Starving Student II Headphone Amplifier

I finished my Staving Student and it is playing music! I use CCS setting to begin with. Before inserting tubes I measured at pin 3 on tube sockets and there was exactly 19.0 VDC.

I have an issue with hum. With Vol. turned down and no input hum gets worse if my hands comes close to the tubes. If I touch chassis it gets worse but if I touch Gnd on either input or output RCAs hum is almost gone. So it must be a grounding issue. Is chassis floating or where is the signal Gnd to chassis supposed to be?

Is it the nut on the vol. pot that makes the gnd to chassis?
I noticed that the cutouts in the front panel is black but 6L6s was pure aluminium so he may get a better galvanic connection.

If I turn up the vol to about middle the hum change from a buzz hum to a more clean not so loud hum and at max vol. it is back to a buzz hum sound.

At vol. set to about 9 o'clock it plays quite loud. My DAC probably has large output. Probably 2Vpp or more.

If I make a connection using a wire (just by hand) from a chassis screw to the gnd of one of the RCA output......it is silent. So now I just have to find out where the "official" Gnd to chassis should be?

I tested using a DMM the connection from the pot nut to chassis and there is no connection. Pot nut has connection to Gnd so I am quite sure if I scrap off the black "paint" or find a shim which can cut through the black layer the hum is gone……..
 

Attachments

  • DSC_3957_00001.jpg
    DSC_3957_00001.jpg
    318 KB · Views: 687
  • DSC_3958_00001.jpg
    DSC_3958_00001.jpg
    303.1 KB · Views: 683
Last edited by a moderator:
I was able to get electrical contact between pot nut and chassis and this reduced the hum a bit and hum is not dependent of vol setting anymore. But I have still a bit of hum. If I lay hands on chassis and wrap my hands around the tubes without touching them hum is gone so I guess the tubes pick up some hum?

I have no Earth ground in the 230 VAC outlets.

I tried to screen one of the tubes and that made it almost silent in both channels. But only if I connect the end of the solder to be chassis. If I screened both tubes like you sometimes see in old tube amps it may make it 100% silent. Seems there is something very high impedance capacitive coupling from the surroundings (eg from my hands)?

I also just checked the two LEDs inside the chassis which controls the CCS. They also lights up nicely.
 

Attachments

  • DSC_3959_00001.jpg
    DSC_3959_00001.jpg
    532.3 KB · Views: 663
Last edited by a moderator:
I tried to screen one of the tubes and that made it almost silent in both channels. But only if I connect the end of the solder to be chassis. If I screened both tubes like you sometimes see in old tube amps it may make it 100% silent. Seems there is something very high impedance capacitive coupling from the surroundings (eg from my hands)?

The plate circuit of the tube is a very high impedance node, especially when using the CCS. So it is susceptible to picking up environmental noise. You are a big antenna, so bringing your hand near the tube couples line voltage noise into the plates.

I think you figured out that the pot frame is connected to ground. The issue with black anodize is that it is non conductive, so to get good contact you may have to remove some of the anodizing around the volume control shaft.

Bonding the chassis to a real ground might help things, but it sounds like you may not have a real ground easily available.

Pete
 
The most likely error is that the grounded end of R11 isn't really soldered to ground. It's easy to get a solder joint that looks good, but actually has not flowed onto the grounded pad. It's connected to a large copper area that sinks heat away, so it is difficult to solder.

If this happens, the gate of Q3 is biased to 48V, so the source (and VT1 pin 3) goes to 48V - Vgs(th) or about 43V.

With power off, use an ohmeter to confirm that R11 is 200k, and R15 is 221k. And measure from the grounded end of R11 (on the resistor lead) to ground to see if it is connected.

Also check for solder bridges on Q3.

Pete

After reflowing R11, VT1 pin3 to headphone ground now measures 20.5 volts.

Plugged in the tubes and headphones but there's a lot of loud crackling after the 20 second turn on. The tube in VT1 glows bright orange while VT2 tube I barely see any orange.
 
The plate circuit of the tube is a very high impedance node, especially when using the CCS. So it is susceptible to picking up environmental noise. You are a big antenna, so bringing your hand near the tube couples line voltage noise into the plates.
I think you figured out that the pot frame is connected to ground. .

Pete


Thank you for the answer. Yes, I removed some of the black "layer" and got electrical contact to the pot. I will experiment a bit more with grounding and also change from CCS to resistor to check this out. A shield around the tubes could also help.

I was thinking of adding a grounding screw to the chassis for easy experiments with grounding. Like on some phono pre-amps for grounding the turntable. First I will switch to resistor "load". Then the 47.5 k should "ground" the plate through the PSU I guess and then reduce the plate impedance. I learned many years ago that a CCS has infinite impedance :) …..it is a good learning project also. That is the funny part of DIY.

I switched from CCS to resistor and in this mode hum is only in one channel (in CCS mode I also think hum was much worse in one channel). It follows the tube. If I switch tubes hum also switch channel. So some tubes seems to be more sensitive for picking up hum than others? I think I may prefer this mode for playback. But I may also "hunt" for a shielding solution of the tubes.


Some of the "hardware" they sell here will probably work?
https://www.surplussales.com/Tubes-Sock-Acc/TubeShields-1.html

The sound is very good from this amplifier used as headphone amp. The subjective impression is that bass seems deeper than when I use the Phone output on my DAC. I found shields from my "local" store in Germany. Think it will look good when installed. Will post pictures and results when installed. It only requires two M3 holes for each shield. For now I use my "solder roll" on the humming tube. 60/40 solder.....have not tried using solder with silver :)
 
Last edited by a moderator:
After reflowing R11, VT1 pin3 to headphone ground now measures 20.5 volts.

Plugged in the tubes and headphones but there's a lot of loud crackling after the 20 second turn on. The tube in VT1 glows bright orange while VT2 tube I barely see any orange.

Can you check the voltage on pin 3 with the tubes installed? Both should be around 19-20V.

The crackling is again usually indicative of a cold solder joint somewhere...

Pete
 
Some overall observations from the build. This was probably one of the more difficult boards I have ever worked on. Some of the pads were so small it was difficult to get the solder to flow well. I tried using Cardas as well as Kester 44 solder, neither was easy to work. Was using a Hakko iron with a fine pointed tip. When completed and ready for the smoke test, no LED's lit up. I went over the solder joints and tried again and success! I checked the voltage from ground to pin 3 of both tubes and got about 18.95 volts. I plugged the tubes in and fired it up and was getting about 17. 5 on one and 17.72 on the other. I have hooked it up to my ACA and ran it for about a half hour using my phone as an audio source. Everything sounded very good, no noise at all.
 
I'm on the notify list. This will be used at work with a pair of HD-600 Sennheiser cans and an extra DAC I have laying around for a source.



I'm not a big fan of the blue LED's lighting up the glass. I'll probably mount them on the opposite side of the PC board. I'll do that before I mount the tube socket. ;)
 
I had the same idea but ended up mounting the blue LED's in the socket. There are not much natural tube glow in this kind of tubes. The "socket LED" seems to have a real purpose and act as a "cathode resistor" to control the bias and lift up some of the grids to a few Volts.
 
...The "socket LED" seems to have a real purpose...
Absolutely! That is why Natural Sound said he is going to mount them downwards, not avoid them altogether! It would be a serious mistake to avoid them!

Natural Sound, I felt the same way when I saw the 'fake' light not being emitted by the tube itself. My only concern about your approach is maintainability. Should something happen to that led, you need to actually desolder the entire tube socket to change a 3 cent piece? Seems like a lot of effort. Perhaps just tape the led with some black tape so it doesn't emit light? Or put a small piece of 'something' on top so it doesn't bleed light into the tube?

Don't know.
 
I am using B&W P7 headphones in the moment. It has its sensitivity specified as 111 dBV/uBar. Is that a high sensitivity or how does compare to modern specifications?
Most uses dB SPL/mW?

Those units are incorect. 111 dBV is something like 350kV, and 1uBar is ~73dBSPL. Perhaps it should be 111 dBSPL/V.

This page: Measurement's report Bowers & Wilkins P7 test and graphs - Reference Audio Analyzer indicates that 0.233mW gets you +94dBSPL. That would be 100dBSPL/mW. Or 116dBSPL/V. Assuming I did the math right.

That is higher than power average efficiency. The impedance is quite low, so it takes very little voltage to drive them.

Pete