I agree with just about all that. At one stage you could buy thick rectangular washers the same size as a TO220 tab to spread the pressure, but they have not been available for years.

I tried M3 PEEK screws on my build, but was unhappy with the very low torque necessary to strip the threads. In the end I used a combination of ceramic insulators (as specified by Kevin) and the really long insulating bushes, and stainless cap screws, nuts and Belville washers. My T2 was one of the first batch made, with casework supplied by Kevin, and I have had zero thermal problems.

1 hour ago, Craig Sawyers said:

I agree with just about all that. At one stage you could buy thick rectangular washers the same size as a TO220 tab to spread the pressure, but they have not been available for years.

I tried M3 PEEK screws on my build, but was unhappy with the very low torque necessary to strip the threads. In the end I used a combination of ceramic insulators (as specified by Kevin) and the really long insulating bushes, and stainless cap screws, nuts and Belville washers. My T2 was one of the first batch made, with casework supplied by Kevin, and I have had zero thermal problems.

Craig, what size belville washer did you use? And did you have the b- washer just sitting on the insulating bush or did you use them on the other side (which would seem to defeat the purpose of the b-washer)?

I just followed the recipies in this http://www.onsemi.com/pub_link/Collateral/AN1040-D.PDF .

Thanks, Craig, I see. So from the underside then. And am I seeing this correctly, that the flared part of the washer faces the nut and not the board?

The flared part of the washer faces a plain washer, and does not bear directly on the board.

Do you notice the "rectangular steel washer"? Unless someone has an obscure source, these are unobtainium.

Like these?

Well spotted. But then I saw the price for one that would do a TO220 https://www.seastrom-mfg.com/washerdetails.aspx?productNumber=5729-9-4

 

At qty 5,000 they go to $0.058 a piece.

If you split that five ways, that's about $60 per 1,000.

First of all, check that part is correct for a TO220 package - I didn't spend a vast amount of time verifying it was the right size.

I was more looking for fun, but yes if anyone was thinking of moving forward the part should be confirmed.

On 9/23/2016 at 1:21 AM, Craig Sawyers said:

Well spotted. But then I saw the price for one that would do a TO220 https://www.seastrom-mfg.com/washerdetails.aspx?productNumber=5729-9-4

 

Yeah, I hadn't looked at that. Isn't it $2.26 for a package of 100 Guess not...

So I had another member borrow my T2 to take a look at it. But I was told that the noise is around the normal level for this amp. Am I just the only asshole that finds an audible noise floor to be a dealbreaker? Or are there T2s out there that are actually dead silent (even in a room with no background noise)?

I could be going deaf (not outside the realms of possibility) but my T2, either through 007's or a 1980's pair of original low bias Lambdas is absolutely silent.

Is there some meaningful way to measure the noise at the output with a scope? To compare what I measure with other T2s would probably answer the question.

Yes I can do that. They key is to measure over that audio bandwidth and the A-weight it. I'll give that a bash tomorrow.

Mine is very quiet after using the battery adjustment method Craig articulated in this thread. Before that the amp had some consistent background noise which was livable but noticeable. 

22 hours ago, s_r said:

So I had another member borrow my T2 to take a look at it. But I was told that the noise is around the normal level for this amp. Am I just the only asshole that finds an audible noise floor to be a dealbreaker? Or are there T2s out there that are actually dead silent (even in a room with no background noise)?

I have never heard a stat amp where you couldn't tell if it was turned on or off - ie every one, where it is absolutely silent, will give some extremely light background noise - whether it was a Stax, Mjolnir kgsshv, t2 or carbon. The difference is barely perceptible, but is easier to detect if you place your hands over your headphone cups. Like Andy, my experience with the T2 is that it is extremely quiet. The Carbon is a touch more quiet, but not significantly so.

5 hours ago, Craig Sawyers said:

Yes I can do that. They key is to measure over that audio bandwidth and the A-weight it. I'll give that a bash tomorrow.

Thanks, I should ask if it'd be something I could repeat with my rigol DS1054Z though.

2 hours ago, blubliss said:

Mine is very quiet after using the battery adjustment method Craig articulated in this thread. Before that the amp had some consistent background noise which was livable but noticeable. 

I tried a while ago to adjust the battery to see if it would have any effect on the noise. It'd sooner go out of whack than change the noise. Unless there's something I'm missing I'm not sure that's the solution.

1 hour ago, GeorgeP said:

I have never heard a stat amp where you couldn't tell if it was turned on or off - ie every one, where it is absolutely silent, will give some extremely light background noise - whether it was a Stax, Mjolnir kgsshv, t2 or carbon. The difference is barely perceptible, but is easier to detect if you place your hands over your headphone cups. Like Andy, my experience with the T2 is that it is extremely quiet. The Carbon is a touch more quiet, but not significantly so.

Right, thanks for clarifying. Without another T2 to compare to it's hard to say whether the noise on mine is higher than what's normal. All I can say is that it's high enough to bother me.

Edited October 6, 20169 yr by s_r

Looking at the spec for the Rigol, you have a choice of 50MHz or a bandwidth limit of 20MHz. Since the noise voltage goes as the square root of the bandwidth, and assuming you have selected 20MHz bandwidth the noise voltage you measure will be 32 times higher than you would get for a 20kHz audio bandwidth. A-weighting applies a correction for the way the ear's sensitivity changes with frequency for quiet sounds.

I consider my T2 to be absolutely silent. I can't hear any noise, might be because my old ears though.

@s_r Did you try that specific adjustment method I mentioned?  From Kerry: 6.55V at 22k resistor using 10k pots while keeping 2k centered, then fine tune with 2k.  It is different from the original adjustment methods.  I do remember having the experience you mentioned, of the adjustments doing nothing for the noise, until I used this other method.

Edited October 6, 20169 yr by blubliss

I think that's how I did it the first time. Just in case I did it that way just now, but no change in the result. 

Today, my DIY T2 has successfully come to life!  After a year on and off working on this amp on my kitchen table, I was able to power up today and get music playing.

All 4 batteries are sitting around 743V.  The +/- voltages for each channel are surprisingly well balanced, and both L and R sit about +10V offset at power-on and gradually get down to about 2-4V offset at steady state.  I measured the heatsink temperature -- looks like it's steady at 124F/51C after about 3 hours continuous burn-in.

A shoutout to GeorgeP for enabling this build--one of my good buddies helped me reached out to you about a year ago for the case/amp board/torroid set.

Thanks KG, Spritzer, and everyone else who posted all the useful information on every aspect of the amp on this thread.  I learned a lot through the process of building this amp!

One interesting thing I found while testing the PSU was that when I had both HV toroids hooked up and powered, the combined inrush blew through my 3.15A slow blow fuse (at 120VAC from the wall).  I take it that the inrush of the 2x HV toroids (at least on my toroids) are pulling closer, but not quite, to 4A, since I was able to power up one HV toroid with a 2A fuse that I had left over.  I have a 5A slow blow in there now, and so far no issues!

-Kuen

 

Congratulations!  

Well done. Congrats and enjoy!