I can say with confidence the GMarsh-Wiener comes in at #1 out of all the 3116/3118 out there as having the best SQ for a stock amp out of the box. Other then upgrading to cinemags trafos for the inputs and possibly swapping out the decoupling caps, there isnt really anything to upgrade on this board, its already upgraded. Considering it has the thin film caps at the inputs, Abraxalito's PVCC mod, Am Avoid. dip, no hum, no start up pop or on shut down, on mute its dead quiet with ear pressed against the Heil tweeter. In other words, it has most of all the bells and whistles you would want for the exception its not 2.1. So its a "WIENER!!!" More over, as with all of the 3116 in stereo, I would like to have more power to drive the bigger speakers. Therefore I think the ultimate 3116/3118 offering would be the gmarsh-wiener in a mono block so we can see 125+ per channel. Hint Hint Gmarsh 
Last edited:
3 dogs - Charlie (german shepherd mutt), Lucy (doberman) and Cheddar Smokie (dachshund). Also 2 cats.
Here's Cheddar right now, stapled up, still on cloud 9 from the anesthesia.
Thanks for the good review. I'm charging several times what a chinese card costs for these things, glad to know people are getting some enjoyment out of my card.I can say with confidence the GMarsh-Wiener comes in at #1 out of all the 3116/3118 out there as having the best SQ for a stock amp out of the box. Other then upgrading to cinemags trafos for the inputs and possibly swapping out the decoupling caps, there isnt really anything to upgrade on this board, its already upgraded. Considering it has the thin film caps at the inputs, Abraxalito's PVCC mod, Am Avoid. dip, no hum, no start up pop or on shut down, on mute its dead quiet with ear pressed against the Heil tweeter. In other words, it has most of all the bells and whistles you would want for the exception its not 2.1. So its a "WIENER!!!" More over, as with all of the 3116 in stereo, I would like to have more power to drive the bigger speakers. Therefore I think the ultimate 3116/3118 offering would be the gmarsh-wiener in a mono block so we can see 125+ per channel. Hint Hint Gmarsh
I'd advise against swapping out the decoupling - the caps on there (5 different values) are selected as a set for good wideband impedance from hundreds of Hz (where the TPA's PSRR is excellent, and decoupling not so critical) to well up into the tens of MHz.
Though if you're planning to run the amp hard, it probably can't hurt to put some fat electrolytics after the rectifier on a linear supply - mainly to keep the rails from dropping out and effectively lowering the usable power supply voltage.
I've considered PBTL but you need 2 ohm speakers to make the best of it, and DUG makes a half decent PBTL card already. 2.1 also, but it'll be tough to make an easy crossover design that pleases everyone, and I don't think the TPA3116/8 really have enough grunt to drive a sub.
Heil AMT has a constant ~4 ohm impedance going up in frequency, that is nice, maybe less power is lost outside audioband. If erik777 feels the Dugs are more powerfull it isn't related to pbtl power I would think, but do the dugs pbtl feel more powerfull than gmarsh 3118 btl just like they did compared to Sure3116 btl ?
3 dogs - Charlie (german shepherd mutt), Lucy (doberman) and Cheddar Smokie (dachshund). Also 2 cats.
Here's Cheddar right now, stapled up, still on cloud 9 from the anesthesia.
Fine looking dog. Should be up and running soon.
Regards,
PBTL can in theory make the amp louder by providing a slight bit more headroom, but not by much.
The FETs in the TPA have a Rds(on) of 120mohm typ, which puts 240mohm in series with your load. This reduces the available voltage into the load by a factor of Rload/(Rload+240), or about 0.51dB with a 4 ohm load. In PBTL mode into the same 4 ohms, Rds(on) is effectively halved and the voltage reduction is 0.26dB instead. So a difference in peak power capability of 0.25dB, which isn't a whole lot.
You can get that back by slightly increasing the rail voltage to 24.7V instead of 24.0V.
I received my boards today! Both boards looks intact and I'm getting ready for some listening tests!
I'm hooking one amp up now to my living room speakers and I emmediately notice that I have huge grounding issue.
I am doing the same hook up as with my old tpa3116 but I get hum with the new card. I noticed that the signal - isn't tied to ground on this amp.
I use a 24V power brick with just 2 wires and my rDac uses a 2 wire 6V cheap adaptor. So can I tie the psu grounds together or should I tie the amp's - signal to ground like my other boards? The dac has output caps if that matters..
If somebody could give me a quick answer that would be great! I don't wan't to take any chances with my dac but I'm eager to hear how the amp sounds
Both boards looks intact and I'm getting ready for some listening tests! I'm hooking one amp up now to my living room speakers and I emmediately notice that I have huge grounding issue.
I am doing the same hook up as with my old tpa3116 but I get hum with the new card. I noticed that the signal - isn't tied to ground on this amp.
I use a 24V power brick with just 2 wires and my rDac uses a 2 wire 6V cheap adaptor. So can I tie the psu grounds together or should I tie the amp's - signal to ground like my other boards? The dac has output caps if that matters..
If somebody could give me a quick answer that would be great! I don't wan't to take any chances with my dac but I'm eager to hear how the amp sounds
The signal ground was tied to the power supply ground on your old board, and presumably your DAC's output ground is tied to its power ground - so when you cabled them together, the signal ground between the two boards was grounding the two cards together. The Wiener has differential inputs (the - inputs aren't tied to ground) so that ground path didn't exist anymore, and the power supply grounds of the two devices were floating with respect to each other. The differential inputs do provide some noise rejection, but they can't handle a major ground difference.
If you strap the wiener power ground and DAC power ground together, you provide that ground path and make everything happy. And ground currents will flow through that strap rather than through the signal grounds - that's a good thing.
If you strap the wiener power ground and DAC power ground together, you provide that ground path and make everything happy. And ground currents will flow through that strap rather than through the signal grounds - that's a good thing.
Night and day difference with the psu grounds connected together No detectable hum now! Hm, I thought the dac's ground were floating.. obviously not.
I've also noticed that the chip runs a lot hotter than my modded 3116 chips. At moderate volume it gets almost painfully to touch. As test setup Im running 24V and 87db 6ohm speakers with a 4ohm amp output filter. Not the best way to test this amp but my car setup isn't ready yet.
So is the best way to solder a copper heatsink to the tin blob underneath or glue/paste a heatsink on top? It could problably survive without the heatsink for a short while, but I don't want to take any chances with it since it's already a bit underpowered for my setup.
Gmarsh, do you think the chip could handle a soldering temp of 320C for 3 seconds If I solder a copper sheet to it?
No detectable hum now! Hm, I thought the dac's ground were floating.. obviously not. I've also noticed that the chip runs a lot hotter than my modded 3116 chips. At moderate volume it gets almost painfully to touch. As test setup Im running 24V and 87db 6ohm speakers with a 4ohm amp output filter. Not the best way to test this amp but my car setup isn't ready yet.
So is the best way to solder a copper heatsink to the tin blob underneath or glue/paste a heatsink on top? It could problably survive without the heatsink for a short while, but I don't want to take any chances with it since it's already a bit underpowered for my setup.
Gmarsh, do you think the chip could handle a soldering temp of 320C for 3 seconds If I solder a copper sheet to it?
My chip is not running as cool as when i had it at 14V. Now its running at 23V., and running hotter.
Maybe some adjustments in the CTS 33 (Gain/slav) could temper the heat a bit_ ? If so which of the buttons need switching.
This amp has has all switches in the off position till i figure how to run them.