Tube Lab SE Problem

[Doc Jr 8156]

Hello,
I had my TubeLab SE (with JJ 300Bs) running for more than 3 months now but recently I noticed that after 2-3 minutes of playing music the 300Bs will turn off while tyhe rectifier and the input tubes are still on. Never had this problem before. The 300Bs are biased at 75ma. The input tubes are at 170 volts. B+ at 340 volts. Can someone shed me some light on what is going on here? How can I narrow the origin of this problem? Thanks for your help.

Doc Jr 8156

 

[BillEpstein]

Isn't that the power drive overheating issue? Got a big enough heat sink?

 

[Sendy]

... the 300Bs will turn off while tyhe rectifier and the input tubes are still on...

It means 300B no heater voltage and other tubes yes ? May be defect on socket, may be on tube pin (lead free solder), or defect inside tube.

 

[DigitalJunkie]

IIRC it's an overheating issue with the Sharp regulator used for the filaments. Ya need a bigger heatsink.

 

[Doc Jr 8156]

Thanks Bill and Sendy. Bill, the heatsink in there now had been there for 3 months and never had this problem before but I will investigate that. Sendy, maybe defect in the tube sockets? The input tubes and the rectifier remained lit when the 300B shuts off. The tubes were fine when tested on my other 300B amp though. Will it help if I'll install independent heater transformers? Thanks again.

Doc Jr 8156

 

[Sendy]

If you have another 300B, can you exchange and test amplifier with other piece 300B ?

 

[tomchr]

Doc,

If by "300B shutting off" you mean that the filaments turn off, read on.

There are many possible reasons for why you didn't see the issue earlier but is seeing it now. Maybe the heat sink has more dust on it now. Maybe the 300B filaments have aged and draw slightly more current. It is also possible that the AC line voltage has changed slightly, so now the Sharp regulator has a larger voltage across it and dissipates more power. It sounds like it's a marginal heat sink design to begin with.

Put a bigger heat sink on it. Use thermal grease between the regulator package and the heat sink - or at least a thermally conductive silicone washer. The stuff is available at DigiKey Corp. | Electronic Components Distributor | United States Home Page and elsewhere.

A new heater transformer will not solve the problem. A new heater transformer will not change the power dissipated in the regulator.

~Tom

 

[Doc Jr 8156]

Hello Sendy and tomchr,
Thanks for responding.

Sendy, I will try other 300Bs n there.

Tomchr, the heatsink in there is only 1X1/2 inch square. I'll get a bigger one. On the original sink, I used both silicone wafer and thermal grease. What if i'll change the regulator to a 7 amp one instead of the 5 amp in there? Godspeed.

Doc Jr 8156

 

[rmyauck]

Try Arctic Silver used for computer CPU's etc. Works better than white grease. A computer CPU heatsink with fan may be worth a try. They can be found for free from discarded computers. Larger ones recommended. When cleaning fan and heatsink with compressed air or vacuum cleaner, hold fan so it doesn't spin as dirt seems to get in bearings easier when spinning.
Good luck,

Randy

 

[rmyauck]

I forgot to say you will need power to run fan. DC tube heater might have enough voltage. 5V might be enough and it will be quieter too. I believe fans run on 12V.

Randy

 

[Evenharmonics]

Larger heat sink will be a better choice than having to add a fan.
As mentioned already, check the connection to 300B. It's possible that when things heat up and expand (movement), it may loose connection. I once had a bad solder joint that caused same problem. It looked fine visually but re-soldering fixed it for me.

 

[pointy]

has your home heating system come on.............................the amp may be near a rad

 

[Doc Jr 8156]

Hi Evenharmonics,
Thanks for your input. I'll find a bigger heatsink and will check the connection. I wonder how big were the heatsinks in other TubeLab SE amps using 300Bs. TubeLab SE owners, please post a picture if possible.

Pointy,
I'm in Southern California. Not that cold yet. No need for house heater at the moment.

Godspeed

Doc Jr 8156

 

[Evenharmonics]

Mine is about the same size as what's shown on Tubelab's website for his 300B TSE.

 

[Tubelab_com]

Can someone shed me some light on what is going on here?

IIRC it's an overheating issue with the Sharp regulator used for the filaments. Ya need a bigger heatsink.

There are many possible reasons for why you didn't see the issue earlier but is seeing it now......It is also possible that the AC line voltage has changed slightly

The 300B tubes losing filament power is caused by the Sharp filament regulator shutting down due to thermal overload. The reasons tomchr listed for the delayed onset are valid, and I have a few more, but I am finding more and more issues in line powered equipment caused by dirty power and escalating line voltages.

30 years ago our local power company (FPL) had published specifications for power line quality. I had to check this stuff out when I was doing factory test line support at a big electronics manufacturing plant. The voltage was specced at 115 to 120 volts, the distortion on the power lines was quoted at 3% max, and the frequency was good to +/- 1 second in a 24 hr period (that was the problem for us).

Today I measure anywhere from 120 to 125 volts at 8 to 12% distortion. These numbers may change a lot in a small time period. What does this have to do with a tube amp? A lot. I have been working on the Simple P-P amplifier design. There is one running on my bench daily. The B+ voltage can be anywhere from 325 to 360 volts without changing anything. The distorted line voltage coupled with todays cost reduced power transformers that tend to run on edge of saturation makes the issue worse.

It is likely that the heat sink on your regualtor chip was marginally sufficient and some small change has caused the regulator to go into thermal shutdown. This could be related to power line quality, a degradation in the thermal transfer from the regulator to the heat sink, a degradation in the heat sinks ability to radiate heat, a degradation in the regulator chip itself, or (more likely) a small contribution of several of these factors adding up.

The thermal insulating washers do degrade slowly with time, so does the white grease. A heat sink will loose efficiency if dust collects on it, or the airflow around it changes. The Artic Silver compound mentioned above does not degrade with time (manufacturers claim) like the white grease (degradation shown on Wakefields data sheet).

Try directing air from a small fan across the heat sink temporarilly to see if this is the issue. If it helps, a larger heat sink is probably the answer.

I have found that there is no perfect answer to the exact heat sink size requirements. The thermal resistance of a given heat sink can vary over a very wide range depending on its surroundings. A heat sink depends on the "chimney effect" to generate a small airflow around itself. If there is something near the heat sink blocking the airflow, the heat sink can not function. An extremely large heat sink in a closed box will still eventually overheat. The board shown below used a small 1.5 X 1.5 X 1 inch heatsink on the regulator. It worked fine when the board was sitting on the bench. Once mounted inside this small closed box, I had to add this monster (the green object) which is about 2.5 X 2.5 X 2.5 inches. I also had to drill some ventillation holes in the box.

 

[Evenharmonics]

I drilled a few holes at the bottom and top of chassis where my heat sink is located for air flow and it seems to work well.

 

[Doc Jr 8156]

Thanks Tubelab and Evenharmonics. I think i need a "massive heatsink". Where to I get them? I was perusing DigiKey and I wanted to ask if Blue Ice heatsinks are okay to use. I was thinking of the biggest Blu Ice heatsink to use cause the boards on my amp are in the open anyway. Godspeed.

Doc Jr 8156

 

[rknize]

Any heat sink will work as long as it has significantly more surface area than the "stock" one (and good airflow around it). I had the same issue as you when switching to 300Bs from 45s and was left scratching my head when the amp would fade away until I stood there and actually saw the filaments go out. This was all on a breadboard, so I just clamped a squangle to the little heatsink.

The key is increasing the surface area and setting it up in such a way that air will be drawn from under it and dissipated above it through convection. Can't get more specific than that without seeing your chassis.

 

[Ty_Bower]

I think i need a "massive heatsink". Where to I get them?

Scavenge them off old solid state amplifiers, or from old Pentium II processors.

 

[Doc Jr 8156]

Thanks Russ and Ty. My amp board is on the "open" so the heatsink will be ventilated from all sides. I'm working on a heatsink 3X the size of what I used before and hopefully this will work. Godspeed.

John Revilla (Doc Jr 8156)