ECP Torpedo III [indexed in first post]

Honestly, I'm no expert on a Flir device. However, I do have a contractor working for me right now that's using one on our labs where I work. For surveying a general area to determine hot spots, there is no equal. For determining exact temperatures, there may be a lot more involved than getting the target in the sights. From what little I've read, there are a number of adjustments that must be made to make accurate, specific temperature measurements - not the least of which is emissivity.

This is a great article that describes the potential pitfalls of a Flir device (from a Fluke engineer):
http://www.reliableplant.com/Read/1...-difference-between-apparent,-actual-ir-temps

Whether it's a fundamental inaccuracy in reading specific temps in a complex environment with multiple, dissimilar materials - or the fact that Doug's PCB design for the Output CCS boards has numerous holes that a Flir will see as a points of perfect emissivity, I don't know. However, the article raises a number of issues in using a Flir device for obtaining precision temperatures.

I would note that our contractor had temperature scales for the Flir pics on our laboratory spaces that had no basis in reality. The bottom of the scale (dark blue to black) meant that there were areas in our labs that were down to 40 deg. F., which is pretty much impossible, even at the throat of an air conditioning outlet. So again, looking at the photo itself was an unparalleled tool to pinpoint the hot spots and lack of air distribution in a high-heat load environment, but as for precision temperature measurement - not so much.

I made a number of tests in temperature measurements with heat sinks in two different positions on the CCS boards and without. All measurements were done with the amp completely cased up and running in one instance, up to 3-1/2 hours of running before the temps stabilized. These measurements were made with a Type K thermocouple in direct contact with the plastic body of the Icky chip transistor, and connected to a Fluke 179.

Measurements were made with the amp running and recorded when the temps stabilized (stopped rising) on the Output CCS board that is directly in front of capacitor C5. That one seemed to be the most constricted and developing the highest temp. Although, you might tell us which one for certain. It's possible that we misinterpreted which one you seemed to find was the hottest. Anyway, all measurements were made with the amp completely cased up and with both endplates attached. This particular T3 had all tweaks - Mundorfs, Tube CCS boards, Output CCS boards, and 12AZ7 tubes, running at 110V line power.

Without Heat Sink (stock Output CCS boards) - 106.5 deg. C.

With Heat Sink (thermal tape attachment) directly on Icky chip plastic body - 96.5 deg. C.

With Heat Sink (thermal tape attachment) directly on PCB above the Icky chip, with bottom of heat sink touching the Icky chip metal tab - 89.3 deg. C.

Doug also supplied me with some 5-minute Arctic Silver epoxy. In my opinion, this is the best, permanent solution - only one drop per heat sink would be needed. Everything I've read seems to say the thinner the glue joint, the better. My thoughts are to supply you with both, however. For those who find the epoxy too messy, the thermal tape will do a good job. With two-part epoxy, there should be no problem of it curing if I send it to you in little bits inside ziplocks and you mix it yourself. That's unlike the Dow Corning 748, which started curing on contact with the air.

Anyway, it's a fairly easy and inexpensive enhancement to bring down the temperatures on a long-term basis.

 

Not much meat in the PCB above the chip right? Which heatsinks are these?

I have a bunch of tiny tiny tape on heatsinks but they are missing in action just when I need them.
https://www.amazon.com/Cosmos-Aluminum-Cooling-Heatsinks-cooler/dp/B007XACV8O

 

Thanks, @TomB , now I'm wondering what the temp spec says for the devices on the boards. It's really hard to see the part number in this pic. Transistors can run hot, though usually. I had a transistor short out in a case once and I figure it took 5 minutes for smoke to trickle out before I turned it off. A heatsink pad had lifted off the PCB. But after fixing the short, the transistor still worked. And I bet it got blazing hot.

Regarding the hottest, are you referring to R19? That was my second hottest. I believe it was R21 that was my hottest, which is in the same orientation, right? I'd have to look at it later when I get home to remember for sure.

That's a very interesting article on emissivity. Thank you very much for setting me straight there. It indeed looks like the Flir would report inflated numbers. I guess it can't be taken at face value.

 

The top one is a 10M45. The bottom a LM334.

 

OK, I guess @TomB mentioned the "Icky" chip which is the 10M45... maximum rated junction temperature is 150C. With Tom's measurements showing <110C worst case, I'm not too concerned. I think a 125C, 10 year lifetime is common. I'll probably have a DSHA8 by then anyhow.

 

Sorry. It's an IXYS chip. I call them Icky chips because I can't ever remember the name ... or pronounce it. Actually, the chip that Doug uses is IXCY, signifying it's a TO-252 rather than TO-220. I can't say IXCY, either, but it seems awfully close to "Icky."

 

I'm actually more worried about that cap right next to it.

 

If we're concerned about the cap, then why don't we put Tom's thermocoupler on the cap itself instead of on the opposing face of the CCS board ? The CCS board is oriented with the transistors away from that little electrolytic aluminum cap I think (right?), and my thermal imaging was showing the back side of the CCS board around 20C cooler for what that's worth. Tom's thermocoupler could remove all doubt.

 

These are the cheapo aluminum $8 stick on heatsinks from Amazon. You can actually get 4 heatsinks per board but this pic shows 3. Last one is the LM334 and if you go for it watch out for the transistors closest to them. Fit is OK, maybe not perfect. R20 is the hardest one due to the tight spacing, my precision tools helped in that regard.



Precision tools!



Make sure they are flush for max surface contact.

 

Mr.Sneis,

I gave your post a "like," but the Precision Tools deserved a shout-out post, too!

 

Had yet another awesome meet with @bazelio and @drfindley. Here's a new ad campaign for you: "buy/build a T3, comes with friends!" The meet allowed for a few very interesting comparisons, but relevant to this thread was another adventure in tube rolling on the CCS-modded T3. Here are some impressions, which others can add to.

Whereas we previously favored tubes in the 12AZ7 family, things now were a bit more complicated. These tubes still sound amazing in the T3 amp, and can be found rather cheaply. In fact, we compared RCA 12AZ7 tubes to the Amperex 12AZ7 we all liked in a previous meet, and the RCAs were pretty much the same, if not better. This is awesome news, given that the RCA cost about 1/4 to 1/2 what the Amperex cost, and are readily available on the 'bay (unlike the amperex). The pre-CCS T3 benefited from the 12AZ7 family for it's punch and attack. The post-CCS T3 is insanely dynamic and punchy, and as a result isn't in as much need of those tubes. In fact, tubes seemed to affect the amp way less after the CCS mods than before them. The 12AZ7 tubes provided the most punch, tubes like valvo or telefunken were slightly more neutral but still sounded quite punchy on the T3, likewise the Brimar tubes (probably still my faves) were slightly warmer than neutral, but still punchy on this T3 You see what I'm getting at.... The amp just does what it wants, and what it wants is to convey awesome dynamic and engaging sound.

EDIT: Baz babes thanks for dinner and the engaging conversation as ever

 

Great stuff @brencho . Some additional thoughts and obligatory pic. The highlight of the evening, audio-wise of course, was the opportunity to listen to the Focal Utopia through the T3. Quick impressions of that amazing headphone posted to its thread. Another highlight of the evening was watching @brencho eat 1 dozen meatballs for dinner and then announce he needed an In-And-Out burger after the meet.

So, yeah, we rolled through more tubes, I think for the final time. Those were: Amperex 12AZ7A, Amperex 12AT7, RCA 12AZ7, Raytheon windmill getter 5751, Telefunken 801S, Valvo 6201, Brimar 6060. Brought but not listened to: Sylvania 12AZ7, Raytheon 12AZ7 Utopias were the only headphone used this time.

Some observations: Yeah, as @brencho mentioned, tubes are now having a smaller impact on that final sound signature you get from the T3 since installing the output CCS boards. The T3 seems to have it's own distinct sound signature and just imposes its will now. The two Amperex tubes are so similar, and while the AZ version probably has slightly less coloration and slightly tighter bass, I still think the AT version is a top recommendation of mine for this amp. They are both excellent with their natural timbre and competent extension top to bottom. Beware, though, these tubes are punchy and some might start to feel that a T3 with CCS boards combined with these tubes gets a little fatiguing. They are not what I'd call harsh, ever, but they keep you on your toes at all times. So that's Amperex, next RCA 12AZ7. What a surprise. I don't know if I even spent $20 bucks on the pair. These turned out to be a tube I'd definitely spend more time with. I'd call them a bit more euphonic than the Amperex, probably with a bit less attack (probably in a good way), and still plenty detailed and good tonal balance and reach. Another winner. Then, on the downside for me, the 801s was still as I had remembered - lacking in engagement factor with notably less punch and realism. I didn't care for it yet again, but I can see it possibly working for someone who might want to back off the forcefulness of the CCS output board sound. Also the vaunted 5751 windmill getter Raytheon was a let down. While it seemed to have glorious mids on some tracks, it had a strident harshness in certain regions that I couldn't live with.

 

Terrific impressions.. thank you, fellas! That pic is great.. It's like the SBAF "ring of honor" .. BW, Gungnir Multibit, Mutec, HD650, TIII, Modi Multibit, Vali2, Utopia, etc. lmao.

So, I get the sense that tweaking the TIII's low z output for the Focals the isn't necessary? Any perceivable issues with extension or control/damping?

 

@TomB - thanks for the spacers! Looking forward to trying out the Mundorfs from @bazelio at some point.

 

@FlySweep - no perceptible issues. I was wondering if there'd be any bass rolloff, but I didn't feel like I could hear any. I might still be tempted to socket OL, OR, OG and experiment with a couple resistor values there, but I might not be. I do still think output power loss is a potential concern in doing that. There was volume headroom with the Utopia but not a lot. We were probably listening at 66% (?) on the pot?

@badf00d - I'm slammed. I forced myself to take a break last night and meet with my audio buds, but give me another week to get to the post office with the Mundorfs.

 

Also.. how did the Modi Multibit pair with the TIII? @brencho had some nice thoughts on it.. I'm curious to hear from the rest of the guys at the meet? Anything objectionable (however minor) about the pairing? I sense they'd go well together, but I wonder how the Modi Multibit's soundstage (which, to my ears, leans to the intimate side) works with the TIII (which is said to have a tighter, intimate soundstage, as well)?

Did either of those TIII's with the output CCS boards have the Amtrans resistor in them?

Unrelated to the meet.. how do people like the HD800SDR paired with a fully modded TIII? More specifically, what do people like or dislike about the HD650M/TIII combo vs HD800/TIII combo (DACs being the same, of course)?

 

@bazelio - no worries. I wouldn't have had time to put them in anyway. Just looking forward to hearing what you have been. Your generosity is much appreciated.

 

My hearing is starting to get better so I can actually enjoy listening now.

Subjectively speaking with the heat after about 2 hours... with my 4x per board heatsinks and the tubes spaced up with socket savers + case on, the T3 is less hot than a stock Lyr 2 running LISST and significantly less than Lyr 2 with tubes.

This could change with more time on but not bad at all.

When I ran with the top off, the tiny heatsinks were definitely too hot to touch for too long.

 

The scary part is that, to Mr.Sneis, a plastic fork is a Precision Tool.

 

This just in: The T3 is a hit! Great validation for a great amp!!