SI PCB faux pas?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Just started tearing into my latest shipment of T-amps and found that the first two I liberated from the cases had a strange look to them. It is the look of no lumpy solder flowed on the boards. Its been understood that this was for heatsinking, and so its either been deemed unneccessary or its a manufacturing oversight.

Anyone else seen this on newly acquired boards? Mine are from PE.

amt
 

Attachments

  • solderless1.jpg
    solderless1.jpg
    69.9 KB · Views: 2,043
Ex-Moderator
Joined 2005
i've been doing the THD tests and such on one of these new ones (see the stock SI spectrum thread for all that jazz) and hadn't even known there was a difference with the new ones. twice now i've triggered the thermal limit and the amp has shut off. the first time it happened i thought i killed the amp, but i wasn't doing anything wrong other tahn driving it pretty hard doing the intermodulation distortion test. it seems my modified amp (has the second revision board) has never gone into thermal protection when doing the same tests. granted, it *might* have better ventilation in the aluminum case, but it can't be that much better next to the heatsink of the voltage regulator, which gets pretty damn hot when the amp is driven hard. i would argue that the solder slug is needed for better cooling if you drive your amp hard enough or have a difficult 4 ohm set of speakers. that or some kind of smaller heatsink attached to the area where the solder slug should be.
 
I just looked to see if I could attach one but the chip actually floats above the board. The perforated holes need to be filled with solder until they contact the bottom of the chip. Can a solder bridge be built from the little metal taps that protude from the sides of the chip, w/o damaging something?

And I wonder if tripping the protection has any ill effects on the sound like it seems to with ICs.

amt
 
Ex-Moderator
Joined 2005
amt said:
And I wonder if tripping the protection has any ill effects on the sound like it seems to with ICs.

amt

i ran the THD tests after the protection circuit had tripped and obtained practically the same results as i did before it was tripped. whether or not that correlates to the amp sounding worse i don't know. i haven't actually listened to my test amp because i was just using it for determining baseline measurements to compare the modded amp to.

it would take a LOT of solder to make a solder slug for one of these things. i only have very fine kester 44 rosin core and that would take forever to make the slug assuming my iron could get the thing hot enough for that long... maybe the'yre just cutting back on the cost of solder to make a couple extra bucks per amp :confused:
 
Administrator
Joined 2004
Paid Member
This can't be good. The Tripath docs call for the solder slug for cooling. Obviously the PCB is made for it. Even with the slug, the chip gets to hot to hold.

There must have been a slip up in the factory. Odd that the boards got all the way thru the assembly like that.

AMT, I know that you emailed S.I. about the problem, it will be interesting to see what they have to say, if anything.
 
Mine too

Just opened one of my new last week ones. It has the naked board also.
It sure looks better than the solder slug. The quality of the soldering seems a little better also.
Hope it is not an issue. Maybe the holes in the board will let cooling air to the bottom of the 2024. I cannot tell by looking, but on the old ones does the slug touch the bottom of the chip?

George
 
Hey amt, why don't you try using some of this:

http://www.arcticsilver.com/arctic_alumina_thermal_adhesive.htm

to glue a heatsink to the back of the chip. Should conduct the heat but not the electricity.

Todd

Edit - It may be a mistake but I think its cool you can see through the PCB and see the heat slug on the backside of the chip. Also, it should make it less confusing to trace out the circuit. BTW, in case I wasn't clear I meant use the compound to fill in the hole leading to the heat slug on the back of the chip and then attach the heatsink with the compound.
 
Ex-Moderator
Joined 2005
electrically conductive materials are almost always better at conducting heat as well as electricity. that's why tripath says to use a solder slug as a heat sink on the back of the TA2024 chip. i've seen another chip with a heat sink on the opposite side (cant remember if it's a TA2024 or not). it almost seems that would've been a better one to use for our purposes, but not so for the space constraints of the molded SI chassis. but i guess that's how it goes in the DIY world. it seems as though the 41hz kits will be all the rage very shortly. especially if the board is reworked to include discrete signal path components (hint hint ;) ).
 
Ex-Moderator
Joined 2005
i don't doubt their thermal paste or adhesive, i'm sure it's just as good or better than others. i'm just saying that it literally doesn't have nearly as good thermal conductivity as solder does. on arctic silver's site it says their arctic alumina thermal compound has a thermal conductivity of >4W/m*K. well regular old 60/40 solder has a thermal conductivity of around 57W/m*K, making it a much better conductor of heat. it's probably cheaper than that thermal paste too, that's why tripath says the solder slug is a must to increase the little chip's ability to dissipate power, and also why it has me :confused: that SI said the solder slug wasn't necessary when i've already driven the amp hard enough to go into thermal protection.
 
Well, thats not good...

I dont know the exact reason for the change, but it was said to not be needed. I didnt get the specifics as to R&D etc, but they are sticking with this as far as I know.

I took a good look at the chip and its raised off the board just enough so that I could slip a piece of aluminum under it.

Dont laugh, but maybe this could be a solution of sorts. I will give it a try tonight and see what happens. Its a strip of a coke can and it fits tightly between the bottom of the chip and the board.

amt
 

Attachments

  • cansink2.jpg
    cansink2.jpg
    46.3 KB · Views: 1,181
BWRX said:
on arctic silver's site it says their arctic alumina thermal compound has a thermal conductivity of >4W/m*K. well regular old 60/40 solder has a thermal conductivity of around 57W/m*K,

Wow, I didn't realize it was that much worse than solder. But it should still be better than air and won't have the electrical conductance worries of solder, right?

Todd
 
Ex-Moderator
Joined 2005
todd, yes - air has a much lower thermal conductivity that depends on the ambient temperature and the humidity. googling "thermal conductivity of air" led me to a study that says air at 20 degC and 20% humidity is about 0.026W/m*K.

amt, i laughed but also think that's a pretty clever idea! that's "junkyard engineering" at its finest ;)
 
OK, so what are D1, D2, D3, and D4? I don't have those on my board. Are they necessary for good sonics/ proper funtioning? On my (I believe) version 2 board, those areas are covered with the solder slug.

It's interesting that SI changed their whole solder slug approach. At this price point, they probably don't worry about squeezing the best out of the litle amp, not when they can save a good deal by skipping all the solder. Please keep us posted on any performance differences you find.

Best,
KT
 
Diodes

KT said:
OK, so what are D1, D2, D3, and D4? I don't have those on my board. Are they necessary for good sonics/ proper funtioning? On my (I believe) version 2 board, those areas are covered with the solder slug.

It's interesting that SI changed their whole solder slug approach. At this price point, they probably don't worry about squeezing the best out of the litle amp, not when they can save a good deal by skipping all the solder. Please keep us posted on any performance differences you find.

Best,
KT


Tripath lists them as needed. The early SI's did not have them. The pupose is to prevent back emf coming back from the speaker. This CAN drive the output to below zero.
There are lots of SI out there without them. Never heard that just playing music blew one up. Usually it is hooking up a new power supply, or a modifing the circuit boards. Lots have gone up in smoke after being modified.

George
 
Well Ive run the new board without a case and in the stock case and it does get fairly hot. Im using 96db 8ohm speakers and after several hours of continuous use, Ive yet to trip the thermal switch. But I will admit that at comfortable listening levels, Im probably keeping it within the 5watt range.

My guess at this point is that SI has "de-rated" his amp so as to make it a little less home hifi friendly. Bet the next generation version thats $120 has heatsinking.

Im now playing with a heatsink concept that attaches to both sides of the chip via the slug taps that are visible on each side. The sink will contact them and route over the top of the chip. An upside-down U. It will probably use a screw, running from each side to clamp the sink to the tabs. I checked the amount of heat being dissipated though the plastic case versus the bottom side, where the slug is and Im not sure that its worth mounting a heatsink on it. Pulling heat from the slug is the key to cooling this chip.

amt
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.