jean-paul - I ordered an A8 yesterday as ali is having a sale and managed to get it w/ tax for $48 shipped. I know you did a lot of mods on it which are beyond my scope (70 year old eyes). Are there any that are easier to do that are worth it? I was able to do the basic mods on the BrzHiFi unit, but was frankly shocked that it all worked when powered up.
https://m.aliexpress.us/item/3256803444034368.html I ordered over $90 worth of stuff which was $9 off automatically; Ali also offered an $8 off of $80 coupon. Amazingly they stacked, so it took a few dollars off of each item.
OK, thank’s for picture. They are more optimistic in the lower of these two videos
https://www.infineon.com/cms/en/pro...cs/ma5332ms/#!?videoId=6mUSJP_1Mg1JvPk7MoENJ1
At 3:50 into the video the crazyness begins. Actually 212 watts in and 200 watts out. So only 12 watts to heat the board. Impressive
But guess they have soldered the epad;-)
And in 60 seconds it overheats
Guess also they use kind load.
It will of course be even better with a small heatsink or heatcoupled to a metal plate. Se the distortion raises at 80 watts without.
Last edited:
Did you mean to type Aiyima A8 (ma12070) here or did you mean the A08 (tpa3255) Pro?
since I only use balanced inputs, I yoinked the OPA1656 out of my PA3S amps. No more worrying about it clipping into SE mode and introducing noise. I then swapped the LME4970s in voltage follower stage of my cheap DAC with the 1656 and it made a significant improvement! The 1656 is killer
It is directly tied inverting input to output. I think the 1656 pulled well ahead because I changed the I/V to 1k resistors for 1V:1mA. The dual pllel 9038q2m can output 7.3XmA in current mode per output so the total output is near 7Vrms Hot-Cold on the XLRs. The 49720 starts to fall on its face above 4V output.
I also changed the high pass / DC filter to a .47uF film cap on the bottom of the PCB and the R to 100k (e.g. R2 in that photo taken before the full mods)
Oh, and I reflowed the epads on the PA3s and supplemented the 1k on the volume pot contacts with a 10k to ground. MA12070 is silent with that 10k to ground. Can have floating/open inputs and just dead silence. The heatsinks dont even go above room temperature any more.
I also changed the high pass / DC filter to a .47uF film cap on the bottom of the PCB and the R to 100k (e.g. R2 in that photo taken before the full mods)
Oh, and I reflowed the epads on the PA3s and supplemented the 1k on the volume pot contacts with a 10k to ground. MA12070 is silent with that 10k to ground. Can have floating/open inputs and just dead silence. The heatsinks dont even go above room temperature any more.
So the heat issue is a PA3S only issue?!
My challenge is to find through hole ferrite beads that can handle some current to upgrade existing amplifiers. Can't find any that are suitable.
My challenge is to find through hole ferrite beads that can handle some current to upgrade existing amplifiers. Can't find any that are suitable.
You will have to read the mods and decide if you can do them. The opamps were a nice improvement just like the caps. I think improving the output filter will also bring improvement but I had to deal with quite severe issues so I haven't had any time.
Last edited:
Strange is the 49720 fakes?
4 V should be no trouble if the supply voltage is high enough
I wouldn;t use the word problem specifically. As I have said all along, it is not so much a problem as much as just not to Infineon's recommendations. I trust they know how to extract the maximum performance. I would guess it was not done on the PA3s because it looks like all bottom solder joints are manually executed (judging by the flux residue and lack thereof on the top).
My take is: THe epad is listed as a connection by the IC manufacturer. It is not listed as epad/NC. Infineon documentation shows the epad to be minimum viable:
And the thermal conduction to ambient top/plastic coated areas) is awful. Again, from the datasheet.
It's like running a DAC's AVCC from a noisy LDO. Will it work? Of course. Will you get full rated performance? Absolutely not. But most MFG's probably don't want to pay for wave soldering two sides of a PCB.
I got them directly from Mouser, so definitely not fakes. But I am not outputting 4Vrms. Instead, it is greater than 5.7Vrms or 8Vpp on each leg of the XLR (so the full differential opposing potential difference reaches >16V across the + XLR and -XLR).
Comparing the LME49720:
To the OPA1656 at 8V:
It makes a bit more sense.Both the I/V and the folower are set up on the inverting input. OPA1656 really likes this. If I dropped 2Vpp then they should be about equal. If I dropped 4Vpp then the 49720 should pull ahead. When I modded the output, I thought my DriveRack clone wanted a hot signal (because gain staging should be done as near the source as possible). That proved to be.... a mistaken belief. https://www.audiosciencereview.com/...ese-drivehack-pa2-hopeful-improvements.39071/
I might look for some 700-800 ohm 0603 1% resistors to swap in and see if I can coax out more THD+N numbers. (currently https://www.diyaudio.com/community/threads/es9038q2m-board.314935/post-7180448 ). I did fix the frequency imbalance on the right channel. I thought I harvested 4x 01D marked resistors but I grabbed 2x 01D and 2x 01C so I swapped two 01C with 2x 01C on the right channel. I have since swapped those again with the proper 01D so I'm not artificially elevating the highpass filter of a single channel. .47uF and 1k makes for a 338Hz filter which shows well on the RMS vs Frequency measure I posted. .47uF and 100k, as I intended, has a HPf of 3.38Hz.
Have you tried to install a combination of electrolytic capacitors - Panasonic FM and Rubycon ZL - on the PVDD line? On my Merus hat, this combination sounds nicer than Panasonic alone or Rubycon alone. I used panasonic fm 390/35 and rubycon zl 470/35.
Attachments
Anyone tried one of these yet? I am tempted, but for £40 I feel it would be better to get a half decent TPA3255 implementation..
https://www.aliexpress.com/item/1005001810845146.html?spm=a2g0o.detail.1000014.6.32df20b9RuICrs&gps-id=pcDetailBottomMoreOtherSeller&scm=1007.40000.317745.0&scm_id=1007.40000.317745.0&scm-url=1007.40000.317745.0&pvid=ea38b322-c7a5-4245-88ac-bc35806e9c52&_t=gps-id
cDetailBottomMoreOtherSeller,scm-url:1007.40000.317745.0,pvid:ea38b322-c7a5-4245-88ac-bc35806e9c52,tpp_buckets:668#2846#8115#817&pdp_ext_f={"sku_id":"12000017711302004","sceneId":"30050"}&pdp_npi=2@dis!GBP!28.0!26.6!!!!!@211b5e2b16710498899314298e2042!12000017711302004!rec
Or for that matter the before linked PL-AD-160 board?
https://www.aliexpress.com/item/100...store_pc_home.smartJustForYou_6000601971963.0
https://www.aliexpress.com/item/1005001810845146.html?spm=a2g0o.detail.1000014.6.32df20b9RuICrs&gps-id=pcDetailBottomMoreOtherSeller&scm=1007.40000.317745.0&scm_id=1007.40000.317745.0&scm-url=1007.40000.317745.0&pvid=ea38b322-c7a5-4245-88ac-bc35806e9c52&_t=gps-id
cDetailBottomMoreOtherSeller,scm-url:1007.40000.317745.0,pvid:ea38b322-c7a5-4245-88ac-bc35806e9c52,tpp_buckets:668#2846#8115#817&pdp_ext_f={"sku_id":"12000017711302004","sceneId":"30050"}&pdp_npi=2@dis!GBP!28.0!26.6!!!!!@211b5e2b16710498899314298e2042!12000017711302004!recOr for that matter the before linked PL-AD-160 board?
https://www.aliexpress.com/item/100...store_pc_home.smartJustForYou_6000601971963.0
Attachments
Last edited:
Well, bitten the (ok, affordable) bullet and ordered this module:
https://www.aliexpress.com/item/100...store_pc_home.smartJustForYou_6000601971963.0
Will be interesting to see how it compares back to back with the basic TPA3116D2 that precedes it.
https://www.aliexpress.com/item/100...store_pc_home.smartJustForYou_6000601971963.0
Will be interesting to see how it compares back to back with the basic TPA3116D2 that precedes it.