Not another 300B PSE (plea for help) HUM thread?!

[Audionorth]

Well yes . . . I can't help it.

My 300B PSE amps haven't been satisfactory since completing them a couple of years ago. I've ran searches on this site (and WAD site) for other threads regarding hum issues and tried many different approaches but have yet to acheive success.

Strategies tried:

I've placed a thin layer of insulating foam/rubber and felt between Mains, O/P T's and chassis including rubber and nylon washers to reduce mechanical vibration.

I've upgraded BR1 , BR2 and I am running my 300Bs with DC (pi filters w/ shottky diodes).

(All upgrades/pi-filters were made from information and assistance from members on the old WAD site.)

I've tried changing out tubes (except 300Bs since I don't have an extra set laying around.)

I've turned the unit over and with power on and my trusty gloves moved wires around with no success. My input signal wires are shielded and routed according to the instructions.

Measuring the (ac mV) levels with speaker terminals connected to 8R load:

amp 1
No feedback and shorted input; 15.8mV
No feedback and no shorted input; 2.1mV
With feedback and shorted input; 5.1mV

amp 2
No feedback and shorted input; 14.2mV
No feedback and no shorted input; 4.1mV
With feedback and shorted input; 5.2mV

Some times (not always) the hum coming from the mains T alone (not connected to a speaker) is audible 4-6 feet away.

One thing I recently noticed is that my DC to the 300Bs is about 5.15v instead of 5v. I will certainly fix this but I didn't think that could account for all the hum. I experienced no noticable decline in hum levels going to DC with the 300Bs.

Apart from taking a hammer to them, do any of you have any suggestions. I am not overly knowledgable in this area so please explain your insights in simple (electronics) language so I can hopefully understand and follow your suggestions.

Thanks in advance for your willingness to share your wisdom.

Mike
Anchorage, Alaska

[LGarfirth]

Hi Mike, I had a similar problem on mine with the noise level going up when the inputs were shorted. The problem, it turned out was that I had tried to make the input stage too neat.

The 10k, 100k input network must have the leads cut very short, formed into a triangle and soldered directly to the valve pin and base. On mine the lead are so short the joint is hard against the body of the resistors. The peak of the triangle then joins to the inner portion of the signal cable which is routed along the top corner of the length of the chassis. I used a 12inch long piece of gaffer tape to hold it in place. In order to keep the outer portion of the input signal cable (screen) tidy, I cut that short and soldered a short piece of wire to connect it to the valve base.

At the other end the signal cable fixes directly to the input phono and I found that an additional 100k resistor across the socket made a huge improvement to the noise level. This does have an effect of reducing the gain a little - but providing you are not using a really lossy passive pre should not give you any problems.

Also, I found I never needed to use feedback and there is a significant noise advantage in simply disconnecting the feedback circuit and hard wiring it in the off mode. In addtion this means the output tx can be left floating and removes another route for noise.

If you are running your 300b's DC you can remove the entire hummbucker circuit and replace it for a fixed value network, this helps too and gets rid of those horrible thick wires.

I found adding a regulator to the power supply of the 6AU6 & 12AU7 with a 4700uf smoothing cap. helped as well. This I built on vero board and fixed to the side of the chassis, nothing fancy standard silicon bridge and low dropout 5v reg from maplin - worked like a charm.

In the past I have noticed some builders to replace the 5U4 rectifier with a solid state version and strongly reccommend against it as it can cause all sorts of long term problems, including strain on the output TX and 300b's.

Currently my amps measure >1mV noise and sound wonderful.

I hope this helps, if you need any more help please feel free to drop me a line. email@lincolng.co.uk

[JDA99]

I had a problem with hum and found that some of it was due to mechanically transmitted vibration.
Decoupling the amplifier from the surface it was mounted on reduced it enormously.
You could use vibropods or a suitably sized piece of MDF mounted on squash balls cut in half.

JDA

[Paul Bannister]

I have a pair of silent 300B PSE’s. On the input I have both signal and return leads running inside a screening sheath connected to star earth at the input end. R2 is soldered hard onto valve pin 1 on 6AU6.
All signal and power valves have individual dc. supplies. No special gaskets used on Tx’s. I paid a fair amount of attention to star earthing, although not as much as Morgan Jones advocates - (must try harder next time.)
I still have the humbucker – although it’s redundant in theory and does not appear to make any audible difference, one can still meter the varying voltage on the second 300B and use the reading/humbucker to match to the first 300B. Feedback never fitted and battery bias installed on 6AU6 in initial build.
You could look for a thread on this site for building a simple filter to remove spurious dc voltage on the incoming mains. This may cause your mains tx to run hot and possibly being noisy. It’s a diode(s) and caps arrangement and can sit inside or outside the amp if space is an issue.

[Audionorth]

Thanks for the replies. It may take a bit of time for me to get to these suggestions but I will be implementing them in the near future. I may of course have a follow question or three. . .

Paul, when you said pay "a fair amount of attention to star earthing" what does that entail beyond the kit instructions?

Mike

[Paul Bannister]

I took back each of the earth connections to star points mounted near the choke. MJ talks about grading the connections. I just had separate signal and PSU groups which joined together at the star point.

[LGarfirth]

Hi again Mike,

Although I really like Paul's method for the star earth (and I might even try it myself) I just wanted to say that I had configured the star earth as per the original build instructions and for me, although maybe not the neatest solution seems to work fine.

Regards

Lincoln

[Audionorth]

Just to follow up with this thread, apparently perserverence does pay off. I've sorted out one of the amps (.9mVac and .4mVac with f/b) and I'm currently working on the other. The completed one is virtually undetectable through my speakers (93db).

Thanks for all those who helped me with this and gave me ideas. Especially BB member LaGarfirth. Thanks!

[Max N]

Hi Mike
Are you able to pinpoint what made the biggest difference? Or did you just try all of Lincoln's suggestions together?
Cheers
Max

[Audionorth]

What made the biggest change?

It is difficult to tell because a number of different things were tried but I suspect the biggest was the fact that I had made a few mistakes with the grounding/star earth implementation and once that was made right, in addition to running dc via a pi-filter through BR2 to V2, everything came together.

Additional changes included:

better signal and feedback wire
shorter leads for R1 & R2 to V1
100K on the phono plug
insulated transformers from chassis
tweaked my 300b filters to exactly 5Vdc
moved wires around while amps on and attached to MM

I also upgraded the phono input and the speaker binding posts (not due to noise) but just because I wasn't a big fan of the stock parts.

I realize it would have been more instructive to make changes one by one, but once I started taking the amp apart it was difficult to stop. . . clearly my initial screw up with the earth/ground wiring was causing the biggest problems. I'm in the middle of changing my other amp, so I will see if I can repeat the good results.

Lincoln really helped me sort out the earth problem. He made the mistake of giving me his phone number so my bill from calling the UK from Alaska may be a bit high but the results made it all worth while.

I hope that helps a bit.

M-