300B Hum

[steve]

Well here is yet another thread on 300B PSE hum.

After much fiddling around with gaskets for the mains transformer, and swapping the positions of the insulating nylon transistor insulating washers the only way I can get these buggers to be quiet is to sit them on rubber tap washers without the M5 x 80 mounting screws. Clamping the laminations tightly did not help either.

I also have around 40mV p-p of 50Hz hum (no feedback) on the speaker outputs on each bloc (no earth loops - happens with input tied to ground via 47K testing plugs). If I remove ECC82, hum drops to around 10mV. Whilst this is generally not a problem as the hum can only be heard at a max of 2.5 foot from the speakers, when combined with the TX mechanical noise it can get a bit annoying on quiet passages of music.

Anyone any suggestions on making the amp hum a little quieter?

An interesting point I observed whilst testing. I am sure many of you have your 300B's side by side on a suitable table. I you switch off one side, you can still hear a slight hum on the speaker connected to the bloc that is off. If the blocs are separated by about 1.5 foot the hum goes away - looks like the output TX of the amp which is off is picking up the field of either the mains TX or output TX of the amp that is still is on. May be worth experimenting on moving you amps further apart to ensure best sound quality!

[Richard]

Steve

You seem to have hum from 2 sources, mechanical from the TXs or chokes and electrical from some circuit problem.

It is possible the 2 are connected in that electrical hum could cause a mechanical vibration.

Therefore I'd tackle the electrical hum first. There's no way it should be "around 40mV p-p of 50Hz hum (no feedback) on the speaker outputs on each bloc". 4mV at the speakers would be more than enough and many will look to reduce from that by going dc heating etc.

If you have measured 40mV with a meter then you need to get to the bottom of it. Start with 1 bloc at a time out of the system. Short the input and put 8R on the output. Fire up and check hum across the output with the meter.

If it is now acceptable the problem lies elsewhere with the system or earthing. Swap the 8R for a speaker and measure again to see if your cable/speaker load is upsetting it. If it's still OK add the source to the input, then finally connect the second bloc in the system. This sort of logical path should lead to the problem, possibly earthing.

If it is still 40mV with shorted input and 8R on the output then start looking inside the amp. "If I remove ECC82, hum drops to around 10mV." suggests ht to me. Hum on the ht will be amplified and passed down as signal in an se. Perhaps removing the ecc82 just leaves the hum from the ht on the op stage. Is the hum the same on both blocs? Is the 6au6 dc feed OK and the blue TX wire for its centre tap connected to earth? What about the hum bucker? Is it working? Have you tried swapping valve pairings around to try and get 2 which are closely matched? And the favourite is earthing again - the paths must be exactly as per the build sheet. One member had a problem of a component shorting under the tagboard where he couldn't see it.

Rich

[steve]

Hi Richard,

borrowed a scope from work so I can make some proper measurements. First, now I have a calibrated device its not 40mV but 15mv p-p. (Cheap DVM obviously not good at low volatges). This drops to around 5mV with the ECC82 removed. Removing the 6AU6 does not have any impact on the 15mV p-p output.

The problem is exactly the same with both blocs, I have shorted inputs and have an 8R test load. (Its not an external earth loop problem). Yes the blue CT DC feed for the 6AU6 is grounded and I did follow the wiring routing exactly. Both amps work perfectly apart from the TX and speaker hum. No components touching anywhere they should not be. Checked all earthing, looks fine. Hum bucker works fine and has been adjusted for minimum hum. Valves are matched pairs but moving them round has no impact.
(If I have made a mistake somewhere I have done it twice, but I don't think so - all relevent mains TX wires to star earth on tag then 10R to IEC earth+Earth post. All other earth/tag connections to star earth. Input phono insulated from chassis).

1. This is the output waveform at the speaker O/P across the 8R load SPK.GIF
SPK.gif
2. This shot shows the grid of the ECC82 and the waveform is threfore similar on the grids of the 300B's ECC82.GIF
ECC82IP.GIF
scope setting 5mv/cm time base 5ms, therefore, frequency = 50Hz.

HT ripple is about 100mv 100Hz at Point A (top of output TX). As you will see from the various waveforms in the attachments they are all 50Hz not 100Hz rectified HT.

Point B (6AU6 Anode) no significant ripple.

The mechanical noise is from the mains TX not the choke. Mechanical TX noise goes away when I lift mains TX off chassis.

Any further thoughts?

Thanks..Steve.

[NickG]

I would have said the first plot was 50hz+100hz, and a classic example of the hum pot not being quiet right. Are the 300b's AC heated?

0.1v seems a lot for the B+ line, can someone remind me what the power supply consist of, that would give a clue what the expected ripple should be.

If you are getting 50hz ripple on the B+ thats a very very odd sign. Check the input to the first cap to make sure both the halfs of your rectifier are conducting. Thats the first reason that springs to mind why you are getting 50hz ripple.

[steve]

Hi Nick,

perhaps my wording was a little confusing. "HT ripple is about 100mv 100Hz at Point A (top of output TX). As you will see from the various waveforms in the attachments they are all 50Hz not 100Hz rectified HT."

The HT ripple is 100Hz. The waveforms are parts of the amp circuit not the PSU. I was trying to make the point that the hum heard on the output isn't just rectified HT.

Yes the 300B's are AC heated. I agree could be a combination of 100Hz HT + 50Hz from elsewhere.

The amp diagrams are attached. However, these are old copies. In the PSU diagram R19 is actually 15K not 22K as is shown.
300bcrt.pdf

[russell allen]

I had a problem like this with mine, I was cured by tuning the amps upside down, powered up, standing on books, then moving the ecc82 wiring around with a plastic drinks stirrer. The problem was some coupling of HT and signal wiring caused by the position of some of the cables. This was suggested by Dr John and worked a treat, He told me my wiring was probably too neat, which it was, I removed some cable ties and separated the wiring with the drinks stirrer and suddenly the hum drops from 50mV to 4mV with a little tweaking. This was more than ample reward for messing up my lovely neat wiring. Worth a try

Regards

Russell Allen

[steve]

Hi Russell,

good suggestion - I will investigate this. Have not cable tied many of my bundles together just the heater leads in fact, but I will give it a go and see if there is any influence especially since it seems related in part to the ECC82. Fancy being accused of being too neat