[Beefy]

If he is following the BOM, it would be these kits. Pretty standard steel screw.

 

[mypasswordis]

If the magnetic field is constant due to a constant current (still a function of r), maybe the heatsinks are very slightly vibrating somehow and there's some paramagnetic effects with the magnetic permeability of aluminum being slightly higher than that of air? I'm picking at straws here... will think on it some more.

 

[TimJo]

Quote:

Originally Posted by runeight /img/forum/go_quote.gif And after more testing it is only the two CCS heatsinks at the center of the board that are causing the problems. They were coupling the two channels to each other which is why they looked exactly the same. I am wondering if any of the rest of you will actually see this problem. I bet that some of you, maybe all of you, won't. And it may not appear at all on the tube amp.

Just to be clear, you are referring to Q11R & Q12L - correct?

If that's so, have you tried leaving them floating, but separating each channel by snapping the amp boards apart, to see if this prevents the oscillation?

Quote:

Originally Posted by runeight /img/forum/go_quote.gif I am not really sure how this is working. The currents flowing in the devices are very small and the CCSs are keeping the currents constant. Thus, there shouldn't be any current-created time-varying magnetic fields to induce current in the heatsinks which can then radiate to nearby heatsinks. In fact, I'm not sure that I believe that this is what it is. I cannot, however, think of a logical and sensible explanation at this moment. But I do know that if I let the two heatsinks float the amp pops into oscillation within a second or two. Aluminum is actually a pretty good conductor (used in power transimission lines because of its light weight) so there is no trouble inducing currents in it. Perhaps someone reading can offer an explanation.

Your guess makes perfect sense theoretically. It is interesting though that the plastic bodies become both transmitters and receivers in order to induce the instability.

So is the oscillation likely a harmonic of some common mode current pulse being injected into each channel, based on the switching of the transistors within each CCS pair?

 

[runeight]

Quote:

Originally Posted by mypasswordis /img/forum/go_quote.gif If the magnetic field is constant due to a constant current (still a function of r), maybe the heatsinks are very slightly vibrating somehow and there's some paramagnetic effects with the magnetic permeability of aluminum being slightly higher than that of air? I'm picking at straws here... will think on it some more.

Possible. I don't hear anything if I put my ears close to them but that doesn't mean much and the oscillation continues if I grab the heatsinks.

TimJo, it is Q11R and Q12L. I thought about separating the boards but I don't want to break them apart. Eventually I may have to do this test. It's peculiar though that this does not involve the other set of heatsinks in the middle of the board. I'll post a picture in a bit.

I don't know exactly how it would be happening, but the CCSs are maintaining constant current during the oscillation. At least as measured at the control resistor and every other point inside the CCSs. The only spot on the CCSs where the scope shows oscillation is at the collectors of the HV transistors.

luvdunhill the transistors are attached with screws and thermasil pads, flat washer, lock washer, and nut on the backside.

Gents it is very hard for me to believe that this is EM radiation from the heatsinks. And even harder to believe any kind of capacitative coupling. Yet, I pull the clipleads off and the amps go crazy. Put them on and they are quiet as a mouse.

 

[cetoole]

Pickup of airborn RF, which then gets capacitive coupled to the CCS transistors? Definitely odd though.

 

[TimJo]

As another test, I wonder if you could take a sheet of some permeable metal connected to star ground, and position it between the two heatsinks, and see if by shielding them you can get it to work.

I'm not sure if it would work, or what it would prove, other than more evidence to the theory that it is wave-based coupling between the two heatsinks.

BTW, what kind of frequency are you talking about on the collectors of the KSC5042's?

 

[runeight]

The mystery deepens. I am a bit wrong about the heatsinks. It turns out that grounding either heatsink on a single channel stops the oscillation. And, likewise, having one channel's heatsink grounded while the other floating causes the floating amp to oscillate.

So I am wrong about this being a channel interaction. But it is still related to the heatsinks.

Here's a not-very-flattering photo.

The board is much better looking in person and not on a magenta colored towel that covers the work area.

 

[runeight]

cetoole/TimJo we cross posted. It could be local radiation but then, I think, if it were powerful enough to get through the heatsinks it would get into everything when the inputs are floating. But it doesn't.

TimJo, it's about 20kHz. Very low frequency for parasitic oscillations that might occur, say, from a mosfet with no gate stopper. This is why it seems almost like a multivibrator to me. Also because the oscillation slams the outputs to the rails producing almost a square wave, like the amp is flip-flopping.

Now, with the new information, it could be some kind of coupling to the interntal components in the alleyway between the heatsinks. But, heck, I don't know.

 

[runeight]

One more thing. I've had it running for over an hour now. Offset/balance are extremely stable, less than 1V.

 

[Beefy]

Quote:

Originally Posted by runeight /img/forum/go_quote.gif Here's a not-very-flattering photo.

Looks pretty damn good to me...... very imposing with those big sinks

 

[BoilermakerFan]

Quote:

Originally Posted by runeight /img/forum/go_quote.gif cetoole/TimJo we cross posted. It could be local radiation but then, I think, if it were powerful enough to get through the heatsinks it would get into everything when the inputs are floating. But it doesn't. TimJo, it's about 20kHz. Very low frequency for parasitic oscillations that might occur, say, from a mosfet with no gate stopper. This is why it seems almost like a multivibrator to me. Also because the oscillation slams the outputs to the rails producing almost a square wave, like the amp is flip-flopping. Now, with the new information, it could be some kind of coupling to the interntal components in the alleyway between the heatsinks. But, heck, I don't know.

Is it only the SS version? I was planning to use 2-1/2" sinks, but now I wonder if that would make it worse. However, I'll screw mount mine so they'll be rigid and all will be grounded at the star.

Is there anyway you can remove Q11R/Q12L and bottom mount them? Just to see if it goes away with ungrounded sinks when their 180 degrees apart?

 

[runeight]

Thanks. I think it's ok for me to say that these are really good looking boards. I may have to case one of these up, but then I'll have to buy a set of stats. Egads.

Here is the latest beta list including quads and caps. Please check your stuff and then send me email with your shipping address and prefered method. If you don't say, US packages will go first class or priority mail. Overseas will go whatever the business class is for 5-7day delivery.

alex at cavalliaudio dot com

 

[runeight]

Quote:

Originally Posted by BoilermakerFan /img/forum/go_quote.gif Is it only the SS version? I was planning to use 2-1/2" sinks, but now I wonder if that would make it worse. However, I'll screw mount mine so they'll be rigid and all will be grounded at the star. Is there anyway you can remove Q11R/Q12L and bottom mount them? Just to see if it goes away with ungrounded sinks when their 180 degrees apart?

I haven't built the tube version so I don't know. And I don't have the parts for it now, but I'm betting that because the arrangement of heatsinks on the tube board is different you won't have this problem.

I suspect that bottom mounting would fix the problem because it would get the heatsinks away from all the other components. But I guess I can't say for sure.

 

[runeight]

Oh, and when you send me your shipping address please tell me your handle so I can match to the spreadsheet. I don't think I know everyone of you yet.

 

[BoilermakerFan]

Man, I blew by 1,000th post and didn't even notice!

If anyone is interested, I emailed Jack at Electra-Print about a custom transformer for the Hybrid so I can run a few accessories and power the heaters with DC. Here's the email I sent:

Quote:

Jack I'm interested in a custom power supply transformer for a hybrid tube amp. 120Vac primary or 120/240Vac primary A 500VCT secondary at 75mA - 80mA (assuming that's how to designate 250-0-250 secondary) A 24Vac Aux filament winding, 1A - will be rectified and regulated down to 24Vdc for a small SS attenuator and input switching relay. A 12.6VCT - 1.5A - I plan to run 6.3Vdc to my heaters, but need the 12.6V to rectify and regulate down. The tubes draw about 2A at 6.3Vac. The dimensions are tight on this build so if you could let me know the transformer size as well, I would appreciate it.

And his reply:

Quote:

Brian, This transformer is $103.00 each with standard 120v primary, with dual primary it $133.00 each, standard bell ends vertical mount. Plus shipping. The dimensions or footprint, is 3" x 3 1/2" x 3 7/8" tall.

I have two other customs from E-P and they are beefy. Jack's communication was great and they arrived a week earlier than expected. This transformer will deliver a little more current than the 369JX, plus it will provide power for my OptiVols and relay input switching if I build it in, as well as leads for 6V DC (or 12Vac for any crazy tube rollers) heaters with plenty of current head-room. I plan to keep my SylveXStata amp so I'm ordering only 120Vac primaries, but I know some of you guys need 240Vac or want dual voltage capability... Only $20 more than Mouser and an extra 10-15mA of power, plus saves me space as only one trafo.