 amp for a week now. I really like these op amps and I've been going back and forth with the OPA627AU op amps. Here is what I noted with the OPA1656 op amps:
In comparison to my tube amps, I'll take the tube amps, the MC240 and Dynakit Stereo 35, over the Aiyima amp. Tube amps with my Klipsch Forte II speakers create this incredible scale and image that is life-size and life-like. The Sony VFET amp sounds very much like a tube amp and creates a huge soundstage with incredible depth and detail. The Aiyima TPA3251 and the Texas Instruments TPA3255EVM amps are great, but they cannot create the soundstage depth or the tonal richness of my reference amps. Interestingly, most people would be very pleased with the Aiyima amp's sound until switched to one of the other amps for comparison.
The best sound I have ever heard from the TI TPA3255EVM was when a friend lent me a now-extinct PonoPlayer DAP with the audio output circuitry designed by engineers at Ayre Acoustics and used an ESS Sabre32 ES9018K2M DAC. I used custom-built balanced cables from the PonoPlayer to the RCA connectors on the TPA3255EVM configured in differential mode, and I was amazed how good it sounded. Using a balanced source with the TPA325x amps in differential mode is probably the best scenario. If I had a fully balanced sources and a balanced preamp like an Audio Research LS22, I could really take advantage of the TPA325x amps run in differential mode.
So tomorrow, I will wire in the Coilcraft inductors and do some more listening. daniboun has provided me the contact information with Audio Science Review to have them conduct some testing of the Aiyima amp using different op amps.
I have read the same about XRKaudio's TPA3255 amp as well, but I already have the "TPA3255 / TPA3251 / TPA3245 universal - one for all" boards from doctormord of 360Customs. Having used Class D amps since the Tripath days, doctormord has been a great resource on these Class D thread on optimizing Class D amps, especially those using the Texas Instruments TPA-3xxx amps.
I have been using the OPA1656 op amps in my Aiyima TPA3251 amp for a week now. I really like these op amps and I've been going back and forth with the OPA627AU op amps. Here is what I noted with the OPA1656 op amps:
- Upon initial listening, the clarity and transparency were readily evident.
- As ClaudeG mentioned, the OPA6156 are neutral, but not in some sterile way. I would say they offer no bumps in the midrange like I hear with the OPA627AU
- The bass definition is solid
- I could hear fine details like hall ambience, natural reverb and other spatial cues
- The sound is dynamic with great reproduction of transients like string plucks and cymbal strikes
- The sound is warmer with a more forward midrange. I like vocals and brass instruments in jazz music, and the OPA627AU op amps provide greater presence.
- This may be related to the midrange emphasis -- vocals from multiple performers are better separated and located within the soundstage. With the OPA1656 ops amps, the vocals were all in the center like they were stacked on top of one another, rather than one vocalist behind and the to right, or immediately to the left.
- The OPA627AU op amps offer a more opaque image and maybe this my own personal preference as I like the same from my McIntosh MC240 tube and Pass DIY Sony VFET amps.
- Another detail that I like about the OPA627AU op amps is the cymbal detail. The brass shimmer was more fleshed out than the t-t-t-t-t-t I heard with the OPA656 op amps.
In comparison to my tube amps, I'll take the tube amps, the MC240 and Dynakit Stereo 35, over the Aiyima amp. Tube amps with my Klipsch Forte II speakers create this incredible scale and image that is life-size and life-like. The Sony VFET amp sounds very much like a tube amp and creates a huge soundstage with incredible depth and detail. The Aiyima TPA3251 and the Texas Instruments TPA3255EVM amps are great, but they cannot create the soundstage depth or the tonal richness of my reference amps. Interestingly, most people would be very pleased with the Aiyima amp's sound until switched to one of the other amps for comparison.
The best sound I have ever heard from the TI TPA3255EVM was when a friend lent me a now-extinct PonoPlayer DAP with the audio output circuitry designed by engineers at Ayre Acoustics and used an ESS Sabre32 ES9018K2M DAC. I used custom-built balanced cables from the PonoPlayer to the RCA connectors on the TPA3255EVM configured in differential mode, and I was amazed how good it sounded. Using a balanced source with the TPA325x amps in differential mode is probably the best scenario. If I had a fully balanced sources and a balanced preamp like an Audio Research LS22, I could really take advantage of the TPA325x amps run in differential mode.
So tomorrow, I will wire in the Coilcraft inductors and do some more listening. daniboun has provided me the contact information with Audio Science Review to have them conduct some testing of the Aiyima amp using different op amps.
Last edited:
Whaou, great review and great feedback, thanks a lot for sharing all that, Rich!
I can understand fully your findings and personal preferences, be it op-amps or power amps... very well described and very respectable of course.
Last but not least, well done regarding the ASR testing(s) to come. I wouldn't be surprised to read that the (listeners) prefered Class D option is the one that produces more distorsions... but of the right harmonic kind 🙂
Stay safe and read you soon
Claude
I can understand fully your findings and personal preferences, be it op-amps or power amps... very well described and very respectable of course.
Last but not least, well done regarding the ASR testing(s) to come. I wouldn't be surprised to read that the (listeners) prefered Class D option is the one that produces more distorsions... but of the right harmonic kind 🙂
Stay safe and read you soon
Claude
Thanks ClaudeG. The OPA627AU op amps match up with my taste best, but I can understand why others would prefer OPA1656 op amps. I will send the amp with the op amps to Audio Science Review once I put the Meanwell LRS-200-36 power supply in an extruded Aluminum case with an IEC inlet, mains switch and DC power jack.
As for preferable distortion, I agree that some of the better sounding equipment for me possesses H2, negative phase. That’s how I biased my Sony VFET Class A amp.
rhing, your results and conclusions are the same as mine.
And once you have a differential source you can go a step further - bypass the TPA3255's opamp stage altogether. This is quite easy with the EVM board because direct inputs to the TPA3255 chip are exposed at the J28 header.
You may find that the best opamp is NO opamp!
Sure, but buying a $2000 preamp to optimise a $200 power amp is a bold move!
Instead, maybe see if you can buy your friend's PonoPlayer and use this with the TPA3255EVM as a second system? Just thinking out loud.
Hi Rhing. First : thanks for the feedback about the OPA1656. I am almost sure you should like the Burson V5i from what you wrote about the OPA626 vs 1656. Your feeling makes me think of mine to some extent. My heart / ears swings between the OPA1656 and the Burson V5i. I hesitated to sell my Bursons but in the end they had "something very pleasant" to listen to.
Something I did not understand (sorry again for my rough English) :
You mean you own the Board from Doctormord ? From what I understand, would you have bought the PCB board (ready to solder) ?
I can't wait to see what the Aiyima TPA3251 with the Coilcraft inductors ) yeahhhhh )
PS : I will soon receive my TPA32XX / 360 Customs with PFFB and its improved SMPS LLC power supply
We will have the opportunity to compare a very high-end TPA32XX and our surprising Aiyima TPA3251.😀
Last edited:
Hi Rich,
Quote : As for preferable distortion, I agree that some of the better sounding equipment for me possesses H2, negative phase.
So do I ;-)
Hence not impressed by the ASR reviews in absolute terms, but of course nice to have if FOC.
I was surprised though on your comments re spaciality/soundstage/location of the OPA1656, something that worked at mine's, instead you found it quite packed (love jazz and opera voices)
20y ago, in my mother's DAC section, I fitted an AD827... whereas clearly the better op-amp in that location would have been 2xAD825 in Class A. Reason being that DAC was limited in terms of sonic qualities and the prefered option was a more forgiving op amp. Once I plugged my usual DAC on her drive, the AD825s were a very clear winner, but not on her moderate DAC "considering overall compensation and personal taste" aspects.
Hence me thinking loud... Should the coil mod be efficient, may I humbly suggest you may (depending on the sonic results) benefit from giving op-amps another (very quick) try, just to make sure the findings are the same?
Having said that, in my experience so far indeed Class D amps defo lack that spaciality, flesh and "subtle note ending" and are no match up to now to my conventional amp re musicality... to my ears of course!
Fun though to see though how they improve quickly in the sub 1000$ bracket at a steal's price.
Argh, not easy for me to beat / replace my simple 2SA1302/2SC3281 output stage, hence big hopes in Fab's coming "classical" amps, but that's another topic 🙂
Thanks again for all this Rich!
Claude
Quote : As for preferable distortion, I agree that some of the better sounding equipment for me possesses H2, negative phase.
So do I ;-)
Hence not impressed by the ASR reviews in absolute terms, but of course nice to have if FOC.
I was surprised though on your comments re spaciality/soundstage/location of the OPA1656, something that worked at mine's, instead you found it quite packed (love jazz and opera voices)
20y ago, in my mother's DAC section, I fitted an AD827... whereas clearly the better op-amp in that location would have been 2xAD825 in Class A. Reason being that DAC was limited in terms of sonic qualities and the prefered option was a more forgiving op amp. Once I plugged my usual DAC on her drive, the AD825s were a very clear winner, but not on her moderate DAC "considering overall compensation and personal taste" aspects.
Hence me thinking loud... Should the coil mod be efficient, may I humbly suggest you may (depending on the sonic results) benefit from giving op-amps another (very quick) try, just to make sure the findings are the same?
Having said that, in my experience so far indeed Class D amps defo lack that spaciality, flesh and "subtle note ending" and are no match up to now to my conventional amp re musicality... to my ears of course!
Fun though to see though how they improve quickly in the sub 1000$ bracket at a steal's price.
Argh, not easy for me to beat / replace my simple 2SA1302/2SC3281 output stage, hence big hopes in Fab's coming "classical" amps, but that's another topic 🙂
Thanks again for all this Rich!
Claude
Last edited:
I have followed this thread now since almost its inception in July 2015. I ordered my first TPA3255 Amplifier from Texas back in 2017 - Never arrived because of problems shipping to the UK ?? I now have the Aiyima 07 TPA3255 boxed Amplifier with a Aiyima 04 TPA3251 ordered but still not arrived. I note there are now some newer versions being offered under the Aiyima name but they are clearly cost reduced and should be avoided - They don't appear to have sockets for the Op-Amps !!
I have played with many, many TPA3116 boards and amplifiers over the same period of time and maybe found a couple that were worth listening to !! Keep in mind that most of these products, even in cases cost less than $30 US. !! I have also collected a mass of MeanWell SMPS Power Supplies from 12 volts to 48 volts - Most cost less than $25 on eBay second hand - But clearly some have never been used - Only had to throw one dead one !
The TPA3255 Amplifier has given me the opportunity to play (roll) Op-Amps and so far the OPA1656 sounds the best to my ears (of NE5532, OPA1622) but my OPA627AU chips haven't arrived yet. The OPA1622 sounds good too but I don't like all the extra components required to allow it to operate in an 8 pin socket and it's not DIY friendly - might be OK for a pick and place machine but my eyesight isn't good enough to solder it !!
I guess the point I wanted to make is that I have never spent more than $1,500 US on an integrated Amplifier (Yamaha) so the reference point to my ears is probably a lot lower than most of you on this forum. I think what we have here is something quite amazing for what most of us paid less than $50 US.
I am quite handy with a soldering iron and await with interest what might be the best options for the Input/Signal Capacitors and Output Inductors/Capacitors.
Something that doesn't seem to get much coverage is that this device is basically working/switching at 500KHz Radio Frequencies and I would have thought that positioning of the Output Inductors would be critical as is the length of wires from the power supply as resistance climbs with frequency !? Some designs seem to have all the Inductors close together in a row - Surely there must be some transfer of power between them - equivalent of crosstalk ?
The other question that I have not seen answered is whether the quality of the output is dependant of the level of the input signal. The input to the chip is an ADC so with a low level signal you are only using a fraction of the resolution of the converter. So it seems to me that you are better to set the Supply Voltage to give you enough headroom (Volume) for your normal listening and take advantage of the full bit rate of the ADC - Or have I missed something ??
Just an aside, anyone wanting any OPA1656 chips on DIL8 adapters I have some available !
I have played with many, many TPA3116 boards and amplifiers over the same period of time and maybe found a couple that were worth listening to !! Keep in mind that most of these products, even in cases cost less than $30 US. !! I have also collected a mass of MeanWell SMPS Power Supplies from 12 volts to 48 volts - Most cost less than $25 on eBay second hand - But clearly some have never been used - Only had to throw one dead one !
The TPA3255 Amplifier has given me the opportunity to play (roll) Op-Amps and so far the OPA1656 sounds the best to my ears (of NE5532, OPA1622) but my OPA627AU chips haven't arrived yet. The OPA1622 sounds good too but I don't like all the extra components required to allow it to operate in an 8 pin socket and it's not DIY friendly - might be OK for a pick and place machine but my eyesight isn't good enough to solder it !!
I guess the point I wanted to make is that I have never spent more than $1,500 US on an integrated Amplifier (Yamaha) so the reference point to my ears is probably a lot lower than most of you on this forum. I think what we have here is something quite amazing for what most of us paid less than $50 US.
I am quite handy with a soldering iron and await with interest what might be the best options for the Input/Signal Capacitors and Output Inductors/Capacitors.
Something that doesn't seem to get much coverage is that this device is basically working/switching at 500KHz Radio Frequencies and I would have thought that positioning of the Output Inductors would be critical as is the length of wires from the power supply as resistance climbs with frequency !? Some designs seem to have all the Inductors close together in a row - Surely there must be some transfer of power between them - equivalent of crosstalk ?
The other question that I have not seen answered is whether the quality of the output is dependant of the level of the input signal. The input to the chip is an ADC so with a low level signal you are only using a fraction of the resolution of the converter. So it seems to me that you are better to set the Supply Voltage to give you enough headroom (Volume) for your normal listening and take advantage of the full bit rate of the ADC - Or have I missed something ??
Just an aside, anyone wanting any OPA1656 chips on DIL8 adapters I have some available !
Can't wait to hear what the TPA3251 sounds like with the Coilcraft. I haven't received my amp yet, but I'm already putting a shopping list together.
Hi Lee. Which TPA3251 Amplifier have you ordered ? So many now to choose from !!
I will definitely test the Aiyima TPA3251 amp again with the OPA1656 op amps after installing the Coilcraft inductors.
I would expect doctormord's latest TPA32XX amp to sound very good. I bought his first PCB design revision over two years ago when he first designed it. I will need to build it from scratch, so I will probably use the solder stencil and an oven to reflow the solder to make the work a lot easier. Soldering surface mount components (SMDs) is like performing brain surgery.
My plan is to run the doctormord TPA3255 amps in a pure dual mono amp configuration at 300 watts per channel with balanced and single ended inputs. I have a pair of CineMag CMLI-15/15B input transformers for the single ended inputs to bridge to the TPA3255's differential inputs. I've always had a side interest in sound reinforcement applications. I've worked with Lab Gruppen Class TD amps in live sound reinforcement, and they are amazing.
I may try the Burson Audio V5i-d op amps, but I have so many other projects to complete. I just received my set of capacitors to replace the aging Sprague Bumblebee film caps in my McIntosh MC240. The Mac sounds and works great right now, and I want to keep it that way.
The Aiyima TPA3251. It's in transit at present. It's destined for my office system, which of course, is where I listen to most music. I've just started to get back into warming the soldering iron up and fixing a few things up and this looks loke an easy thing to play with, not much to change.
I installed the Coilcraft MA5173 7uH inductors in the Aiyima TPA3251 amp and the sound quality is further improved in terms of clarity, detail and imaging. The amp with the stock inductors is on the left, and the same amp with the Coilcraft inductors is on the right.
I measured the stock inductors and they are 12.5uH. I calculated cut-off frequency and Quality Factor, Q, for three different inductor values using the free TI Class D LC filter designer tool:
..L............C........f(cut-off)............Q
7uH........ .68uF....72.9kHz...........0.623
10uH...... .68uF....61.0kHz...........0.522
12.5uH... .68uF....53.7kHz...........0.474
Below are the plots for L = 7uH, 10uH and 12.5uH (C = 0.68uF):
Depending on your speaker system's impedance, the Coilcraft MA5172 (10uH) might be a better inductor to use. TI's TPA3251EVM amp uses 7uH.
I measured the stock inductors and they are 12.5uH. I calculated cut-off frequency and Quality Factor, Q, for three different inductor values using the free TI Class D LC filter designer tool:
..L............C........f(cut-off)............Q
7uH........ .68uF....72.9kHz...........0.623
10uH...... .68uF....61.0kHz...........0.522
12.5uH... .68uF....53.7kHz...........0.474
Below are the plots for L = 7uH, 10uH and 12.5uH (C = 0.68uF):
Depending on your speaker system's impedance, the Coilcraft MA5172 (10uH) might be a better inductor to use. TI's TPA3251EVM amp uses 7uH.
These plots are generated under lab conditions using low inductance dummy resistors. As real drivers are anything else but low inductance resistors these plots are pretty useless for high frequencies in the real world.
just my 2c
just my 2c
The upshot of these plots is use 10 uH and any speaker between 4 and 8 Ohm will be just fine.
Thanks a lot for this detailled feedback, Rich!
Eventhough LS aren't dummy loads, it gives IMHO a pretty good idea what coil to chose for your own LS, indeed provided you know their complex impedance (modulus and phase) above 10kHz.
There isn't a lot of music content / density above 10k, but from 10 to 20kHz being a wide band in terms of overall perception it can quite affect our perception in terms of thiness and brillance. It does though IMHO also impact our perception re agressivity and flesh on the bone at lower octaves, so worth considering eventhough the TI calculation is a simple model - at least it is very well detailled.
Rich, what magnitude would you give to your coil mod vs op amp rolling? Is that quite majour, or more a small step? I do understand that op amp rolling was performed whith the old coils, so less clarity and differences between op amps I presume, but nevertheless, would you rate that coil mod a majour must-do step or is it more fine tuning?
Further, I always felt (to my ears) Class D was lacking in terms of 3D imaging and of end of note truncating - whereas details and control were quite OK. I have the feeling reading your comparo with state of the art amp it may also be the case for you. Now that you have a better imaging, how close does it get re soundstage and placement? Is majour field covered, or the gap still quite there - possibly chip or technology limited?
Sorry for the numerous questions Rich - it is your fault, your posts are really that interesting 🙂
Last but not least, I really wonder how much 3D image that tiny and presumably very cheap vol pot is robbing... cheap pots do that very often in my experience...
Thanks again for all this interesting stuff
Stay safe
Claude
Eventhough LS aren't dummy loads, it gives IMHO a pretty good idea what coil to chose for your own LS, indeed provided you know their complex impedance (modulus and phase) above 10kHz.
There isn't a lot of music content / density above 10k, but from 10 to 20kHz being a wide band in terms of overall perception it can quite affect our perception in terms of thiness and brillance. It does though IMHO also impact our perception re agressivity and flesh on the bone at lower octaves, so worth considering eventhough the TI calculation is a simple model - at least it is very well detailled.
Rich, what magnitude would you give to your coil mod vs op amp rolling? Is that quite majour, or more a small step? I do understand that op amp rolling was performed whith the old coils, so less clarity and differences between op amps I presume, but nevertheless, would you rate that coil mod a majour must-do step or is it more fine tuning?
Further, I always felt (to my ears) Class D was lacking in terms of 3D imaging and of end of note truncating - whereas details and control were quite OK. I have the feeling reading your comparo with state of the art amp it may also be the case for you. Now that you have a better imaging, how close does it get re soundstage and placement? Is majour field covered, or the gap still quite there - possibly chip or technology limited?
Sorry for the numerous questions Rich - it is your fault, your posts are really that interesting 🙂
Last but not least, I really wonder how much 3D image that tiny and presumably very cheap vol pot is robbing... cheap pots do that very often in my experience...
Thanks again for all this interesting stuff
Stay safe
Claude
Great ! thanks amigo )
do you have pictures of the PCB with a bottom view? thanks
As Claude already asked : is the tweak worth it in the end?
Last edited: